Rational Choice and Culture

Document Sample
Rational Choice and Culture Powered By Docstoc
					         Genoese non-Jewish merchant class)                                                             and how much a piece of work affects
         that employed one-shot contracts, was
                                                        Rational Choice and                             our prior beliefs. Whether our beliefs are
         more accepting of occasional betrayal,         Culture                                         falsified (as with the paradoxes of so­
         and demanded less extensive informa­           George Tsebelis                                 cial choice, starting with Arrow's Theo­
         tion about potential partners; Among the       University of California. Los Angeles           rem) or corroborated (as with evidence
         many rich (and, to me, Weber-like) im­         tsebelis@ucla.edu                               that the composition of committees in
         plications ofGreif's argument is this: that                                                    Congress affects policy outcomes), im­
         the first, more 'traditional,' culture will    "Why did the dinosaurs disappear?" I            portant work significantly affects our pri­
         have been more stratified and will have        asked my three year old son. He did not         ors. It makes us understand something
         offered less opportunity for social mo­        understand that it was a rhetorical ques­       that we did not understand before, in­
         bility.                                        tion and answered with conviction, "Be­         forms us of something that we did not
            Whether one finds Greif's particular        cause they died." There are lots of ar­         know, changes our minds about how the
         approach congenial or not, it illustrates      guments in political science stated with        world works or reinforces beliefs that
         the three things that any social-scientific    equal seriousness and with similar tau­         are otherwise diffuse and/or unjustified.
         treatment of culture must do:                  tological qualities.                               This is the yardstick that I will apply in
            1.	 regard all actors as governed by           Consider the answers to the question         the remainder of my argument, so the
                 the same fundamental laws ofbe­        of why people vote. Several decades             reader should try to evaluate the crite­
                 havior (and not invent one social      ago, this was not a question in the minds       rion at this point. Does the work that you
                 science for Americans, another         of political scientists. At that time, em­      like provide significant added value, and
                 for Russians, yet a third for Ira­     pirical studies concluded that people with      the work that you dislike little or none?
                 nian fundamentalists, etc.);           higher income or more education were            When we hear fascinating presentations,
            2. understand culture not as a set          more likely to vote than people oflower         can we not readily summarize the main
                 of individual propensities but as      socioeconomic status, but all of these          points while with trivial work we leave
                 a coordinator of strategies and        studies took voting (as well as non-vot­        the room saying, "So what?" If this is
                 expectations among independent         ing) for granted. Only after the seminal        the case, then added value is the most
                 but mutually reliant actors, i.e. as   work ofOlson did political scientists (par­     important evaluative criterion, and we are
                 a social institution akin to lan­      ticularly those subscribing to the rational     on safe ground when we apply it to any
                 guage;and                              choice research program) ask the ques­          field of work, including the variants of
            3. offer a coherent explanation of          tion, "Why do people vote?" There is no         rational choice analysis that deal with po­
                 what sustains, and hence also of       shortage ofrational choice 'explanations'       litical culture.
                 what can change, culture even          of voting. I focus on one in particular,           The purpose of any 'analysis' includ­
                 among fully socialized adults.         which claims that people vote because           ing rational choice is not to say what
            Again, some of the earlier students of      they derive satisfaction from the act of        happens, but to explain why known
         culture tried to do exactly this. Gabriel      voting. There are several variants ofthis       events or empirical regularities happen.
         Almond's pioneering 1956 article on po­        'psychic income' approach: 'consump­            For example, we know that plurality elec­
         litical culture in the Journal ofPolitics,     tion value,' 'D term' and other versions.       toral systems are associated with two
         or David Laitin' s Gramscian attempt of        One thing is certain: having heard such         party systems (Riker has traced state­
         the mid-1980s to explain the absence of        explanations, the enquiring mind knows          ments of this association back some 150
         religious conflict among the Yoruba of         nothing new. While it may be true that          years). An analysis (known as
         Niger (Hegemony and Culture (Uni­              people vote because they like to, this does     'Duverger's law') explains that this as­
         versity ofChicago Press, 1986», sought         not constitute an explanation (rational         sociation is not accidental, but due to two
         truecovering-law explanations ofcultural       choice or any other kind) of voting. The        effects: the mechanical (that plurality
         variety,survival, and change. It is a thread   reason is that the added value introduced       electoral systems favor big parties) and
         that too much of mainstream culturalist        by the statement "people vote because           the psychological (that voters who un­
         theoryhas subsequently ignored and that        they like to" is nil, or very little.           derstand the mechanical effect will avoid
         now must be re-addressed. Currently, the                                                       'wasting' their votes on small parties).
         most promising way ofdoing so is as part       Added Value As An Evaluative Crite­             Duverger did not use rational choice ter­
         of a multiple equilibrium story among ra­      rion                                            minology, but the essence of his argu­
         tional actors.                                 I submit that the major criterion by which      ment is that voters perform expected
                                                        we should evaluate scientific work is not       utility calculations and don't vote for par­
                                                        whether it is theoretical, empirically valid,   ties which have a low probability ofwin­
                                                        consistent with what we know already            ning. Duverger's account has significant
                                                        or iconoclastic. These are important cri­       added value, because he persuasively
                                                        teria, but they are subordinate to whether      explains the mechanism underlyinga long

          APSA-CP Newsletter                                                   15	                                               Summer 1997


recognized but little understood associa­     basis of the government proposal or of        this way or 'reacted spontaneously' are
tion. He organizes our beliefs about the      the corresponding committee report, and       ad hoc. Even if true, these assertions
world such that we expect plurality elec­     so on. Institutional approaches to poli­      don't explain anything, and worse yet
toral systems to lead to two party sys­       tics (rational choice or not) focus on in­    they often seem invented just to save the
tems. This new prior is so strong that        stitutions as the independent variables       model. We do not account for behavior
when we find countries where the asso­        that explain human action.                    by identifying the categories of actors
ciation breaks down (like Canada or In­          Other constraints may be imposed by        (whether our categories are ethnic
dia), we need to explain why these vio­       the choices of other actors. For ex­          groups, genders, races or even ideolo­
lations ofDuverger's law occur.               ample, the government may not admit a         gies). Arguing that 'Italians' have 'sub­
                                              parliamentary amendment, or a witness         ject' culture, as Almond and Verba do,
Culture, Rational Choice and Added            may be treated as hostile (different rules    even if accurate, does not constitute an
Value                                         will apply to her) or you may be late for     explanation of their behavior, merely a
What is the role of culture in rational       an appointment because ofa traffic jam.       relabelling.
choice analysis? There are three signifi­        Finally, constraints may be imposed by        I am afraid that this is too often the
cantly different ways that culture ap­        some person's beliefs, ideology or cul­       pattern in scholarly articles, not only
pears in rational choice arguments. The       ture. I may be not be allowed to bear         those that belong to the rational choice
first (and most frequent) is to use cul­      arms by my religion. I may believe (as        tradition. For example, explanations of
ture as a constraint along the equilib­       Christians did long ago) that asking for      voting on the basis of 'party identifica­
rium path, the second is to use culture       interest on a loan is immoral because it      tion' have an unpleasant tautological ring
as information for equilibrium selection      is equivalent to charging for time, which     to them (people with Democratic identi­
(in both these cases, culture is used as      is a gift from God. Such restrictions (for    fication vote Democratic). Similarly, ge­
an independent variable) and the third is     the people who believe in them) are no        netic explanations (Germans behave dif­
to use culture as a dependent variable.       less real than those in the previous cat­     ferently than Italians because they have
I will argue that there is an hierarchy of    egories. I may not violate the law be­        always done so) certainly don't push the
added value among these three variants.       cause the police are present, or I may        limits ofimagination. So tautologies are
The first approach - even in the best         obey because I believe that compliance        not exclusive to rational choice analysis.
case - provides little added value, the       has an inherent moral or transcendental       What is particular to poorly wrought ra­
second adds significantly in our under­       value. My beliefs predict my behavior         tional choice articles is that they dress
standing of the world and the third best      as well as institutional constraints do; in   up tautological arguments with a ratio­
combines rational choice and culture.         some cases, they may provide more ac­         nal choice vocabulary. But familiar vo­
1. Culture as independent variable            curate predictions.                           cabulary does not mean that we under­
Rational choice analysis assumes that in­        While cultural accounts of human ac­       stand the phenomenon better - as the
dividuals are goal-oriented and try to        tion may be true, there is a significant      initial example ofvoting indicates.
maximize the achievement oftheir goals,       difference in their explanatory value. Ex­       I have tried to distance myself from
given existing constraints. The basic con­    planations by culture or ideology may be      this use of culture in rational choice
cept for rational choice analysis is 'equi­   trivial. Under what conditions will a cul­    analysis in Nested Games. This kind of
librium.' Equilibrium is a situation from     tural explanation be trivial, as opposed      analysis gives rational choice approaches
which no rational actor has an incentive      to non-obvious? The crucial difference        a bad name among scholars who study
to deviate. If a rational actor had an in­    is whether culture is used to define an       culture. They justifiably believe, after
centive to deviate, then she would not        actor's choice directly (as a constraint      reading a tautological' explanation,' that
select that option and we would not ob­       on an actor's behavior) or to define an       they have learned nothing new. Fortu­
serve that outcome as an equilibrium.         actor's response to the constraints of        nately, this is not the only intersection of
  From the above discussion, it is obvi­      other actors (in which case culture is        culture and rational choice.
ous that the selected actions depend on       used as information for equilibrium se­       b. Culture as an equilibrium selection
the existing constraints. What is the na­     lection).                                     mechanism
ture of these constraints? Some of them       a. Culture as a constraint along the          Andre Malraux was General de Gaulle's
may be imposed by existing institutions.      equilibrium path                              Minister of Culture. He contributed not
For example, in parliamentary systems         Suppose that a model predicts some            only to the content of the General's
most of the bills considered by the par­      (maximizing) set of actions, but that         speeches but also to the selection of the
liament are introduced by the govern­         some actors do not follow the prescribed      time and place that the General deliv­
ment. Similar rules may define what kind      behavior. The cultural explanations that      ered his speeches. He made his selec­
of amendments (if any) are permitted,         they did not think of it, did not have the    tions to maximize the cultural impact of
who is recognized from the floor,             cognitive capacities, were prohibited by      each speech. Today, media consultants
whether discussion will be made on the        their ideology or culture from acting in      advise candidates to package themselves

  APSA-CP Newsletter                                                16                                               Summer 1997
    in order to have maximum impact, given          information about the culture of the           duces added value is that the selection
    the preferences, biases, stereotypes,           Council affect the proposal ofP+C? Yes,        of strategy is not intuitively obvious. The
    beliefs,ideologies and cultures ofthe pub­      because unless they make a proposal            actor did not select this course ofaction
    lic. The scholarly works of Tarrow and          that makes every member of a qualified         because of some constraint on his own
    Popkin describe how revolutionary lead­         majority better off than any possible          beliefs or capacities, but because of the
    ers study and incorporate the culture of        unanimous decision, some potential mem­        information he possessed about the other
    their followers into their strategies. In all   ber of the qualified majority will success­    actors' cultures and, therefore, their
    of these examples, beliefs about the cul­       fully make that unanimously favored pro­       likely courses of action.
    ture of other people affect the optimal         posal.                                         2. Culture as a dependent variable
    strategies of some actors. Culture helps           Follow my reasoning in the figure. The      An even more interesting way of look­
    answer the question of why particular           Council can unanimously approve any­           ing at culture is as a dependent variable.
    actors select particular courses of ac­         thing in the area [SQ, SQ'] where SQ'          In general, the assumption ofrationality

1   tion as optimal.
       I emphasize that cultural reasons may
    explain not only the selection of strat­
    egies, but also their avoidance. One of
                                                    is symmetric to the status quo with re­
                                                    spect to the ideal point of the voter piv­
                                                    otal for unanimity (1). The new P+C pro­
                                                    posal Y makes the qualified majority piv­
                                                                                                   and the use of game theory does not re­
                                                                                                   strict the number ofequilibria very much.
                                                                                                   Indeed, under conditions of incomplete
                                                                                                   information (the norm in politics) or re­
    the procedures used by the European             otal member of the Council (3) just shy        peated play (also quite frequent), equi­
    Union for legislative decisions is the co­      ofindifferent between it and SQ' (which        libria are infinite and the real question is
    operation procedure. The important fea­         he can get by unanimity). So the restric­      how to select among them. For example,
    ture of this procedure is (I simplify here      tion ofthe off-equilibrium beliefs ofP+C       while the outcome of a confrontation
    for the sake of the argument) that a pro­       leads us to the selection of equilibrium       with complete information may never be
    posal emanating from the Commission             Y rather than X.                               a war (a point raised as an argument
    and the Parliament can be accepted by              Note that had the empirical literature      against rational choice analyses by un­
    theCouncilwith a qualifiedmajority,while        on the Council found that one country          informed critics), war becomes a pos­
    it can be modified only with unanimity.         was always in the minority and that the        sible equilibrium with incomplete infor­
       Consider the figure below. The status        others never tried to incorporate it into      mation.
    quo (SQ) is outside the area defined by         their bargains (same institutional rules but      One way of understanding' cultures'
    the ideal points of the members of the          different culture), X would emerge again       is as such manifold equilibria. In this
    Council (the numbers), while the Parlia­        as the predicted equilibrium.                  conceptualization, different equilibria
    ment (EP) and the Commission (C) that              The reason that this use ofculture pro­     come from different antecedent condi-
    make the proposal fall on the other side
    of the Council. The required qualified
    majority in the Council is 5 o£1.                          Figure. Equilibrium Selection on the
       What proposal will the Parliament and                   Basis of Different Off-Equilibrium Beliefs
    the Commission (hereafter denoted P+C)
    make, knowing that the Council cannot                           1234567
    modify their proposal except by unanim­
    ity? P+C will make the proposal X that
    makes the pivotal member of the Coun­
    cil (3) almost indifferent between X and
    the status quo. Indeed, in this case 3 will
    prefer X to the status quo, and so will all             SQ          SQ'          Y             x            C EP
    the members to his right (4,5,6 and 7).
1   Note that in this example we did not use                   SQl=lSQ', SQ'3>3Y, SQ3>3X
I   the power of the Council to modify the
       Suppose now that we learn from the                      Restricting off-equilibrium beliefs to a
    empirical literature on the European                       consensus council leads to the selection
    Union that the Council is a consensus­
    oriented body which tries to reach deci­
                                                               of Y as the equilibrium. Restricting off­
    sions by unanimity whenever possible,                      equilibrium beliefs to a divided council
    and that most of the time there are no                     leads to the selection of X as the
    formal votes. (This information is readily
    available in the EU literature.) Does this
                                                               equilibrium. See the text for details.
      APSA-CP Newsletter                                                  17                                                Summer 1997 I
tions. If one wants to explain why cer­         rencies) equally.                              ever, is not just political economy, nor are
tain rural cultures practice female infan­        This is an example where 'culture' is        all forms ofpolitics highly institutional­
ticide, she may assert that parents con­        the equilibrium corresponding to a series      ized - especially in the developing world.
sider their children to be assets or liabili­   of exogenous conditions (sequence of              Consider how different ethnic conflict
ties; ifphysical strength leads to survival,    moves, impatience). Rubinstein selects         is from legislative or bureaucratic battles
parents will keep male babies.                  the unique perfect equilibrium from the        over economic policy. Ethnic conflicts are
   Another example will make my point           infinitepossible equilibria,and this is what   a form ofmass politics marked by highly
more clearly. Suppose that two people           gives power to his result. It may, how­        risky or costly forms ofbehavior in which
are to divide a dollar. Any division ofthe      ever, be the case that the set of perfect      ethnic partisans not only kill but are will­
dollar that leaves no residual is an equi­      equilibria is infinite, in which case ana­     ing to die. Just as it is hard to explain ­
librium. If the amount to be divided is         lysts will look for some additional refine­    given rational calculations of cost and
significant, disputes among individuals         ment that further restricts the predicted      benefit, why people vote - it is also hard
can last forever.                               outcome.                                       to understand - with tools of rational
   Up to the 1970s, bargaining was a                                                           choice - why so many people in the
branch of cooperative game theory in            Conclusions                                    world demonstrate ethnic fervor or em­
which criteria of 'fairness,' symmetry          Cultural studies produce a wealth of in­       brace nationalism. From an individualper­
and mathematical elegance produced              formation about how different people ­         spective, the instrumental benefits of par­
different solutions. Ariel Rubinstein bril­     from Africa to Capitol Hill - think and        ticipating in nationalist mobilization are
liantlyproduced a game which simulated          behave. If these reports yield beliefs,        obvious only under two strict conditions:
real bargaining: Player One makes an            behaviors and rituals that we did not pre­     (a) when nationalists are already close
offer for a division ofthe dollar to Player     viously recognize, then they produce           to capturing power and much can be
Two. If Player Two accepts, the game            added value. Their existence does and          gained, or anticipated losses cut, by join­
ends; if not, he makes a counter-offer to       should alter the way we analyze these          ing the bandwagon; or (b) when law and
player one. If Player One accepts the           societies. Rational choice does not have       order have broken down, ethnic animosi­
counter-offer, the game ends; ifnot, the        anything to offer to such studies, but         ties have soured group relations, and
game goes continues until the two play­         much to learn from them. Repetition of         even neighbors oflongstanding belong­
ers agree. To bring the game to an end,         these studies with a rational choice vo­       ing to a different ethnic group can't be
Rubinsteinendowed his players with 'im­         cabulary helps neither tradition. Ratio­       trusted, creating a "security dilemma" for
patience' - that is, a preference for the       nal choice contributes by incorporating        individuals (Posen, 1993) and making
game to end sooner rather than later. He        these cultural findings into the rational      preemptive violence against neighbors of
thereby calculated a unique perfect equi­       calculations of actors. Evert better, it       a different ethnic group an exercise in
librium as a function of who makes the          enables researchers to understand the          personal security (Hardin, 1995).
first offer and the levels of impatience        reasons why particular cultural patterns          These extreme conditions constitute a
of the players. If we call the level of         emerged as equilibria from the wide va­        rather small proportion of the universe
impatience (the time discount factor) of        riety ofpossible behaviors.                    of ethnic conflict. The former Yugosla­
each player d, the final division of the                                                       via, Rwanda and Burundi are not typi­
dollar gives the first player x=( I-d)/(l­                                                     cal; they are simply the most dramatic
                                                Cultures and Modes                             and gruesome cases of ethnic conflict.
   What is interesting in this approach is      of Rationality                                 Violence may be common in ethnic con­
that ifboth players are infinitely patient      Ashutosh Varshney                              flicts, but a complete disintegration of the
(d, and d2 tend to 1), the final outcome        Harvard University                             state is not. The latter breakdown has
is x=l/2. So the familiar Western habit         varshney@cfia.harvard.edu                      seldom marked ethnic conflicts in Asia,
of splitting the difference evenly can be                                                      Europe and North America. This does
derived as the equilibrium outcome of a         Rational choice theory has made re­            not mean that there are no risks or costs
game if both players are infinitely pa­         markable contributions to two subfields        associated with participation in ethnic
tient. The same outcome results if the          ofcomparative politics. It has deepened        mobilizations in societies where the state
players are not infinitely patient, but         our understanding of political economy         has not collapsed. Risks of incarcera­
equallypatient and equally likely to move       issues - especially the politics of eco­       tion, injury and death remain, but in the
first. If in a different society men made       nomic growth and distribution. And its         absence of state disintegration, ethnic
the first move, the split ofthe dollar would    ability to explain behavior in highly insti­   conflicts don't produce security dilem­
not be symmetric but would favor men.           tutionalized settings - as in the rule-gov­    mas. By and large, the situation from an
I suppose (although I do not know it for        erned universe of a Western bureau­            individual perspective can be summa­
a fact) that in some cultures men and           cracy, legislature and executive - has         rized as follows: the benefits of partie i­
women do not split dollars (or other cur­       been strikingly impressive. Politics, how­     pation - a better job, a political office ­

:APSA-CP Newsletter                                                   18                                                Summer 19971