Transcript Modeling Decision-Making - uni

Remarks on a Political Modeling
Strategy for Social Systems
Detlef Sprinz
PIK - Potsdam Institute for Climate Impact Research
Overview
• Goals
• Options
• Conclusions
2
Goals
• Provide a (Set of) Tool(s) to
– provide a framework for the analysis of nontrivial decisions
– among multiple actors
– on various levels of analysis or aggregation
– to replicate past decisions and
– explore the likely outcomes of major decisions
for the future
3
Options
• Basic Options
–
–
–
–
–
case study
statistical analysis
(cooperative and non-cooperative) game theory
optimization tools (incl. CBA)
simulation
• form of quasi-experimentation
4
Options
• Example: Political Simulation Model of
Bueno de Mesquita
• Assumptions
– Rationality of Actors
• competing actors
• maximize expected utility under limited time
horizon
• “voting” on issues
• no level of analysis problem
– Ability to Provide Input Data
5
Options
• Goal: Decision-Making
– offers and counteroffers
– “produce” winning coalitions (if possible) or
“fabricate” majorities
• Outcome of Vote
– determined by
• median voter theorem
• veto player
– fall back: status quo
6
Options
• Inputs by Player
– general power or influence (resources) of each
actor
– stated policy position
– salience (priority) of each actors
• Simulation Terminates When
– expected benefits of further negotiations
< expected benefits of further negotiations
– discounting the outcomes over time
7
Options
• Advantage of BdM Policy Forecaster
– utilizes case experience similar to case study
analyst in a fully structured way
– provides forecasts and simple dynamic of
decision-Making
– rigorous tool rather than guesstimate
• accuracy is imperfect (ca. 90% correct predictions)
– permits sensitivity analyses
8
Options
• Some Challenges & Limitations
– lack of simultaneously including related agenda
items (problem of uni-dimensional political
space)
– high-quality data inputs
• not easily available even for many European
countries
9
Options
• How to Use it for Social Systems I
– context of climate change
– three rounds of simulations
• Germany
• European Union
• global
– replications of some decisions where we know
the outcome
– predictions about some decisions where we do
not yet know the outcome (e.g., post-Kyoto) 10
Options
• How to Use it for Social Systems II
– beyond climate change
– in context of European Climate Forum
• with select target groups (self-selected)
• replication
• short-term forecasting (easy “verification”) &
learning
• longer-term forecasts
11
Conclusions
• A Proposed Sequence of Activities
– build a backbone at PIK incl. endowment
– review other options in the family of simulation
models
– outreach to a limited set of relevant
inter/national institutions
– engage a small set of internationally renowned
scholars
• Bueno de Mesquita, Jacobson et al.
• Underdal, CICERO et al.
12