4.37M

Category:

management

Models of decision making

1.

SCHOOL OF MANAGEMENT
Giancarlo Vecchi
POLICY ANALYSIS COURSE 2022
4. MODELS OF DECISION MAKING

1. DECISION MAKING MODELS
Solution A
Solution B
Problem
Solution C
Solution D
To decide (from latin de-caedere) means “to cut off”
It is impossible to observe the very moment in which
a decision is made (when the alternatives are cut off)
The decision is a process
Giancarlo Vecchi

3.

2. Main Variables of a D-M Model
A decision-making model is a short
description of the main features of the
decision-making process:
the properties of the decision-maker
who is he/she?
his/her cognitive capabilities
how do he/she think?
how solutions are searched and
assessed
how the final choice is made
Giancarlo Vecchi

4.

3.
a) The rational model
Giancarlo Vecchi

5.

3.
A decision is a process in which the actor:
Knows all his/her objectives
Is able to rank the objectives in an order of
preferences (prioritisation)
Knows all the alternatives courses of action
Is able to measure costs and benefits of
every alternative
Chooses the alternative maximising the
benefit/cost ratio (optimizing approach)
Giancarlo Vecchi

6.

3. The rational choice elements
Completeness – all possible courses of action can
be ranked in an order of preference
To be able to do that, you need a lot of
information about the state of the world and
the outcome of your actions. In the rational
choice model, actors have the sufficient level of
information.
Transitivity – if action A is preferred to B, and
action B is preferred to C, then A is preferred to
C (also, local non-satiation, independence of
irrelevant alternatives – IIA)
An actor will select a course of action that yields
the highest utility (under resource constraints):
maximization
Giancarlo Vecchi

7.

3.
The rational model is based on
• the idea of economic rationality as it
developed in economic theory
• The idea of bureaucratic rationality as
formulated by sociological theories of
organization and industrial society
• BUT: IS THE MAIN MODEL OF
SCIENTIFIC RESEARCH: ITS BASES
ARE ON THE KNOWLEDGE OF
EXPERTS, SO THEY PLAY A
FUNDAMENTAL ROLE
Giancarlo Vecchi

8.

3.
Methods for rational decision-making
Cost-Benefit Analysis
Multicriteria Analysis
Environmental Impact Assessment
Operational Research
etc
Giancarlo Vecchi

9.

3.
Rational decision-making.
Reasons for its success
Making technique prevail over politics
Overcoming the bureaucratic behavior of
public servants
Computation capabilities
Giancarlo Vecchi

10.

3.
Conditions for operating with the
rational model
Well-structured problem (without ambiguity)
Non contradictory goals
Possibility of setting goals before means
Availability of causal theories: if A then
(always) B, in a invariable/changeless context
condition the behaviour of B is explained by
the behaviour of A (and only A)
Giancarlo Vecchi

11.

3.
Single decision-maker (or well coherent
group)
Time availability
Giancarlo Vecchi

12.

3.
• STRATEGIES IF YOU, AS AN EXPERT
OR A PROMOTER OF A PROJECT,
NEED “MORE KNOWLEDGE”
a) Claim for more time before the final decision,
underlying the future costs of a sub-optimal
solution
b) Claim to study good practices around the world
c) Claim for valid and consistent basic information
d) …
Giancarlo Vecchi

13.

3.
Too pretentious?
This type of rationality has been defined as:
Olympic (Simon)
Synoptic (Lindblom)
Comprehensive
As we often cannot defeat uncertainty, we have to
practice other methods that are less pretentious,
such as:
a) the bounded rationality model
b) the incremental model
c) the garbage can model
Giancarlo Vecchi

14.

4.
b) The bounded rationality
model (or procedural
rationality)
Giancarlo Vecchi

15.

4.
Herbert Simon (economist, political scientist,
sociologist of organizations) observed that very
often in real life decisions are made by people
that are limited by:
limited knowledge
limited attention span: problems must be dealt
with on a serial, on-at-a-time basis, since decision
makers cannot think about too many issues at the
same time; attention shifts from one value to
another;
limited memory, limit on the storage capacity of
the human mind;
Giancarlo Vecchi

16.

4.
habit and routine: human beings and
organizational behaviour structured in routines to
manage in a experienced way the same type of
situations
organisational environments which frames the
process of choice
time availability
consequences that cannot be known, so that the
decision-maker relies on a capacity to make
valuations;
etc.
Giancarlo Vecchi

17.

4.
Therefore, the ideal of perfect (substantial)
rationality must be dismissed in favour of
limited (procedural) rationality
An individual is rational if the behaviour is
purposive, intentional, i.e. directed at
realising the goals of expressed values
In this situation the criterion of choice is not
optimising but satisficing, i.e. reaching a
decision that is good enough
Giancarlo Vecchi

18.

4.
This conclusion can be reached also through the
observation that:
most decisions in political settings are taken by
a plurality of actors (coalitions)
the solution is often reached through a
sequential assessment (and not a parallel
investigations with the comparison among all
the available alternatives) that stops when the
first “good enough” solution is found
Giancarlo Vecchi

19.

4.
Satisfaction criterion and sequential assessment/a
Satisficing is the strategy of considering the
options available to you for choice until you
find one that meets or exceeds a predefined
threshold—your aspiration level—for a
minimally acceptable outcome.
Giancarlo Vecchi

20.

4.
Satisfaction criterion and sequential assessment/b
Sequential assessment: if you find the first
option meets your minimum aspirations, and if
you have not time to search for many
alternatives, you can stop your decision
making process, without compare it with other
available alternatives
If the first isn’t good in a sufficient level, you will
analyze the second option, and compare it
with your desired minimum level of aspiration,
repeating the ‘satisfaction procedure’.
This a sequential, linear, procedure
Giancarlo Vecchi

21.

4.
STRATEGIES IF YOU ARE THE DESIGNER
• Be sure that your proposal reached the agenda
(it is part of the group of the solutions that will be
discussed)
• Claim to present your proposal as the first, and
• Define before the decisional process, the rules
about ”How to decide” with the aim to analyze
firstly your proposals
Giancarlo Vecchi

22.

4.
Interlude: the Condorcet/Arrow
paradox
A given order of preferences is transitive
when
• if choice A is preferred to choice B
and
• choice B is preferred to choice C
then
• choice A is preferred to choice C
Giancarlo Vecchi

23.

4.
Example: choosing energy policy
Giancarlo Vecchi

24.

4.
A committee is appointed to advise on
the best energy policy, composed by
Prof. Red – economist – maximises
efficiency
Mr. Green – environmentalist –
maximises sustainability
Dr. White – engineer – maximises
technology
Giancarlo Vecchi

25.

4.
First choice
Prof. Red
Mr. Green
better coal
than nuclear
power
better solar
energy than
coal
better nuclear
power than
solar energy
power than
solar energy
better coal
than nuclear
power
better solar
energy than
coal
better coal
than solar
energy
better solar
energy than
nuclear power
better nuclear
power than
coal
Second best better nuclear
therefore
Giancarlo Vecchi
Dr. Black

26.

4.
To learn
It is possible that collective preferences are
non-transitive (or cyclical)
Kennet Arrow has enlarged the paradox to
the “impossibility theorem” stating that it is
impossible to aggregate preference into a
single social welfare function under the
conditions of democracy
Giancarlo Vecchi

27.

5.
c) The incremental model
Giancarlo Vecchi

28.

5.
Charles Lindblom observed that often policy
making implies different groups with different
values and goals.
They are at the same time partisan and
interdependent.
In other words they have:
structurally conflicting goals
but they need each other
(example: majority/opposition,
politicians/bureaucrats, central
state/regional governments)
Giancarlo Vecchi

29.

5.
Decision-makers do not follow the
synoptic method
Means (often) determine goals
(interdependence)
Only few alternatives are explored
and assessed
The assessment consists in a
comparison between the expected
change and the status quo
Giancarlo Vecchi

30.

5.
An alternative is preferred if there is
an agreement among decisionmakers that this alternative is good
(i.e. acceptable)
The decision process is a partisan
mutual adjustment
Decisions are thus incremental, i.e.
departing as little as possible from the
status quo
Giancarlo Vecchi

31.

5.
• In the incremental model the typical
decisional situation is not the one in which
all the actors seek to find the solution to a
specific problem.
Giancarlo Vecchi

32.

5.
• On the contrary very often there is
- someone interested in solving the
problem,
- someone else only interested in
imposing their solution,
- someone interested in participating in
the process but not really keen in
solving the problem or adopting a
given solution, but only interested to
improve its role, or exchange the current
support with the support of the actors in
another decisional arena
Giancarlo Vecchi

33.

5.
Lindblom, in the book “The intelligence
of democracy” (note the double
meaning), shows that this
“confusion” is really what democracy
is all about
Giancarlo Vecchi

34.

5.
Objections to incrementalism
A conservative method
Not all interests are equally powerful
BUT
it is useful to explain decision making
processes at a micro-level
Giancarlo Vecchi

35.

5.
STRATEGIES IF YOU ARE THE DESIGNER
You should prepare designs that have:
• one ‘core contents’, i.e. things that you do not
want to negotiate
• contents that you could change if a negotiation
will be required to reach a decision
• analyse the policy field to know the actors that
could be your partner and those that could be
your opponents, to understand why and which
changes you can accept
Giancarlo Vecchi

36.

6.
The garbage can model
(Michael Cohen, James March e Johan
Olsen + John Kingdon)
Giancarlo Vecchi

37.

6.
The garbage can model stresses the
anarchical nature of decision-making
processes as “loose collection of
ideas” as opposed to rational
“coherent structure”. Actors discover
preferences through action, rather
than act out of preferences.
Understandig is poor (problems are
complex), trial and error learning
operates, and participation is fluid.
Giancarlo Vecchi

38.

6.
Decision-making arenas are collection
of choices which looks for problems
and issues and seeks decisional
situations in which they may be
advanced. Solutions looks for
problems. Choices thus compose a
“garbage can” into which various kind
of problems and solutions are
dumped by participants as they are
generated.
Giancarlo Vecchi

39.

6.
Ambiguity, not only uncertainty
Factors of ambiguity
Actors’ preferences are not always fixed
and consistent
- rather they change during the
development of the process and, due
their interests and beliefs, not
necessarily coherent;
Actors’ participation is fluid
- people move in and out
- their attention is a scarce resource (it
depend on the period)
Giancarlo Vecchi

40.

6.
(Factors of ambiguity)
Solutions can come out also when
problems are not present yet
- there are problems in search of a
solution and solutions in search of a
problem (because they are supported
by different actors, because we can
use an available solution to deal with a
new problem, adapting it…)
Problems and solutions can be presented
in several decisional venues, when actors
see choice opportunities
Giancarlo Vecchi

41.

The four streams
Actors (A)
Problems (P)
Solutions (S)
Choice Opportunities (O)
O are garbage cans in which A throw P and
S. The final decision will depend on the
casual matching of P and S.

42.

How garbage can model works
A
A
S
S
P
S PS S
P PS
P
S
O
P
S S
P
P P
P
O
P
P
P
P
S
S
P
S
P
S P
PS S
O

43.

6.
In a garbage can process there are
exogenous, time-dependent arrivals of choice
opportunities, problems solutions and decision
makers. The logic of ordering is temporal
rather than hierarchical or consequential.
Problems and solutions are attached to
choices, and thus each other, not only
because of their means-end linkages but also
because of their simultaneity. (Cyert & March,
1992:25)
Giancarlo Vecchi

44.

6.
The garbage can model argues that
• there are situations in which some
issues will have a solution attached
to them, and the coupled and
decided
• others will not,
• other solutions may be roaming
around looking for an issue to which
attach themselves.
Giancarlo Vecchi

45.

6.
• ‘Policy entrepreneurs’ are actors interested in
reach a decision, and in the coupling
between problem and solutions:
“people who are willing to invest resources of
various kinds in hopes of future return in the
form of policies they favour”.
• They are crucial to the survival and success of
an idea: ideas must be technically feasible,
compatible with dominant values and able to
anticipate future constraints.
Giancarlo Vecchi

46.

6.
STRATEGIES IF YOU ARE THE DESIGNER
You should have many proposal designed, a sort of
projects’ basin; it will be useful when a window
opportunity opens, and you will be ready to present
an appropriate proposal
You should know many decisional venues, to check
when a decisional processes starts there;
You should learn, during the process, if you should
change something to improve the opportunities for
your project or/and start a new partnership
Giancarlo Vecchi

47.

7.
To sum up
Giancarlo Vecchi

48.

7.
Model
Who decides
Evaluation Criteria
Individual
Predictive rationality:
Optimum
Individual/coalition
Retrospective Rationality:
Satisfaction
(eg: no losses, to win something)
incremental
Partisan interdependent
actors
Retrospective Rationality:
Mutual adjustment
(negotiate, constructing a coalition to
reach a decision)
garbage can
Changing actors and
positions depending on
arenas, problems and
learning from interactions
Retrospective Rationality:
Contingency
Entrepreneurship and capability to
connect problems and solutions in
open policy windows
rational
bounded rationality
Giancarlo Vecchi

49.

7.
In the table going from the top to the bottom
decreases the prescriptive value and
increases the descriptive value
The prescriptions of the rational model are
strong and clear, but they are very difficult
to apply, while the prescriptions of the
garbage can model are universal but weak
and unclear
Which model? It depends on the situation
we have to tackle
Giancarlo Vecchi

50.

8. Uncertainty and decision making
Designs as theories
If I adopt the measure x at the time t1 I will
get the result y at the time t2
But how can I know whether a causal link
between x and y exists?
This is the key problem of uncertainty
Giancarlo Vecchi

51.

8. Simple, complicated, and complex
problems
S
From Glouberman & Zimmerman 2002, p.2
Giancarlo Vecchi

52.

8. Wicked problems
Wicked problem: complex problems, that
characterize the activity of many social
professions: management, planning, etc.
Wicked: complex, interdependent, intractable,
conflict-prone
Examples?
Giancarlo Vecchi

53.

8.
Wicked problem: complex problems, that
characterize the activity of many social
professions: management, planning, etc.
Wicked: complex, interdependent, intractable,
conflict-prone
e.g. environmental/ecological problem, climate
change, poverty, obesity/food,
health/pandemic, metropolitan
management, etc.
Giancarlo Vecchi

54.

8. Wicked problems
Modern society is too pluralistic to tolerate artificial solutions
imposed on social groups with different attitudes and
values, and this pluralism undermines the possibility of
clear, agreed solutions.
The finite problems tackled by science and engineering are
seen as relatively ‘tame’ or ‘benign’ in the sense that their
elements are definable and solutions are verifiable
By contrast, modern social problems are generally ‘ill-defined’
and resistant to an agreed solution. They are therefore
‘wicked’, relying on political judgements rather than
scientific certitudes (Rittel & Webber, 1973: 160).
Giancarlo Vecchi

55.

6.
From: Alford & Head 2017, 402
Giancarlo Vecchi

56.

8.
Two responses to uncertainty
Defeating uncertainty
more information, more solid theories, better
method
planning, rationality, etc.
Living with uncertainty
we must accept that often we have to make a
decision in the darkness, and try cope with it
incrementalism, negotiation, partisan adjustment,
trial and error
Giancarlo Vecchi

57.

8.
Decisional strategies:
• Authoritative: one strong leader should decide with
authority. BUT: wicked problems by their nature are usually
beyond the cognitive capacity of any one mind to diagnose or
comprehend
• Competitive: fostering competition between societal actors to
come up with understandings of the problem and potential
advances in dealing with it. BUT: risk generating heightened
conflict that consumes resources and delays solutions.
• Collaboration: public consultation or participation in decisionmaking. BUT: how to pull this together into a coherent account
• Expert authority: again the nature of the problem is beyond
the thinking abilities of even the most erudite expert.
Giancarlo Vecchi

58.

8.
OPPORTUNITY:
• Cooperation: we need strong cooperation among
different types of actors to work with different
resources. We need technical expertise, but it is
not the only type of capability required; also
needed is the capacity to lead, organise and
manage the cooperative processes, the exchange
of information, the implementation of responses,
etc.
Giancarlo Vecchi

59.

Thompson’s matrix
Goals
Certain
1
Rational approach
Computation
Uncertain
Certain
(agreement)
3
Trial and
error,
experiments
Uncertain
(disagreement)
2
Bargaining
4
Chaotic
Redefining
problem

60.

8.
Thompson’s Matrix/2
• The matrix shows the approaches to follow on the
basis of two variables: technology (means) and
goals; and the level of uncertainty.
• Technology: do we know how to solve the
problem?
• Goals: do we agree about the characteristics of
the problem (the problem definition)?
Giancarlo Vecchi