Content uploaded by Robert Gaßner
Author content
All content in this area was uploaded by Robert Gaßner on May 12, 2014
Content may be subject to copyright.
Content uploaded by Robert Gaßner
Author content
All content in this area was uploaded by Robert Gaßner
Content may be subject to copyright.
Methods of Future and Scenario Analysis
Hannah Kosow
Robert Gaßner
Overview, Assessment, and Selection Criteria
Methods of Future and Scenario Analysis
German Development Institute (DIE)
The German Development Institute is a multidisciplinary research, consul-
tancy, and training institute for Germany’s bilateral development coopera-
tion as well as for multilateral development cooperation. On the basis of in-
dependent research, it acts as consultant to public institutions in Germany
and abroad on current issues of cooperation between developed and devel-
oping countries. In a 9-month training course, the German Development
Institute prepares German and European university graduates for careers in
the field of development policy.
Hannah Kosow is a researcher at the Institute for Futures Studies and
Technology Assessment (IZT), Berlin. Hannah studied social and political
sciences at the University of Stuttgart and at the Institut d’Etudes Poli-
tiques, Bordeaux with a main focus on technological and environmental so-
ciology, political theory and public and political communication. Since
2006, she has worked at IZT in the areas of technology assessment, user ac-
ceptance, risk assessment, and participatory methods, as well as futures
analysis and scenario analysis. Hannah’s research focuses on new tech-
nologies in health care and on methodological and empirical questions re-
lated to participatory and future oriented approaches.
E-Mail: h.kosow@izt.de
Robert Gaßner, a psychologist by training, is a senior researcher at the In-
stitute for Futures Studies and Technology Assessment (IZT), Berlin. Since
1985, Dr. Gaßner has worked in the field of interdisciplinary technology as-
sessment and technology design. In recent years, his research has included
work on sustainable development and general and methodological ques-
tions of futures research, particularly using scenario planning and other par-
ticipative approaches. Dr. Gaßner also serves as a facilitator for futures
workshops, future search conferences and scenario meetings. From 2001 to
2005 he acted as scientific advisor for the German Foresight Dialogue “FU-
TUR”.
E-Mail: r.gassner@izt.de
Studies
Deutsches Institut für Entwicklungspolitik 39
Methods of future and scenario analysis
Overview, assessment, and selection criteria
Hannah Kosow
Robert Gaßner
DIE Research Project “Development Policy: Questions for the Future”
Bonn 2008
Studies / Deutsches Institut für Entwicklungspolitik
ISSN 1860-0468
Kosow, Hannah: Methods of future and scenario analysis : overview, as-
sessment, and selection criteria / Hannah Kosow ; Robert Gaßner. DIE Re-
search Project “Development Policy : Questions for the Future”. – Bonn :
Dt. Inst. für Entwicklungspolitik, 2007 – (Studies / Deutsches Institut für
Entwicklungspolitik; 39)
ISBN 978-3-88985-375-2
Dt. Ausgabe u. d. T.: Methoden der Zukunfts- und Szenarioanalyse :
Überblick, Bewertung und Auswahlkriterien. – Berlin : Institut für Zu-
kunftsstudien und Technologiebewertung, 2008. – (WerkstattBerichte Nr.
103). – ISBN 978-3-941374-03-4
© Deutsches Institut für Entwicklungspolitik gGmbH
Tulpenfeld 6, 53113 Bonn
+49 (0)228 94927-0
+49 (0)228 94927-130
E-Mail: die@die-gdi.de
http://www.die-gdi.de
!
"
Contents
Summary 1
1 Introduction 5
2 Overview of scenario methods 8
2.1 Procedure and Sources 8
2.2 Basic principles 10
2.2.1 What is a scenario? 10
2.2.2 Basic assumptions: The understanding of the future
which is implicit in scenario methods 13
2.2.3 To what end can scenarios be used? 18
2.2.4 When are scenarios inappropriate? 21
2.3 Methodological commonalities and differences 22
2.3.1 The range of the field of scenario methodology 22
2.3.2 General phases of the scenario process 24
2.3.3 The basic characteristics of scenarios 30
2.3.4 Scope 35
2.3.5 Criteria of quality and process criteria 38
2.4 Three ideal-typical scenario techniques 42
2.4.1 Scenarios on the basis of trend extrapolation 44
2.4.2 Systematic-formalized scenario techniques 50
2.4.3 Creative-narrative scenario techniques 61
2.4.4 Interim assessment: Scenario techniques in overview 75
2.4.5 Excursus: Techniques of scenario transfer 79
2.5 Method combinations 83
2.5.1 Scenarios and modelling methods and/or simulations 83
2.5.2 Scenarios and Delphi surveys 87
2.5.3 Scenarios and roadmapping 90
3 Dimensions of selection for the application
of scenario methods in the development policy field 93
3.1 Basic questions on the application of scenario methods in DP 94
3.2 Defining the underlying conditions 96
3.3 Selection of a concrete scenario technique in DP 97
4 “Check-List” for the selection of suitable scenario
methods for the DIE project “Development Policy:
Questions for the Future” 110
Bibliography 115
Figures
Figure 1: Funnel-shaped span of possible developments
of individual factors 15
Figure 2: The scenario funnel 16
Figure 3: The general scenario process in five phases 25
Figure 4: Trend extrapolation, forecast, “business as usual” (BAU) 46
Figure 5: Diagram of a trend variation with TIA 49
Figure 6: Constant expansion of the “funnel into the future”
through systematic-formalized scenario techniques 54
Figure 7: Widening of the “funnel into the future” by means
of creative-narrative scenario techniques
(shown here in simplified form) 61
Figure 8: Description of a complete permutation, taking
population development as an example
(in simplified form) 62
Figure 9: Sample description with morphological analysis 68
Figure 10: Sample description of the analysis of key factors
and normative dimensions in the context of
normative-narrative scenarios 73
Figure 11: Transitions between “ideal-typical” scenario techniques 78
Figure 12: Backcasting 82
Figure 13: Example: Quantification of scenarios in the
modeling process 85
Figure 14: Example: Input of Delphi-results in a scenario process 88
Figure 15: Example of backcasting from conceptual futures
in the roadmapping process 92
Figure 16: Diagram of the multidimensional “DP scenario field” 103
Figure 17: Basic organizational dimensions of a scenario project 110
Tables
Table 1: Schematic comparison of explorative and
normative scenarios 32
Table 2: Comparison between quantitative and qualitative scenarios 34
Table 3: Tabular explanation of the influence matrix 52
Table 4: Consistency matrix 56
Table 5: A cross-impact matrix 58
Table 6: The morphologic box 67
Table 7: Overview of different scenario techniques 76
in the scenario process
Table 8: The phase of backcasting 83
Abbreviations
AS Active Sum
BMBF Bundesministerium für Bildung und Forschung /
Federal Ministry of Education and Research
BMZ Bundesministerium für wirtschaftliche Zusammenarbeit
und Entwicklung /
Federal Ministry for Economic Co-operation
and Development
CIA Cross-impact analysis
DIE Deutsches Institut für Entwicklungspolitik /
German Development Institute
DP Development Policy
GTZ Deutsche Gesellschaft für Technische Zusammenarbeit /
Association for Technical Cooperation
IZT Institut für Zukunftsstudien und Technologiebewertung /
Institute for Future Studies and Technology Assessment
MA Morphological analysis
PS Passive Sum
SRI Stanford Research Institute
TIA Trend impact analysis
Methods of future and scenario analysis
German Development Institute 1
Summary
The future context that development policy will have to respond to is
both complex and uncertain. This study provides an overview and eval-
uation of methods of futures research and scenario analysis methods in
particular in order to identify how these methods might be applied to
research and policy advising in the development policy arena. Al-
though scenario analysis methods have been applied in a variety of
contexts, the literature on these methods has to date provided limited
guidance on how to select appropriate scenario techniques and how to
evaluate scenario exercises. This study addresses this shortcoming by
outlining three main categories of scenario techniques (scenarios based
on trend extrapolation, systematic-formalized scenario techniques, cre-
ative-narrative scenario techniques) and discussing common applica-
tions and strengths and weaknesses of these varied approaches.
A scenario can be defined as a description of a possible future situation,
including the path of development leading to that situation. Scenarios
are not intended to represent a full description of the future, but rather
to highlight central elements of a possible future and to draw attention
to the key factors that will drive future developments. Many scenario
analysts underline that scenarios are hypothetical constructs and do not
claim that the scenarios they create represent reality.
This study outlines several functions that scenarios can serve. First,
scenarios can be used to generate knowledge about the present and the
future and to identify the limits of that knowledge. Second, scenario
analysis can serve a communicative function, since scenario develop-
ment is often based on an exchange of ideas between people with dif-
ferent perspectives. Scenarios may also be used as a public communi-
cation tool to draw attention to specific issues. Third, scenarios can aid
decision makers in formulating goals. Finally, scenarios can provide a
tool for examining the potential effectiveness of organizational strate-
gies.
Although there are many different kinds of scenario analysis tech-
niques, the scenario process unfolds in a broadly similar manner across
these varied approaches. The first phase of the scenario process deals
with the identification of the scenario field by establishing the precise
Hannah Kosow / Robert Gaßner
2German Development Institute
questions to be addressed and the scope of the study. In the second
phase, researchers identify the key factors that will have a strong influ-
ence over how the future will unfold. The third phase then examines
what range of outcomes these key factors could produce. This phase is
followed by a fourth phase that involves condensing the list of central
factors or bundling key factor values together in order to generate a rel-
atively small number of meaningfully distinguishable scenarios. The fi-
nal phase of the scenario process can be labelled “scenario transfer”
and involves applying the finished scenarios for purposes such as strat-
egy assessment.
The techniques used in the scenario process depend on the general ori-
entation of the scenario exercise. Scenario analyses can be distin-
guished on the basis of whether they are normative or exploratory in
nature, with normative scenarios aiming to chart paths to desirable fu-
tures and exploratory scenarios aiming to identify possible develop-
ments regardless of their desirability. Scenario analyses may also be ei-
ther quantitative or qualitative in nature. The advantages and disadvan-
tages of these alternative orientations are discussed in this study.
This analysis proposes several criteria that can be used to assess the
quality of scenario exercises, many of which can also be used to eval-
uate other forms of research. Scenarios can be judged by their plausi-
bility, internal consistency, comprehensibility and traceability, distinct-
ness, and transparency.
The choice of an appropriate scenario technique depends on the goals
of the research project and the context in which this research takes
place. This study outlines a number of key questions that researchers
should ask prior to undertaking a scenario analysis and on this basis de-
velops a checklist for the selection of suitable scenario analysis meth-
ods in the development policy field.
Researchers should for example be careful to identify whether project
goals require the articulation of multiple alternative futures rather than
making predictions on the basis of readily available data. At the outset
of a scenario process, it is also critical to identify the target audience
and to specify the nature of organizational resources that can support
the scenario development effort. In the development policy field, there
Methods of future and scenario analysis
German Development Institute 3
are several foreseeable goals of conducting scenario exercises. These
goals may be exploratory or in contrast related to establishing concrete
targets to achieve. Scenario exercises may also serve to encourage net-
working among actors or to sensitize external actors to critical issues.
In some cases, the goal of promoting internal networking suggests that
greater attention should be placed on the design of the scenario process,
while scenario exercises aimed at sensitizing external actors should pay
special attention to the manner of description of the scenarios them-
selves.
The study stresses that researchers seeking to apply scenario methods
should carefully consider how they can best manage the complexity of
the subject matter scenario exercises attempt to deal with in a manner
that fits with their existing organizational resources. Important deci-
sions that researchers need to take relate to the geographical, thematic,
and chronological scope of the scenario project, as well as to the selec-
tion of the participants that will be involved in the process.
In conclusion, the study offers a short list of key recommendations for
applying scenario methods to examine questions for the future of de-
velopment policy. The selection of appropriate methods should follow
from an exhaustive delineation of goals and priorities of the scenario
project. Researchers should avoid a purely quantitative approach and
acknowledge the normative elements of questions related to the future
of development policy. Rather than conducting a global scenario exer-
cise, it is also advisable for researchers to divide the scenario analysis
into a number of smaller, more focused, projects. Finally, the study em-
phasizes that the ultimate target audience for scenario analyses regard-
ing the future of development policy should be involved in the scenario
generation process in order to strengthen the legitimacy and overall ef-
fectiveness of such an undertaking.
Methods of future and scenario analysis
German Development Institute 5
1 Introduction
The future of development policy – like everything else involving the fu-
ture – is full of complexity; developments and shifts in mutual interactions
at many levels on the world stage follow courses which are at times unbro-
ken, but also at times disruptive. In addition, the future of development pol-
icy is of its very nature characterized by uncertainty and unpredictability.
Whereas the potential for numerous, fully different paths into the future is
always present, it is also the case that final selection of a single future di-
rection and/or the emergence of a single future course automatically ex-
cludes certain alternatives while simultaneously, in most cases, opening up
a multitude of other possibilities for moving into the future. For this reason
it makes sense to speak in the plural of the “possible futures” of develop-
ment policy. In turn, these “futures” of development policy are themselves
marked by ambivalence, inasmuch as different possibilities for develop-
ment themselves will be – or can be – evaluated quite differently depend-
ing on the standpoint of the viewer.
In the field of study and consultation regarding development policy (DP),
it has mostly been the case that questions related to the future have received
little explicit attention. Nevertheless, a study of the “futures” of DP appears
highly relevant in light of the ever-increasing complexity and unpre-
dictability of the framework conditions of DP, including, for example, glob-
alization, climate change, the dynamics of energy and raw materials mar-
kets, the risks and conflicts of maintaining political security, and techno-
logical revolutions, and in view of the internal transformation processes to
which DP itself is subject, it becomes important to reflect on decisions be-
ing made today as a means of orienting DP in such a way as to make it vi-
able for the future.
Futurology, i.e. “the scientific study of possible, probable and desirable fu-
ture developments, the options for shaping them, and their roots in past and
present” (Kreibich 2007, 181), offers a set of instruments and a rich store
of methods for the generation of orientational and future-oriented knowl-
edge. Kreibich names the following methods (Kreibich 2006, 12):
“Trend analysis and trend extrapolation; envelope curve analysis; rele-
vance tree techniques; morphological methods; analogy techniques; in-
put-output models; techniques involving questionnaires; surveys of ex-
perts and interview techniques; cost-benefit analysis; cross-impact analy-
Hannah Kosow / Robert Gaßner
6German Development Institute
sis; innovation and diffusion analysis; construction of models and simula-
tion techniques; brainstorming; Delphi methods; scenario methods; role-
playing; creativity methods; future workshops.“
The present study will undertake to investigate how this body of method-
ological knowledge of futurology can be made fruitful for those who carry
out research and provide advisory services in the context of DP. In the
process, the primary focus here will be on scenario methods. The reason:
work with scenarios is central to futurology and one of its most widely used
methods (cf. Steinmüller 2002b, 3). It constitutes one of its most compre-
hensive and complex approaches, and often integrates within itself differ-
ent methodological manners of tackling issues, such as scientific tech-
niques, evaluation techniques, decision-making techniques, event-shaping
techniques, and participative techniques (cf. Grunwald 2002, 226).
Viewed historically (cf. among others Steinmüller 2000, 37 ff.; Mietzner /
Reger 2004, 48 ff.), it has been customary since the 1950s to develop sce-
narios in the context of strategic military planning. At the end of the 1960s,
however, companies like General Electric and Royal Dutch Shell began for
the first time to use scenarios and, in this context, to develop the first ener-
gy scenarios. Scenarios came into the eye of the general public on the ba-
sis of computer simulations with the report of the Club of Rome on “Lim-
its to Growth” (1972). Today, scenarios are used in all sorts of contexts.
Among their primary fields of application are strategic planning in compa-
nies, municipal and land-use planning, political consultancy, and global
scenarios concerning the future of energy or the climate. Numerous differ-
ent scenario techniques have been developed for the various fields of ap-
plication.
The present study has two goals: first, to present to the German Develop-
ment Institute (DIE) a qualified overview of methods used in futurology
and, in particular, to present scenario methods which could be used in the
area of development policy. The intention was to widen the range of possi-
ble methods within the DIE for dealing with the future avenues of DP.
The other goal is to present a study which can also serve as a practical
“handbook” within the context of the DIE project “Development Policy:
Questions for the Future” by making it possible to support the method-
ological design of this project and/or to concretize the manner in which
such scenarios might be applied within the framework of this project.
Methods of future and scenario analysis
German Development Institute 7
Against this background, Chapter 2 develops a structured overview of the
field of scenario methods. It takes as its point of departure a clarification of
concepts, basic principles, along with both the aims and limitations of sce-
nario methods. This is followed by a description of the general process
common to many scenario techniques, which in turn leads to an introduc-
tion of the criteria used to characterize and evaluate different scenario ap-
proaches, namely the basic characteristics of scenarios (including explo-
rative vs. normative, quantitative vs. qualitative), their scope (geographical,
chronological, and thematic) and criteria to evaluate their quality. Follow-
ing that, scenario techniques are grouped into three ideal types with re-
spective pros and cons: scenarios on the basis of trend extrapolation, sys-
tematic-formalized scenario techniques, and creative-narrative scenario
techniques. The techniques of scenario transfer are also presented in an ex-
cursus, along with sample sketches of some hybrid method designs in
which scenarios are combined with other methods of futurology: modeling
methods and/or simulations, Delphi surveys, and roadmapping techniques.
Chapter 3 in turn proposes a set of criteria and decision-making processes
which might make it possible to select appropriate scenario approaches for
carrying out research and providing advisory services in the field of DP. To
this end, numerous dimensions of selection are discussed and outlined, us-
ing DP as an example in each case, thus making it possible to formulate pre-
liminary recommendations for organizing scenario work in this field. In the
process, both basic questions and their underlying conditions are taken up
with regard to the selection of scenario methods, after which concrete con-
siderations regarding the organization of a scenario process in the context
of DP are presented.
Chapter 4 concludes the preceding reflections with a “checklist” type re-
sumé of dimensions for selection, along with the decision-making issues
which are involved in the DIE project “Development Policy: Questions for
the Future“. It can be used for methodically working out a concrete scenario
process, including a determination of targets, resources and scenario con-
tents.
2 Overview of scenario methods
We begin by describing the procedure used in analyzing literature for this
study (2.1). Then the basic principles of scenario methods are explained;
this involves, among other things, the definition of scenarios and the un-
derstanding of the future upon which they are based (2.2). With this as ba-
sis, the field of scenario methods is presented in its full range: first common
elements in the general phases of the scenario process are identified; then
criteria and dimensions which are relevant in characterizing and evaluating
different scenario approaches are introduced, i.e. the basic characteristics,
the scope, and the criteria of quality (2.3). Following this, three groups are
presented in ideal-typical fashion on the basis of their dimensions, each in-
volving different scenario techniques. Additionally, the techniques of sce-
nario transfer are described (2.4). Finally, some examples of method de-
signs are sketched in which the scenarios are combined with other methods
of futurology (2.5).
2.1 Procedure and sources
This study is based on a study of the literature published to date concerning
(national and international) research and the status of experience gathered
with scenario methods. The point of departure of this research project was
a very broad understanding of the term “scenario methods”, that is, each
and every method which deals with scenarios. In the process, it was for the
moment irrelevant to the search for source literature whether that literature
dealt with the development, analysis, evaluation, or application of scenar-
ios. That is, it was of no great importance what position of importance the
scenario occupied within the respective research processes discussed in the
literature or, for example, what point of departure was used or what inter-
im or final results were obtained. This broad-based search strategy was ex-
pedient, firstly because it embraced all the different individual scenario
techniques and secondly because it included conventional combinations of
methods as well. In the process, recourse was had to research and literature
databases, along with the Internet, library catalogues, and cross-references
in the literature. In addition, the search was enhanced by surveys of experts
as well as utilization of the resources and experience already present at the
Institute for Future Studies and Technology Assessment (IZT).
Hannah Kosow / Robert Gaßner
8German Development Institute
Methods of future and scenario analysis
German Development Institute 9
Our search of the literature revealed the following preliminary situation re-
garding sources:
Basically, and first of all, literature was found concerning the various ap-
proaches to a discussion of methods (e.g. Mietzner / Reger 2004; van Not-
ten et al. 2003; Greeuw et al. 2000); a second body of literature concerned
experiences gathered from practical application (e.g. Shell International
2003). Here there were often detailed descriptions of individual techniques,
above all the rather formalistic approaches prevalent in the 1990s (e.g. von
Reibnitz 1991; Mißler-Behr 1993). Many of these descriptions were direct-
ed above all to the application of scenario methods in companies (e.g. van
der Heijden 1996; Gausemeier / Fink / Schlake 1996). In addition, propos-
als became common from the mid-1990s on for systematic overview re-
ports (e.g. Steinmüller 1997), along with volumes of collected essays which
attempted to give an overview of the field (e.g. Wilms 2006a). On the oth-
er hand, a nearly endless number of scientific studies were found concern-
ing the actual application of scenarios, together with collections of reports
on completed scenarios and scenario texts1. There is also a wide spectrum
of information on the offer of scenario services provided above all to en-
terprises.
What is not found, however, is a comprehensive or even consistent, theo-
retical and methodical substantiation for scenario methods. By themselves,
the methodological procedures of many studies are thought through only
partially or not at all; moreover, the methodological procedures of existing
studies are not always transparent. In addition, comprehensive, detailed
“toolkits” for the practical implementation of scenario methods are almost
universally absent. Fundamental sets of instructions for the selection of ap-
propriate scenario techniques are nowhere to be found; the same is true of
generalized evaluation criteria in the sense of “best practices” (cf. Mietzner /
Regner 2004, 60). It is on the whole conspicuous that when scenario meth-
ods are discussed they are more a matter of internal experience and knowl-
edge of the ins and outs of advisory services than of detailed and published
methods which are available to all (cf. Mietzner / Reger 2004, 60).
1 For example, an updated version of the “State of the Future” Reports appears annually
and documents the work of the AC/UNU “Millennium Project” (In 2007: Glenn / Gor-
don 2006). Among other things, it contains an ongoing annotated bibliography which al-
ready contains more than 650 scenario sets.
Hannah Kosow / Robert Gaßner
10 German Development Institute
2 Here the spectrum ranges from textually formulated outlines to quasi-literary descrip-
tions. Also, other medial forms of presentation (e.g. audiovisual, film) are possible (cf.
Steinmüller 2002b, 8).
With regard to source material, this situation leads to the following conse-
quences for this study: The current status of discussion concerning methods
will be used together with a study of the status of practical application, mu-
tually supplementing each other in order to permit the most complete
overview possible. In the process, the discussion will fall back on “classi-
cal” scenario methods and their application on the one hand, while on the
other hand frequent use will be made of the knowledge gained by the IZT
through practical experience, since this makes it possible here to clearly
comprehend the methods and method combinations used.
Nevertheless, a study of the literature yields a good overview of the field of
scenario methods; a few fundamental clarifications of this will first be giv-
en in the following.
2.2 Basic principles
2.2.1 What is a scenario?
“Scenario” is “a fuzzy concept that is used and misused, with various
shades of meaning” (Mietzner / Reger 2004, 50). It is also, so to speak, a
fashionable word which has come to be widely used in journalistic and
everyday language. The term “scenario” is also often used to describe the
future course of events regarding a single variable, e.g. “in the scenario of
a global warming of 3°C“. In the context of futurology, however, scenarios
can also represent far more complex products which include the interac-
tions of a plethora of variables (cf. Eurofound 2003, 88). Here too, howev-
er, “scenarios” may refer on the one hand to texts (with different degrees of
comprehensiveness and detail) (cf. Steinmüller 2002b, 7)2while on the oth-
er hand the term “scenario” may also refer to modulations of a quantitative
model (cf. Steinmüller 2002b, 6). Even within the field of futurology, there
is a multiplicity of proposals for definition. This multiplicity is directly con-
nected with the multiplicity of extant scenario methods themselves; this is-
sue will be discussed further during the course of this study.
Methods of future and scenario analysis
German Development Institute 11
Within the discussion of methods, however, it is possible to identify a ba-
sic understanding which is implicitly shared – at least by a majority of the
authors – concerning that which is to be understood under the term “sce-
nario“.
Ascenario is defined by many authors as3
– a description of a possible future situation (conceptual future),
– including paths of development which may lead to that future situation.
In contrast to a conceptual future, which merely represents a hypothetical
future state of affairs, a scenario describes the developments, the dynamics,
and the moving forces from which a specific conceptual future results (cf.
e.g. Greeuw et al. 2000, 7; Gausemeier / Fink / Schlake 1996, 90; Götze
1993, 36).
The aim behind scenarios is to generate orientation regarding future devel-
opments through an observation of certain relevant key factors. Three
things are to be noted in the process:
Firstly, a scenario is not a comprehensive image of the future; rather, its true
function consists in directing attention to one or more specific, clearly de-
marcated segments of reality.
“[Scenarios] are hypothetical sequences of events constructed for
the purpose of focusing attention on causal processes and decision
points.” (Kahn / Wiener 1967, 6)
In the process, various factors and events are deliberately included – and
others excluded – and brought into certain constellations in relation to one
another. The idea behind this work of “composition” is not to work out a
description of the “future” as such; rather, the function of a scenario con-
sists in placing the focus of attention squarely on certain interesting aspects
by means of a future-oriented involvement with a specific area of study.
Secondly, it is to be noted that the selection and combination of key factors
with regard to a future time horizon is also a construct. That is, certain fac-
tors and events are deliberately taken to be relevant or are ignored, and
3 This definition is found explicitly for example in von Reibnitz (1991, 14); Gausemeier /
Fink / Schlake (1996, 90); Götze (1993, 36); Steinmüller (2002b, 6).
Hannah Kosow / Robert Gaßner
12 German Development Institute
these are then brought into play and set in a context of interrelationship with
one other in light of certain assumptions. However, they can also be re-
structured in another way at any time. In the process, assumptions con-
cerning the relevance of factors for the period under study or even the man-
ner in which they interact with one another are suggested more or less by
the available data; however, these assumptions also require on the one hand
a well-founded body of knowledge, particularly knowledge of an experien-
tial nature, and are grounded on the other hand for the most part in subjec-
tive and thus invariably normative assessments. Quite apart from the fact
that scenarios do not represent the future as a whole, they also do not rep-
resent the future “as such“, but rather as a possible, future-oriented con-
struct of certain key factors.
Connected with this, thirdly, is the fact that every such scenario-construct
is based on assumptions about how the future might one day look: what di-
rection certain trends might take, what developments might remain con-
stant, and which ones might change during the course of time (UNEP 2002,
320):
“Scenarios are descriptions of journeys to possible futures. They
reflect different assumptions about how current trends will unfold,
how critical uncertainties will play out and what new factors will
come into play.“
These assumptions are indicative of comprehensive mental outlines and
models of the future, “mental maps or models that reflect different per-
spectives on past, present and future developments” (Rotmans / van Asselt
1998, quoted by Greeuw et al. 2000, 7). Such mental constructs are often
implicitly present in thoughts about the future; they can – and must – then
be made explicit, at least in part, via the building of scenarios.
In the process, attention must be given to the fact that scenarios have no
claim to reality and therefore do not provide a “true” knowledge of the fu-
ture; rather, they merely supply a hypothetical construct of possible futures
on the basis of knowledge gained in the present and past – a construct
which includes, of course, probable, possible and desirable future develop-
ments.
With regard to differences in the generalized definition of scenarios, one as-
pect stands out: the distinction between scenarios and prognoses. The con-
cept “scenario” is often used in contradistinction from the concept of “prog-
Methods of future and scenario analysis
German Development Institute 13
nosis” and that of “prognostics“, with all its negative connotations (cf. e.g.
Greeuw et al 2000, 7; Steinmüller 1997, 49 ff.). Prognoses are statements
about future developments which may be expected. In contrast to prophe-
cies these statements are supported by a basis of knowledge, as in the sta-
tistical extrapolation4of present and past trends (cf. Grunwald 2002, 181).
Some authors explicitly exclude prognoses, i.e. predictions based on the ex-
pected “extension” of present-day developments into the future, from the
concept of a scenario. They emphasize that it is precisely the nature of sce-
narios not to offer prognoses but rather in essence to take into account the
possibility of several alternative futures. In contrast, however, concepts like
“prognosis“, “outlook“, “forecast“, “prognostics” and “trend extrapolation”
are often equated on the one hand with scenario approaches in the areas of
market research and consultation. On the other hand, however, it must also
be recognized that classical techniques of prognosis, along with traditional
forecasting techniques, have made their way into scenario methods and are
enhanced by, although not completely replaced, by the latter. They can well
be said to represent a partial aspect of scenario approaches (cf. Steinmüller
2002b, 7).
As already indicated here, different concepts of the future and/or of knowl-
edge of the future underlie the different conceptions of what a “scenario”
is. One task of the following reflections will be to depict these different
concepts.
2.2.2 Basic assumptions: The understanding of the future
which is implicit in scenario methods
Scenario methods are used in the construction of different possible models
of the future; their purpose is to generate a body of orientational knowledge
which can serve as a compass for lines of action in the present. However,
various views or ways of understanding the relationship between the future
and the present and past are possible. Stated in ideal-typical form, three dif-
ferent views can be distinguished (cf. Grunwald 2002, 178 ff.). In turn, the
respective understanding of the future has a decisive effect on the way in
which we attempt to grapple with the future from our present position (cf.
van der Heijden 1996, 21 ff.):
4 See Section 2.4.1 for more on trend extrapolation.
The first view: “the future is predictable“: whatever will come to pass in
the future can (in principle at least) be calculated from our knowledge of
the present and past. The more knowledge we gather in the present, the
more certain is our prognosis of the future course of events. This view of
the future leads those who use it to rely above all on a statistical trend ex-
trapolation. According to this paradigm, the future is viewed as predictable
and controllable.
The second view: “the future is evolutive“. In this manner of viewing
things, our present knowledge is taken to be inadequate for predicting fu-
ture developments; the future follows a chaotic, uncontrolled, and random
path. This paradigm assumes that a purposeful control of the course of fu-
ture events is impossible; instead, emergent strategies and an “intuitive
muddling through” are the appropriate manner of dealing with future cours-
es of events.
The third view: “the future is malleable“. In this view, the course of future
events is not predictable, but neither is its development fully chaotic. The
development of the future is open to intentional manipulation and can thus
be influenced (at least in part) by our actions. This paradigm puts its trust
in strategies of intervention aimed at shaping the future, with an emphasis
on the role of those who take action, along with their goals and decision-
making processes in shaping the future.
Viewed historically, futurology has gone through various phases (cf.
Kreibich 2006, 4 ff.) which are closely coupled with these different ways of
understanding the future. In turn, a gradual evolution of paradigms has tak-
en place from the origins of futurology to the present day and parallel to the
changes which have taken place in our understanding of the future. This
evolution has consisted on the one hand in a shift away from purely quan-
titative techniques to more qualitative and/or combinative techniques which
are often more appropriate for dealing with the complexity of future (cf. al-
so Mietzner / Reger 2004, 61). On the other hand, a general shift is also rec-
ognizable from “forecasting” (i.e. prediction) to “foresight” (i.e. a look
ahead) (cf. Mietzner / Reger 2004, 60; Cuhls 2003).
As these paradigms have continued to evolve, the direction of development
of scenario methods has been more and more away from an exclusively an-
alytic-descriptive prognosis with its accompanying optimism to a more
complex view of the future (cf. Kreibich 2006, 6 f.). In view of its multi-
Hannah Kosow / Robert Gaßner
14 German Development Institute
Methods of future and scenario analysis
German Development Institute 15
plicity, however, the scenario method cannot be univocally ascribed to any
one of the above-mentioned three forms of understanding the future.
Rather, the understanding of the future which is basic to the scenario tech-
nique is marked above all by the fact that its point of departure is not any
single inevitable future but rather a set of numerous different possible fu-
tures. The concept of a “scenario” represents the idea of a single possible
future and therefore always refers implicitly to the possibility of other al-
ternative futures.
The so-called “funnel model” has established itself as a means of illustrat-
ing this open-endedness and multiplicity of the future and the possibility of
anticipating it by means of scenarios (see Fig. 1 and Fig. 2).
The basic idea behind this description5is that the farther we gaze from to-
day’s standpoint into the future, the more the number of possible develop-
Figure 1: Funnel-shaped span of possible developments of
individual factors
Source: IZT description, in accordance with Minx / Böhlke (2006, 19);
Gausemeier / Fink / Schlake (1996, 91)
==HHLLWW
W
W
WW
66
D
D
FF
GG
EE
HH
5 These graphical representations go back to von Reibnitz (1991, 38) and have been taken
over by many others (cf. e.g. Geschka / Hammer 1984, 242; Götze 1993, 40; Gausemei-
er / Fink / Schlake 1996, 91; Minx / Böhlke 2006, 19).
Hannah Kosow / Robert Gaßner
16 German Development Institute
Figure 2: The scenario funnel
Source: IZT description in accordance with von Reibnitz (1991, 38)
ments increases; the room for possibilities opens in funnel fashion into the
future. In this way, an expanding space emerges for possible future devel-
opments rather than merely one single possible future.
Looking from the present into the future, the range of possible develop-
ments on the part of individual aspects and/or factors – in this case factors
a to e, becomes ever greater. Figuratively speaking, a “funnel” of various
conceivable salient characteristics opens out for every individually ob-
served aspect of the future (as indicated by the growing cross-sections as
time goes on).
Taken together, all these individual “factor funnels” form the total space of
joint possible futures for all these aspects. In the field of scenario methods
it is common to speak of the “spread” of the scenario funnel (cf. Fig. 2).
This perspective of an infinitely spreading space of possible future devel-
opments is the genuine primary characteristic of scenario methods and sets
them apart from other methods.
The outer limits of the funnel symbolize the range of future developments
which are left out of consideration (for example because these develop-
ments are regarded as impossible).
DD
FF
DD
EE
EE
FF
66
66
==HHLLWW
W
W
WW
66
Methods of future and scenario analysis
German Development Institute 17
In the field of scenario methodology, a specific future point in time on this
scenario funnel is chosen for observation (cross-section at time ts). Various
different scenarios – here S1 and S2 – are then used to depict the space
within which possible developments may unfold. To this end, possible
courses of events for the various factors are selected for each scenario, and
these are then “condensed” into larger scenarios (as indicated here by ar-
rows a1, b1 and c1 for the first scenario and a2, b2 and c2 for the second
scenario).
The selection of factors and factor values required for the construction of
scenarios depends on what the respective researcher is interested in finding
out. From the broad range of possible developments, for example, it is pos-
sible to single out for description probability scenarios (i.e. those which in-
clude probable developments), or to condense scenarios into extreme sce-
narios (e.g. best-case, worst-case scenarios) or even wish scenarios (cf.
Steinmüller 1997, 53 with reference to Godet 1993, 56).
It must be remembered in any case that the scenario concept is based on the
fundamental assumption that numerous different alternative futures are al-
ways possible and that scenarios have the purpose of spanning the space to
be filled by possible futures.
It must also be remembered that there are different schools of thought re-
garding the use of scenarios, each of which views and applies scenarios dif-
ferently according to its own understanding of the future. For example var-
ious approaches rely with different degrees of emphasis on a certain pre-
dictability of the future (and thus on that which we can presume to know).
These approaches also differ from one another in the way in which they
very randomly include developments and discontinuities (i.e. that which we
do not yet know or cannot know) in their thinking. Finally, the approaches
differ in the degree to which they take the unpredictabilities of the future as
an occasion for shaping the future.
The tension created by these three poles – i.e. the tension between knowl-
edge of the future, the limits of this knowledge, and the possibility of in-
fluencing the future – is a constitutive characteristic of scenarios and sce-
nario methods. This field of tension includes and demarcates not only the
goals and functions of scenarios but also their limitations.
Hannah Kosow / Robert Gaßner
18 German Development Institute
6 Greeuw et al. (2000, 9), for example, distinguish between an “information function” and
an “advisory function in the decision-making process“, and Gaßner / Steinmüller (2006,
134 ff.) and Steinmüller (1999, 696 ff.) differentiate even further and distinguish in ad-
dition a communication function and a goal-setting function.
2.2.3 To what end can scenarios be used?
Scenarios are used to attain different goals and thus meet the need for dif-
ferent functions (cf. e.g. Steinmüller 2002a, 44; Greeuw et al. 2000, 9).6
As a whole, it is possible to lay out the range of these functions in ideal-
typical manner in four dimensions: first an explorative and/or scientific
function, secondly a communicative function, thirdly a function of target
concretization and creation, and fourthly a decision-making and strategy
formation function.
The explorative and/or knowledge function
Scenarios have a knowledge function and this on more than one level.
Above all, they have an explorative function inasmuch as they serve to sys-
tematize and deepen the existing understanding of contemporary develop-
ments, conditions and influences. Inasmuch as they build upon an assess-
ment of future relevant factors, they force those who use them to make ex-
plicit existing (implicit or even subconscious) basic assumptions about fu-
ture developments (Shell International 2003, 12). They also serve to focus
attention on possible paths of development, salient characteristics, and the
interactions of key factors, along with the range of possible eventualities
(cf. Braun / Glauner / Zweck 2005, 33 f.).
In the process, however, scenarios serve not only to produce and/or to deep-
en our knowledge but also to reveal the limits of that knowledge, i.e. the un-
predictabilities, the gaps, dilemmas, and the points of uncertainty (cf.
Greeuw et al 2000, 9; Braun / Glauner / Zweck 2005, 33 f.).
It is possible with the aid of scenarios to achieve a transformation effect (cf.
Tegart / Johnston 2004, 35 ff.). That is, an initially unknown future envi-
ronment which is characterized by a spectrum of possible developments, “a
range of futures” (Tegart / Johnston 2004, 33 [referring to Courtney 2001])
can be transformed into a future environment in which developments are
Methods of future and scenario analysis
German Development Institute 19
assembled into scenarios, so that clearly distinguishable alternative or “al-
ternate futures” (ibid.) are recognizable.7
In addition, scenarios can also widen the scope of our reflections and im-
prove their accuracy concerning alternatives beyond the limits of conven-
tional paradigms (Greeuw et al. 2000, 7):
“Scenarios are perhaps most effective when seen as a powerful
tool to broaden perspectives, raise questions and challenge con-
ventional thinking.“
Scenarios likewise make a special contribution to science inasmuch as they
frequently make it possible to combine qualitative and quantitative knowl-
edge (Greeuw et al. 2000, 9):
“Scenarios are in principal powerful frameworks for using both
data and model-produced outputs in combination with qualitative
knowledge elements.“
“Scenarios allow for looking ‘far and wide” (Barré 2004, 116; quotation
marks in the original text); they provide support for more long-term and
more system-oriented observations than other approaches (cf. Barré 2004,
116).
The communication function
Secondly, scenarios have a communication function, and this in turn on sev-
eral different levels:
On the one hand, they can themselves be generated as part of communica-
tive processes and thus serve to stimulate a discourse in which they help to
promote a common, shared understanding of a problem while also promot-
ing an exchange of ideas and the integration of different perspectives con-
cerning a topic. In this way, they can bring a focus to communication
processes while improving them, thus contributing to better cooperation
7 Tegart /Johnston base their thoughts on the classification of Courtney and distinguish on
the whole four levels of uncertainty. The task and the possibilities of futurology, argue
these authors, is to attempt a reduction of uncertainty by taking the respective degree of
uncertainty in each case into consideration (cf. Tegart /Johnston 2004).
Hannah Kosow / Robert Gaßner
20 German Development Institute
while creating a network among the different persons who are actively in-
volved (e.g. among experts from different areas or between theoreticians
and those involved in the practical application of ideas) (cf. Gaßner / Stein-
müller 2006, 134).
On the other hand, scenarios can also be used to generate communication
and to inform about topics and priorities, thus expanding the understanding
of topic areas (cf. Eurofound 2003, 88), thus casting light on problem situ-
ations and enriching debate about these matters. In particular, the most il-
lustrative scenarios are preferred for use in public communication.
The goal-setting function
Thirdly, scenarios serve as aids in the development or concretization of
goals to be kept in mind. They direct attention to the personal positions of
those involved (cf. Minx / Böhlke 2006, 18). With the help of scenarios it
is possible to deal with the questions “Where do we want to go from here?”
and “What do we hope to achieve?” Scenarios can be used to develop nor-
mative ideal images of the future or to aid in reflections about the desir-
ability of future developments.
The decision-making and strategy formation function
Fourthly, scenarios are employed in the processes of arriving at decisions
and carrying out strategic planning inasmuch as they mediate points of ori-
entation to those carrying out the planning (Braun / Glauner / Zweck 2005,
34). On the basis of scenarios it is possible to work out options and indica-
tors for taking action (cf. Eurofound 2003, 88). Moreover, they also make
it possible to evaluate decision-making processes, actions to be taken, and
strategies. Usually, this work is done with numerous different alternative
scenarios which are then compared with one another (cf. Eurofound 2003,
88) in order to illustrate different future developments and to let the conse-
quences of various developments and/or decision-making processes play
out against a virtual backdrop. In this way, scenarios serve to test the relia-
bility, robustness, and effectiveness of policies (cf. Eurofound 2003, 88).
In addition to these variegated functions of scenarios, it also appears advis-
able to keep in view the limitations of that which can be achieved with
them.
Methods of future and scenario analysis
German Development Institute 21
2.2.4 When are scenarios inappropriate?
First of all, it is important to emphasize that scenarios are not a kind of uni-
versal methodological tool; there is no one scenario approach which can
provide all four of the functions described above at one and the same time.
On the contrary, scenarios are applied specifically and at times with clear-
ly different points of emphasis in order to reach different goals.
Secondly, it is important to repeat that although it is quite possible for sce-
narios to be based (among other things) on prognostic knowledge, they are
nevertheless not to be viewed as “hard and fast” predictions (e.g. Greeuw
et al. 2000, 7). It is much more the case that scenarios are projections which
– for example in thought experiments – combine and answer various “What
would happen if” questions. The factual prognostic value of scenarios
should therefore not be overestimated. Scenarios can at most reveal ranges
of developments; in the rule, however, they make no claim to hit the mark
with precise predictions. Scenarios in this sense never depict true and nec-
essarily impending futures but always only possible ones. Scenarios also
make no claim to be self-fulfilling; rather, their task is to direct attention to
the development of various factors and how these interact with one anoth-
er (Eurofound 2003, 89). It nevertheless occurs time and again that scenar-
ios are misunderstood as representing the only possible future, even when
numerous other scenarios are present as alternatives. The fact is that they
can only serve as “indicative of a spectrum of possibilities” (Eurofound
2003, 89). The selection and construction of scenarios always implies that
other scenarios could have been constructed and selected.
A further limitation of scenarios is to be found in our own cognitive limita-
tions in thinking about the unknown and the uncertain. Even though sce-
narios should have the function of breaking through old thought structures,
human beings nevertheless often tend to follow and extend well-beaten
paths. The problem in doing so can be illustrated by the metaphor of a
drunkard who, thinking he needs only bright light, searches for his house
key under a street lamp at night, even though he has already lost it – in the
dark – somewhere else; that is, whenever we are unable to process infor-
mation because it is lost to us in darkness, we prefer to turn to the “known
suspects“. For this reason scenarios can run the risk of being marked by
thoughts which show little innovation, which in their orientation are very
much extrapolations of existing trend vectors, which are allegedly “objec-
tive knowledge“, and which thus overlook the presence of inconsistencies
Hannah Kosow / Robert Gaßner
22 German Development Institute
and the possibility of less likely developments (cf. Greeuw et al. 2000, 7
and Braun / Glauner / Zweck 2005, 34).
Because of their focus on the future, scenario methods do not use the crite-
rion of the falsifiability of scientific theories; this is because scenarios make
no claim to insights in the sense of the natural sciences. At the same time,
however – and in spite of ever-present and changing boundary conditions –
futurologic research – including scenario methods – always remains subject
to the criteria of good scientific work, such as logical consistency, a clear
description of scope, an explanation of premises, and transparency (cf.
Kreibich 1996).
The following presentation of points common to concrete scenario ap-
proaches and points at which they differ from one another is based on the
basic principles which are generally understood under the term “scenario“,
together with a description of the underlying understanding of the future
and the aims and limitations of scenario approaches.
2.3 Methodological commonalities and differences
The following Section 2.3.1 begins by sketching and systematizing the field
of scenario methods as a spectrum. Then the general course of a scenario
process will be outlined as a basic foundation common to many scenario
approaches (2.3.2). Then the differences within scenario approaches are
discussed against the background of their basic characteristics (2.3.3) and
their differences in scope (2.3.4.). The picture is then enlarged by introduc-
ing the criteria of “good” scenarios and the issue of process criteria (2.3.5.).
2.3.1 The range of the field of scenario methodology
The scenario method does not exist as such; rather, “scenario methodology”
is rather a comprehensive term which in actual practice covers the most var-
ied possible assortment of approaches, techniques, and research and work-
shop designs. The term “scenario methods” represents a methodological
concept encompassing a canon of approaches with different degrees of
complexity.
For purposes of systematization, it is only logical to carry out a study of dif-
ferent methodological levels (see also with regard to the following Stein-
müller 1997, 40 ff.). In the process, scenario approaches can be regarded as
Methods of future and scenario analysis
German Development Institute 23
a complex set of methods which invariably consists of numerous different
methodological steps or phases.
Different techniques may be applied within the framework of a practical
scenario process. The sequence of steps or phases comprising the concrete,
salient characteristics of a scenario method is determined by the selection
of a specific scenario technique. A synonym commonly used when speak-
ing of scenario techniques is “scenario analysis” (cf. Mißler-Beehr 1993,
8). Another concept sometimes used synonymously is “scenario manage-
ment“, which emphasizes the aspect of the strategic application of scenar-
ios on the part of decision-makers (cf. Gausemeier / Fink / Schlake 1996,
14). The concepts of “multiple scenario analysis” (MSA) and “scenario-
writing” are also widespread. Many different approaches are to be found on
the level of these scenario techniques (cf. e.g. Steinmüller 1997, 40); for
their part, they employ a multiplicity of instruments and/or supplementary
techniques in order to work out the inner design of the individual steps.
At the same time, scenario method techniques with all their procedures and
instruments do not stand alone in a “methodless” space, but rather have re-
course to techniques and instruments which are also applied in other types
of methodological design (e.g. trend analysis, actor analysis, cross-impact
analysis etc.). In fact, they are often coupled in research designs with other
independent methods. As a result, one for example finds method combina-
tions involving modeling methods, Delphi-methods, or road-mapping tech-
niques.
What is the reason for this multiplicity of approaches, and why is there no
clearly defined canon of methods for scenario techniques?
– Firstly, many different scenario techniques have been developed due to
the growing spread of scenario use in different application contexts (cf.
e.g. Blasche 2006, 66; Eurofound 2003, 88). Among the fields of ap-
plication are e.g. business enterprises, city and land-use planning, and
research and advisory services (e.g. global scenarios affecting the en-
vironment or energy uses) with their correspondingly different as-
sumptions and standards. Many areas of science and practical applica-
tion today use scenario techniques. The individual forms of these tech-
niques, however, may vary widely depending on those who commis-
sion or instigate the respective scenario and on the respective develop-
mental roots of these techniques.
Hannah Kosow / Robert Gaßner
24 German Development Institute
– Secondly, and this is presumably the primary reason for the multiplici-
ty of methods, the spectrum of goals and functions has grown con-
stantly since the first emergence of the scenario concept.
– Thirdly, different schools of thought and paradigms have influenced
work with scenarios and have infused different perspectives into the
field of scenario methods by bringing in patterns of thought and cre-
ative techniques from the natural sciences.
– Fourthly, scenarios may have widely varying positions of importance
in projects and research processes depending on the concrete, salient
characteristics involved. Scenarios may not only be end product of a
project (scenario generation), but equally also its point of departure
(scenario evaluation) or even its interim product (scenarios as an inter-
mediate step toward further processing and transfer) (cf. Eurofound
2003, 90).
– Fifthly, the concept of a “scenario technique” subsumes on the one
hand fully different approaches, while on the other hand different labels
may also exist for intrinsically similar approaches inasmuch as differ-
ent “scenario service suppliers” use them merely to give prominence to
their own approach and set it off from the others (cf. Steinmüller 1997,
40).
“Scenario methods” are thus a point of confluence for different approaches
whose origin is not alone scientific and/or theoretical but rather often – and
quite the contrary – deeply shaped by their practical implementation. Above
all, scenario methods represent applied knowledge, with theoretical under-
pinnings which may vary in importance from one situation to another; de-
pending on the concrete practical situation, this knowledge is and must be
always adapted practically (and pragmatically as well). For that reason, the
present study has the aim of identifying the characteristics of different key
variants of scenario methods.
2.3.2 General phases of the scenario process
In spite of all the multiplicity of scenario techniques, it is possible never-
theless to identify a more or less important “lowest common denominator”
on the basis of typical phases. This means that there is a widespread com-
mon consensus about the general course taken by them. However, the indi-
vidual phases take on very different shapes in the various techniques.
Methods of future and scenario analysis
German Development Institute 25
Various proposals have been made for delineating and designating these
phases.8The most abstract of these (cf. e.g. Mißler-Behr 1993, 9) is a divi-
sion into the three phases of analysis, prognosis and synthesis. This division
places emphasis on the special characteristic of scenario techniques in that
they offer both analytic and synthetic functions. The term “prognosis“,
however, may be misleading (as already mentioned earlier). For that reason
the following, somewhat more concrete division will be used: the scenario
process goes in ideal-typical fashion through the five phases of 1) identifi-
cation of the scenario field, 2) identification of key factors, 3) analysis of
key factors, 4) scenario generation, and, if necessary, 5) scenario transfer
(cf. Fig. 3).
Figure 3: The general scenario process in five phases
Source: IZT
8 Cf. e.g.
– The five phases of Gausemeier / Fink / Schlake (1996): scenario preparation, analysis,
prognostics, formation, and transfer.
– The four phases of Burmeister / Neef / Beyers (2004); Dießl (2006): monitoring, ana-
lysis, projection, transformation.
– The four phases of Phelps / Chan / Kapsalis (2001): defining the scope, database con-
struction, building scenarios, choosing strategic options.
– The eight phases of Steinmüller (2002b): problem analysis, scenario field identifica-
tion, projection, consistency checks, scenario building, analysis of distruptive events,
impact analysis, scenario transfer.
66FFHHQQDDUULLRR
WWUUDDQQVVIIHHUU
66FFHHQQDDUULLRRIILLHHOOGG
LLGGHHQQWWLLIILLFFDDWWLLRRQQ
..HH\\IIDDFFWWRRUU
LLGGHHQQWWLLIILLFFDDWWLLRRQQ
66FFHHQQDDUULLRR
JJHHQQHHUUDDWWLLRRQQ
..HH\\IIDDFFWWRRUU
DDQQDDOO\\VVLLVV
Phase 1 Phase 2 Phase 3 Ph ase 4 Phase 5
Hannah Kosow / Robert Gaßner
26 German Development Institute
Phase 1: Identification of the scenario field
The first step in every scenario process is to define precisely for what pur-
pose scenarios are to be developed. “What specifically is the issue here“?
What is the topic? What problem is to be dealt with? How is the scenario
field to be defined? What must be integrated? And of equal importance:
Where are the limits, that is: what is to be left out of consideration? This
thought corresponds for the most part with the definition of the object to be
researched and the definition of topics in other research designs; in its de-
gree of concreteness, however, it even goes to some extent beyond them.
This phase sets the perspective to be selected for the period under study (cf.
Gausemeier / Fink / Schlake 1996, 132 ff.). At the beginning of the scenario
process, comprehensive decisions regarding relevancy are to be made re-
garding the boundaries of the field which will be taken under study. For ex-
ample, will a self-contained field of organization – such as a business en-
terprise, a clearly defined area of technology, or an organization like the As-
sociation for Technical Cooperation (GTZ) be observed, meaning its inter-
nal factors alone? Or will mostly external factors, that is, the world imme-
diately around it, be taken under study? Such “surroundings” scenarios may
well include the widest possible variety of dimensions: environmental, eco-
nomic, political, technical, and cultural factors. Or will the internal arena
and the surroundings, together with their interrelationships be taken for
study as a system, resulting quite deliberately in so-called “system scenar-
ios“? One example of this would be: “What impact do contemporary events
in politics, the environment, the economy, etc. have on the GTZ, and what
impact does the GTZ have on the world immediately surrounding it?” In
addition, this phase may also include a “peeling away” of non-essential top-
ics contained in the scenario in order to confine it to certain points of em-
phasis. To come back to our example, it is not the GTZ which will be ob-
served in this case, but rather, for example, gender issues within the GTZ.
Phase 2: Identification of key factors
The next phase involves working out a description of the scenario field in
terms of its key factors, or “descriptors“, as they are sometimes called.
These are the central factors which together form a description of the sce-
nario field while also having an impact on the field itself and/or serving as
means for the field to have an impact on the world around it. Key factors
are thus those variables, parameters, trends, developments, and events
Methods of future and scenario analysis
German Development Institute 27
which receive central attention during the further course of the scenario
process.
Identification of these key factors requires knowledge of the scenario field
as such and its interactions with the various key factors.
The process of actually identifying the key factors within the framework of
scenario processes differs very much from one case to another.The required
information about key factors is sometimes fed into the scenario process
through a very intensive preliminary period of empirical and theoretical
analysis (often in the form of desk research); sometimes however, it is also
generated in participatory fashion through workshops or through rounds of
surveys. The first procedure attempts above all to establish a sound theo-
retical foundation for each scenario and relies upon in-depth analysis; the
second focuses above all on establishing a foundation for each scenario via
the intuitive and implicit knowledge of those involved while also relying on
an ability to overview and the power of synthesis. And whereas in the first
case the concrete selection decisions are central (What factors are to be fo-
cused on, and why?), the second case focuses more on the synergy which
results from the composition of those who participate and on procedural
support for the development of a sense of “ownership” among the partici-
pants and the resulting interdisciplinary legitimacy of the later scenarios.
Phase 3: Analysis of key factors
This brings us to the step which is especially typical of scenario techniques
and sets them apart from other methods: the widening scenario “funnel” in
which individual key factors are subjected to analysis to find what possible
future salient characteristics are conceivable in each case. An individual
“funnel opening into the future“, so to speak, widens out for each factor
inasmuch as those salient characteristics are selected which are to become
part of the budding scenario.
Although this step can be carried out in numerous ways, it always contains
intuitive and creative aspects; these are essential for visualizing the various
future developments of any key factor.
Phase 4: Scenario generation
Scenarios are generated by singling them out and condensing them from the
“cross section” of the scenario funnel whose opening extends to the select-
ed projection point in the future. This is where consistent bundles of factors
Hannah Kosow / Robert Gaßner
28 German Development Institute
are brought together, selected, and worked up into scenarios. However, ma-
jor differences in method are also found at this step. The process by which
the “condensation” into scenarios takes place may extend from narrative lit-
erary procedures all the way to formalized, mathematical techniques (cf.
Chap. 2.4).
In addition, a sorting out of scenarios is required in many scenario tech-
niques. Even though many scenarios are often theoretically conceivable,
the number of scenarios which can be processed cognitively is limited.
Practical experience has shown that the number of scenarios which can be
meaningfully distinguished from one another and are thus open to interpre-
tive processing lies around 4 to 5 scenarios at a maximum for any one sce-
nario field (cf. Eurofound 2003, 89).
This process of selection may take place, for example (cf. Henrichs 2003),
according to the following rule-of-thumb: as many as are required to cover
an adequate number of perspectives and possible futures, but as few as pos-
sible, in order to avoid fatigue and to ensure that the process remains man-
ageable.
A meta-study of European and global scenario studies (with special focus
on the areas of the environment and energy) has shown that in actual prac-
tice such research frequently singles out four scenarios according to the cat-
egories in the table below (cf. Greeuw et al. 2000, 89). In the process, the
intensity of actions to be taken and/or policies may be varied on the one
hand as a means of studying different future possibilities for taking action
or avoiding it; on the other hand, assumptions concerning the possible fu-
ture development of surrounding factors may be varied in order to antici-
pate the different contexts of such action:
A further possibility of differences in the construction of scenarios is illus-
trated by the following example.
The end product of this phase: finished scenarios.
Scenario “Wait and See” “Just Do it” “Doom Monger” “Carpe Diem”
type
Under- No or only few Many new Negative Positive
lying new actions actions development of development of
logic external factors external factors
Methods of future and scenario analysis
German Development Institute 29
In the narrower sense, the scenario process is completed after these four
phases. Of central importance in all four phases, however, is that a series of
selection steps be taken and backed up with reasons.
Example: “A tale of four futures“; Outlook 2002-2032 (UNEP 2002, 328 ff.)
(Abstracts of Scenarios)
Markets first
... Most of the world adopts the values
and expectations prevailing in today’s
industrialized countries. The wealth of
nations and the optimal play of market
forces dominate social and political
agendas. Trust is placed in further glo-
balization and liberalization to enhance
corporate wealth, create new enterpri-
ses and livelihoods, and so help people
and communities to afford to insure
against – or pay to fix – social and en-
vironmental problems. Ethical inve-
stors, together with citizen and consu-
mer groups, try to exercise growing cor-
rective influence but are undermined by
economic imperatives. The powers of
state officials, planners and lawmakers
to regulate society, economy and the
environment continue to be overwhel-
med by expanding demands.
Policy first
... Decisive initiatives are taken by go-
vernments in an attempt to reach speci-
fic social and environmental goals. A
coordinated proenvironment and anti-
poverty drive balances the momentum
for economic development at any cost.
Environmental and social costs and
gains are factored into policy measures,
regulatory frameworks and planning
processes. All these are reinforced by
fiscal levers or incentives such as car-
bon taxes and tax breaks. International
‘soft law’ treaties and binding instru-
ments affecting environment and deve-
lopment are integrated into unified
blueprints and their status in law is up-
graded, though fresh provision is made
for open consultation processes to al-
low for regional and local variants.
Security first
…This scenario assumes a world of
striking disparities where inequality
and conflict prevail. Socioeconomic
and environmental stresses give rise to
waves of protest and counteraction. As
such troubles become increasingly pre-
valent, the more powerful and wealthy
groups focus on selfprotection, creating
enclaves akin to the present day ‘gated
communities’. Such islands of advan-
tage provide a degree of enhanced secu-
rity and economic benefits for depen-
dent communities in their immediate
surroundings but they exclude the dis-
advantaged mass of outsiders. Welfare
and regulatory services fall into disuse
but market forces continue to operate
outside the walls
Sustainability first
…A new environment and development
paradigm emerges in response to the
challenge of sustainability, supported
by new, more equitable values and in-
stitutions. A more visionary state of af-
fairs prevails, where radical shifts in the
way people interact with one another
and with the world around them stimu-
late and support sustainable policy
measures and accountable corporate be-
haviour. There is much fuller collabora-
tion between governments, citizens and
other stakeholder groups in decision-
making on issues of close common con-
cern. A consensus is reached on what
needs to be done to satisfy basic needs
and realize personal goals without beg-
garing others or spoiling the outlook for
posterity.
Hannah Kosow / Robert Gaßner
30 German Development Institute
9 See e.g. Steinmüller (1997) or Mietzner / Reger (2004) for an overview and discussion
of the various attempts at identifying typologies.
10 Van Notten et al. (2003) have presented an “updated scenario typology” in an attempt to
fill this gap with a new typology. They classify the different characteristics of scenarios
into three main groups: 1) Target, 2) Process design, and 3) Scenario content.
Optional: Phase 5: Scenario transfer
This phase involves a description of the further application and/or process-
ing of scenarios which have been generated. However, it is counted explic-
itly as part of the scenario process proper only in the case of a certain few
scenario techniques. Here again, there is a wide range of possibilities for us-
ing finished scenarios, e.g. in impact analyses, actor analyses, strategy as-
sessment and development, etc. (cf. Section 2.4.5 for a discussion of the
techniques of scenario transfer.)
Following this description of general points in common in the scenario
process, the focus will now be directed to points of difference among the
various scenarios, and criteria will be presented for distinguishing and char-
acterizing the various scenario approaches.
2.3.3 The basic characteristics of scenarios
The literature contains some proposals for identifying characteristics and
typologies among the multiplicity of scenarios.9However, no typology has
yet been presented which covers all approaches, that is, none which can be
detailed enough to clearly and simultaneously characterize the widest vari-
ety of approaches in depth. Most of the existing characterizations thus re-
main either very generalized or are so specialized that they fail to cover the
entire spectrum of different approaches (cf. van Notten et al. 2003).10 They
take the form of pragmatic categories rather than well-founded typologies
(see Mietzner / Reger 2004, 52 for a discussion of this deficit).
For this reason, only a few basic features which permit a basic characteri-
zation of many approaches and are regularly used in the scenario literature
will be presented here. They are, to begin with, the contrary pairs of “ex-
plorative” vs. “normative” and “qualitative” vs. “quantitative“. An addi-
tional important aspect is the question to what extent scenarios can include
possible future actions to be taken (i.e. “reference” scenarios vs. “policy”
scenarios) or can integrate “surprises” and/or discontinuities.
Methods of future and scenario analysis
German Development Institute 31
Explorative vs. normative approaches
The literature frequently divides scenario techniques basically into the “ex-
plorative” and the “normative” (cf. e.g. van Notten et al 2003; Alcamo
2001; Greeuw et al. 2000; Steinmüller 1997 etc., etc.). These two poles al-
so stand for two basic, ideal-typical stances regarding scenario method
techniques.
When used in connection with techniques, the appellations “explorative”
and/or “descriptive” designate sets of possible events regardless of their
desirability (Greeuw et al. 2000, 8). Such techniques pose “What-would-
happen-if” questions and take the present as their starting point. They then
use considerations regarding developments, driving forces, and possible
consequences to work out a conceptual future (cf. Eurofound 2003, 8). The
primary function of such techniques is to lay bare the unpredictabilities, the
paths of development, and the key factors involved: “What do we know and
what do we not know“? (the “explorative” and/or the “knowledge” func-
tion). They are employed, for example, like “simulators” in order to go
through the consequences of possible decisions and actions which might be
taken.
Normative scenarios, on the other hand, assimilate values and interests (cf.
Greeuw et al. 2000, 8). They pose questions either about the desirability of
conditions in the future “What do we want the future to be like? Where do
we want to go with it?” and/or questions which take possible futures as
their point of departure: “How can we get there? What must happen in or-
der for it to become reality?” (cf. Eurofound 2003, 88). This second type
of normative scenario clearly looks back from a future point in time toward
the present. Its function is to work out the process by which a specific (de-
sired) state of affairs can be attained. It is used to demonstrate how certain
goals can be achieved. Normative scenarios have a goal-setting function
and a strategy-developing function.
However, scenario techniques differ from one another not only in whether
they are “unprejudiced” in their study of possibilities or (preferably) attrac-
tive objects of desire, but also and in addition in whether they attempt to de-
termine the probability of future developments (cf. Steinmüller 1997, 53).
This is sometimes attempted in explorative scenarios, but only seldom in
normative scenarios, since the latter assume that the probability of devel-
opments can be influenced to a major degree by taking an active part in
Hannah Kosow / Robert Gaßner
32 German Development Institute
Table 1: Schematic comparison of explorative and normative scenarios
Source: The IZT with elements borrowed from Henrichs (2003);
Greeuw et al. (2000); Steinmüller (1997)
shaping future developments. The following table summarizes explorative
and normative scenarios by placing them in juxtaposition with one another
(cf. Table 1).
However, this dichotomous characterization of scenario approaches also
has its difficulties. Firstly, selective decisions must be made at many points
of the scenario process when a scenario is being constructed (i.e. decisions
regarding not only the definition of the scenario field, but also the relevance
of key factors, the determination of key factor characteristics to be studied,
and the condensation of factors into individual scenarios). For this reason,
scenarios are always – at least implicitly – normative. However, the differ-
ent approaches either view this normativity to varying degrees as open or
deal with the scenarios reflexively (cf. van Notten et al. 2003). Secondly, it
has become common in actual contemporary practice to use both explo-
rative and normative scenarios in combination, especially when the aim is
to develop strategies (ibid. and Steinmüller 2002b, 13).
Explorative Normative
Procedure Explores possible future Identifies desirable futures
developments with the or investigates how to
present as point of departure arrive at future conditions
Function Explorative and/or Target-building function
knowledge function and/or strategy
development function
Implementation Study of factors and Definition and
unpredictabilities, test of concretization of goals
possible actions to be taken and/or, if appropriate,
and/or decision-making identification of possible
processes ways to reach a goal
Central What? How?
question – What if? – How is it to come about?
– How do we get there?
Inclusion of Possible Indirect, part of plausible
probabilities shaping and planning
Methods of future and scenario analysis
German Development Institute 33
Qualitative vs. quantitative approaches
Scenarios and scenario techniques are also distinguished by the type of in-
formation which they can and should assimilate and/or transport. Are qual-
itative descriptions alone used, or are quantitative data employed? Or will
the users use estimates to quantify qualitative data? Different instruments
of analysis are used for the identification and analysis of key factors, and
different techniques are employed for the generation of scenarios depend-
ing on whether quantitative or qualitative data are required, meaningful,
and available. Quantitative knowledge is used, for example, in topic areas
like demography and economics, whereas on the other hand cultural, insti-
tutional or political dimensions often tend to be recorded qualitatively.
The methodological decision for proceeding either qualitatively or quanti-
tatively has direct consequences regarding the possible degree of formal-
ization of the scenario technique to be used. To put it provocatively and ide-
al-typically, quantitative approaches have recourse to mathematical models,
qualitative approaches on the other hand have recourse to narrative and/or
literary techniques.
The two approaches also differ in the manner in which they select and study
the respective key factors. Quantitative scenarios make it necessary to ar-
rive at a firm definition of a reduced number of factors, whereas qualitative
scenarios make it possible to achieve an intrinsically more meaningful ob-
servation of details and nuances without the need of definitively including
or excluding key factors.
Another difference between these approaches is the chronological horizon
which they are capable of describing meaningfully. Quantitative approach-
es can be used above all for short, at most medium-term perspectives; qual-
itative approaches, on the other hand, can be employed especially when al-
legedly “hard” quantitative knowledge suffers a loss of plausibility during
the course of longer-term observation.
In actual scenario practice, however, this dichotomous characterization of
scenario approaches has only conditional relevance, since scenarios today
are often based on a hybrid approach in which both qualitative and quanti-
tative data are gathered and translated from qualitative to quantitative
knowledge (quantification) or from quantitative to qualitative narrative
knowledge (as in the textualization of key bundled key factor characteris-
tics into scenario texts).
Hannah Kosow / Robert Gaßner
34 German Development Institute
Table 2: Comparison between quantitative and qualitative scenarios
Source: IZT description with reference to van Notten et al. (2003);
Alcamo (2001, 10); with additions by the IZT
The following two distinctions of types of scenarios lie on quite another
level; common to both, however, is the basic question of how to deal with
future changes, that is, with change and unpredictability: 1) Is it also the
aim of scenarios to study possible new actions to be taken, along with de-
cision-making processes? 2) Are surprises, i.e. unexpected, sudden and pos-
sibly even dramatic events also to be taken into consideration in the devel-
opment of scenarios?
“Reference scenarios” vs. “Policy-scenarios“
Reference scenarios and/or “baseline-scenarios” (Gausemeier / Fink /
Schlake 1996; Steinmüller 2002b) project contemporary developments
continuously into the future, i.e. they assume that no new decision-making
processes or actions whatever are to be initiated. Their logic is “Business
As Usual“, and for this reason they are often described in brief as “BAU-
Quantitative Qualitative
Implementation
When quantitative knowledge
When qualitative knowledge
– is required – is required
– and present – or quantitative knowledge
– and/or quantification is is not present
possible – or quantitative knowledge
is not present
Topic areas e.g. demography, economic e.g. institutions, culture,
development politics
Impact on the Tendency to a high degree Tendency to a low degree
degree of of formalization of formalization
formalization
The ideal-typical Modeling methods Narrative and/or literary
scenario technique
techniques
Manner of Firm definition of a narrowly Intrinsically sensory
selecting key limited number of factors observation of details and
factors nuances, possible without a
stringent selection of factors
Chronological Short to medium-term Medium to long-term
projection space
Methods of future and scenario analysis
German Development Institute 35
scenarios“. Their goal is first to explore what will happen “If we continue
as up to now“. Secondly, these scenarios serve as reference-scenarios in
comparison with scenarios which study the possible alternatives for decid-
ing on how to act and what actions are to be taken. Such “policy-scenarios“,
as they are called, or “alternative scenarios” explicitly integrate new deci-
sion-making processes or actions to be taken in order to simulate and test
the possible options for action and their consequences.
Integration of discontinuities
However, the manner of proceeding with reference-scenarios and alterna-
tive scenarios also involves the danger of failing to take the unexpected in-
to account and, as a result, tending to develop more “conservative“, i.e. less
creative conceptual future. Greeuw et al. (2000, 8) and van Notten et al.
(2003) come to the conclusion that most current scenario studies take only
incremental changes into account while overlooking discontinuities almost
completely.
For this reason, it is important to seek methodological possibilities for in-
tegrating the element of chance and/or discontinuities into future develop-
ments. Among the approaches which integrate the improbable, the undesir-
able, or even the “unthinkable” aspects of development are Problem Event
Analysis (cf. e.g. Gausemeier / Fink / Schlake 1996) or the so-called “wild
cards” (cf. e.g. Steinmüller / Steinmüller 2003 and Section 2.4.5, with the
Excursus: Techniques of Scenario Transfer).
2.3.4 Scope
Scenarios can differ widely in scope. This affects, for example, their selec-
tion of a chronological horizon, their geographical scope, and their cover-
age of themes.
In general scenario techniques are faced with the fundamental challenge of
reducing complexity sufficiently to permit a process of synthesis. Their
aim, after all, is to keep numerous different factors simultaneously in view
in order 1) to observe their interactions and 2) to be able to develop overall
images of future situations. This process of synthesis, however, is always
limited by the cognitive abilities of those involved in the scenario. This al-
so means, for example, that global scenarios cannot include hundreds of
key factors since processing them cognitively in a meaningful way would
then be impossible. In many respects, interrelationships are to be found in
Hannah Kosow / Robert Gaßner
36 German Development Institute
11 As, for example, in the IZT project “Forest Visions 2100“.
the process of weighing pros and cons between the various scopes and be-
tween the scopes and the degree of abstraction and/or depth of detail in sce-
narios.
The chronological horizon and/or observation period
Scenarios are constructed with chronological horizons of varying breadth.
The periods to be studied may be short-term (up to 10 years), medium-term
(up to 25 years) and long-term (more than 25 years) (cf. Kreibich 2006, 3;
van Notten et al. 2003).
In addition (cf. Blasche 2007, 89), static observations from a point in time
in the future are possible; we then speak of “static scenarios” and/or “end
state scenarios” (van Notten et al. 2003). Again, the dynamics of develop-
ment throughout a number of different stages in time in the future may be
observed, in which case the scenarios are dynamic / sequential scenarios
and/or “chain scenarios” (van Notten et al 2003). In this case, numerous dif-
ferent stages may be selected during the study of very long periods of time,
for example in the case of developments up to the year 2100, and then sce-
narios may be developed extending first up to 2020, thereafter on this basis
up to 2050, and only subsequently up to 2100.11
Geographic scope
Scenario concepts may be formed with varying geographical scopes.
Greeuw et al. (2000, 9 f.), for example, distinguish four different geo-
graphical points of reference for scenarios:
– The global level
– The international and regional level
– The national level
– The sub-national and regional level
In addition, the local level may be adduced as a fifth level.
Thematic coverage
Scenarios may of course also be distinguished – depending on the problem
to be dealt with – by their thematic pattern (cf. Greeuw et al. 2000, 9 f.).
Methods of future and scenario analysis
German Development Institute 37
Some, like “issue-based scenarios” (van Notten et al. 2003), focus on indi-
vidual themes (e.g. “sustainability“); others observe individual sectors
and/or social fields (e.g. “the environment“, “energy” or “water“), while
“institution-based scenarios” direct their attention to the special area of in-
terest of an organization or institution (van Notten et al. 2003). In addition
to a generalized classification of the levels of observation and viewpoints
of scenarios, a classification into macro-, meso- and microlevels is some-
times also used (cf. e.g. Mietzner / Reger 2004, 52).
Integration
With regard to the chronological, geographical and thematic scope of sce-
narios between the poles of depth of detail and degree of abstraction, there
exists a basic problem: a very wide chronological, geographical or themat-
ic scope is achievable only through a high degree of abstraction, general-
ization, or aggregation. Theoretically, on the other hand, various possibili-
ties exist for achieving a wide range through the integration of different lev-
els; these possibilities are currently being tested more and more in actual
practice (cf. van Notten 2003).
For example, some researchers attempt on the geographical level to inte-
grate the global, regional and local levels, rather than regarding them in iso-
lation from one another. Numerous different approaches exist for achieving
such integration:
– All three levels are observed concurrently (i.e. parallel or iteratively).
– Beginning with global scenarios, top-down regional and, ultimately, lo-
cal scenarios are developed (including feedback loops to the next high-
est level, as the case may be).
– Beginning with local scenarios, bottom-up regional and global scenar-
ios are developed (including, as the case may be, feedback loops).
– Double integration: for example, European scenarios constructed in the
European Commission’s VISIONS Project („Integrated visions for a
Sustainable Europe“) included global trends on the one hand while al-
so being linked with scenarios on the level of the European regions
(bottom-up and top-down integration) (cf. Greeuw et al. 2000, 9 f.).12
12 See e.g. Rotmans et al. (2000) for further information about this project and the integra-
tion techniques used in it.
Hannah Kosow / Robert Gaßner
38 German Development Institute
13 Concerning the criteria for judging the quality of scenarios, Greeuw et al. (2000, 7)
among others name the following: internal consistency, plausibility, and sustainability.
Kreibich (2007, 183) names the following general criteria of quality in futurology: logi-
cal consistency, openness to evaluation, terminological clarity, simplicity, definition of
range, explanation of premises and boundary conditions, transparency, relevance, practi-
cal manageability, and fruitfulness (i.e. in terms of gain in knowledge, orientation, inno-
vation, motivation etc.) .
Wilson (1998) names the following: plausibility, differentiation, consistency, decision-
making utility, and challenge.
Heinecke / Schwager (1995) name the following: tangibility (clearness, cohesion with
the object of investigation, suitability, transparency), closeness of the content (flawless-
ness: no invalid assumptions, plausibility, completeness, finding of cohesions, descripti-
on of development, information content: precision, universality, utility), relevance
(function of decision, function of orientation, relevance in different planning processes
and analysis of problems, forecast, assessment and decision); constitution and proportion
of scenarios among themselves (dissimilarity, registration of all future situations, homo-
geneous forms and statements, stability).
It is conceivable that these integration strategies might be used analogous-
ly for the integration of different chronological dimensions and different
thematic fields.
2.3.5 Criteria of quality and process criteria
The standards taken as a basis for evaluating scenarios and scenario tech-
niques are often based on the same criteria as those of good research. But
there are also more scenario-specific criteria for the evaluation of scenar-
ios. In the following, some process criteria will be mentioned.
The literature proposes some criteria as central in evaluating the quality of
scenarios and scenario processes, independently of the respective goal and
type of the scenario process. Although scenarios are always hypothetical in
nature, this by no means makes them arbitrary. Therefore a good scenario
should have the following characteristics:13
Plausibility
In relation to scenarios (cf. e.g. Greeuw et al. 2000; Wilson 1998), plausi-
bility means that the possibilities of development which are presented must
at least be regarded as possible developments. That does not mean, howev-
er, that these developments are also probable or desirable (the manner of
proceeding differs here depending on the respective goal and technique).
Methods of future and scenario analysis
German Development Institute 39
The path to the futures and images which are described must thus be con-
ceptually feasible and may not be regarded as impossible.
Consistency
“Consistency” with regard to scenarios (cf. e.g. Greeuw et al. 2000; Wilson
1998; Gausemeier / Fink / Schlake 1996; Steinmüller 1997) means that
paths to the futures and images within a scenario must be consistent with
one another, i.e. their aspects may not be mutually contradictory or even go
so far as to exclude each other for reasons of logic and plausibility. A sce-
nario on the topic of water, for example, is therefore inconsistent if it as-
sumes an abatement of research and developmental efforts in the area of
drinking water technology while simultaneously assuming major technical
progress in the processing of drinking water.
Consistency and plausibility are the decisive conditions for assessing sce-
narios as credible (cf. Steinmüller 1997, 62).
Comprehensibility & traceability
In relation to scenarios (cf. e.g. Greeuw et al. 2000; Heinecke / Schwager
1995), comprehensibility means that the developments and conceptual fu-
tures which are presented must be traceable. This in turn means on the one
hand that they must be detailed enough to be comprehensible, while not
combining so many dimensions and key factors on the other hand that they
suffer a loss of comprehensibility due to their complexity.
Distinctness
Distinctness, i.e. the quality of being clearly distinguishable (cf. e.g. Wilson
1998; Heinecke / Schwager 1995), means that the selected, alternative sce-
narios differ from one another clearly enough that they can be interpreted
and compared with one another as separate and distinct sketches of the fu-
ture.
Transparency
During the process of their development, scenarios go through an entire se-
ries of assumptions and choice decisions, e.g. in answer to the central ques-
tion of which key factors are to be studied and how possible salient char-
acteristics in future are to be defined and determined. As a means of in-
creasing the degree of verifiability and legitimacy, the assumptions made
and the processes by which decisions are reached should be laid open: Who
Hannah Kosow / Robert Gaßner
40 German Development Institute
decided or carried out what, why, how? (cf. e.g. Greeuw et al. 2000; Stein-
müller 1997; Kreibich 2007)
The criterion of transparency appears particularly important for doing jus-
tice to the criteria of qualitative science. While it is true that such process-
es are neither reproducible nor falsifiable, such process reflexivity can en-
sure that a considerable degree of intersubjective verifiability is attained.
In addition, the following applies: “Scenarios always, either implicitly or
explicitly, embody perceptions and judgements.” (Greeuw et al. 2000, 9)
This means that a reflexive manner of proceeding which incorporates its
own value-stamped, normative positions can greatly increase the trans-
parency of scenarios even in “descriptive-analytic” procedures.
It is also important here to note that this transparency can differ among the
persons or groups for whom the scenario is intended (cf. Braun / Glauner /
Zweck 2004, 34). For example, a scientific presentation of lists of variables
will be transparent for a specialized audience but may possibly be opaque
for the general public. Vice versa, scenarios formulated in a more “popular”
way are sometimes rejected by groups of specialists as being too deficient
in “transparency“.
Above and beyond these generalized criteria, individual authors also pro-
pose more scenario-specific criteria; of these, the following two appear rel-
evant for certain scenario approaches.
Degree of integration
Since scenarios generally do not focus on detailed issues but are rather em-
ployed to study the causal relationships between different dimensions and
factors, a further criterion of a good scenario is the question of the extent to
which it integrates the interactions of developments on different levels (cf.
Greeuw et al. 2000, 10). For example, does it take note of and study the
causal relationships between social, economic, ecological and institutional
developments? Important in this regard is not only vertical integration, that
is, the chain of cause and effect within a topic area and/or sector, but also
horizontal integration, that is, the interaction of different sectors and
themes. In most scenario fields, moreover, an interdisciplinary approach is
indispensable in the process of scenario development in order to achieve a
certain degree of integration (ibid.).
Methods of future and scenario analysis
German Development Institute 41
Quality of reception
In addition to the above-mentioned criteria of quality, with their heavily sci-
entific bent, it is also to be noted that a good scenario should also be “read-
able“; it should not become a “torture” for the reader to fight his way
through it (cf. Gaßner 1992, 230 f.). For this reason it is also important in
working out the concrete formulation of scenarios to pay heed to their more
unobtrusive building block factors. For example, Gaßner names “the pow-
er of fascination [...], of implication, of esthetic dimensions, and ‘enjoy-
ment qualities’ like suspense and humor” (Gaßner 1992, 230) as possible
means for improving the readability of scenarios and increasing their cre-
atively stimulating impact (Gaßner 1992, 230).
“Process criteria” which are directed toward developmental interrelation-
ships thematize questions such as those of participation and the time and ef-
fort involved.
Participants
Scenario processes also differ in the types of persons who participate in
their development or evaluation. Depending on the degree of involvement,
three rough types of participants may be distinguished:
– Scientists / consultants
– (Internal and/or external) experts or persons actively involved, stake-
holders with a personal interest
– “Those affected“: citizens, consumers, employees, etc.
Some scenarios are created through “desk research” (van Notten et al.
2003) by individual scientists or teams of scientists. In such cases, the de-
gree to which the groups have an interdisciplinary composition is of im-
portance. Other, more “participative” scenario processes, make sure to get
different directly or indirectly affected stakeholders and experts involved,
such as the person commissioning the scenario or even external persons
with practical expertise in the widest sense. Again, other scenario process-
es, for example, involve the “man on the street” as the one “potentially af-
fected“, with the knowledge gained from his or her daily experience of life
(in the sense of the “everyday expert“) and the goals which he conceives of
as normative.
Hannah Kosow / Robert Gaßner
42 German Development Institute
14 There is a much larger group of variously named scenario techniques. Götze (1993), for
example, describes twelve approaches. However, these approaches clearly differ only
partly in their manner of proceeding. Different proposals have been made in the litera-
ture for the typological features of scenario techniques, e.g. “hard” vs. “soft“, “deduc-
tive” vs. “inductive” (cf. e.g. Götze 1993, 385 ff.; Heinecke 2006, 187 ff.; Heinecke /
Schwager 1995, 17). The classification into three ideal-typical groups selected here is
rooted in the basic division into “formal” and “intuitive” scenario techniques (cf. e.g. van
Notten et al. 2003; Götze 2006). This classification in turn can be broken down even fur-
ther – in our opinion – through the inclusion of a quite independent group of scenarios
on the basis of trend extrapolation.
Time and effort involved
In general it must be concluded that scenario processes are work-intensive
and time-consuming; that is, they require time, money and personnel re-
sources (cf. Mietzner / Reger 2004, 61 f.; van Notten et al. 2003). Whereas
it may be possible to manage the evaluation of an already finished scenario
in half a day, the generation of scenarios requires in the rule at least a num-
ber of days, if not months. The time and effort involved in a scenario
process increases proportionately to the degree of inclusion and integration;
this in turn has to do with the number of developments and key factors un-
der study, the breadth of the geographical space, the chronological horizon,
and the number of participants. In addition, scenario processes also differ
very much in the quantity of materials and the number of techniques which
find implementation in them (ranging from pencil and paper to computer
software). A further central factor is the question of how much prior work
and knowledge has already been carried out or established and how much
is still required.
2.4 Three ideal-typical scenario techniques
Three ideal-typical groups of scenario techniques which basically differ
from one another and, in doing so, are good representatives of the entire
spectrum of scenario techniques, will be presented in the following. In the
process, widely accepted variants of one and the same basic type will also
be treated.14
Taking the general course of scenario processes as a background, we will
systematize the different scenario types here in the form of five phases (cf.
section 2.3.1). The first phase of the scenario process, i.e. “selection of the
Methods of future and scenario analysis
German Development Institute 43
scenario field” takes a very similar course in most cases, quite independ-
ently of the concrete scenario technique which is applied. For that reason,
this step will be excluded from the description of the different techniques.
Nevertheless, the importance of this phase must again be emphasized, since
the entire subsequent focus and course of the scenario process, including in
some cases selection of the scenario technique which will later be applied,
depends upon it.
An evaluation of different scenario techniques with regard to their respec-
tive strengths and weaknesses is possible only in relation to an evaluation
criterion. Strengths and weaknesses are always dependent upon the func-
tion and goal of a methodological approach. It is thus possible to evaluate
in particular the appropriateness of an approach either within a specific
knowledge context or in relation to a concrete goal of the respective sce-
nario approach. For this reason it is imperative to clarify the following pri-
or to every scenario process:
– Whether the scenario technique reflects the most appropriate selection
of methods,
– what goals and functions are to be achieved or carried out with scenario
techniques,
– what basic assumptions about the predictability and unpredictability of
the future and our ability to shape it are taken as a foundation.
When compared directly with one another, individual scenario techniques
prove to be different in what they are capable of and in their respective lim-
itations. These will be discussed in the following as the “advantages” and
“disadvantages” of the individual techniques.
Scenarios constructed on the basis of trend analysis and trend extrapolation
will first be presented as fundamental techniques (2.4.1). Then the group of
systematic-formalized scenario techniques will be presented (2.4.2), fol-
lowed thirdly by a discussion of the group of creative-narrative scenario
techniques (2.4.3). Such a tabular juxtaposition of ideal types has the ad-
vantage of clearly showing their basic form of approach. In actual practice,
however, these techniques are often characterized by a plethora of
crossover areas and hybridization (2.4.4). Therefore a subsequent excursus
will go into some of the techniques of scenario transfer and show how it is
possible to implement it as a follow-up to the creation of a scenario (2.4.5).
Hannah Kosow / Robert Gaßner
44 German Development Institute
2.4.1 Scenarios on the basis of trend extrapolation
The first scenario techniques to be treated here as examples are those in
which the respective scenario is supported primarily and even exclusively
by trends which already exist or have already existed and by their projec-
tion into the future. The heart of this technique consists of trend analysis
and trend extrapolation. This will be followed by an explanation of how
these techniques are normally used within the respective scenario technique
in order to work out the “most probable” scenarios or reference scenarios
as a basis for contrasting one alternative scenario with another. Additional-
ly, the technique of trend impact analysis (= ‘TIA’) will also be presented;
it can be employed to examine the alternative courses which events may
take during a trend.
This technique is based on the basic assumption that the most appropriate
way to visualize future developments is the extrapolation of existing devel-
opments.
Trend analysis and trend extrapolation
A “trend” in this causal relationship is to be understood as a development
over a period of time, that is, a long-term vector of development in which
the waxing or waning of an interesting factor takes place (e.g. the develop-
ment of average life expectancy). When understood in this way, “trend” is
firstly not congruent with the everyday use of the word, in which (short-
term) in-vogue phenomena are termed “trends“. Secondly, it is also neces-
sary to distinguish this understanding from so-called “trend research“,
which sees trends as “economically relevant manifestations of the new”
(Pfadenhauer 2006).
The point of departure for trend analysis is an observation of trends which
is supported by the collection of – as far as possible – long-term informa-
tion and data.
Once identified, trends are projected into the future, that is, future courses
of events within the individual trends are subjected to calculation by means
of statistical techniques (given the availability of quantitative data) or de-
scribed (given the availability of qualitative data). The instrument which
serves as a basis for this is referred to as “trend analysis” and represents
an independent, frequently employed method of futurology (cf. Strategic
Methods of future and scenario analysis
German Development Institute 45
Futures Team 2001, 5). Trend analysis is used within a plethora of applica-
tion contexts, even independently of work with scenarios.
Quantitative trend analyses are used above all in areas like demography,
economics, and technology, provided that solid collections of data which
extend far enough into the past are available (cf. Strategic Futures Team
2001, 5).
A typical procedure is the collection and processing of data, the identifica-
tion of logical or systematic processes of development, and the statistical
projection of these into the future (cf. Steinmüller 2002b, 26). Such ex-
trapolations can take place as calculations ranging in form from processes
of linear logic to complex S-curves (Gordon 1994a, 3).
Such calculations have the advantage of being relatively uncomplicated
and requiring little effort; they are verifiable on the logical-intersubjective
plane, and it is possible to subject them to statistical reliability validity test-
ing (cf. Strategic Futures Team 2001, 6).
However, a major disadvantage of such quantitative extrapolations is that
they communicate a sense of greater objectivity than they are capable in
fact of delivering (cf. Gordon 1994a, 3). Their identification of trends is
therefore always based upon interpretation and decisions of selection. In a
data series, for example, they often arrive at numerous different possibili-
ties concerning how a trend, that is, the structure of a development, may be
visualized. This in turn has consequences for the projection of the respec-
tive trend into the future. Moreover, selection of a period for observation
and analysis (short time period vs. longer time period, i.e. the “day trader”
vs. the “historian” perspective) and the criteria for visualization and analy-
sis can have considerable influence on proper recognition of the trend:
when the analysis period is too short or the increments used to measure a
long wave movement are too detailed, for example, the result may be erro-
neous interpretation of a factor which in turn may then mistakenly be
shown in a model as either constant or continuously rising.
On the other hand, there are also many developments which cannot be
meaningfully operationalized and projected in quantitative fashion; for this
reason qualitative trend analysis is often used in combination with and as
an adjunct to quantitative analysis.
Hannah Kosow / Robert Gaßner
46 German Development Institute
Figure 4: Trend extrapolation, forecast, “business as usual” (BAU)
Source: IZT
Qualitative trend analysis (cf. Strategic Futures Team 2001, 7 f.) is em-
ployed when no quantitative data are available and/or quantitative delin-
eation of the respective trends is possible but inadequate. This is often the
case when the development of “softer” factors such as social aspects (stan-
dards and values) or institutional and political aspects is to be followed.
One example of such a qualitatively delineated trend is the change in val-
ues in western industrialized societies. The typical procedure is to define
factors which are important because of their influence and to provide them
with a theoretical underpinning as a means of arriving at the most thorough
possible understanding of these factors and then to further strengthen this
foundation of support with all available information in order to accurately
describe future developments as such.
“Most probable” scenarios and reference scenarios
When scenarios are constructed on the basis of trend extrapolation, this
does not automatically mean that a scenario funnel opens up; in many in-
stances, on the contrary, only a single development comes under observa-
tion, namely that which is assumed to be most probable (cf. Fig. 4).
The result of such extrapolations, especially those carried out on the basis
of quantitative trend analysis, may therefore be a single scenario, the “trend
77LLPPHH
WW
WW
66
Trend A
Methods of future and scenario analysis
German Development Institute 47
scenario” (cf. Gausemeier / Fink / Schlake 1996, 114). Such a scenario is
often called an “outlook“, a “prognosis“, a “forecast” or a “spotlight” rather
than a “scenario“.
One example of a current study which is based on numerous different trend
extrapolations and which has drawn up a single reference image of the fu-
ture in “spotlights” for each of the many different factors is “2018 – The 10-
Year Future” (Rodenhäuser / Daheim / Uerz 2008). In this case, this ap-
proach is implemented with the specific goal of generating not alternative
scenarios but rather individual trend scenarios. Also typical of this kind of
sample study is the relatively narrow chronological time horizon which is
taken under observation with this approach.
Within the field of scenario methods, this manner of proceeding is criticized
by many authors as inadequate, since it assumes too strongly that the future
consists merely in a prolongation of the past, thus making it completely cal-
culable; it is the height of implausibility, they assert, merely to assume a
continuation of existing trends for the decades to come (cf. Greeuw et al.
2000, 8; Gordon 1994a, 1). Minx and Böhlke compare this to the attempt
to drive an automobile merely by looking into the rear view mirror (cf. ibid.
2006).
For this reason, although mere trend extrapolation does indeed provide a
possible basis of knowledge, it is frequently flanked by other approaches
and techniques (cf. Strategic Futures Team 2001, 5). Thus quantitative trend
analysis often forms only the point of departure for scenario work; it is then
first the inclusion of qualitative trend analysis which brings the possibility
into view of thinking in terms of numerous different alternative develop-
ments.
It is often the case that a most probable trend scenario is constructed as a
reference scenario with which other scenarios can be compared. Taking
the contemporary status of knowledge of trends, actions, and developments
as a starting point, this scenario is then used to paint a future in which no
new developments or actions whatever are assumed.
However, this practice, too, is not without problems regarding the con-
struction of “reference” scenarios. The interpretation and operationalization
of contemporary developments and actions is always selective, meaning
that it runs the danger of ignoring new developments which, although dif-
ficult to perceive, nevertheless do in fact exist. In addition, such a scenario,
Hannah Kosow / Robert Gaßner
48 German Development Institute
with its policy of “take no action” and its assumption that there will be no
major changes, suggests greater certainty about the future course of events
than can actually be attained due to the fact that it is often termed “most
probable” (cf. Greeuw et al. 2000, 8). One need think only of still unrecog-
nized saturation effects or ceiling effects.
The exclusive implementation of scenarios which are based on trend ex-
trapolation is appropriate only in the observation of very stable trends
which can be extrapolated with a relatively high degree of certainty (e.g.
geological or demographic developments) or in the case of relatively short
horizons of study (1–3 years) (cf. Strategic Futures Team 2001, 4). This
technique emphasizes those aspects of the future concerning which rela-
tively certain knowledge exists.
Trend analysis and trend extrapolation can yield interesting pointers for
scenario techniques while also serving as basic principles. However, they
fail to do justice to the basic idea of scenarios whenever they take only a
single possible salient characteristic of a future trend into consideration.
This disadvantage of scenarios based on trend extrapolation can, however,
be improved, e.g. via trend impact analysis, which makes it possible to
study the different possible courses of events within a trend.
Trend impact analysis
Trend impact analysis (TIA) originated in the 1970s and was developed in
order to compensate for a weakness of extrapolations in that they fail to
take unexpected future events into account.
The method was originally quantitative in character; it served as a means of
analyzing the influence of future events on the development of trends (cf.
Gordon 1994a, 1). This technique can be implemented in scenario methods
in order to carry out an outward unfolding of individual key factors, i.e. to
define the different possible values of various factors.
In its classical form (cf. Gordon 1994a, 2), the method starts by calculating
a “surprise-free” course for a trend (extrapolation) (cf. Fig. 5, “Trend de-
velopment A“). Then a survey of experts is used to define a set of future
events, each of which, should it come to pass, can bring about a significant
change in the course of the trend. Finally, alternative courses of events
within the trends are calculated by taking these possible future events into
account along with their (estimated) probabilities and the strength of their
Methods of future and scenario analysis
German Development Institute 49
respective influences. The result is, so to speak, an expansion of the “fun-
nel into the future” of individual trends. This technique combines a very
formalized manner of proceeding with explicitly creative elements
(cf. Mietzner / Reger 2004, 54).
Within the framework of comprehensive scenario processes, this technique
is well-suited for creating variations in development with respect to indi-
vidual key factors with the aid of an assumption of future events, thus de-
termining different salient specifications. Following that, the key factors
which have been thus varied can be combined with other factors which
have remained constant in order to see what might/could happen in the sce-
nario field given different courses of events for central key factors.
The advantage of TIA is that it displays a spectrum of possible future de-
velopments for individual factors rather than merely a single, individual
possible development as in the case of pure extrapolation. This procedure
makes it possible to anticipate future events and to study their impact on the
path taken by trends. It then becomes possible to estimate which of the
Figure 5: Diagram of a trend variation with TIA
Source: IZT
77LLPPHH
Historical data
Extrapolation
WW
WW
66
Trend A
Trend development b
-
Trend development a
(baseline)
Trend development c
Hannah Kosow / Robert Gaßner
50 German Development Institute
15 In contrast, consistency analysis and cross impact analysis are well suited for the analy-
sis of interactions, see Section 2.4.2.
events assumed to occur in future might have the (relatively) greatest im-
pact.
The main disadvantage of TIA is that any definition of the sets of future
possible events is always subjective and cannot pose any claim whatever to
reliability. The reason: the probabilities and effects of these events always
remain merely estimates and are dependent upon the assessments of ex-
perts. In addition, the individual future events in question are regarded in
isolation from one another, as if no mutual influence would exist between
different events and trends.15 In particular, this method requires a solid ba-
sis of data; it cannot be used in its classical form if detailed and reliable
time series are unavailable (cf. Mietzner / Reger 2004, 54).
These disadvantages are perhaps the reason why TIA tends to be used only
rarely in scenario work (cf. Bradfield at al. 2005, 801). One practical ex-
ample named by Gordon (1994, 8) is the calculation of different possible
developments in the area of crude petroleum consumption.
In summary, the following may be said: scenarios which are based on
(quantitative) trend extrapolation often form the point of departure for oth-
er, more comprehensive scenario techniques. TIA can be used to form an
estimate of the relative influence of different events on the paths taken by
trends as well as on the respective scenario field.
In the following, a closer look will be taken at comprehensive scenario
techniques which take trend extrapolation and their variation as a basis for
carrying out a spread of the future environment by varying the course of
events of numerous different key factors rather than merely individual
trends. In the process, the different salient characteristics of key factors will
be selectively combined with one another.
2.4.2 Systematic-formalized scenario techniques
This group of scenario techniques is basically characterized by the fact that
it begins with a clear definition of key factors, then varies them and com-
bines them with one another in order to arrive at a widening scenario fun-
nel and generate different scenarios within it. This is all carried out in a sys-
Methods of future and scenario analysis
German Development Institute 51
16 The term “systematic-formalized scenario techniques” is found e.g. in Heinecke (2006,
187 ff.) and Heinecke / Schwager (1995, 17), where this group is juxtaposed with the
more “intuitive” techniques. These techniques go back, for example, to the tradition of
the Batelle Institute (Frankfurt) and are linked among others with the names of von Reib-
nitz (1991) and Geschka / Hammer (1984); the techniques are presently employed, for
instance, in the scenario techniques of SCMI (Gausemeier) and Z_Punkt (Burmeister).
17 This technique for the identification of key factors can – in contrast to its description as
“paper computer” – also be supported by software such as the program MICMAC (pri-
marily developed, among others, by Michel Godet).
tematic and formalized manner.16 These are in general explorative scenario
techniques which acquire their data in part both quantitatively and qualita-
tively. On the basis of these techniques, however, it is also possible to de-
velop normative scenarios.
Within the framework of these techniques, the identification of key fac-
tors (Phase 2) is as follows: 1) The influencing factors are identified. These
may be trends in the sense described above or qualitatively described de-
velopments and events, actions or persons actively involved. 2) Then these
influencing factors are regarded as a whole, i.e. with regard to their com-
bined effect. For this purpose the individual factors are juxtaposed in order
to identify their respective mutual interrelationships. The central question
during this whole process is: “How do the different factors behave in rela-
tion to each other?“
Impact analysis
Often the so-called “paper computer“17 of Vester is used as a means of sys-
tematically identifying the interactions and the dynamics of factors Vester
(2002, 226 ff.; cf. also Wilms 2006, 51 ff.; Blasche 2006, 74 ff., and
Table 3).
This is done by listing the factors already identified in a matrix of columns
and rows, in both cases in the same order of succession; in this way, each
factor is juxtaposed with each of the others. For each pair of factors, the
question is then asked, “To what extent does a direct relationship take effect
between these factors?” (cf. Wilms 2006b, 51). To quantify the influence,
the following scale is often used: 0 = No influence; 1 = Weak relationship;
2 = Medium relationship; 3 = Strong relationship. All combinations are
Hannah Kosow / Robert Gaßner
52 German Development Institute
Table 3: Tabular explanation of the influence matrix
Source: the authors following Blasche (2006, 74)
18 AS>PS
evaluated, and the center diagonal of the matrix remains empty. It is then
possible to calculate the sums of the lines and columns (regarding what fol-
lows cf. also Blasche 2006, 75 f.), which then serve as a measure of the de-
gree of networked interrelationships. The “line sum” of any factor repre-
sents the so-called “Active Sum” (AS) and indicates how strongly that fac-
tor affects other factors. The “column sum” of a factor, on the other hand,
represents the so-called “Passive Sum” (PS) which shows how strongly that
factor is influenced by other factors.
In this way, every factor is evaluated according to the relationship between
its active and its passive sum. It is customary in this regard to make a divi-
sion into:
– Active and impulsive factors (high AS, low PS)18. That is, the factor in-
fluences the problem field more than it itself is influenced. Such fac-
tors are termed effective “levers” or “switches” provided they are con-
currently steerable factors upon which it may be possible to have an ef-
fect through intervention.
Impact
On
Of
Factor A Factor B Factor C Factor D Active Sum (AS)
Factor A 3 3 1 7
Factor B 0 3 2 5
Factor C 1 1 2 4
Factor D 3 3 1 7
Passive Sum (PS) 4 7 7 5
Methods of future and scenario analysis
German Development Institute 53
– Reactive or passive factors (high PS, low AS)19. That is, the factor is
influenced more strongly than it itself influences. These factors repre-
sent useful indicators for the observation of a situation.
– Critical or. dynamic factors (high AS, high PS)20. That is, the factor has
a strong influence on the field but is itself subject to a strong influence.
These factors are linked with a network of other factors and are not to
be lost sight of at any time.
– Buffering or lazy factors (low AS, low PS)21. That is, the factor has on-
ly a weak influence on the field and is itself influenced only weakly.
Such factors have only a relatively inconsequential link with the net-
work of other factors; on the whole, they are rather isolated.
On the basis of this description and with the help of a so-called “priority
matrix“, it later on becomes possible, among other things, to calculate ef-
fective points of intervention as well (cf. Wilms 2006b, 54 ff. for details of
the procedure); that is, a search is made for active factors which can also do
justice to the criterion of openness for direct change by a person actively in-
volved (steerability) with only a brief period for change.
This manner of evaluating factors makes it possible to “filter out” those
which are to be tracked during the further course of events of the scenario
process, thus permitting a visualization of key factors in the narrower sense,
i.e. those factors assessed as active or critical in character. The basic as-
sumption behind this is that “lazy” and passive factors are to be assumed to
be either stable or that – precisely as functions of the active factors – they
need not be individually studied because they are linked only at second
hand with the network of other factors via the critical factors. A further ba-
sic assumption which is required is that the network of relationships among
the key factors at the moment of projection of the scenario will remain pre-
cisely as it is at present.
This process of characterization is used to select central factors which rep-
resent key factors in the narrower sense, often with the pragmatic goal of
19 AS<PS
20 AS*PS> (n-1)/2
21 AS*PS< (n-1)/2
observing from ca. 10 to a maximum of 20 factors during the further course
of events.
One example of such impact analysis is found, for example, in Gause-
meier / Fink / Schlake (1996, 191 ff.) concerning the future of individual
consumption behavior.
The next step, i.e. the analysis of key factors (Phase 3), begins by deter-
mining alternative possible future values of the selected key factors. For-
malized and mathematical as the procedure may be otherwise, this step al-
ways includes subjective elements. The scenarios may appear more “con-
servative” and/or “creative” later on, depending on how broad or narrow
the range of salient characteristics is assumed to be. (To take as an example
the key factor “development of the oil price“: possible future developments
might be 100 dollars and 150 dollars – or should we even visualize an oil
price of 200 dollars and more?). Via this step, the boundary lines of the
“funnel into the future” are then defined: which values are conceivable,
which others are unthinkable?
Hannah Kosow / Robert Gaßner
54 German Development Institute
Figure 6: Constant expansion of the “funnel into the future” through
systematic-formalized scenario techniques
(shown here in simplified form)
Source: IZT
77LLPPHH
W
WWW66
Methods of future and scenario analysis
German Development Institute 55
Various instruments are available for the following step of variation, and in
particular, the combination of salient characteristics. Two frequently em-
ployed techniques are consistency analysis and cross-impact analysis; these
two will be briefly described in the following
Consistency analysis
Consistency analysis (cf. Heinecke 2006, 190 f.) is used to unfold the range
of possibilities regarding the different conceivable values of all key factors
and to decide which combinations behave consistently toward one another
and may thus play a role in the construction of consistent scenarios. This is
decisive for the credibility and in particular for the interpretation of (cf.
Gaßner 1992, 230) of any scenario.
The technique begins by determining the various possible values of all key
factors; in the process, at least two possible values are assumed in the rule
for each factor: e.g. a rise in average temperature due to a climatic change
of one degree and another change of 4 degrees. The probability of this oc-
curring is, however, not explicitly taken into consideration in the process. It
may then be possible to construct a very large number of different sets of
“bundled” characteristics in the sense of so-called “raw scenarios” when all
factors with their different values are taken together. Depending on the
number of key factors and the selected number of respective values, a very
large number may be reached very quickly. For example, a total of
1,048,570 combinations is possible in the case of 20 factors with 2 values
each (cf. Heinecke 2006, 191). However, not all combinations of values are
equally credible, so that a ranking procedure can be used to select those sets
of factor characteristics which are particularly consistent.
A consistency evaluation is first carried out for each pair (cf. Gausemeier /
Fink / Schlake 1996, 255 ff). All factor values are juxtaposed in each case
with all other factor values. That is, simply put, key factor “A” is compared
with key factor “B” regarding both a) and b), etc. As in the “paper comput-
er“, this is achieved by combining each value of each factor with each val-
ue of every other factor (cf. Table 4). The consistency of each combination
is often assessed on a scale of 1 to 5 where 5 = Strong consistency (strong
mutual support); 4 = Weak consistency (mutual support), 3 = Neutrality or
independence from one another, 2 = Weak inconsistency (mutual opposi-
tion) and 1 = Strong inconsistency (complete opposition).
Hannah Kosow / Robert Gaßner
56 German Development Institute
It is then possible to carry out an evaluation of the consistency of various
“bundles” (cf. Gausemeier / Fink / Schlake 1996, 257 ff.), i.e. a calculation
of the consistency of all theoretically possible bundles of factor character-
istics. For this purpose, a consistency “unit of measurement” is calculated
(that is, a sum of consistency values for the individual pairs of characteris-
tics), which in turn makes a ranking procedure possible for the various
“bundles“. In addition, it is also possible to exclude those bundles which are
either completely inconsistent or contain too many weakly inconsistent
pairs.
Consistency analysis is used, for example, by the Department of Future
Analysis at the German Army Center for Transformation as a means of gen-
erating scenarios regarding topics related to national security. It is also fre-
quently recommended and employed for the implementation of scenarios
within enterprises (cf. e.g. Gausemeier / Fink / Schlake 1996).
The advantage of this instrument is that inconsistent pairs of factors can be
excluded from consideration, thus reducing the total number of possible fu-
ture factor “bundles” (cf. Gausemeier / Fink / Schlake 1996, 260).
Table 4: Consistency matrix
Source: Abstracted from Gausemeier / Fink / Schlake (1996, 258)
Factor A Factor B
Factor C Factor D
How do lines and col-
umns (i.e. "a" and "b"
characteristics) Value
Aa)
Value
Ab)
Value
Ba)
Value
Bb)
Value
Ca)
Value
Cb)
Value
Da)
Value
Db)
Value Aa)
Factor
AValue Ab)
Value Ba) 2 4
Factor
BValue Bb) 5 2
Value Ca) 5 2 2 5
Factor
CValue Cb) 3 4 5 2
Value Da) 4 3 1 3 5 2
Factor
DValue Db) 3 4 3 4 4 2
interrelate?
Methods of future and scenario analysis
German Development Institute 57
22 In order to avoid a confusion of terms: Consistency Analysis today is itself already being
called “Cross-Impact Analysis“, even when no probabilities are taken into consideration.
23 It was first used as a promotional gift from the Kaiser Aluminum and Chemical Com-
pany.
On the other hand, this type of consistency check does not take probabili-
ties into account. Their consideration is often not desired at all in the sce-
nario process, for example when extreme event developments are to be
studied (see above and also Gausemeier / Fink / Schlake 1996, 260).
The disadvantage of this instrument is that the calculation of units of meas-
urement for the consistency of factor “bundles” is possible only with the
help of computers (except in the case of a very small number of factors and
values (5 factors, each with 2 values, result immediately in 320 possible
bundles) (cf. Gausemeier / Fink / Schlake 1996, 257). This of course re-
duces the transparency and visual validity of the procedure. In addition, the
number of factors and values which can be taken into account is signifi-
cantly reduced.
When probabilities are also to be studied with this technique, a plausibility
check is also added in the rule. Here cross-impact analysis is appropriate
(see below).22
Cross-Impact Analysis
Cross-Impact Analysis (CIA) was developed in 1966 Theodore Jay Gordon
and Olaf Helmer, initially as a game (“Future”) (cf. Gordon 1994b, 1).23 To-
day this instrument has a plethora of uses, alone or in combination with oth-
er methods, and it has also advanced – e.g. in the tradition of Michel Godet
and the Batelle-Institute – to the status of a typical scenario technique (cf.
Mietzner / Reger 2004, 54).
Cross-Impact Analysis is used to present the causal relationships among
probabilities of different possible future events, to analyze them, and to take
into account their mutual consequences. It is used in the scenario technique
above all to analyze plausibility. “Plausibility” here means that probabili-
ties are taken into account additionally on the basis of the consistency
check (cf. Gausemeier / Fink / Schlake 1996, 259).
The basic logic followed by this analysis is that future developments de-
pend on the interaction of future events. Much as in the case of consisten-
Hannah Kosow / Robert Gaßner
58 German Development Institute
Table 5: A cross-impact matrix24
Source: The authors according to Gordon (1994b)
cy analysis, described just above, the different values of future develop-
ments are observed and their interactions are examined.
At the beginning, the future possible values of the key factors are deter-
mined. In cross-impact “language“, these are called “events“.
Then an “event” probability (initial probability) is estimated for each
event. This is done by regarding each event in isolation from, i.e. inde-
pendently of, other events.
In a third step, conditional probabilities are calculated on the basis of the
following central question: “If event A occurs, how great is the probability,
possibly influenced by it, that event B will occur?” The result is then dis-
played in a cross-impact matrix (see Table 5).
The conditional probability (for the following cf. Gausemeier / Fink /
Schlake 1996, 264) shows the probability of event A (e.g. reduced mobili-
ty) in case event B occurs (e.g. a rise in the oil price). In addition, a joint
probability can also be calculated (that is, the probability that both event A
and event B will occur (rising oil price and concurrently reduced mobility).
These interlinked probabilities can be worked out mathematically with the
aid of linear optimization based on the event probabilities of future events
and their consistency values.
24 The table is to be read as follows: event B has an initial probability of 0.4 (40%). If event
A occurs, the probability of B rises to 0.50. Event C has an initial probability of 0.75. If
event A occurs, the probability of C rises to 0.85, etc.
... Event probability changes as follows:
If this event
occurs ... Initial probability Event A Event B Event C
Event A 0.25 0.50 0.85