Institutions Making Behaviour Predictable

1
Institutions Making Behaviour Predictable
2
The Puzzle

• social interaction works
• it presupposes that Interaction Partner (IP)
• predicts reasonably well, how Actor (A) will
  behave
• yet psychology demonstrates
• almost unlimited plasticity of human behavioural
  programs

3
(Puzzle)

• two possibilities
• wrong assumption
• IP can do with less predictability
• predictability must be defined differently
• predictability provided otherwise
• topic here
• claim by institutional intervention

4
Puzzle Policy Problem ?

• if not explanatory power
• understanding an overlooked function of existing
  institutions
• if yes task for institutional design
• hunch
• massive decontextualisation (e.g. Internet)
• welfare from handling greater complexity
• both call for new institutions

5
Organisation of Talk

• Psychological Predictability Problem
• Rational Choice Responses
• Behaviourally Informed Responses
• ( dropping the rationality assumption for A)
• Behaviourally Informed Responders
• ( dropping the rationality assumption for IP)

6
I. Psychological Predictability Problem

• Level of Analysis
• Plurality of Mental Tools
• Cognition
• Motivation

7
1. Level of Analysis

• actual interest of IP As behaviour
• economics, rational choice (behaviourism)
• where does behaviour come from ?
• preferences (at random)
• restrictions
• digging deeper I
• where do preferences come from ?
• better understanding behavioural dispositions
• digging deeper II
• where do behavioural dispositions come from ?
• better what translates into at the symbolic
  level

8
(level of analysis)
9
(level of analysis)
10
(level of analysis)
11
(level of analysis)
12
effects on the predictability problem

• descriptively
• no linear relation between levels
• each of them has his autonomy
• prescriptively
• understanding / modelling one level is not enough
• ? how does it translate into behaviour ?

13
2. plurality of mental tools

• illustration three tasks
• grasping the coffee mug in front of me
• driving to work in the morning
• designing my next article
• is not done the same way mentally
• ? plurality of mental tools

14
(plurality of mental tools)

• many models for classifying the plurality
• in a sketchy way
• radically simple decisions rules (heuristics)
• routines, habits
• deliberate decision making
• criteria
• simple complex
• contextual decontextualised
• subconscious - conscious

15
3. cognition

• ensuing predictability problems
• attention
• can be governed by heuristics
• perception
• cue based decision rule
• ? mental representation
• e.g. mental accounting
• reference point

16
(cognition)

• elaboration
• relying on schemata
• blending
• giving meaning by
• conceptual sex

17
4. motivation

• economics assumes
• exclusively by interest
• mentally organised in a well-behaved utility
  function
• actually
• much less orderly
• motivation often elicited ad hoc
• more motivators than utility

18
(motivation)

• alternative motivators
• happiness
• self esteem
• fairness
• envy
• clarity
• preferences for institutions
• intermitting emotions

19
II. Rational Choice Responses

• rationality does only get you so far
• descriptively
• prescriptively
• but
• clears the ground
• provokes sharper questions

20
Rational Choice

• Information Asymmetry
• Game Theoretic Analysis of the Problem
• Unilateral Response by IP
• Unilateral Response by A
• Joint Response
• Third Party Intervention
• Sovereign Intervention

21
1. Information Asymmetry

• two interpretations
• problem of indeterminacy
• no attempt by IP to generate additional
  information about A
• no attempt by A to transmit additional
  information
• low predictability treated as behavioural risk
• more precisely as exogenous risk

22
(information asymmetry)

• problem of information asymmetry
• IP / A also consider generation of information
• inspection
• mechanism design
• Bayesian updating
• in case the additional information only allows
  better educated guesses

23
2. Game Theory
24
assumptions

• IP does not know natures move
• A knows IPs move
• gg gt g gt 0 gt -l
• ? defection is As best response
• 2g gt gg-l
• ? g, g is not only fair, but also efficient

25
illustration

• detrimental actor homo oeconomicus
• beneficial actor keeps given promises
• technically not opportunistic

26
equilibriae
27
implications

• in this game
• a rational IP is better off,if he learns As
  type
• investment in type revelation pays for IP
• but also for beneficial A
• pooling ? separating equilibrium

28
3. Unilateral Response by IP

• assumed fully rational
• direct route
• inspection
• indirect route
• mechanism design
• bringing A in a situation where even a
  detrimental A is better off revealing type
  information

29
4. Unilateral Response by A

• assumed rational reaction to his own exogenously
  given type
• detrimental A will not move
• any prior beliefs of IP lt 100 detrimental
  actors in the population are to his advantage
• beneficial A will want to move
• how much so depends on his beliefs about IPs
  prior beliefs

30
(Unilateral A)

• direct route
• revelation such that IP can verify
• indirect route 1
• direct mechanism
• indirect route 2
• bail out
• A pays gg-g to IP, if IP credibly commits to
  cooperation (and thus A gets g for sure)
• rational, if g gt gg-g

31
5. Joint Response

• stalemate between IP and beneficial A
• who bears the cost of type revelation
• both prefer revelation to non-revelation
• but both would rather have the other pay
• battle of the sexes

32
investment game
33
investment game
34
investment game

• bargaining
• would be an easy way out, if IP knew that A is
  beneficial
• since IP does not,settlement range is much
  smaller
• only unilateral options of IP can be exploited
• burden-sharing only, if A pays upfront
• and if As contribution lt g

35
6. Third Party Involvement

• gives more options for making statements /
  promises credible
• in particular in policing IPs obligation to
  invest under a burden-sharing agreement

36
7. Sovereign Intervention

• trivially helpful
• enforceable rule that punishes As defection
• Pigouvian tax gt gg-g
• but only due to the rigorously simple model
• in particular assuming away that sovereign can
  only exploit verifiable information
• about defection
• about efficient tariff

37
8. Summary

• radically simplified problem
• can always be solved by sovereign intervention
• in some cases joint solutions by the parties
• sovereign intervention fails in more realistic
  cases
• even if dichotomy of beneficial and detrimental
  actor is maintained
• even more so in light of the almost infinite
  number of reasons for low predictability

38
III. Behaviourally Informed Responses

• dropping the (limited) rationality assumption for
  A
• keeping the assumption for IP
• Limited Ability to Comply
• Exploiting Generic Behavioural Knowledge
• Changing the Character of the Task

39
1. Ability to Comply

• rational choice
• takes for granted that actors are able to comply
• but assumes they are unwilling to do so, unless
  it is in their interest(opportunism)

40
(Ability to Comply)

• actually, for a host of psychological reasons,
  actors can lack the ability to comply, e.g.
• non-conscious mental tools at work
• cognitive limitations
• rational beliefs about beliefs
• different mental models ? both play a different
  game

41
2. Exploiting Behavioural Knowledge

• One Level ? Two Level Approach
• Availability
• Strategies
• Mechanisms
• Institutions

42
a) One ? Two Level Approach

• One Level Approach
• enriching the information asymmetry approach
• Two Level Approach
• Level 1 impact on behavioural disposition
• Level 2 exploiting the changed disposition

43
b) Availability

• of Two Level Approach
• on the spot
• occasionally
• in particular switch to conscious decision
  making
• learning
• development / enculturation
• mutation

44
(Availability)

• implications for the choice of actors
• on the spot
• (in principle) all actors
• learning
• typically only with greater social embeddedness
• development / enculturation
• at most weakly directed
• ?limited unilateral or short-term bilateral
  options

45
c) Strategies

• Exploiting the Plurality of Mental Tools
• Cognitive Route
• Motivational Route

46
i) Exploiting the Plurality

• each type of mental tools can be emphasised or
  deemphasised
• rationalisation
• or better stressing reason based choice /
  narrative reasoning
• routinization
• patterning behaviour
• automatization

47
ii) Cognitive Route

• can try to match all elements of cognition
• e.g.
• attention management
• framing
• e.g. offering a reference point

48
iii) Motivational Route

• can exploit all motivational factors,
• especially those transcending interest, e.g.
• appeal to self-esteem
• exploiting punishing sentiments

49
d) Mechanisms

• mostly indirect, e.g.
• imposing formality
• including obligation to rely on intermediaries
• imposing or shaping context
• discourse and communication

50
e) Institutions

• three examples
• market
• rationalizer (within limits)
• professionalisation
• routinizer
• drivers licence
• automatizer

51
3. Changing the Character of the Task

• such that unpredictability no longer matters
• frequent advice from
• rational choice theoreticians
• behaviorism
• parallel to climate change negotiations
• mere strategy of adaptation
• no attempt at mitigation

52
drawbacks

• expensive
• institutions must be foolproof
• minmax
• or impose an arbitrary degree of uncontained risk
• illusory
• would be a mere standardisation of situation
• but personality and situation interact

53
(drawbacks)

• apparent attractiveness results from artificially
  narrowing down the analysis
• standard suggestion market participants act as
  if rational
• ? results from the fact that market shapes
  behaviour, not just situation

54
IV. Behaviourally Informed Responders

• two opposite directions of analysis
• detrimental effect
• limited ability to exploit behavioural knowledge,
  even if available
• beneficial effect
• better than rational

55
1. additional complication

• in particular cognitive limitations
• excessive complexity
• too many variables
• non-linearity
• counter-intuitive character of findings
• intermittent self-esteem
• seeing at happiness, rather than utility
• adaptive, rather than strategic

56
2. better than rational ?

• example as in rational choice part
• powerful cheater detection mechanism
• strong search heuristics
• exploiting subtle nuances, like body language
• credible, due to punishing sentiments

57
limitations

• existing mental tools are well adapted to stable
  social interaction
• are well embedded in existing social institutions
• change in context new task for institutional
  design
• context B, rather than A
• quicker changing contexts