1. No category

Multi-Dimensional Threat Perception and State Repression: An

Multi-Dimensional Threat Perception and State Repression: An Inquiry into Why States Apply
Negative Sanctions
Author(s): Christian Davenport
Source: American Journal of Political Science, Vol. 39, No. 3 (Aug., 1995), pp. 683-713
Published by: Midwest Political Science Association
Stable URL: http://www.jstor.org/stable/2111650 .
Accessed: 30/03/2013 23:09
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact [email protected].
.
Midwest Political Science Association is collaborating with JSTOR to digitize, preserve and extend access to
American Journal of Political Science.
http://www.jstor.org
This content downloaded from 141.213.172.160 on Sat, 30 Mar 2013 23:09:32 PM
All use subject to JSTOR Terms and Conditions
Multi-DimensionalThreatPerceptionand
State Repression:An InquiryInto Why
StatesApplyNegativeSanctions*
ChristianDavenport,University
of Houston
Theory:
Regimesrespond
todomestic
withpolitical
threats
Theprecise
repression.
natureofthedomestic
threat
itself,
is subjectto discussion.
however,
Hypothesis:
Staterepression
is a function
of eithera unidimensional
conception
ofdomestic
threats
(i.e.,wherethereis oneattribute
ofpolitical
conflict
considered
bytheregime)or one thatis multidimensional
in character
(i.e.,wherethereare
severalattributes
conditioned
considered),
bycertain
political-economic
characteristics:democracy,
economicdevelopment,
coercivecapacity,dependency
and
laggedrepression.
Methods:A pooledcross-sectional
timeseriesanalysisof53 countries
from1948
to 1982.
Results:Threedifferent
aspectsof politicalconflict
(conflict
frequency,
strategic
variety,
anddeviancefromcultural
norm)arestatistically
in theirrelasignificant
to repression,
themultidimensional
tionship
supporting
of domestic
conception
threats.
thedegreeto whichthegovernment
Additionally,
is democratic
signifiof relationships
cantlyaltersthepattern
betweenpoliticalconflict
andrepressive
behavior.
Introduction
Implicitwithinmoststudiesthatassess theimpactofconflicton repression' is the assumptionthatgovernments
generallyrespondto only one
*Alldataemployed
within
thisanalysismaybe obtained
from
ICPSR.Documentation
necessarytoreplicate
theanalysescan be obtained
All statistical
fromtheauthor.
analyseswere
conducted
withSAS, withsomeexploratory
within
analysesbeingconducted
STATA version3.1.
'To avoidredundancy
I willuse "conflict"interchangeably
with"dissent,""dissident
behavior"and "politicalopposition."
Repression
is defined
as government
regulatory
actiondirected
againstthosechallengingexisting
powerrelationships.
Thisis similar
toGoldstein's
definition
(1978,1983)where,
"politicalrepression
consists
ofgovernment
actionwhichgrossly
discriminates
againstpersonsororganizations
viewedas presenting
a fundamental
challenge
toexisting
powerrelaor keygovernment
tionships
policies,becauseof theirperceived
politicalbeliefs"(1978,
xvi).
To decreaseredundancy
within
thetext"repression"
willbe usedinterchangeably
with
"repressive
behavior"and "negativesanctions."
AmericanJournalof PoliticalScience,Vol. 39, No. 3, August1995,Pp. 683-713
? 1995bytheBoardof Regentsof theUniversity
of Wisconsin
System
This content downloaded from 141.213.172.160 on Sat, 30 Mar 2013 23:09:32 PM
All use subject to JSTOR Terms and Conditions
684
ChristianDavenport
of theevents.2
thefrequency
behavior-particularly,
aspectof dissident
as the numberof conflicteventsdirected
Accordingto thisliterature,
efforts
oftheregimetoconso do therepressive
againstthestateincrease,
is believedto be
The exactnatureof therelationship
trolthisbehavior.3
bothdirectandlinear.4
andresponseis
perception
threat
of unidimensional
Thisassumption
Ifaccepted,
itleads
repression.
ofpolitical
tothestudy
important
extremely
arenotonlymyopicinfocus,payingattention
us toconcludethatregimes
butalso
behavior,
to onlyone aspectof dissentin its use of repressive
diversemanifestorelatively
withregard
tohowtheywillrespond
uniform
would
politicalconflict
Within
thisperspective,
ofdomestic
unrest.
tations
in theTiananmen
Squareincito similarly
(withrepression)
be responded
ofMay 1970in
dentofJune1989inChinaandintheKentStateincident
in termsof
wereperhapscomparable
theUnitedStates.Thesesituations
notwithregard
butdefinitely
eventsthatoccurred
thenumber
ofconflict
the
thatwereappliedby dissidents,
strategies
of different
to thenumber
relative
ofconflictual
activity
selectedforprotest
orthemagnitude
targets
thethreat
posed
to thesecriteria,
Withreference
to previousexperience.
in eachcase.
to theregimewas different
of
andextantcases,theassumption
In lightoftheexamplesprovided
diffithreat
andresponseshouldbe somewhat
unidimensional
perception
level.Considering
availableworkon
cultto acceptevenon an intuitive
madewithin
severalarguments
also counters
thissubject,theassumption
(Stohl
thisissue,numerous
authors
itself.
concerning
Directly
theliterature
1986;
andLopez 1984;Eberwein1987;Gurr1986a,1986b;Ziegenhagen
rethatrepressive
Franks1989;HooverandKowalewski1992)maintain
to twofactors:1) different
according
sponsesto dissentvarysignificantly
ofthe
and2) thestructure
oftheconflict
behavior
encountered,
attributes
andso forth).
type,economicdevelopment
politicaleconomy(i.e.,system
but
phenomena
arenotunidimensional
domestic
threats
To theseauthors,
acrossconto differently
defined,
andresponded
multidimensional,
rather
texts.
considered
areactually
byregimes
ofpoliticalconflict
Whatattributes
do thedifferent
threats?
Whatrelationship
ofdomestic
intheir
classification
Somehaverepretheseanalyses.
within
is everemployed
ofconflict
2Onlyonemeasure
countsdetermine
as well, whereoverallfrequency
sentedthisvariabledichotomously
or not(Cingranelli
regulation)
ofbehavioral
(worthy
is significant
or notconflict
whether
1992;Poe andTate 1992).
issue
structural
as a long-term
thisrelationship
thatI amaddressing
3I wouldhighlight
tothatdiscussedbyLichbachandGurr(1981).
onesimilar
logistical
andnotas a short-term
existas well(Duvall
relationships
thatnon-linear
whosuggest
4There
aresomeauthors
et al. 1978).
andShamir1980;Jackson
This content downloaded from 141.213.172.160 on Sat, 30 Mar 2013 23:09:32 PM
All use subject to JSTOR Terms and Conditions
WHY STATES APPLY NEGATIVE SANCTIONS
685
aspectsofpoliticalconflict
havewithrepression?
Whatpolitical-economic
characteristics
affect
theidentification
ofthreatening
andthegovbehavior
ernment'
s repressive
responses
toit?Using53 countries
forthetimeperiod
1948 to 1982,thepresentstudyaddressesthesequestions.The analysis
itselfis madeup of severalcomponents.
To begin,I identify
fourattributes
ofpoliticalconflict
possiblyconsidered by regimes: 1) basic frequencycountsof events,2) thepresence of
violence,3) thevarietyof strategiesemployedbydissidentsand 4) current
behavior relativeto the culturalnorm.5This approachencompassesthe
mostoftenexamineddimension
of conflict
(frequency)
as well as three
additional
attributes.
Thesecondcomponent
ofthestudyaddresses
thedeterminants
of "repressive
propensity,"
i.e., thedegreeto whicha given
regimeclassifiesbehavioras threatening
and respondswithrepression.
Drawnfrom
theliterature,
thesecontextual
factors
includesystem
type(i.e.,
coercivecapacity,economicdevelopment
democracy),
and dependency.
Thethird
component
ofthestudy
is anempirical
witha pooled
investigation
cross-sectional
time-series
design.Here,in thecontextof a multivariate
model,I analyzetworelationships:
1) theimpactdifferent
aspectsofpolitical conflict
haveonrepressive
and2) thecontribution
ofvarious
behavior,
statecharacteristics
to theapplication
ofrepression.
The conclusion
summarizestherelevant
of bothof theseanalysesand identifies
findings
the
practicalimplications
thatfollowfromthem.Specifically,
I addressthe
importance
ofthestudy
forfuture
ofrepression
investigations
andalsohow
theseresultsimprove
ourunderstanding
ofdomestic
politicalprocessesin
general.
PoliticalConflictand DomesticThreatPerception
Mass politicalbehaviordirected
againstthestate,itspoliciesand its
practicesis generally
viewedas beinga threat
to thosein authority.
Its
candisrupt
manifestation
anditcanundermine
theposition
society,
ofthose
within
power.In an effort
toregulate
thisthreat
andreducethepossibility
thattheseoutcomesmight
takeplace,governments
oftenuse politicalrepression:e.g.,censorship,
politicalrestrictions.
Thisbehavioris expected
to neutralize
and/or
increasethecostsofthebehavior
politicalopponents
to sucha largeextent
thatitis no longerdeemeda worthwhile
of
strategy
protest.
Therehas beena greatdeal of theoretical
discussionaboutthis
causallinkage(Gurr1986a,1986b,1989;Tilly1978;Oommen1990;HooverandKowalewski1992;Eberwein1987;Lopez 1986)as wellas empirical support
(MarkusandNesvold1972;Hibbs1973;Goldstein
1978;1983;
5Deathsbycollective
arenotusedherebecauseofthepossiblecontamination
protest
of politicalrepression.
This content downloaded from 141.213.172.160 on Sat, 30 Mar 2013 23:09:32 PM
All use subject to JSTOR Terms and Conditions
686
ChristianDavenport
andAllen
1991;Alfatooni
Muller1985;Davis andWard1990;Davenport
199 1).6
theanalis widelyacceptedintheliterature,
thisrelationship
Although
interms
ofscope.Generlimited
onithavebeensomewhat
ysesconducted
events
ofconflict
has beenon thenumber
ally,thefocusofpastresearch
threat
and
analysesofdomestic
empirical
thattakeplace,and,as a result,
(alternaresponseare usuallybasedon measuresof frequency
repressive
however,
thispractice,
Carefully
considering
toas intensity).
tivelyreferred
ofpoorconceptumaybe morea reflection
oneseestheuseofthismeasure
theseriousness
Regimesmayidentify
alizationthanplausiblealternatives.
ofconflict
withthefrequency
inaccordance
threats
notsimply
ofdomestic
aspectsofconflict
in termsofseveraldifferent
butrather
eventsidentified
1972;Zie(Sorokin1957;ChadwickandFirestone
viewedsimultaneously
of these
combinations
genhagen1986;Lichbach1987).Whenparticular
themagnitude
of threat
perceivedand thelikeliare apparent,
attributes
when
Correspondingly,
willbe appliedis increased.
hoodthatrepression
ofthreat
themagnitude
is apparent,
oftheseattributes
combination
another
behavioris less
perceivedis deemedminimaland theuse of repressive
likely.
I addressthreeaspects
alternative,
thismultidimensional
Considering
in additionto frequency:
1) thepresenceof violence,2) the
of conflict
fromculturally
and 3) deviations
usedby dissidents
variety
of strategies
is important
tothestudy
Eachofthesemeasures
acceptedlevelsofdissent.7
of what
becausetheyenhanceour understanding
of politicalrepression
ofthesefactors
Consideration
a threat
toa regime.
constitute
might
actually
is applied,forourattention
lightuponwhyrepression
shedadditional
might
as
ofactsthatchallengethosein authority
thenumber
towards
is directed
withintheseacts,the
manifested
well as theoveralllevelof aggression
and
thatareappliedbypoliticalopponents,
strategies
number
ofdifferent
involved.
mobilization
thediverseprocessesofdissident
(For example,
increasespoliticalconflict
6Thereare thosewho findthatrepression
chalpolitical
toregulate
is notintended
1986).Thisdoesnotmeanrepression
Ziegenhagen
it simplymeansthatit is notalwayseffective.
lenges.On thecontrary,
that
behaviorthatis simply"out of vogue,"i.e.,threats
'I do notidentify
dissident
thisis
notadvocatedby theregime.Primarily
butare ideologically
do notpose a threat,
I maintain
data.On thecontrary,
thiswiththeexisting
inidentifying
becauseofthedifficulty
logistical
to numerous
pay attention
thatgovernments
assumption
themoreconstraining
ideologies
behavior.
Thisacceptsthatevenifparticular
identified
withantisystemic
factors
itis stillpossibleforthosechallenging
totheregime,
areacceptable
bydissidents
advocated
These"pobehavior.
forantisystemic
tolerance
thegovernments
tooverstep
thegovernment
correct"groupsmaysimplyhavea widerberththanone thatespousesideologies
litically
stillremains.
butthepointon stateresponsiveness
thatarenotacceptable,
This content downloaded from 141.213.172.160 on Sat, 30 Mar 2013 23:09:32 PM
All use subject to JSTOR Terms and Conditions
WHY STATES APPLY NEGATIVESANCTIONS
687
Identifying
DifferentAspects of ConflictBehavior
attribute
I consideris thepresenceofviolence.This
The first
conflict
haveidentified
substantive
becausemanyauthors
characteristic
is important
conflict
behaviorwithregard
differences
betweenviolentand nonviolent
on negativesanctions
to itseffect
(Hibbs1973;DuvallandShamir1980;
oftheseanalysesfindthatviolent
theresults
Guptaetal. 1993).Generally,
ofconflict
response
from
theregime.
provoke
a greater
repressive
strategies
Two reasonsareconsistently
offered
forthis.
of
amount
ofdissentusuallyentailthegreatest
First,violentstrategies
dissidents
mightkill
Usingthesestrategies,
threat
to thepoliticalsystem.
ofsocialchaosorprovoke
members
ofthecitizenry,
createa situation
those
in theantisystemic
behaviorto move
who are notalreadyparticipating
fromnonbecauseitcan no longerprovidesafety
againstthegovernment
todeterthesepossible
stateactors(MasonandKrane1989).In an attempt
theregimewouldbe expectedto increasetheuse ofrepression.
outcomes,
is morelikelywhenconbehavior
The secondreasonwhyrepressive
When
ofusingrepression.
theperceived
legitimacy
flictis violentconcerns
andthreatening
to othercitidissidents
areportrayed
as beingdangerous
to attempt
behavioral
regulaposition
zens' lives,theregimeis in a better
behavior
tion.Thecitizenry
wouldbe morelikelytosee thegovernment's
Thiswouldtend
as legitimate,
andtheymayevencall forit themselves.
of itsapplication
to the
to increasethelikelihood
becauseless resistance
policywouldbe apparent.
particular
government
is much
Whenconflict
is notviolent,
on theotherhand,thesituation
In thiscontext,
different.
posedbythose
perhapsunableto see thethreat
citizens
being
theregime,
negatively
torepression
mayrespond
challenging
behavior:
or
protest
i.e.,theymayjointheprotestors
appliedto nonviolent
lose faithin thepoliticalsystem.8
Thisresistance
wouldtendto decrease
ofthe
thelikelihood
of negativesanctions
becausetheoveralllegitimacy
we wouldexpect
comeintoquestion.
As a consequence,
government
might
theuse ofrepression
tobe muchlowerwhenconflict
was nonviolent
(ifit
wouldbe appliedat all).
the secondconflictattribute,
strategic
variety,
regimes
Concerning
ofdifferent
shouldpayattention
tothenumber
strategies
employed
bythose
it.9As opposedto simplypayingattention
to thenumber
of
challenging
eventsoccuror thepresenceof violencewithinconflict
timesparticular
thefocusoftheregimeis directed
towards
thenumerous
means
behavior,
theirobjectives(i.e.,
evencounton thisresponseto further
8Somesocialmovements
thecivilrights
movement).
wherecertain
types
schemeemployed
bygovernments
9Theremaybe someweighting
is notexamined
here.
thanothers.
Thispossibility
ofeventsaregivenmoreattention
This content downloaded from 141.213.172.160 on Sat, 30 Mar 2013 23:09:32 PM
All use subject to JSTOR Terms and Conditions
688
ChristianDavenport
of dissentusedbytheopposition.
Severalauthors,
underdifferent
labels,
discussthisgeneralconcern:Tilly's(1978) "repertoire"
of dissident
behavior,Eckstein's(1980) "alternative
channels"ofpoliticalparticipation,
Ziegenhagen's(1986) "dominant"and "general" variety,
Lichbach's
(1987) "tacticalchoices"andMorris'(1993) "multipletactics."
As conceived,therelationship
of strategic
varietyto repression
is
If thenumber
of strategies
straightforward.
brought
againsttheregimeis
low,thenthemagnitude
ofthethreat
is minimal.
Ifonlyoneor
perceived
a fewstrategies
ofconflict
arebeingusedbydissidents,
theregimewould
be abletobringall ofitscoerciveexpertise
tobearon thislimited
activity.
Thisdiminishes
theoverallamount
ofrepressive
behavior
necessary
toprotectthestatusquo anddecreasestheamountof repression
thatwouldbe
applied.
If thenumber
of strategies
brought
againsttheregimeis high,on the
otherhand,thenthemagnitude
ofperceived
threat
is increased.'?
Here,the
is confronted
witha diversity
government
of antisystemic
each
behavior,
withitsownimpactuponthedomesticpolitical-economy.
For example,
generalstrikes
whileguerrilla
warfare
mayparalyzeindustrial
production
mayterrorize
thepopulace.Undertheseconditions,
theprocessofbehavioralregulation
is "stretched
thin"becausedifferent
ofdissent,
strategies
different
anddifferent
outcomes
mustbe addressed
targets,
simultaneously.
In an effort
to counterthisincreasedvariety,
should
politicalrepression
increaseas well."Analyzedwithin
twoempirical
studies(Tilly1978;Ziehavegenerally
beensupported
genhagen
1986),thesecausallinkages
quite
well.
Thefinalattribute
ofpoliticalconflict
hereconcerns
addressed
cultural
limits
ofdissident
behavior.
"Cultural
limits"refer
totheparticular
amount
ofpoliticalconflict
thattheregimewillallowtotakeplacebeforeitapplies
Theparameters
repressive
behavior.12
ofthiscodecouldbe communicated
means:i.e.,explicit
enumeration
within
theconstitution
bynumerous
(Volofpolitical/civil
libergyes1978;Franks1989,6 and147),massperception
thisrelationship
is conceivedofas beinglinearin formalthough
'"Generally
different
responses(i.e., nonlinear
relationships)
havebeenidentified
(Jackson
et al. 1978;Duvall
andShamir1980).
in repression
"Frequency
is thusequatedwithregulatory
effort
(Duvalland Shamir
1980;Ziegenhagen
1986;Eberwein1987;HooverandKowalewski1992).
'2Normal
politicalbehaviorand dissent(or antisystemic
behavior)is thusnotdeterminedbysomeobjectivecriteria
(as suggested
byDogenharat
andDay [1983]),butrather
ofthosechallenging
andthoseprotecting
byaninterplay
thestatus
quo.Forusefuldiscussion
of theprocessby whichdifferent
components
of thepoliticaleconomyestablishnormal/
anddeviant/unacceptable
see Wilkens(1964) fortheimpactofculture
acceptable
behavior,
andFranks(1989) fortheimpactof thelegalsystem,
regimetypeandideology.
This content downloaded from 141.213.172.160 on Sat, 30 Mar 2013 23:09:32 PM
All use subject to JSTOR Terms and Conditions
WHY STATES APPLY NEGATIVESANCTIONS
689
ties(Gibbs1981)orpastexperiences
withdissentandrepressive
response
(Dedrick1978;Liska1992).Regardless
ofhowthecodeis communicated,
however,as longas conflict
stayswithinthisrealmof acceptability,
the
withrepression.
doesnotrespond
government
Indeed,giventheacceptance
theregimedoes notevenperceivethatit is being
of manifested
conflict,
threatened.
If,however,
thecultural
limitis violated,i.e., an event(or groupof
normofbehavioral
events)is abovethecultural
thenthesituation
deviance,
In thiscontext,
changesmarkedly."3
manifest
conflict
sendsa messageto
theregime
thatitis directly
beingchallenged
andpolitical
repression
would
likelybe increased
as a meansofreestablishing
theculturally
defined
paof acceptablebehavior.'4
rameters
Addressedimplicitly
withinnumerousstudies(Lichbachand Gurr
1981; Muller1985;Blalock1989)andexplicitly
within
one(Franks1989),
therelationships
described
abovehavenotbeeninvestigated
systematically
in anyempirical
I haveno statistical
sense.Consequently,
support
forthe
relationship
identified
above. The literature
has providedenoughinformationforat leasta tentative
of thisrelationanalysisto be conducted
ship.
I proposethatdomestic
thepreviousdiscussion,
Summarizing
threats
arecomposedoffourelements:
the
ofviolence,stratefrequency, presence
andcultural
deviance.The greatest
gic variety
amountofthreat
to thereis expectedwhenconflict
gimeandthemostlikelyuse ofrepression
freand
are
is
violent
behavior
and
conflict
rates
are
above
quency variety high,
norm.Withinthissituation,
thecultural
thegreatest
measureofdomestic
threat
is notedbecauseeach variableis approaching
itshighestvalueon
theattribute
scale. Whenconflict
and varietyare low, on the
frequency
otherhand,behavior
is nonviolent
is belowtheculturandexisting
conflict
ofacceptable
then
the
amount
ofdomestic
allydefined
boundary
deviance,
threat
is low.Domesticthreats
andtheuseofrepression
perceived
expected
areminimal
becauseeachvariableis approaching
itslowestvalueon the
attribute
scale.
'3Foran interesting
discussion
see Pfuhl(1986).
'4Injustifying
thisgovernment
response,
theremarks
ofoneChileanleaderareparticucertain
larlyinformative:
"Socialenvironment
anda correct
judicialorderrequire
restrictions
on individual
notonlytopreserve
thepersonal
freedom
ofothers,
butforthecomliberties,
of thesamehierarchy.
mongood . . . Norareall rights
Evenamongnatural
rights,
some
aremorefundamentally
thanothers.
important
Theymayusuallyall be exercisedsimultabutthisis impossible
whensocietybecomessick.The lattersituation
neously,
is precisely
a symptom
of politicalabnormality
an exceptional
requiring
juridicalregimein whichthe
in orderto ensurethefreeexercise
exerciseof somerights
is limited
or can be suspended
of otherimportant
ones" (Lopez 1986,84).
This content downloaded from 141.213.172.160 on Sat, 30 Mar 2013 23:09:32 PM
All use subject to JSTOR Terms and Conditions
690
Christian
Davenport
ExplainingVariance in DomesticThreatPerception
and the Use of State Repression
thecausallinkageidentified
Understanding
above,one shouldnotassumethattheapplication
of repressive
behaviorin responseto domestic
threats
wouldbe comparable
acrossall regimes.As Duvall and Shamir
(1980, 160) suggest,
"therepressive
character
ofthegovernment
(i.e.,the
likelihood
thatrepression
wouldbe applied)is greater
accordingly
as the
government
perceivesmoresituations
as threats,
or whenitviewsthreats
as moreseriously
andhenceis disposedtorespond
threatening,
morecoerto vary,however,
cively.9"15Whatcauses theperceptionof domesticthreats
predisposing
theregimetowardstheuse ofpoliticalrepression?16
FundaI wouldarguethedifference
mentally,
is determined
byfourstatecharacteristics:systemtype,coercivecapacity,
economicdevelopment
anddependency.Drawnfrom
theliterature,
eachofthesefactors
is important
because
theidentification
theyfacilitate
ofdifferent
contextual
elements
thatimpact
thegovernment
perceptions
of domesticthreats
as well as thelikelihood
thatrepressive
behaviorwouldbe applied.By examining
eachofthesein
turnanditsrelationship
to repressive
thesignificance
propensity,
ofthese
variouscharacteristics
shouldbecomeclearer.
Political-Economic Contexts,ThreatPerception
and Political Repression
Whenthepresenceof democracy
increaseswithina nation-state,
the
likelihoodof threats
beingperceivedby thegovernment
and repression
being appliedis decreased.Supportedby numerousempiricalstudies
(Hibbs1973;Ziegenhagen
1986;Henderson
is usu1991),thisrelationship
to thefactthatdemocracies
are generally
allyattributed
morelegitimate
forms
ofgovernment
as wellas moretolerant
ofdissident
behavior.
These
factors
makethegovernment
less likelyto perceivethreats
fromconflict
(however
conceived)becausetheyenjoyrelatively
highlevelsof "diffuse
support"(Easton1975).17Additionally,
theywouldbe less likelyto use
becausetheregimewouldbe prepared
repression
to allowcertainlevels
'5Thisargument
aboutperception
andrepression
is sharedin manydifferent
contexts:
Buzan(1983) in boththeinternational
and domesticrealms,Hermanand Herman(1989)
in theinternational
realmandOommen(1990) in thedomestic
realm.
'6AsLopez (1986, 75) suggests,"[therepressive]
policyof government
unfoldsin
circumstances
farmorecomplexthan[simply]
torule. . . Rather,
suchan action
challenges
wouldappearto be a 'fulldecision,'inthatitoccursas a choicewithin
a particular
setting
andone weighted
in lightof otheravailablechoices."
'7Thiswouldalso tendtodecreasethelikelihood
thattheregimewouldbe challenged
in anysubstantive
fashion.
This content downloaded from 141.213.172.160 on Sat, 30 Mar 2013 23:09:32 PM
All use subject to JSTOR Terms and Conditions
WHY STATES APPLY NEGATIVESANCTIONS
691
ofdomestic
unrest
totakeplace,thusrelaxing
thenatural
tension
thatexists
whenmassprotest
behaviordoes occur.18
The secondfactorconsidered,
coercivecapacity,
is also important
to
theregime'sperception
ofdomestic
threats
andtheiruse ofrepressive
behavior.Two explanations
inparticular
havebeensupported
within
numerouscase studies(DallinandBreslauer
1970;WalkerandLang1988;Zwick
1984)andempirical
analyses(Hibbs,1973;Ziegenhagen
1986;Davenport
1991).Thefirst
addresses
a morestrategic
issuewhilethesecondaddresses
an organizational
matter.
Concerning
thefirst
explanation,
byproviding
theregimewiththeresourcesforrepression,
thecostsofitsimplementation
weredecreased
(Laswell, 1941;Randle1981; Goldstein1983;Gurr1986a;Davis and Ward
1990;WalkerandLang 1988).'9Thisreducedcostmaderepressive
strategies moreattractive
to thosein authority
becausemoreareasand more
individuals
couldbe subjectto theregulatory
efforts
of theregime.This
in turnincreasedthelikelihood
thatrepression
wouldbe appliedagainst
politicaldissidents
becausetheapplication
ofrepression
itselfwas deemed
morefeasible.
Withreference
tothesecondexplanation,
oncerepression
was applied
andcoercivestructures
wereputinplace,someformofbureaucratic
"law
of instrument"
wouldgo intoeffect(Gurr1986a,1986b,1989;Seligson
1987;Mitchell
andMcCormick
1988).Inthemidstofthislaw,thecoercive
seekstojustify
itscontinued
apparatus
existence
andincreaseitsaccessto
resources
(socialstatus,
monetary
allocations,
politicalinfluence
andtechnologicalequipment).
As a resultof thispursuit,
theorganization
would
becomemoreattentive
to different
behavior(i.e.,
aspectsof antisystemic
frequency,
variety,
etc.),andit wouldconsistently
lobbythegovernment
torespondrepressively.
Bothactionsassisttheorganization
in meeting
its
existenceandaccessto resources)
objectives(i.e.,continued
andalso increasethelikelihood
thatnegativesanctions
wouldbe applied.
The thirdpolitical-economic
characteristic
withan impacton repres'8Inthiscontext,
thesituation
is placedintosomewhat
of a "gray" area.As Franks
(1989) suggests,
therearetwodimensions
thatregimes
employtounderstand
politicalconflict
anddomestic
threats.
Thedimensions
ofconcern
are:legal-illegal
andlegitimate-illegitimate.Thesituation
addressed
byEastonrefers
tothelegal-legitimate
Whenconflict
category.
is in thiscategory,
it is unclearwhena threat
existsandwhena regimewouldbe justified
in applying
As opposedtothissituation,
repression.
however,
within
theillegal-illegitimate
quadrant
all behavior
is unacceptable
totheregimeandrepression
is "justifiable"
inevery
case. Here,theperception
ofthreat
andthejustification
forapplying
behavior
is
repressive
morestraightforward.
'9Thisdiminishes
theoverallpriceofitsimplementation,
as itbecomescheaprelative
to otheroptions.
This content downloaded from 141.213.172.160 on Sat, 30 Mar 2013 23:09:32 PM
All use subject to JSTOR Terms and Conditions
692
ChristianDavenport
sivepropensity
is economicdevelopment
(Banks1985;Henderson
1991;
Mitchelland McCormick1988). Whendevelopment
of theeconomyis
high,threats
areless likelytobe perceived
andrepression
is less likelyto
be used(Goldstein1983).Sincebasichumanneeds(including
healthcare,
housing,
food,economicopportunity,
etc.)havea greater
likelihood
ofbeingmetwithin
thiscontext,
important
elements
ofsocietyareprobably
not
calledintoquestion.Repression
is notexpectedbecauseexistingpower
relations
arenotdirectly
threatened.
Indeed,despitethepresence
ofantisystemicbehavior,
shouldhavea negahighlevelsofeconomicdevelopment
tiveeffect
on threat
perception
andrepressive
response
becausetheregime
does notwishto antagonize
dissidents.
Lowerlevelsofeconomicdevelopment,
on theotherhand,present
a
different
situation
entirely.
As underdeveloped
economieshave a much
moredifficult
timeproviding
basichumanneeds,ifconflict
doestakeplace,
thenthelatenthostility
felttowards
theregimehas a potential
forescalating.Inaneffort
todeterthispossibility,
all aspectsofantisystemic
behavior
shouldincreaserepression
(Sloan 1984;Gurr1986a,1986b).Theresponse
ofthegovernment
wouldthendissuadeanymovements
beingtakenagainst
itand,furthermore,
wouldmaintain
thepolitical-economic
system
thatallowedtheunderdevelopment
to existin thefirst
place.20
A majorpremise
intheliteraThelastfactor
considered
is dependency.
tureis thatdomestic
increases
theneed
penetration
bytheglobaleconomy
of certainpolitical-economic
(and desire)for protection
relationships
within
thestate(Jackson
etal., 1978;Herman1982;Petras1986;Alfatooni
andAllen1991).Since,independently
orinconjunction,
laborcouldstrike,
couldrevoltordisenfranchised
couldrebel,
politicalopponents
capitalists
all politicalconflict
mustbe prevented
hereforitcoulddeleteriously
affect
bothproduction
as wellas extracted
Thesearebothoutcomes
unacprofit.
fromtherelationship.
ceptableto thosewhobenefit
In an effort
to detertheseactivities
fromtakingplace and allowthe
benefits
of dependency
to be maintained,
theuse of repressive
behavior
is high.Responding
wouldlikelybe enhancedwhendependency
to even
thesmallestamountof conflict,
howeverconceived,sucha positionnot
international
anddomestic
an oponlyprotects
capital,butitalso protects
elitethatoftencontrols
thegovernment
pressive
political-military
(Herman
levelsofstatistical
these
1982;Petras1986).Metwithvarying
significance,
andWilhavebeensupported
relationships
byseveralauthors
(Timberlake
20Theinverted
"U" variant
of thisargument
is notaddressedhereas it was refuted
convincingly
byMitchell
andMcCormick
(1988).As a consequence,
onlythelinearrelationshipis addressed.
This content downloaded from 141.213.172.160 on Sat, 30 Mar 2013 23:09:32 PM
All use subject to JSTOR Terms and Conditions
WHY STATES APPLY NEGATIVE SANCTIONS
693
liams1984;Alfatooni
andAllen1991;Davenport
1991).21Whilethemeaningandaccuracyoftheseresults
haverecently
beenbrought
intoquestion
(Lopez and Stohl1989),thebasic formulation
of thehypothesis
is still
believedto be legitimate.
What Do We Know about the RelationshipbetweenDomestic
Threatsand PoliticalRepression?
Fromtheavailableliterature,
I haveidentified
twoimportant
andrelatedissues.First,different
conceptions
of domestic
threat
maybe maintainedbyregimes,
causingthemtoconsider
certain
aspectsofconflict
more
worthy
ofrepression
thanothers.
Second,numerous
factors
mayaltergovernment
ofdomestic
perception
threats
thelikelihood
directly
affecting
that
theregimewoulduse repressive
behavior.
bothof theseissues
Although
are essentialto anycomprehensive
understanding
of negativesanctions,
theliterature
hasaddressed
eachissueinonlya limited
sense.Specifically,
I identify
threelimitations
withexisting
studies.
The firstproblemconcernstheassumption
of unidimensional
threat
As mentioned
perception.
earlier,
generally
onlyoneaspectofconflict
(frequency)is usedas a determinate
ofstaterepression
(Hibbs1973;Alfatooni
and Allen 1991; Cingranelli
1992; Poe and Tate 1992). This constrains
ourconception
ofdomestic
threats
andeliminates
thepossibility
thatother
wouldbe considered
aspectsof politicalconflict
by regimeswhenthey
decideto use repressive
behavior.
The secondproblemconcerns
theomissionof important
explanatory
variables.
Forexample,
threestudiesignoretheeffects
ofsystem
typewhile
forother
variables
accounting
(DuffandMcCammant
1976;DavisandWard
1990;Alfatooni
andAllen1991),threeignorecoercivecapacity
(Davisand
Ward1990;Alfatooni
andAllen1991;Henderson
1991),threeignoreeconomicdevelopment
(Hibbs1973;Ziegenhagen
1986;Alfatooni
andAllen
1991),fourdo notincludedependency
(Hibbs1973;DavisandWard1990;
Henderson
1991;Poe andTate1992),andonestudydoesnotevenaddress
politicalconflict
ofthediversity
in
(Henderson
1991).22 As a consequence
themodelspreviously
empirical
examined
treatment,
areleftmisspecified,
andI am unableto systematically
or
compareresultsacrossexaminations
evenreacha consensusas to whatshouldbe givenattention.
2"Someempirical
analysesfoundnorelationship
atall (Mitchell
andMcCormick
1988).
The nonempirical
workrelevant
to thesecausalrelationships
has beenfarmoreextensive
(Herman1982;StohlandLopez 1984;Petras1986).Mypresentation
hasdirectly
followed
theirsuggestions.
22Within
thestudies
considered,
thenumber
ofcountries
andthenumber
ofyearsexaminedwerealso subjectto extensive
variation.
This content downloaded from 141.213.172.160 on Sat, 30 Mar 2013 23:09:32 PM
All use subject to JSTOR Terms and Conditions
694
Christian
Davenport
Thethird,
andlast,problem
withtheliterature
is thattheeffects
ofthe
variablesare generally
explanatory
notexaminedinteractively
withconflict,
i.e.,theydo notanalyzerepressive
behaviorwithinsituations
ofdomesticthreat.23
The effects
examinedare usuallyadditivein naturewith
variousstatecharacteristics
andconflict
eachcontributing
intorepression
dependent
ofthepossibleconfounding
effects
ofdissenton theothervariables.Thisis a majorshortcoming
becauseresearchers
areleftunableto
determine
whatconditions
enhanceor diminish
theapplication
ofrepressivebehaviorwhena regimeis directly
beingthreatened
bymassprotest.
On thecontrary,
we areleftinferring
particular
relationships
fromderived
results.
The mainpurposeofthepresent
analysisis to addtothisliterature
by
I putforth
correcting
forthesedeficiencies.
a comprehensive
examination
oftherelationship
betweenfourdifferent
aspectsofpoliticalconflict
and
I also assesstheimpactofthemostoftencitedattributes
staterepression.
of thedomesticpoliticaleconomy.Theseattributes
are initially
used as
in theexamination
controls
ofdirecteffects
forcomparison
withprevious
results.
Theseattributes
arethenusedinteractively
whenI investigate
government
behaviorwithin
unrest.
periodsofdomestic
Data and ResearchDesign
My analysesare basedon yearlycountry-level
datafor53 countries
from1948to 1982.The dataaretakenfromtwoprimary
sources:Taylor
and Jodice'sWorldHandbookofPoliticaland Social Indicators(1983) and
Banks' Cross-NationalTimeSeries Data Archive(1992). Each measureis
addressed
below.
The dependent
is operationalized
variable,repression,
by Taylorand
Jodice's(1983) indicator
ofnegative
sanctions.
include
Negativesanctions
defined
censorship,
as limitations
and/or
intimidation
ofthepopularmedia,
as wellas politicalrestrictions,
defined
as limitations
and/orintimidation
of individuals
andpoliticalparties.The eventshavebeencodedannually
from1948 to 1982 by consulting
theNew YorkTimesindexas well as
variousregionalpublications.24
Thevariableitselfis expressed
as a natural
log afteraddingone to thebase value.
23Theone exception
hereis Hibbs(1973). He examinedone interactive
relationship
betweensocialmobilization
andinstitutionalization.
24Although
numerous
problems
havebeenfoundwiththisdata(Goldstein1986),itis
stillheralded
as thebestmeasureavailableforrepresenting
at behavioral
attempts
control
cross-nationally
overtime
(seeGoldstein
[ 1986]andTaylorandJodice[ 1983]forcomments).
Thisacceptability
is alsoidentified
bythepervasive
useofthisdatawithin
numerous
anlayses
of repression
and politicalconflict
(Hibbs 1973; Duvalland Shamir1980;Muller1985;
Ziegenhagen
1986;Davis andWard1990;Alfatooni
andAllen1991).
This content downloaded from 141.213.172.160 on Sat, 30 Mar 2013 23:09:32 PM
All use subject to JSTOR Terms and Conditions
695
WHY STATES APPLY NEGATIVE SANCTIONS
are
fourmeasurements
aspectsofpoliticalconflict
To capture
different
ofthefirst
measurewhichis a basicfrea variant
used.25
Threerepresent
guerevents:antigovemment
demonstrations,
quencycountoffourconflict
andriots.CollectedfromBanks(1992),these
rillawarfare,
generalstrikes
on a yearlybasis.
arepresented
concerns
thepresenceofvioThe secondmeasureofpoliticalconflict
I employtheuse ofa dithisparticular
attribute,
lence.To operationalize
chotomous
variable.
Theactualcodingschemeitselfis quitesimple.When
orriotsoccur(bothinstances
ofviolentdissident
behavwarfare
guerrilla
arenotbeing
strategies
ior),violenceis codedas one.Whentheseparticular
however,
violenceis codedas zero.Thisfollowsinthetradition
employed,
(1980) and
ofRummel(1963),Tanter(1966),Hibbs(1973),Zimmerman
Gupta et al. (1993).26
ofstratevariety:
i.e.,thenumber
addresses
strategic
Thethird
measure
The valuesforthisvariablearederivedfromobgiesusedbydissidents.
conflict
events(antigovemment
demonstrathenumber
ofdifferent
serving
a givenyear.
riots)thatoccurwithin
guerrilla
warfare,
tions,generalstrikes,
within1948,1949and
present
typesofconflict
Here,I countthedifferent
betweenzero
is a variety
scoreineachyearvarying
so forth.
Whatresults
wereapwereapplied)andfour(whereall strategies
(whereno strategies
plied).
measureof politicalconflict
addressesdeviancefromthe
The fourth
thisvariabletwostepsaretaken.First,the
cultural
norm.To operationalize
foreach
1982)is identified
meanofthewholetimeperiod(1948 through
inanygivenyear,
toconflict
country.
Second,thismeanscoreis compared
aboveandbelowthemean.If
variableofconflict
a dichotomous
deriving
25Pearsoncorrelationsbetweenthe variousmeasuresare all providedbelow.
Variable
1) Frequency
2) Violence
3) Strategic
Variety
4) Cultural
Deviance
1
2
3
1.0
.42
.57
1.0
.80
1.0
.45
.62
.70
4
1.0
of conflict
behavior(Rummel1963;
twodimensions
authorshaveidentified
26Many
1980;Guptaet al. 1993):somehavelabeledthese
Tanter1966;Hibbs1973;Zimmerman
tolabel
(Guptaetal. 1993),others
havepreferred
dimensions
"violent"and"non-violent"
ofthe
them"rebellion"and"turmoil"(Hibbs1973;LichbachandGurr1981).Regardless
totheeffect
arethesamewithregards
nameprovided,
theexpectations
however,
particular
on politicalrepression.
This content downloaded from 141.213.172.160 on Sat, 30 Mar 2013 23:09:32 PM
All use subject to JSTOR Terms and Conditions
696
ChristianDavenport
conflict
is abovethemean,thena scoreof one is designated
andconflict
is considered
threatening.
If conflict
is belowthemean,however,
thena
scoreof zerois designated
andconflict
is deemedacceptable.
To identify
whatfactors
influence
repressive
propensity,
fourcomponentsofthedomestic
political-economy
arealso examined.
Theseinclude
democracy,
coercivecapacity,economicdevelopment
and dependency.
Each is discussedbelow.
To operationalize
democracy,
Banks' (1992) politicalpolyarchy
and
pluralism
variableis used.The measureitselfencompasses
severalfactors
including
theeffectiveness
of thelegislature,
thecompetitiveness
of the
nominating
procedure,
thenumber
ofpoliticalpartiesandthedegreethat
politicalpartiesareexcludedfromthenominating
procedure
(Banks1992,
15).27 Thisfollowsin thetradition
of Cutright
(1963),Bollen(1983) and
Gurr(1989).
Coercivecapacityis measured
byeachcountry's
defenseexpenditures
relativeto totalexpenditures.
Capturing
theoverallsignificance
and preofthecoerciveapparatus
paredness
toenactdesignated
policies,thisparticularattribute
hasreceiveda greatdealofattention
fromnumerous
scholars
themilitary's
addressing
effecton staterepression
(Laswell 1941,1962;
Huntington
1964;Thee 1977;Randle1981;WalkerandLang 1988).28
Energyconsumption
percapitais usedas theindicator
foreconomic
development.
GNP percapitahas beenemployedin numerous
Although
studies(Alkerand Russett1964; Hibbs 1973; Dye and Ziegler1988),I
haveoptedto use an alternative
strategy
giventheanalysisof Summers
andHeston(1988) thatidentified
thisvariableas beinghighlyunreliable.
Thispractice
hasbeenemployed
in severalrecentexaminations
ofrepressivebehavior(Henderson
1991;Poe andTate 1992,1994).
To concludeourdiscussion
oftheexplanatory
variables,
thelastcharacteristic
of thepoliticaleconomyto be measured
concernsdependency.
I use Taylorand Jodice's(1983)
Amidstseveraldifferent
possibilities,29
ofexportspecialization
indicator
to represent
thisvariable.Specialization
27Eachofthesecomponents
is measured
on a yearlybasis.
28Previously,
I haveusedthesizeofthemilitary
as a percentage
ofthetotalpopulation
additively
withthemeasure
ofdefenseexpenditures.
Considering
Hanneman
andSteinbeck
(1990) as wellas further
exploring
someworkofmyown,I foundthattheformer
was not
significant
whilethelatter
was.On thebasisofthisinformation,
I usedefenseexpenditures
relative
to totalexpenditures.
29For
example,OECD investment
has beenapplied(Timberlake
andWilliams1984)
as well as Snyderand Kick's (1979) "blockmodel" whichcombinestradeflows,treaty
memberships,
military
intervention
and diplomatic
relations.
Theseare notusedbecause
theydisregard
important
aspectsofthedomestic
economy
andbecausetheywereonlyavailable fora fewyears.
This content downloaded from 141.213.172.160 on Sat, 30 Mar 2013 23:09:32 PM
All use subject to JSTOR Terms and Conditions
WHY STATES APPLY NEGATIVESANCTIONS
697
Table 1. EquationsInvestigated
EQUATION
and Response
ThreatPerception
1. Unidimensional
Repression= a + B1 Democracy+ B2 Coercivecapacity
+ B4 Dependency
+ B3 Economicdevelopment
+ ?
+ B5 Lagged Repression+ B6 Conflictfrequency
EQUATION
and Response
ThreatPerception
2. Multidimensional
Repression= x + B1 Democracy+ B2 Coercivecapacity
+ B4 Dependency
+ B3 Economicdevelopment
+ B5 LaggedRepression+ B6 Conflictfrequency
+ B7 Presenceof violence+ B8 StrategicVariety
+ B9 Deviancefromculturalnorm+ ?
fallwithin
thedegreethatexportcommodities
is calculatedbyexamining
decrease,
As thesecategories
ofcategories.30
smallnumber
a comparatively
is saidto increase.
dependency
effects
twodifferent
investigated,
theempirical
questions
Concerning
timeseriesdesign.First,adare analyzedwitha pooledcross-sectional
have
thatthevariousaspectsof politicalconflict
thepossibility
dressing
betweenthefourcharacterisdirecteffects
on repression,
effects
different
for
controlling
areexamined,
sanctions
threat
andnegative
ticsofdomestic
in
andareprovided
Two equationsareanalyzed,
fourstatecharacteristics.
betweenthefretherelationship
equationinvestigates
Table 1. The first
fordemocracy,
controlling
and theuse of repression,
quencyof conflict
Thisaddresses
coercivecapacityanddependency.
economicdevelopment,
The sechypothesis.
threat
perception/response
thebasicunidimensional
of
betweenthefourattributes
therelationship
ond equationinvestigates
ofviolencewithin
thepresence
ofconflict,
(i.e.,frequency
politicalconflict
dissidentbehavior,"strategicvariety"and deviancefromthe cultural
forthesamefourstatecharaccontrolling
norm)andtheuse ofrepression,
of multidimensional
used above.This addressesthehypothesis
teristics
andresponse.
threat
perception
I hypothesized
thatgovexamined,
ofthesecondrelationship
In terms
configuration
aredetermined
bytheparticular
toconflict
ernment
responses
I emthisrelationship
To investigate
inquestion.
ofthepoliticaleconomy
that
thosestatecharacteristics
process.First,I identify
ploya three-step
yearshad
missing
30Asdataexistedfortheyears1950to 1975,byfiveyearintervals,
to be interpolated.
This content downloaded from 141.213.172.160 on Sat, 30 Mar 2013 23:09:32 PM
All use subject to JSTOR Terms and Conditions
698
ChristianDavenport
havestatistically
significant
andtheoretically
substantive
effects
on political repression.
This information
is providedfromtheempirical
analysis
conducted
within
thefirst
partoftheexamination,
notedabove.Second,I
dividethesampleintothreecategories
(high,medium,
andlow),representing different
valuesof the particular
characteristic(s)
foundsignificant.
Third,therelationship
betweenthesignificant
conflict
variable(s)andrepressionis reestimated
foreachcategory.
Thisallowsmetocompareand
contrast
different
stateresponses
todomestic
threats
determined
bytheparticularstructure
ofthepoliticaleconomybeingconsidered.
A finalpointmustbe madeaboutthespecificresearchdesignemployed.As pooledtimeseriesdesignshavebeenfoundto suffer
fromhetas wellas autocorrelation
eroscedasticity
(Stimson1985;Ostrom1990),I
mustconcernmyselfdirectly
withproperestimation
in thefaceof these
To addressthesedifficulties,
problems.
I haveemployed
tworemedial
strategies.First,country
dummies
areincorporated
intoall equations
tocontrol
forcase specificerror(heteroscedasticity).
Second,I use a laggeddependentvariabletocontrol
fortheoretical
reasons(see below).Thisfrequently
eliminates
withautocorrelation
problems
(Beck1992;Poe andTate1994).31
Withregardtothetheoretical
severalauthors
havemainjustification,
tainedthatthereis a highdegreeofbureaucratic
inertia
involvedwiththe
of negativesanctions
application
(Gurr1986a;Poe andTate 1994).Once
is used,theargument
repressive
behavior
becomehabituated
goes,regimes
toitthrough
theprocesses
ofresource
mobilization
tothecoer(allocations
civeapparatus)
andbehavioral
touse negative
disposition
(thewillingness
The rateof repressive
sanctions).
behaviorexperienced
one year(i.e., at
timet-1)is expectedto havean effect
on thisbehavior
thefollowing
year
(i.e., at timet).32Acknowledging
thisdelay,theuse of a laggedvariable
worksquitewellin modeling
therepressive
decision-making
process.
ofthetworemedial
Gaugingtheempirical
validity
strategies,
postcorrectiondiagnostics
willbe conducted
on eachequationestimated.
To addressheteroscedasticity,
I plottheestimated
squaredresidualsagainstthe
estimated
variablefromtheregression
linesobtained
within
the
dependent
above.Thisfollowsthesuggestions
equations
ofGujarati
provided
(1978).
To addressautocorrelation,
I calculatetheDurbinh-statistic,
of
a variant
theLagrangeMultiplier,
and followthedecisionrulethata statistically
3ITherawdatagenerally
followanAR1 process.
Thisparticular
conclusion
wasreached
4 wereestimated
afterEquations1 through
withan ARIMAmodelandthendiagnosed
with
thestrategy
detailedin McClearyandHay (1980).
thatthislaggedeffectmaybe arbitrary,
I am subjecting
32Understanding
thiscausal
relationship
to morerigorous
in another
investigation
study.Forthisanalysis,however,
I
believetheassumption
to be morethanacceptable.
This content downloaded from 141.213.172.160 on Sat, 30 Mar 2013 23:09:32 PM
All use subject to JSTOR Terms and Conditions
WHY STATES APPLY NEGATIVESANCTIONS
699
Table 2. ExaminingtheUnidimensional
ThreatPerception
and ResponseHypothesis
Independent
Variables
Constant
Democracy
Coercive
capacity
Economic
development
Dependency
Conflict
frequency
Laggedrepression
Number
ofcases
R2
Durbin
h-statistic
error
forthe
Standard
equation
PoliticalRepression
Equation1
Equation2
1.78(.14)**
-.07(.01)**
.02(.25)
-.00(.00)**
.03(.00)**
.04(.00)**
.45(.02)**
1785
.63
1.40(.10)**
-.08(.01)**
-.00(.00)**
.03(.00)**
.04(.00)**
.48(.01)**
1801
.63
-.81
-.60
.72
.74
* p < .05; **p < .01.
areinparentheses.
Standard
errors
number
indicates
thatthepresence
ofautocorrelation
cannotbe
significant
to detectfirstrejectedas a strong
possibility.
Particularly
recommended
orderserialcorrelation
witha laggeddependent
variable(Beck 1992,56;
willbe provided
witheach
Harvey1990,275), thevalueforthisstatistic
I nowmoveto theempirical
itself.
equationestimated.
investigation
Findings
in Table2 andconducting
thevarious
theresultsprovided
Observing
I findthatthefirstmodelexamined(Equations1 and 2) is
diagnostics,
notdeleteriously
affected
are
byeither
heteroschedasticity
(i.e.,no outliers
orautocorrelation
is notsignificant).
identified)
(i.e.,theDurbinh-statistic
oftheexisting
Theresults
fromthismodelaregenerally
literasupportive
betweencoercivecapacityand political
ture,exceptthattherelationship
ThisI believeis generally
is notfoundto be significant.
attribrepression
is morepoliticalthan
utedto-the
factthedecisiontouse negative
sanction
or strategic.33
military
as a definitive
refutation
ofthecoercivecapacity-repression
33Ido notviewthisfinding
theissueis farfrom
withtheliterature
wouldattest,
hypothesis.
As anyonewhois familiar
(Laswell 1941;Walkerand Lang 1988) as well
resolved.First,a greatmanytheoretical
thatsupport
the
as historical
case studies(Thee 1977;Randle1981)havebeenconducted
(Hibbs 1973;
studieshavebeenconducted
relationship.
At thesametime,fewempirical
Hibbs(1973)
ofthisconceptvarysignificantly:
Ziegenhagen
1986).Second,measurements
whileZiegenhagen
(1986) considconsidered
internal
forcespersquarekilometer,
security
This content downloaded from 141.213.172.160 on Sat, 30 Mar 2013 23:09:32 PM
All use subject to JSTOR Terms and Conditions
700
ChristianDavenport
Table 3. ExaminingtheMulti-dimensional
ThreatPerception
and ResponseHypothesis
PoliticalRepression
IndependentVariables
Constant
Democracy
Economicdevelopment
Dependency
Conflictfrequency
Presenceof violence
Strategicvariety
Deviancefromculturalnorm
Laggedrepression
Numberof cases
R2
Durbinh-statistic
Standarderrorforthe
equation
Equation 3
1.26(.10)**
-.07(.01)**
-.00(.00)**
.03(.00)**
.01(.00)**
.06(.05)
.12(.03)**
.21(.05)**
.46(.01)**
1732
.66
-.72
.70
Equation 4
1.27(.10)**
-.07(.01)**
-.00(.00)**
.03(.00)**
.01(.00)**
.14(.02)**
.22(.05)**
.46(.01)**
1797
.65
-.58
.70
*p < .05; **p < .01.
arein parentheses.
Standard
errors
Droppingthisvariableand re-estimating
a reducedformoftheequation
in theireffectson
(Equation 2), fourvariablesare statistically
significant
repressivebehaviorand explain63% of thevariance.Commensuratewith
previousexpectations,thelevel of democracyand economicdevelopment
are foundto decrease the use of repressionwhile dependency,lagged repressionand politicalconflictarefoundto increaseitsapplication.This last
variableis particularly
to thepresentanalysisbecause it directly
important
threatperceptionand response.
supportsthehypothesisof unidimensional
Notedearlier,thisis by farone of themostoftencitedrelationships
identifiedwithinthe literature.
Is a multidimensional
of domesticthreats,however,
representation
more useful in explainingvariancein negativesanctions?Includingthe
of politicalconflictintothebasic equationsome very
otherthreeattributes
resultsare found(Equations 3 and 4 see Table 3).
interesting
ingoverneredthesizeofthemilitary
relative
tothetotalpopulation,
military
representation
mentas wellas thevariable
employed
within
thisanalysis(military
sectorallocations).
Third,
theempirical
examinations
thathavebeenconducted
do notutilizethesamecontrolvariI would
ables.As oneofthefewempirical
attempts
toinvestigate
thisrelationship,
therefore,
tentative.
suggestthatmyresultsaremerely
This content downloaded from 141.213.172.160 on Sat, 30 Mar 2013 23:09:32 PM
All use subject to JSTOR Terms and Conditions
WHY STATES APPLY NEGATIVESANCTIONS
701
heteroscedasticaffected
byeither
Again,themodelis notdeleteriously
ecoAll of thecontrolvariables(i.e., democracy,
ityor autocorrelation.
significant,
inthe
arestillstatistically
anddependency)
nomicdevelopment
to thepreviousequationin
and roughlycomparable
expecteddirection
variable.Thisincreasesourconfitermsoftheirimpacton thedependent
of the
dencein theresults,
as one can see thatthevariouscharacteristics
consistent
andstableeffects.
exhibitrelatively
political-economy
increased
from
theprevious
TheR2ofthisequation
hasonlymarginally
one (Equation2), i.e., from63% to 65%. This causes me to be a little
oftheotherattributes
ofconflict.
Conductabouttheimportance
skeptical
ornottheinclusion
ofthesevariables
whether
inga jointF-testtodetermine
significant,
I findthatthecriticalvalueof 3.0 (at the.05
is statistically
level)is exceededbythederivedvalueofthetest.Inclusionoftheaddito the
is foundto add significantly
tionalconflictvariables,therefore,
of politicalrepression
and thuswill be retainedwithinthe
explanation
model.
result
themostimportant
uponthislastpointandaddressing
Following
I findthatthreeattributes
ofpolitical
conflict
arestatistiforthisdiscussion,
in theirrelationship
to repressive
behaviorandthattheir
callysignificant
on thedependent
variableis in theexpecteddirection
(Equation4).
effect
conflict
andcultural
devianceare
strategic
variety
frequency,
Specifically,
Fromthis,we can conclude
all positively
relatedto negativesanctions.
willbe directly
enhancedwhen
and politicalrestrictions
thatcensorship
ofdifferent
ofconflict
whenthenumber
strategies
theintensity
increases,
acare diverseand whendeviancefromculturally
appliedby dissidents
of dissentis manifest.
Thesefindings
support
directly
ceptedboundaries
of multi-dimensional
threat
andresponse.
thehypothesis
perception
is notsignificantly
relatedtopoliticalrepression
attribute
One conflict
ofthisvariable
ofviolence.Theinsignificance
(inEquation3), thepresence
leads to the
is probably
to thefactthatitspresencegenerally
attributed
control
stateof otherstrategies
of behavioral
including
implementation
executions.
These
stratearmedattacksandpolitical
terrorism,
sponsored
thisparticular
because
attribute
gies wouldbe relieduponin confronting
As a
addressed.
counter
the
threat
dimension
being
theymoreappropriately
of
the
violence
of
this
"logisticalcorrespondence," presence
consequence
as measured
here,butitwould
wouldnotbe relatedto negativesanctions
be relatedto otherformsof stateregulation.
in positiveeffects
is
uniformity
Atthesametimethatthemanifested
additionI
must
the
different
conflict
across
observed
however,
attributes,
variablediffertosomeimportant
Eachconflict
diversity.
allypayattention
of
its
the
value
effects
as
determined
slopecoeffiby
repression
entially
This content downloaded from 141.213.172.160 on Sat, 30 Mar 2013 23:09:32 PM
All use subject to JSTOR Terms and Conditions
702
ChristianDavenport
cient.34
Thissuggests
thatregimes,
although
similarly
affected
bydifferent
aremoresensitized
aspectsofconflict,
tothepresence
ofcertain
attributes
thanothersin itsapplication
ofrepressive
behavior.
Fromtheresults
ofEquation4, cultural
devianceis showntohavethe
largest
impactonpoliticalrepression
at +.22, followed
bystrategic
variety
(+.14) andthenconflict
frequency
(+.0 1). Thesedifferent
effects
areimportant
becausetheyidentify
thatregimesare moreinclinedto respond
repressively
todeviancefromthecultural
normandmulitiple
of
strategies
masspoliticalbehavior.
In thesesituations,
theregimehastoconfront
conflictthatis inviolation
ofitscodeofacceptable
dissentas wellas confront
different
ofpolitical
strategies
eachwithitsownmethod
conflict,
ofrecruitmentandimpactuponthedomestic
politicaleconomy.
Conflict
exhibitsmuchless of an effecton negativesancfrequency
tions.Oneinterpretation
ofthisfinding
is thatthismeasure
doesnotcapture
enoughinformation
abouttheparticular
threat
In fact,
beingconfronted.
themoresignificant
aspectsofpolitical
conflict
appeartoplacethedesignationofthreatening
behavior
within
therealmofmorestrategic
issues:i.e.,
strategic
and culturaldeviance.Commensurate
variety
withpreviousresults,therefore,
politicalrepression
willstillbe appliedin responseto an
increased
number
ofconflict
events.Thisresponse,
however,
occursat almostnegligible
levelsrelativeto otheraspectsofpoliticalconflict.
These
findings
are crucialin thattheybeginto highlight
theinnerworkings
of
therepressive
decision-making
process.Indeed,theyidentify
thatregimes
aremuchmorediscriminating
thanoriginally
believedwithregardtowhat
a threat
andalso towhattheywillrespond
theyconsider
towithrepressive
action.
Addressing
thesecondrelationship
ofinterest,
I nowconsidertheeffectsofdemocracy
on repressive
whilecontrolling
forthedopropensity,
mesticthreats
encountered.
Therearetworeasonsforthissingular
attention
onsystem
from
type.First,
4 democracy
Equations1 through
is consistently
foundtoexhibit
thegreatest
ofallthestatecharacterisimpactonrepression
ticsidentified.
Whilebotheconomicdevelopment
anddependency
aresigtheirsubstantive
nificant,
effects
arequitesmall.35
We wouldthusbe most
informed
abouttherelevance
ofpolitical-economic
context
byexclusively
todemocracy.
payingattention
Thesecondreasonforconcentrating
onthis
characteristic
theoverallamount
concerns
particular
ofattention
giventoit
betasthesameresults
werederived.
Thesecanbe obtained
34Utilizing
fromtheauthor
uponrequest.
ofthethesevariables,
35Indeed,
giventheeffects
thedegreetowhichtheir
valueswould
haveto be changedbeforea meaningful
impacton politicalrepression
wouldbe revealed
is somewhat
unrealistic
to expect.
This content downloaded from 141.213.172.160 on Sat, 30 Mar 2013 23:09:32 PM
All use subject to JSTOR Terms and Conditions
WHY STATES APPLY NEGATIVESANCTIONS
703
As mostauthors
andhumanrights.
bythosewhostudypoliticalrepression
byincreaswouldbe reducedsignificantly
behavior
suggest
thatrepressive
examining
howthisparticofthepoliticalsystem,
ingthegeneralopenness
beingchallenged
functions
whentheregimeis directly
ularrelationship
howregimeopenwouldbe useful.Thiswouldallowus toclearlyidentify
of domesticstress,an exwithinmoments
ness and tolerancefunction
historical
contextas well as
withinthepresent
tremely
likelyoccurrence
unexamined.
one thathas beenpreviously
each
I estimate
threeequations,
To investigate
thismediating
effect,
concerns
only
from
Thefirst
equation
derived
thebasicmodelusedabove.36
is madeup ofthose
Thiscategory
thoseregimesthatarenotdemocratic.
forthefullandno opposition
parties
thathaveno legislature
governments
examplesincludeAfghanistan,
Illustrative
timeperiodunderexamination.
The secondequationaddressesthoseregimesthatI
BulgariaandJordan.
regimesarethosethatvacillatebeTransitional
havelabeledtransitional.
categories.
These regimesare
and nondemocratic
tweenthedemocratic
sometimes
nondemocracies
and sometimes
sometimes
fulldemocracies,
includeArgentina,
Ghanaand
between
thetwo.Examplesofthiscategory
thoseregimes
thatarefulldemocracies.
equationconcerns
Nepal.Thethird
forthefull-time
a competperiod,
haveeffective
legislatures
Theseregimes
Examparties.
procedure
andno exclusionofopposition
itivenominating
CostaRica andtheUnitedStates.
pleshereincludeAustralia,
The resultsfromthethreeequationsare veryrevealingwithregard
As found,
to howdifferent
respondto domesticthreats.
politicalsystems
onlytwoaspectsofponondemocracies
(Equation5 see Table4) consider
Notethatthese
liticalconflict
whentheydecidetouse politicalrepression.
oftheothertwo
arehigher
thanthatreachedbyeither
ratesofapplication
Thissupports
theclaimmadeearlierthatthesepoliticalsystems
regimes.
to theirsecurity
andmorelikelyto use
withthreats
arehighlyconcerned
ofhowthesethreats
areconceptualized.
regardless
repression,
Withina modelthataccountsfor75% of thevariancein repressive
is foundto be themostimportant
behavior(Equation5), strategic
variety
innonandpolitical
restrictions
theuseofcensorship
variableinprovoking
is increasedfrom
democracies
at +.25. The effectof conflict
frequency
within
thebasicequation(i.e.,from.01 to .05),butitis still
thatidentified
of conflict
identified.
The regime's
less thaneitheroftheotherattributes
enthatstrategic
hereis quiteunderstandable
variety
considering
response
interactive
variables
becauseofa highdegreeofmulticollinearity
36Icouldnotestimate
I thusoptedto dividethedataset
andtheircomponents.
betweentheinteractive
measures
interactive
Thisallowedmetoidentify
intothree
andestimate
eachindependently.
categories
of multicollinearity.
effects
andat thesametimeavoidtheproblem
This content downloaded from 141.213.172.160 on Sat, 30 Mar 2013 23:09:32 PM
All use subject to JSTOR Terms and Conditions
704
ChristianDavenport
Table 4. ExaminingRepressiveResponseswithinDifferent
RegimeTypes
Political Repression
IndependentVariables
Constant
Economicdevelopment
Dependency
Conflictfrequency
Strategicvariety
Deviancefromculturalnorm
Laggedrepression
Numberof cases
R.2
Durbinh-statistic
Standarderrorforthe
equation
Equation 5
Equation 6
Equation 7
(Nondemocratic (Transitional (Full Democratic
Regimes)
Regimes)
Regimes)
1.75(.24)**
-.00(.01)**
.01(.O1)
.05(.02)**
.25(.08)**
.11(.13)
.45(.03)**
423
75
1.59
.51
1.83(.21)**
-.00(.00)**
.03(.00)**
.03(.00)**
.09(.03)**
.24(.07)**
.48(.02)**
687
.57
-1.73
.64
2.65(.25)**
-.00(.00)**
.05(.02)**
.01(.00)**
.09(.04)**
.22(.08)**
.25(.03)**
687
.73
-.11
.55
*p < .05; **p < .01.
Standard
errors
arein parentheses.
tailsa significant
amountof coordination
betweendifferent
components
of theopposition.
Thisthreat
is especiallypotentgiventhefactthatthe
hasbasicallyeliminated
government
all other
forms
ofpopular
participation
andhas claimeda monopolization
on legitimate
politicalexpression.
Deviancefromthecultural
normis notsignificant
withinnondemocraticregimes.One explanation
forthisfinding
mightbe thattheoverall
repressiveness
of theseregimesis too highto facilitate
large-scale
challengesfromthepopulace.Indeed,ofall theregimetypesthisone has the
lowestnumber
of deviations
fromthecultural
norm.Withinthiscontext,
dissidents
wouldprobably
notusethisstrategy,
the
dramatically
increasing
oftheirconflict
intensity
so as notto incurthefullwrathofthe
behavior,
government.
Instead,
theywouldmorelikelyincrease
thediversity
ofstrategiesusedagainsttheregime.
Thiswouldallowthemtodecreasetheimpact
ofrepression
experienced
byanyonestrategy,
and,atthesametime,allow
themto challengetheregimein a substantive
manner.
Withregardto thecontrolvariablesof thisequation,laggedeffects
ofrepression
arestillfoundto be statistically
Thisrevealsthe
significant.
importance
giventobureaucratic
inertia
andtheroleplayedbythecoercive
inmaintaining
theuse ofpoliticalrepression
apparatus
afterithasalready
beenapplied.Theeffects
ofeconomicdevelopment
alsoremain
significant.
Evenmoreinteresting,
is thatdependency
is notrelated
however,
torepres-
This content downloaded from 141.213.172.160 on Sat, 30 Mar 2013 23:09:32 PM
All use subject to JSTOR Terms and Conditions
WHY STATES APPLY NEGATIVE SANCTIONS
705
at oddswiththe
appearing
sense.37
Initially
in anystatistical
sivebehavior
First,the
forthisfinding.
thereare twoexplanations
existingliterature,
betherelationship
capture
mayinaccurately
ofdependency
measurement
Second,theuse
politicaleconomy.
factors
andthedomestic
tweenexternal
that
nondemocratic
regimes
within
ofrepression
maybe so institutionalized
Bothof these
to changeswithintheglobalmarket.
it no longerresponds
wouldneedto be exploredfurther.
possibilities
from
different
is somewhat
justidentified
ofrelationships
The pattern
category
(Equation6). Here,exthosewe obtainwithinthetransitional
we findthatall attriplaining57% ofthevariancein politicalrepression,
variable.
on thedependent
in theireffects
aresignificant
butesofconflict
havetopoliticaldisthattheseregimes
Thisrevealstheoverallsensitivity
devihereis cultural
ofconflict
identified
attribute
sent.Themostimportant
varietyis the
variableat +.24. Strategic
thedependent
ance,impacting
is again
frequency
variable,at +.09, and conflict
secondmostimportant
at +.03. Whendissentexceedsthe
totheregime,
factor
theleastimportant
increase
regimes
therefore,
transitional
levelofacceptability,
government's
than
extent
toa fargreater
andpoliticalrestrictions
theiruse ofcensorship
conservative
Thisrelatively
considered.
attributes
anyoftheotherconflict
fromoverextending
itselfwhileat
thecoerciveapparatus
strategy
prevents
theparticular
challengebeingconthesametimeallowingit to confront
fronted.
is
thecontrolvariablesforthisequation,pastrepression
Considering
of
regime
all
three
the
impact
significant, largest
foundto be statistically
importhistomeanthatinertia
playsan extremely
types.Onecaninterpret
of
behavior
to
the
amount
repressive
in
with
tantrole theseregimes regard
itis quitelogical
arerentwithuncertainty,
thatis applied.As thesesystems
Indeed,this
to
of
sanctions.
the
use
negative
thattheybecomehabituated
is probably
the
apparatus
the
that
coercive
more
fact
is even
thecase given
the
varirest
of
control
around.
The
of
the
one
onlypoliticalorganizations
expectations,
previous
are
Supporting
ables
significant.
againstatistically
anddetheuse ofrepression
is foundto diminish
economicdevelopment
is
to
increase
its
found
application.
pendency
forfulldemocracies
theresults
investigation,
To concludetheempirical
fromTable 4, theseresults
are derived(Equation7). Easilyidentifiable
The mostnoticeable
category.
arequitesimilarto thoseofthetransitional
amountofexplainedvariance
betweenthetwois theincreased
difference
modelaccountsfor57% ofthevariance
The transitional
within
thelatter.
theequation,
thesameresultsareobtained.
thisvariableandre-estimating
37Dropping
theresultsfromthisequationare availablefromthe
Deletedforpurposesof parsimony,
authoruponrequest.
This content downloaded from 141.213.172.160 on Sat, 30 Mar 2013 23:09:32 PM
All use subject to JSTOR Terms and Conditions
706
ChristianDavenport
within
thedemoinrepression
whereas73% ofthevariancewasexplained
the
cultural
devianceis foundtomanifest
craticequation.
Fromtheresults,
greatest
impactonrepression
at +.22. Thismakessensebecausewe would
topoliticalconflict
wouldrespond
ata significant
expectthatdemocracies
danger."Thissituation
would
rateonlywhenthereis a "clearandpresent
almostbe essentialto theregimeso thatitcouldmaintain
politicallegitithreat
itself.Strategic
macy,and,at thesametime,deal withthedomestic
variable,
increasing
repression
variety
is againthesecondmostimportant
is theleastimportant
factor
to theregime,
at +.09, andconflict
frequency
impacting
thedependent
variableat +.01.
is stillsignificantly
tocurrent
behavior
related
politiLaggedrepressive
is a greatdealsmaller
itseffect
thanthatidentified
cal repression,
although
Thissuggeststhatdemocratic
regimes
withintheotherpoliticalsystems.
tobureaucratic
inertia
andlesslikelytobe caughtwithin
arelessresponsive
As thesegovernments
ofhabitually
behavior.
thethrongs
appliedrepressive
tothecoerciveapparatus
andmorecontobe lessresponsive
areexpected
thisreflects
wellupontheabilityofthese
cernedwithpoliticallegitimacy,
regimesto curtailrepressive
policiesoncetheyhavebeenappliedandreEconomicdevelopment
is also
ducetheimpactof thelaw of instrument.
initseffect
onrepression
andintheexpected
direcstatistically
significant
todepenWithregard
negligible.
tion,buttheimpactitselfis stillrelatively
increasesthelikelihood
directly
dency,thedegreeof tradeconcentration
wouldbe applied.Thisis in factthelargesteffect
thatpoliticalrepression
within
all ofthedifferent
examined.
politicalsystems
Supporting
exhibited
betweenfree-trade
liberalism
numerous
arguments
abouttherelationship
andtheuseofpoliticalrepression
(Herman1982;Petras1986;Pion-Berlin
in establishing
aremoreinterested
andprotecting
1989),fulldemocracies
thanthoseregimes
certain
economicrelationships
thatareless open.With
wethusfindan interesting
thisresult,
spinona commonly
explored
hypothI wouldsuggest
esis.As thisappearstogo againsta greatdealofliterature,
be conducted.
thatadditional
investigation
Conclusion
Thepresent
twodistinct
andimportant
relationstudyhasinvestigated
of domestic
characterization
ships.First,it compareda unidimensional
threats
toa multidimensional
oneandassessedtheimpactofbothonpolitical repression.
(democSecond,it examinedhowone statecharacteristic
onrepressive
Both
theimpactofdomestic
threats
behavior.
racy)mediated
arediscussedbelow.
Fromtheresultsoftheempirical
analysis,I foundthata multidimenof domesticthreats
in acwas morecomprehensive
sionalrepresentation
of
When
this
threats
was
forrepressive
behavior.
representation
counting
This content downloaded from 141.213.172.160 on Sat, 30 Mar 2013 23:09:32 PM
All use subject to JSTOR Terms and Conditions
WHY STATES APPLY NEGATIVESANCTIONS
707
theamount
ofvarianceexplained
employed
threeaspectsofpoincreased,
liticalconflict
(conflict
frequency,
strategic
variety
anddeviancefromculturalnorm)werefoundtobe statistically
significant,
andthederived
causal
effects
wereall intheexpected
direction.
Uponexamination,
eachvariable
was foundto exhibita different
effect
on thedependent
variable.Cultural
devianceand strategic
werefoundto manifest
thegreatest
variety
impact
on politicalrepression,
whileconflict
frequency
exhibited
onlya marginal
effect.
I concludethatregimes
Onthebasisofthisfinding,
notonlyobserve
different
aspectsof politicalconflict
withregardto thedegreeof threat
butthattheyalsorecognize
perceived,
theneedtoapplypolitical
repression
at different
ratesto counter
thesethreats.
Oneimportant
thatfollowsfrom
thisfinding
is thatinvestiimplication
gationsof repression
shouldno longermaintain
exclusiverelianceupon
basicfrequency
recountsto measurepoliticalconflict.
On thecontrary,
searchers
shouldemploya multidimensional
representation
of domestic
Thiswouldbetter
threats.
enablethemto understand
therepressive
decimodeltherelationship
besion-making
processandmorecomprehensively
tweendissident
behaviorandpoliticalrepression.
I foundit erroneous
Movingto thesecondrelationship
investigated,
toassumethattheissueofmultidimensional
threat
andresponse
perception
wasa universal
one,appliedequallyacrossdifferent
typesofpolitical-economiccontexts.
As found,all regimesdo notrespondto domestic
threats
inthesamemanner.
Themostimportant
characteristic
observed
wassystem
type,i.e.,democracy.38
To investigate
thisrelationship
thesamplewasdividedintothreedifferentcategories
of thisparticular
variable:i.e., nondemocratic,
transitional
anddemocratic
The causalrelationship
betweenconflict
regimes.
andrepression
was thenre-estimated
foreachcategory,
revealing
severalimportantdifferences
acrossthesedifferent
system
types.Nondemocratic
governmentswerefoundto be threatened
bystrategic
variety
anddeviancefrom
thecultural
norm.Responding
to bothattributes,
theseregimesincreased
at
rates
of
repoliticalrepression relatively
Transitional
high
application.
gimes,ontheotherhand,werefoundtobe threatened
all
three
by
attributes
ofpoliticalconflict.
normyieldeda high
Onlydeviancefromthecultural
repressive
whilestrategic
responseratefromthegovernment,
and
variety
conflict
increased
atrelatively
frequency
repression
smallratesofapplication.Democratic
followed
a similar
regimes
generally
ofthreat
pattern
per38Recall,
theempirical
results
showedthatdemocracy,
economicdevelopment
anddewerestatistically
In termsof themagnitude
of theslopecoefficient,
pendency
significant.
was foundtobe substantively
however,
onlydemocracy
I thusoptedto concenimportant.
trateon thisparticular
characeristic
forfurther
investigation.
This content downloaded from 141.213.172.160 on Sat, 30 Mar 2013 23:09:32 PM
All use subject to JSTOR Terms and Conditions
708
Christian
Davenport
ceptionand responseto thatidentified
withinthetransitional
category.
Thereare,however,
twoimportant
differences
betweenthesetworegimes
withrespectto thisfinding.
First,theamountofexplainedvariancebythedemocratic
modelwas
significantly
Within
theequation
higher.
fortheseregimes,
73% ofthevariancewas accounted
forwhileintheequationfortransitional
regimes
only
57% of thevariancewas explained.Thisdifference
I believereflects
the
within
generaluncertainty
whichtransitional
exist.Since
politicalsystems
theobjectives
andcapabilities
oftheseregimes
areconstantly
beingaltered
overtime
(in somecases severaltimesindifferent
a greatdeal
directions),
is lefttochance.As a resultofthisindeterminacy,
thereis lessofan opportunity
toestimate
causalrelationships
becausetheempirical
properly
strategiesthemselves
generally
assumethatsomeconsistency
within
theparticularunitsbeingexamined.
Theoppositesituation
exists,however,
whenthe
is consistently
Inthiscontext,
regime
democratic.
I believeitis mucheasier
to estimate
state-societal
relationships.
The seconddifference
between
thetworegimes
concerns
theexplanationbehindthelow repressive
responseto different
aspectsof dissent.
Within
I believethatthelowrepressive
democracies,
response
to strategic
variety
andconflict
is determined
frequency
bythecommitment
ofthese
regimes
totolerate
certain
aspectsofdissident
behavior.
As longas conflict
theseboundaries,
itwillnotsubstantively
stayswithin
provokea reaction
In transitional
fromthegovernment.
politicalsystems,
I believe
however,
thesituation
tobe muchdifferent.
Within
thiscontext,
I do notbelievethat
therelatively
is determined
lowrateofresponse
bypoliticaltolerance.
On
thecontrary,
therepressive
responseof theregimeis probably
attributed
to theirlow levelofregulatory
capacity:i.e.,thecapacityoftheregimeto
monitor
actsof politicalconflict
andrespondto theseactsin an efficient
manner.
theselimitations,
theregime
wouldgenerally
allowcerAccepting
tain attributes
of politicalconflictto manifest
themselves
withoutreto it withrepression.
Whentheseconflict
sponding
attributes
reachedextremely
highvalues(whentheculturally
defined
ofacceptable
parameters
dissenthadbeenviolated),
thencensorship
however,
andpoliticalrestrictionswouldbe appliedat relatively
wouldallow
highrates.Thisstrategy
theregimeto protect
thestatusquo, and,at thesametime,prevent
the
regulatory
apparatus
fromoverextending
itselfuntilithad to do so.
As a consequence,
we thusfindthattheconsideration
of system
type
servesa crucialroleintheprocessofdomestic
threat
andrepresperception
siveresponse.Systemtypeidentifies
therelativesensitivity
thatdifferent
will haveto domesticthreats.
Nondemocratic
governments
governments
areapparently
themostsensitive
to multidimensional
threats
(in termsof
theirresponserates),and democratic
as well as transitional
governments
This content downloaded from 141.213.172.160 on Sat, 30 Mar 2013 23:09:32 PM
All use subject to JSTOR Terms and Conditions
WHY STATES APPLY NEGATIVESANCTIONS
709
reasons).Regimetypealso
different
(forentirely
appeartobe lesssensitive
can be. As found,stable
ourpredictions
allowsus to gaugehoweffective
providethemostaccurateresults
andstabledemocracies
nondemocracies
providetheleastaccurate.This should
governments
whiletransitionary
within
different
sanctions
negative
in understanding
guideus significantly
contexts.
political-economic
of
ourconception
The workon thisissueis farfromdone.Although
ofwhatcontextoa regimeandourunderstanding
a threat
whatconstitutes
numerous
questotheseprocessesareclearer,
tualvariablesareimportant
totheperimportant
ofpoliticalconflict
tionsemerge.Areotherattributes
affect
repressive
factors
threats?
Do othercontextual
ceptionof domestic
governplaya roleinchanging
oftransition
Does thedirection
propensity?
Howlongdoesittake
andtheuseofrepression?
ofthreats
mentperception
to stabilize,
response
andrepressive
perception
fortheprocessesofthreat
thatthesequesThelistgoeson.Understanding
estimation?
better
allowing
I leavethese
study,
tionsextendwellbeyondtheobjectivesofthepresent
andmodelbuildinvestigation
Theprocessofempirical
analysis.
forfuture
ingcontinues.
Manuscriptsubmitted26 April 1994.
Final manuscriptreceived15 October 1994.
REFERENCES
SanctionandCollectivePolitical
Afra,andMichaelAllen.1991."Government
Alfatooni,
ofPolitiStates:A TimeSeriesAnalysis."Journal
inPerifery
andSemiperifery
Protest
cal and MilitarySociology 19:29-45.
In World
ofTrendsandPatterns."
1964."The Analysis
andBruceRussett.
Alker,Hayward,
Handbook of Political and Social Indicators.New Haven: Yale UniversityPress.
Banks, Arthur.1992. Cross-NationalTime-SeriesData Archive.Binghamton:Center for
Social Analysis.
Analysisof HumanRights."
A Statistical
of Oppression:
Banks,David. 1985. "Patterns
AmericanStatisticalAssociation: Proceedings of the Social Science Section 62:154-
63.
ApTheCase ofPresidential
DynamicSpecifications:
1992."Comparing
Beck,Nathaniel.
proval." In Political Analysis:An Annual Publicationof the MethodologySection of
theAmericanPolitical Science Association,ed. JamesStimson.Ann Arbor:University
of MichiganPress.
Blalock, Hubert.1989. Power and Conflict:Towardsa General Theory.NewburyPark,CA:
Sage.
Bohrnstedt,George, and David Knoke. 1988. Statisticsfor Social Data Analysis.2nd ed.
Itasca,IL: F.E. PeacockPublishers.
ofPoliticalDemocracy."
Measurement
1980."IssuesintheComparative
Bollen,Kenneth.
AmericanSociological Review 45:370-90.
This content downloaded from 141.213.172.160 on Sat, 30 Mar 2013 23:09:32 PM
All use subject to JSTOR Terms and Conditions
710
ChristianDavenport
. 1983. "WorldSystemPosition,Dependency,
and Democracy:Further
Issuesin
theMeasurement
andEffects
ofPoliticalDemocracy."
American
SociologicalReview
48:468-79.
Buzan,Barry.1983.People,Statesand Fear. ChapelHill: University
of NorthCarolina
Press.
andJohn
Chadwick,
Firestone.
Robert,
1972."Instrastate
Socio-Political
andEmpirical
Conflict
Model."Research
PaperNo. 20. Center
forComparative
PoliticalResearch:
Binghamton
University.
Cingranelli,
David. 1992."DemocracyandHumanRightsIn Less DevelopedCountries."
Presented
attheannualmeeting
oftheAmerican
Political
ScienceAssociation,
Chicago.
Cutright,
Phillips.1963. "NationalPoliticalDevelopment:
Measurement
and Analysis."
AmericanSociological Review 28:253-64.
Dallin, Alexander,and George Breslauer. 1970. Political Terrorin CommunistSystems.
Stanford:
Stanford
Press.
University
Davenport,
Christian.
1991."The Dynamics
of StateRepression:
A Cross-National
TimeSeriesExamination
ofDomesticPoliticalProcesses."Ph.D.Diss.Binghamton
University.
. n.d."AssessingtheMilitary's
Influence
on PoliticalRepression:
An Examination
of DifferentHypotheses." Journalof Political and MilitarySociology. Forthcoming.
Davis,David,andMichaelWard.1990."TheyDance Alone:DeathsandtheDisappeared
in Contemporary
Chile." Journalof ConflictResolution34:449-75.
Dedrick,
Dennis.1978."DevianceandSanctioning
within
SmallGroups."SocialPsychology41:94-105.
Dogenharat,Henry,and Alan Day, eds. 1983. Political Dissent: An InternationalGuide to
Dissident, Extra-Parliamentary,
Guerilla and Illegal Political Movements.Detroit:
GareResearchCo.
Duff,Ernest,and JohnMcCammant.1976. Violenceand Repressionin LatinAmerica.New
York:FreePress.
Duvall,Raymond,
and MichaelShamir.1980. "Indicators
fromErrors:Cross-National,
Time-Serial
Measuresof theRepressive
Disposition
of Governments."
In Indicator
Systemsfor Political, Economic and Social Analysis.ed. Charles Taylor. Cambridge:
GunnandHainPublishers.
Oelgeschlager,
andMichaelShamir.1983."Governance
Duvall,Raymond,
byTerror."In ThePoliticsof
ed.Raymond
Terrorism,
DuvallandMichaelStohl.3rded. NewYork:MarcelDekker.
Dye,Thomas,andHarmanZiegler.1988."SocialismandEqualityin Cross-National
Perspective." PS 21:45-56.
Easton,David.1975."A Re-Assessment
oftheConceptofPoliticalSupport."British
Journal of Political Science 5:435-57.
WolfDieter.1987. "DomesticPoliticalProcesses."TheGlobusModel: ComEberwein,
puter Simulation of WorldwidePolitical and Economic Developments,ed. Stuart
Bremer.
Boulder:WestviewPress.
Harold.1980. "Theoretical
Eckstein,
to Explaining
Approaches
CollectivePoliticalViolence." HandbookofPolitical Conflict:Theoryand Practice,ed. Ted Gurr.New York:
FreePress.
Franks,
C.E.S., ed. 1989.Dissentand theState.New York:OxfordUniversity
Press.
andKevinFitzgibbons.
Faucher,
Philippe,
1989."DissentandtheStateinLatinAmerica."
In Dissentand theState,ed. C.E.S. Franks.New York:OxfordUniversity
Press.
Gibbs, Jack. 1981. Norms,Deviance and Social Control.Westport:GreenwoodPress.
Goldstein,RobertJ. 1978. Political Repressionin ModernAmerica: From 1870 to thePres-
ent.Boston:Schenckman/G.K.
Hall.
This content downloaded from 141.213.172.160 on Sat, 30 Mar 2013 23:09:32 PM
All use subject to JSTOR Terms and Conditions
WHY STATES APPLY NEGATIVESANCTIONS
711
. 1983. Political Repressionin 19thCenturyEurope. Totowa, NJ:Barnes and Noble
Books.
HumanRights
of UsingQuantitative
Data in Studying
. 1986. "The Limitations
Abuses." Human RightsQuarterly8:607-627.
New York:McGraw-Hill.
Gujarati,
Damodar.1978.Basic Econometrics.
Coercionof DissiGupta,Dipak,Hanrder
Singh,andTom Sprague.1993. "Government
dents: Deterrenceor Provocation."Journalof ConflictResolution37:301-39.
A Theoretical
ofStateViolenceandTerror:
AnalyGurr,
Ted. 1986a."The PoliticalOrigins
sis." In GovernmentalViolenceand Repression:An Agendafor Research,ed. Michael
Press.
StohlandGeorgeLopez,eds. New York:Greenwood
fora General
ofRepression
andRebellion:
Foundations
Patterns
. 1986b."Persisting
Theoryof Political Coercion." In PersistentPatternsand EmergentStructuresin a
Karns,ed. New York:Praeger.
WaningCentury,
Margaret
of PoliticalScience.Boulder:Centerfor
. 1989.PolityIl Codebook.Department
ofColorado.
Comparative
Politics,University
andPoliticalInsta1990."Military
Involvement
Hanneman,
Robert,
andRobinSteinbeck.
bility:An EventHistoryAnalysis 1940-1980." JournalofPolitical and MilitarySociology 18:1-23.
Harvey,Andrew. 1990. The EconometricAnalysisof Time Series. 2nd ed. Boston: Philip
andMIT Press.
Allan,HemelHempstead
. 1993. Time Series Models. Cambridge:MIT Press.
theUse ofPoliticalRepression."
Journal
Affecting
Henderson,
Conrad.1991."Conditions
of ConflictResolution35:120-42.
Herman,Edward. 1982. TheReal TerrorNetwork:Terrorismin Fact and Propaganda. Bos-
ton:SouthEnd Press.
andCharlesHerman.1989."WhoMakesForeignPolicyDecisionsand
Herman,
Margaret,
How: An EmpiricalInquiry." InternationalStudies Quarterly33:361-88.
Hibbs,Douglas.1973.Mass PoliticalViolence.New York:Wiley.
Hoover,Dean, and David Kowalewski.1992. "DynamicModelsof Dissentand Repression." Journalof ConflictResolution36:150-82.
Huntington,
Samuel.1964.TheSoldierand theState.New York:RandomHouse.
Century.Oklahoma:
. 1991. The ThirdWave: Democratizationin theLate Twentieth
of OklahomaPress.
University
andCoerBruceRossett,
DuncanSnidalandDavidSylvan.1978."Conflict
Jackson,
Steven,
cion in DependencyStates." Journalof ConflictResolution22:627-57.
ofProtest
1965to 1989."
Waves:WestGermany,
Koopmans,
Ruud.1993."The Dynamics
AmericanSociological Review 58:637-58.
onViolence."American
StateandSpecialists
Journal
Laswell,Harold.1941."The Garrison
of Sociology 46:455-68.
Patterns
ofMilitary
1962."The Garrison
StateHypothesis
Today." In Changing
New York:FreePress.
Politics,ed. SamuelHuntington.
andDissent."Journalof
Lichbach,Mark.1987. "Deterenceor Escalationin Repression
ConflictResolution31:266-97.
Process:A Self-Generating
Model."
Lichbach,Mark,andTed Gurr.1981."The Conflict
Journalof ConflictResolution21:3-29.
ofNewYork
andSocialControl.
Albany:StateUniversity
Liska,Allen.1992.SocialThreat
Press.
toStateViolenceandState
Ideologyas anImpetus
Lopez,George.1986."NationalSecurity
Terror." In GovernmentViolenceand Repression:An Agendafor Research, ed. Mi-
chaelStohlandGeorgeLopez.New York:Greenwood
Press.
This content downloaded from 141.213.172.160 on Sat, 30 Mar 2013 23:09:32 PM
All use subject to JSTOR Terms and Conditions
712
ChristianDavenport
Lopez, George,and Michael Stohl. 1989. Dependence,Developmentand State Repression.
New York:Greenwood
Press.
Markus,
Gregory,
andBettyNesvold.1972."Government
Coerciveness
andPoliticalInstability:AnExploratory
StudyofCross-National
Patterns."
Comparative
PoliticalStudies 5:231-43.
Mason,T. David,andDale Krane.1989."The PoliticalEconomy
ofDeathSquads:Toward
a Theoryof theImpactof State-Sanctioned
Terror."International
StudiesQuarterly
33:175-98.
McAdam, Douglas. 1982. Political Process and theDevelopmentof Black Insurgency.Chi-
ofChicagoPress.
cago: University
McCleary,Richard,and RichardHay, Jr.1980. Applied TimeSeries Analysisfor theSocial
Sciences.BeverlyHills:Sage.
Mitchell,
Neil,andJamesMcCormick.
1988."Is U.S. AidReallyLinkedtoHumanRights
in Latin America?" AmericanJournalof Political Science 32:231-39.
Mitchell,Neil, and JamesMcCormick.1992. "ExplainingHumanRightsViolations."
Presentedat the annualmeetingof the AmericanPoliticalScience Association,
Chicago.
Morris,Aldon.1993."Birmingham
Confrontation
An Analysisof theDyReconsidered:
namics and Tactics of Mobilization." AmericanSociological Review 58:621-36.
Muller,
Edward.1985."IncomeInequality,
andPolitical
RegimeRepressiveness
Violence."
AmericanSociological Review 50:47-61.
Oommen,T.K. 1990. Protestand Change in Social Movements.NewburyPark,CA: Sage.
Ostrom,Charles. 1990. TimeSeries Analysis:RegressionTechniques.2nd ed. Sage Univer-
in theSocial Sciences.BeverlyHills:
sitySeriesPaperon Quantitative
Applications
Sage.
Petras,James.1986."PoliticalEconomyof StateTerror:
Chile,El SalvadorandBrazil."
Crimesand Social Justice27:88-109.
Pfuhl,Erdwin.1986.TheDevianceProcess.2nded. Belmont,
CA: Wadsworth
Publishing.
Pion-Berlin,David. 1989. The Ideology of State Terror:Economic Doctrineand Political
Repressionin Argentinaand Peru. Boulder: Lynne Rienner.
inthe1980's:A Theoryand
andFreedom
Poe,Steve,andC. NealTate.1992."Repression
PooledCross-Sectional
at theannualmeeting
of theAmerican
Analysis."Presented
PoliticalScienceAssociation,
Chicago.
in the 1980's: A
Poe, Steve,and C. Neal Tate. 1994. "Repressionof PersonalIntegrity
Global Analysis." AmericanPolitical Science Review 88:853-872.
andRepression."
VII:61-144.
Randle,Michael.1981."Militarism
Alternatives
John.1987."ViolenceandRepression:
Factors
inDevelopment
PlanRichardson,
Neglected
ning."Futures7:651-68.
ofConflict
R.J.1963."Dimensions
Behavior
Within
andBetween
Nations."YearRummel,
book of the Societyfor General SystemsResearch 10:183-211.
in LatinAmerica:
TheCurrent
Seligson,Mitchell.1987."Democratization
Cycle." InAuthoritarians
and Democrats: RegimeTransitionin LatinAmerica,ed. JohnMalloy and
MitchellSeligson.Pittsburgh:
of Pittsburgh
University
Press.
and Enforcement
in LatinAmerica."In
Terrorism
Sloan,John.1984. "StateRepression
The State as Terrorist:The Dynamics of GovernmentViolence and Repression,ed.
MichaelStohlandGeorgeLopez.Westport:
Greenwood
Press.
andMethodological
intheAnalysis
David.1976."Theoretical
Problems
ofGovernSnyder,
mentCoercion and Collective Violence." Journalof Political and MilitarySociology
4:277-94.
intheWorldSystemandEcoPosition
Snyder,
David,andEdwardKick.1979."Structural
This content downloaded from 141.213.172.160 on Sat, 30 Mar 2013 23:09:32 PM
All use subject to JSTOR Terms and Conditions
WHY STATES APPLY NEGATIVESANCTIONS
713
InteracAnalysisof Transnational
nomicGrowth,1955-1970:A MultipleNetwork
tions." AmericanJournalof Sociology 84:1096-1126.
Sorokin,Pitirim.1957. Social Change and CulturalDynamics.Boston: PorterSargeant.
James.1985. "Regressionin Space and Time:A Statistical
Essay." American
Stimson,
Journalof Political Science 29:914-47.
3rded. New York:MarcelDekker.
Stohl,Michael.1988.ThePoliticsofTerrorism.
Stohl,Michael, and George Lopez. 1984. The State as Terrorist:The Dynamicsof Governmental Violenceand Repression.Westport:GreenwoodPress.
Samuels.1986.StateViolation
GeorgeLopez,andStephen
Stohl,Michael,DavidCarleton,
of Measurement."
HumanRightsQuarterly
of HumanRights:Issuesand Problems
8:592-606.
of
Comparisons
Robert,and AlanHeston.1988. "A New Set of International
Summers,
1950-1985."TheReview
for130Countries,
andPriceLevelEstimates
Real Product
of Income and Wealth34:1-25.
Nations."
ofConflict
Behaviorwithin
andbetween
Tanter,
Raymond.1966."Dimensions
Journalof ConflictResolution10:41-64.
Tarrow,Sidney. 1989. Democracy and Disorder: Protestand Politics in Italy 1965-1975.
Press.
England:Clarendon
Oxford,
Taylor,Charles,and David Jodice.1983. WorldHandbookofPolitical and Social Indicators
Press.
III. New Haven:Yale University
Relations."
inContemporary
International
andMilitarization
Thee,Mark.1977."Militarism
Bulletinof Peace Proposals 8:296-309.
Reading,MA: Addison-Wesley.
to Revolution.
Tilly,Charles.1978.FromMobilization
andGovPoliticalExclusion,
Michael,andKirkWilliams.1984."Dependence,
Timberlake,
Evidence."American
SociologicalReview
ernment
SomeCross-National
Repression:
49:141-46.
Volgyes, Ivan, ed. 1978. Social Deviance in Eastern Europe. Boulder: WestviewPress.
intheThirdWorld:
StateSyndrome
andIvyLang.1988."The Garrison
Walker,Stephen,
Research Note." Journalof Political and MilitarySociology 16:105-16.
andNormalofDeviance:Deviance,Opportunity
Leslie.1964."A General
Theory
Wilkens,
ity." In Social Deviance: Social Policy,Actionand Research,ed. Leslie Wilkens.New
York:Prentice.
Ziegenhagen,Eduard. 1986. The Regulationof Political Conflict.New York: Praeger.
Ekkart.1980."Macro-Comparative
Researchon PoliticalProtest."In HandZimmerman,
book of Political Conflict,ed. Ted Gurr.New York: Free Press.
inthePhillipines."
In TheStateas TerrorandRepression
Zwick,James.1984."Militarism
ist: The Dynamicsof GovernmentalViolenceand Repression,ed. Michael Stohl and
Greenwood.
GeorgeLopez.Westport:
This content downloaded from 141.213.172.160 on Sat, 30 Mar 2013 23:09:32 PM
All use subject to JSTOR Terms and Conditions

 Download  Report

Paperzz.com

Your Paperzz

© Copyright 2026 Paperzz