✅🐀 Snidal -- Rational Choice
⭐Summary
Argument 💫
Challenges to rational choice theory are valid but should be used to improve RC rather than dismiss it entirely.
RC is a broad framework that can be applied to multiple paradigms and used either normatively or positively.
RC benefits from being general/broad and simple because it’s so widely applicable and flexible
Rational choice should be proud of how far it’s come but also acknowledge its own distortions and constraints (intellectual humility).
Perspective 👁️
Rationalist
obviously a fan of RC
a very “our approach should depend on the situation” kind of guy
tries to bridge gap between normative and positive uses of RC
argues RC is a broad framework that can be used within all IR paradigms
argues math and verbal (anecdotal, deductive) analyses are equally important, supporting each other and compensating for each other’s flaws
Very Wohlforth in the claim that rationalism is not a monolith
Very Snyder in the claim that “we can blend multiple approaches to get better analysis”
Key Questions ❔
how can one address criticisms of RC from both within RC (internal) and outside of it (external)?
What is RC?
BASIC PREMISE: RC is a theory that describes how goal-oriented actors make decisions under constraints
simple and general — an umbrella, not a specific theory
analyzes the effect of various constraints (technology, domestic interest groups, other actors’ decisions)
it can be normative and/or positive, empirical and/or deductive, mathematical and/or verbal
how empirical (see “Key Vocab”) is RC?
sometimes it’s empirical (testable), sometimes it isn’t
it’s hard to test “rational choice” as a theory because it is so broad, so you can only test applications of it
is it necessary to formalize how people apply RC?
no, many people way back in the when understood RC without mathematical modeling (ex. reciprocal strategies were expressed as “tit for tat” or “eye for an eye”
however, formalization (using math) and verbal theory (anecdotal, pragmatic analyses) complement each other and are best used together
Using math reveals things that verbal theory misses, and vice versa
how does RC address change over time?
accounts for how nonstate actors (from companies to nonprofits to even terrorists!), domestic interest groups (electorate/selectorate the leader is accountable to/derives power from), historical traumas, shifting national goals can change a nation’s preferences over time
however, it doesn’t always need to do this because national struggles (war, shortages, economic turmoil, wealth inequality, immigration) recur over time because states are “like units” (structured differently depending on time and place, but structured to deal with many of the same problems)
how does RC explain who actors are and what they want/like? does it do this effectively?
see above
it can’t fully explain what actors want/like because it can’t see inside their heads, so it will always be limited and ineffective in some ways
with RC one has to deduce what actors want/like from the decisions they make
how does RC address normative concerns, or the idea of “what should be done” in international politics?
RC is already kind of normative: “based on the actor’s goals, what should they do to achieve those goals given the circumstances”
but the author argues that RC should be more normative
RC shouldn’t just be used to describe how interests lead to goals, it should describe how interests should be harnessed to achieve the best possible goals (which is itself a normative statement)
Key terms 🔤
rationalism / rational choice: a broad theoretical framework that mainly focuses on what decisions goal-seeking actors make under constraints (limits)
substantive: involving important, serious issues
methodology: rules and methods people use to do something
formalization: using math models to represent and validate theory
“hard” methodological approaches: math (quantitative), see “formalization”
“soft” methodological approaches: not math (qualitative)
empirical: provable by observation or experiments/testing
dynamism: change over time
normativism: belief in “what should be done”
exogenous: outside the model (taken as given, held constant)
endogenous: within the model (needs to be explained)
equilibria / equilibrium / equilibrium analysis: assumes that each variable within a model has no effect on the others and that they are all in balance (equilibrium)
“aggregate as actor”: simplifying the aggregate (total) of decisions among many individuals into a single decision by one actor
e.g nation as actor, despite nations being made up of many decision-making groups
📝Notes
I. Challenges to Rational Choice (internal summary)
RC key framework => drawn criticism
Key Vocab
substantive: involving important, serious issues
methodology: rules and methods people use to do something
formalization: laying out set, formal rules and frameworks for a thing. making something “official.”
the author specifically uses formalization to mean “using math to represent theory by making models of real things”
“hard” methodological approaches: math (quantitative)
“soft” methodological approaches: not math (qualitative)
empirical: provable by observation or experiments/testing
dynamism: change over time
normativism: belief in “what should be done”
Key Questions
how can one address criticisms of RC from both within RC (internal) and outside of it (external)?
What is RC?
how empirical (see “Key Vocab”) is RC?
is it necessary to formalize how people apply RC?
how does RC address change over time?
how does RC explain who actors are and what they want/like? does it do this effectively?
how does RC address normative concerns, or the idea of “what should be done” in international politics?
internal criticisms
RC too technical
RC not empirical enough
external criticisms
debate between rationalists and constructivists
emphasizes some problems while deemphasizing other equally important ones
cannot explain actors, interests, institutions, or change
rebuttal - rationalism is rooted in realism but widely applicable
II. What Is Rational Choice
“individual goal-seeking under constraints”
actors pursue what they want/like despite challenges
focuses on:
actors = their goals, their capabilities
constraints= technological constraints, institutional constraints, constraints imposed by others’ choices
It is a methodology (a way of doing things) rather than a theory (a way of explaining or predicting things)
It is an “umbrella” describing many things, not one fixed framework
similar to Wohlforth’s analysis of realism!
Goals
Under rational choice, actors’ goals could be self-interested OR normative / ideological
Usually used positively (in ways that describe “what is”) but can also be used normatively (in ways that describe “what should be”)
“goal seeking” isn’t the only thing explaining behavior, but it is one of the key drivers of/ foundations of behavior
Equilibrium
rational choice doesn’t specify what things need to be held constant (exogenous) and what things need to be explained (endogenous)
Commitments
Simplicity
the fewer assumptions (inferences), the better
Generalization
abstract and broad, not specific and concrete
provides widely applicable insights
a lot of “rational choice” debates are actually debates about things that aren’t rational choice, even though they started with a rational choice framework
key terms:
goal-seeking = pursuing goals / desires
constraints = limits on behavior
self-regarding = self-interested
ideational = based on ideas
hyper-rational = very rational, more rational than an actual human probably would be
positive = based on “what is”
assumption = to suppose something is true
anomic = having no norms or values, “gone rogue”
exogenous = outside the model (taken as given, held constant)
endogenous = within the model (needs to be explained)
equilibria / equilibrium / equilibrium analysis = claims that each variable within a model has no effect on the others and that they are all in balance
constitutive = having to do with the parts that make up something
generalization = inferring broad trends from specific scenarios
tautologies =
broad, flexible
not substantial or causal, but explanatory
III. Methodological Debates and Challenges for Rational Choice
internal critiques
about methodology = “how to do rational choice”
Rational Choice and Formalization
What is the role of math models in rational choice theory?
“Rational choice has an elective affinity for formalization” = people who use rational choice like to do math and use math to represent the world\
implicit question: “do you need to use math to use rational choice?”
no, you don’t need lots of math modeling [formalization] to do rational choice, many people have used other methods (game theory, anecdotal) as long as rational choice has been around and even before then. however, if correctly applied, mathematical modeling can make rational-choice analyses better ✌
some people knew things before math could explain it (ex. tit-for-tat), but explaining things with math can create a better understanding of those things and change people’s minds about what an effective course of action really looks like
just because things are “counterintuitive” doesn’t make them less true
using math helps better analyze arguments & assumptions
Schelling — math isn’t sufficient on its own to explain IR and verbal theories of military strategy are dead
use math to build foundations for verbal arguments
How to balance math (hard) and verbal (soft) approaches in RC by creating a successful model
Simple and elegant: model shouldn’t be so overcomplicated it overshoots reality
Synthesizes math and verbal logic: basic logic of the model should be in verbal, not numerical, terms; but that verbal logic should be supported by mathematical evidence
Testable: model should succeed under real-world testing/observation (it should be empirical)
Empirical evaluations of RC arguments
Deductive: conclusions from abstract reasoning over existing evidence
Empirical: conclusions drawn from observation and testing
“RC is not empirical enough!”
it’s nearly impossible to observe things like actor motivations, preferences, and beliefs
it’s hard to test something so general and insubstantial, testing rational choice is like testing “statistics”
But RC is empirically useful
gives examples of empirically useful IR research
RC can’t explain everything. No theory can. So what can RC explain?
motivation based on consequence
motivated based on “appropriateness”
sometimes people like to “curve fit” / “rationalize” their models (tailor their data to fit what they already believe about it) instead of actually testing them
but rational choice can help prevent this
using math helps disprove “rationalized” arguments
emphasis on simplicity helps reveal what is elegant, well-founded theory and what is complicated mental gymnastics
emphasis on generalizable conclusions/insights limits cherry-picking specific situations
“fixed preference” = in any given scenario, an actor will want [thing].
criticism that this ignores change over time
arises from balancing broad theory with testing
big problem to assume “preferences caused by behavior” / “norms caused by behavior”, because behavior causes preferences and those two things are not interchangeable (a tautology)
two ways to have quality qualitative testing of RC models
case studies
“analytic narratives” - evaluating case studies within a RC framework, using RC to describe actors, preferences, choice, constraints, and other variables
can help describe trends over time
can help bridge the gap between valid logic and necessary simplicity
can explain things both inside the case and across cases through RC
Balancing analytic narratives (Elster)
actors within the model shouldn’t be given the ability to play 8d chess; there should be room for irrationality, miscalculation, etc. because that is how it is in reality
“aggregate as actor” - simplifying groups made up of a variety of individuals into a single individual decision-maker
within an actor’s calculations, rational choices/strategy-based decisions can share space with moral concerns
IV. Substantive Challenges for Rational Choice
Incorporating Dynamics and Change
equilibrium (holding certain things constant) = big obstacle to understanding change, as in reality everything is constantly shifting
“equilibrium path” = status quo, prevailing general conditions
when is it constant, when does it shift?
example: “anarchy,” a core assumption of realism
author argues RC works best concerning change when applied to cooperation and institutions
why do RC analysts hesitate to analyze change? two assumptions/reasons
because states are “like units” facing similar problems yesterday, today, and tomorrow
power struggles are constant, financial issues are constant, immigration issues are constant. History repeats itself
because large-scale change happens very slowly
Endogenous Actors and Preferences
traditional assumption: actors and preferences remain constant, what changes are the things that constrain those actors/preferences
because you can’t see preferences, but it’s easier to see constraints
fixed preferences = concrete, testable arguments; dynamic preferences = untestable, inconsistent arguments
but preferences do change!
fixed-preference analysis is good sometimes, but not all the time
RC & Domestic-level interest groups
audience costs
two-level games
commitment problems
preference change comes from changes in underlying constituencies
i.e changes in policy/decision-making come from changes in opinion by the people the policymaker/decision-maker is accountable to/derives power from
authoritarian = “selectorate” / elites
democratic leader = “electorate” / general voting public
RC & International actors
explaining the role of nonstate actors (NGOs, MNCs, terrorist groups) in international politics
RC and Actors’ Movements to Change Each other
ADD MORE ACTORS
Changing membership within international institutions
conditions for membership can change based on the issue
firms/NGO might participate in solving one issue but not another
Actors trying to change other actors
Human rights NGOs trying to make a brutal police state less repressive
RC and Actors’ Movements to Change Their Own Preferences
change in information = change in preferences
change in institutions =change in preferences
Normative and Policy Analysis
Instead of “based on actors’ interests, what will happen”, RC should look at “how can we align actors’ interests with what ‘should’ happen”
it is normative: a theory about how people “should” behave in each situation
but it could be more normative
addressing resource inequalities within cooperation & how those inequalities create “winners and losers”
can help incorporate interests into normative analyses, preventing idealistic miscalculation
can help narrow discussions about the role of international law
V. Conclusion
Math (“soft)” and verbal (“hard”) methods work with each other in RC, rather than being separate from / superior to each other
Flaws in RC are because
RC analysts cannot perfectly predict how the world works
Inability to explain change can be an asset in one situation and a liability in another
Limitations regarding normative/strategic preference ordering, change over time are what make RC effective, it would be too broad otherwise
rational choice should be proud of how much it’s revealed about IR, but also realistic about its flaws and limitations in doing so