u.s. foreign policy and regime instability

8/6/2019 U.S. Foreign Policy and Regime Instability

1/73

U.S. FOREIGN POLICYAND REGIME INSTABILITY

James Meernik

May 2008

This publication is a work of the U.S. Government as denedin Title 17, United States Code, Section 101. As such, it is in thepublic domain, and under the provisions of Title 17, United StatesCode, Section 105, it may not be copyrighted.

Visit our website for other free publication downloadshttp://www.StrategicStudiesInstitute.army.mil/

To rate this publication click here
http://www.strategicstudiesinstitute.army.mil/http://www.strategicstudiesinstitute.army.mil/pubs/display.cfm?pubID=845http://www.strategicstudiesinstitute.army.mil/pubs/display.cfm?pubID=845http://www.strategicstudiesinstitute.army.mil/


2/73

ii

*****

The views expressed in this report are those of the authorand do not necessarily reect the ofcial policy or position of theDepartment of the Army, the Department of Defense, or the U.S.Government. This report is cleared for public release; distributionis unlimited.

*****

Comments pertaining to this report are invited and should beforwarded to: Director, Strategic Studies Institute, U.S. Army WarCollege, 122 Forbes Ave, Carlisle, PA 17013-5244.

*****

All Strategic Studies Institute (SSI) publications are availableon the SSI homepage for electronic dissemination. Hard copiesof this report also may be ordered from our homepage. SSIshomepage address is: www.StrategicStudiesInstitute.army.mil.

*****

The Strategic Studies Institute publishes a monthly e-mailnewsletter to update the national security community on theresearch of our analysts, recent and forthcoming publications, andupcoming conferences sponsored by the Institute. Each newsletteralso provides a strategic commentary by one of our researchanalysts. If you are interested in receiving this newsletter, pleasesubscribe on our homepage at www.StrategicStudiesInstitute.army.mil/newsletter/.

ISBN 1-58487-355-8


3/73

iii

FOREWORD

This Letort Paper examines the relationshipbetween U.S. military ties with foreign states and theextent to which the depth of these ties inuences thelevel of political instability and violence in those states.Many pundits and scholars have criticized U.S. foreignpolicy for its reliance on military means of inuenceand have argued that other foreign policy tools, suchas economic aid, cultural exchanges, and diplomacycan better promote American interests. Yet, fewscholars have chosen to evaluate empirically whetherthe military relationship encourages or discouragespolitical instability and violence in these nations. Theauthor, Dr. James Meernik, analyzes these issues ina systematic and objective fashion and nds that therelationships between a U.S. military presence, U.S.

military aid, the use of military force, and other factorsare much more complex and subtle than many havebelieved.

The Strategic Studies Institute is pleased to publishthis analysis as a contribution to the debate on thisissue.

DOUGLAS C. LOVELACE, JR.DirectorStrategic Studies Institute


4/73


5/73

v

SUMMARY

The United States utilizes a vast arsenal of foreignpolicy tools to induce, compel, and deter changes inother nations foreign policies. As the quantity andquality of such activity increases, the U.S. footprintin such nations grows deeper and wider. The U.S.presence may range from a diplomatic mission to amassive invasion force. The United States may seek touse its presence to openly compel change in a regimespolicies; it may employ its leverage to quietly inducepolicy modication; or it may use a combination ofsuch strategies. And while the regime and citizensof one nation may welcome the United States and itslargess, others may nd such relationships a threatto the nations honor and sovereignty. To the extenta deeper and broader foreign policy relationship (as

measured by a U.S. military presence; U.S. foreign aidrelationship; the discrete use of military force; and asubstantial similarity in foreign policy preferencesbetween the United States and another government)contributes to stability and friendship, U.S. interestsare realized. But does a broad and deep military andforeign policy relationship with the United Statesalways succeed in realizing these interests?

Why would a cooperative relationship withthe United States precipitate political and societalinstability in the host nation? First, the U.S. relationshipwith the friendly or client regime may undermine thepopular legitimacy and sovereignty of the governmentor interfere with local, political processes. Second,political ties with the United States often impact local

economic conditions. Whether it is economic ties perse the United States is seeking to advance throughopening markets, providing economic assistance, or


6/73

vi

promoting U.S. multinational corporation interests, orit is the economic spillover effects from a U.S. military

presence, local market conditions are bound to beinuenced by the actions of the worlds largest economywithin the local borders. Third, the local populationmay also be opposed to the broader U.S. foreign policygoals with which U.S. ofcials are seeking acquiescenceor cooperation. Specic U.S. interests will also provokeantagonism as the populations of other states takeexception to the ends or the means of U.S. foreignpolicy, and to their regimes degree of identicationwith such interests.

On the other hand, U.S. foreign policy means andends are intended and designed to promote positiverelations and maintain stability in those nations withwhom the United States seeks to foster amicable andcooperative relationships. A strong U.S. presence can

promote multiple, positive conditions. First, to theextent that a U.S. presence promotes both internal andexternal security for a nation, it provides the protectionand stability a state needs to develop economically andpolitically. U.S. friendship can deter interstate rivalsfrom overtly aggressive behavior and can dissuadeinternal political rivals from sowing unrest. Second,to the extent a U.S. military presence or U.S. militaryaid alleviates the need for a government to expendresources on its own security, a regime is betterable to utilize freed up resources on economic andsocial development that should further the nationsprosperity. Third, a U.S. military presence and militaryaid can stimulate the local economy and provide jobsfor many nationals who are involved in businesses that

contract with and supply the U.S. military, and canopen avenues of opportunity for citizens to take part ineducational, economic, and military interactions withthe United States.


7/73

vii

I use statistical analyses to evaluate the extent towhich indicators of a U.S. foreign policy relationship

predict the level of terrorism, domestic instability, andwar in other nations. I nd a statistically signicantrelationship between several of the indicators of U.S.foreign policy and instability in foreign countries.The closer the relationship between a country and theUnited States, as measured by many of these indicatorsin most of the estimates, the more likely nations wereto experience various forms of instability. Yet, the sizeof the impact of U.S. foreign policy was not alwaysstrong. Of all the measures of ties to U.S. foreign policy,the one that demonstrated the strongest and mostconsistent effects in the estimates was U.S. militaryaid. The greater the amount of military aid received bya foreign government, the more at risk it becomes forinstability, including terrorism, riots, assassinations,

anti-government demonstrations, and civil wars.The other measure of U.S. foreign policy

relationships that exercises a strong, albeit somewhatinconsistent impact on regime instability is involvementin a militarized dispute with the United States. Whenthe United States has used military force in or towarda foreign regime in the previous year, the predictedincidence of terrorism and civil wars tends to increasein the following year. Uses of force may inspire anti-American sentiment, embolden regime opponents totake violent action against the government (especiallyin cases where the United States is taking action againstthe regime), or may simply indicate the prevalence ofuncertainty and trouble in a nation.

We nd less evidence that a large U.S. military

presence contributes in any signicant manner,at least so far as is apparent in these analyses, toregime instability. The effects of the size of the U.S.


8/73

viii

military presence on the indicators is either small,statistically insignicant, or both. The last U.S. foreign

policy indicator considered is the extent to whicha nations voting record in the United Nations (UN)General Assembly mirrors that of the United States.As a states voting record in the UN more closelyresembles the United States, the incidence of variousforms of instability, including riots, anti-governmentdemonstrations, assassinations, and government crisesincreases. Anti-Americanism is often de rigueurin manynations, and thus making public pronouncementsagainst U.S. foreign policy objectives almost seems tobe reexive in many capitals around the world.

The nature of a countrys political system alsoplays a much more crucial role. We see throughout theanalyses that as constraints on the executive branchof government increased, the incidence of terrorism,

riots, anti-government demonstrations, assassinations,government crises, and civil wars all increased. On theother hand, political competitiveness serves to decreasethe likelihood of riots, anti-government demonstra-tions, government crises, assassinations, and civil wars.Viewed from the perspective of domestic tranquility,the most effective form of government would appearto be one with a strong executive and robust politicalcompetition. Economic prosperity appears to decreaseinstability. The greater a nations per capita grossdomestic product, the lower the predicted incidence ofriots, anti-government demonstrations, assassinations,and civil wars. We also see, however, that morepowerful states are more likely to experience acts ofterrorism, riots, assassinations, and anti-government

demonstrations, but are less likely to be involved incivil wars. These states typically have large economies,large populations, and large militaries. Their major


9/73

ix

power status among the nations of the world may makethem inviting targets for disaffected groups within

their borders and terrorists from both the outside andinside, but not to the point at which intrastate warbreaks out.


10/73


11/73

1

U.S. FOREIGN POLICYAND REGIME INSTABILITY

INTRODUCTION

A superpower, like the United States, and its foreignpolicy actions typically produce substantial con-sequences throughout the worldit is the 800 poundgorilla whose every move carries with it deep and widerepercussions. Hegemons are able to supply a numberof public and private goods that earn them many allies,especially among those states with whom they alreadyenjoy a coincidence of interests. The hegemons verydominance, however, creates conicts of interest,disputes, and challengers to its role regardless of thesubstance of its interests. U.S. foreign policy, broadlyspeaking, subsumes all actions taken by the U.S.

Government that are directed toward inuencing theconduct of world affairs in order to make the UnitedStates more secure and prosperous. When framed inthe broadest and most inclusive terms, the UnitedStates seeks to inuence other nations to adopt policiesand take actions that more closely reect Americaninterests. The more approximately other nations aligntheir foreign policy preferences to those of the UnitedStates, presumably, the better able is the United Statesto realize its preferences. Inuencing other regimes tomove toward the U.S. preferred position encompassesa substantial part of its foreign policy. To the extent thatthe United States is able to effect such changes throughdiplomatic inducement and deterrence and via othermechanisms that utilize its soft power, it preserves

foreign policy resources for use in situations thatrequire more forceful applications of U.S. power. And


12/73

2

while the U.S. Government sometimes deliberately,and sometimes inadvertently, takes actions that result

in other nations moving their policies away from U.S.preferences, nonetheless the overall goal in worldaffairs remains: to make more nations act globallymore like the United States.

The United States possesses a plethora of tools atits disposal to effect change or seek inuence in foreignregimes, including the stationing of U.S. military forcesin foreign nations, the use of military and economicassistance, and the discrete use of military force. Scholarshave researched the extent to which these tools havehelped promote U.S. foreign policy objectives, suchas democratization, improvements in human rights,and economic development. While their conclusionsstress the limited impact the United States, or anyother nation for that matter, can have on such specic

objectives, scholars have yet to investigate the extentto which U.S. foreign policy tools are associated withbroader U.S. foreign policy goalsmost specicallyregime stability. U.S. foreign policy relations dependupon the stability of those nations with which it seeksgood relations. Those nations the United States seeksto inuence that experience civil unrest, terrorism, andwar are unlikely to be capable of maintaining positiveand productive relations with the United States, to saynothing of democratizing or improving their humanrights practices. Thus, a fundamental goal of U.S. foreignpolicy must necessarily be to help such regimes remainpeaceful, stable and free of terrorism. But, to whatextent do U.S. foreign policy relations with nations helpimprove the likelihood that these basic requirements

of an effective foreign policy are realized? Does theuse of military force or the stationing of U.S. troopswithin such nations or the use of foreign assistance


13/73

3

contribute to peaceful and stable regimes that are freeof terrorism? Or do these foreign policy tools in some

manner make the realization of these objectives moredifcult? The purpose of this monograph is to evaluatethe degree to which U.S. foreign policiesthe stationingof U.S. military personnel; the use of military force;the provision of foreign assistanceas well as a moregeneral similarity of foreign policy interests betweenthe United States and a foreign regime stabilizes ordestabilizes such nations.

BACKGROUND

The United States has utilized a vast arsenal offoreign policy carrots and sticks to induce, compel,and deter changes in other nations foreign policies.Traditionally, U.S. foreign policy research focuses

on the degree of success the U.S. Government hasachieved when seeking specic objectives such asimprovements in human rights conditions, democraticchange, United Nations (UN) General Assembly voting,trade policies, and a host of other goals. The resultsof such studies have been mixed, with some ndingevidence that the United States can induce nationsto vote more closely according to its preferences inthe UN General Assembly,1 and that U.S. militarizedactions do sometimes lead to advances in democracy intarget nations.2 Yet, other researchers have concludedthat U.S. foreign assistance has little impact on humanrights3 or democratization,4 and that the use of militaryforce to promote democracy rarely works.5 And eventhose researchers who have found evidence of linkages

between U.S. foreign policy actions and target nationbehavior conclude that such relationships are oftenneither strong nor direct. Thus, evidence of positiveinuence must be treated carefully.


14/73

4

In this analysis, I seek to examine more broadly theeffect of U.S. foreign policy actionsdo they increase

or decrease regime stability? The prime directive ofdoctors, as well as many other types of practitioners,is held to be, do no harm. That dictum may well bedirected toward U.S. foreign policydo U.S. foreignpolicy actions lead to harmful outcomes within thosestates with which the United States has establishedrelationships? Are nations in which the UnitedStates maintains economic, political, and securityrelationships more likely to experience adverse politicalevents that may be a result of the U.S. presence? Morespecically, are the peoples of these nations likely toengage in disruptive or violent behavior because oftheir opposition to the U.S. presence or the degree ofU.S. inuence over their government? Or does theprovision of foreign assistance, the use of military

force, or the stationing of U.S. military personnel ina nation promote greater stability and lead to fewerinstances of civil unrest, terrorism, and war? Doesa close relationship with the United States provideforeign governments with the resources, assistance,and support they need to protect themselves from suchthreats?

Before delving into the relationship between U.S.foreign policy and domestic unrest in other nations, Ishould make clear that as in all studies of inuence,we must be extremely cautious in ascribing cause andeffect status to phenomena whose precise relationshipwill often be simplied in our models and remainobscured in our data. Such caveats must be assumedin all such analyses, but I state them unequivocally

here at the outset.The U.S. Government engages in a plethora of

actions to seek, maintain, and promote relations with


15/73

5

other nations to better realize its broader foreign policyaims. As the quantity and quality of such activity

increases, the U.S. footprint in such nations growsdeeper and wider. The U.S. presence may range froma diplomatic mission to a massive invasion force.The United States may seek to use its presence toopenly compel change in a regimes policies; it mayquietly leverage policy modication; or it may usea combination of such strategies and everything inbetween. And while the regime and citizens of onenation may welcome the United States and its largess,others may nd such relationships a threat to thenations honor and sovereignty. I assume that in mostcases the United States prefers a stable governmentand acquiescent population in such target nations,although it may foment unrest against some regimeswhose policies it nds particularly objectionable. I

further assume that when unrest and violence do occur,they tend to undermine U.S. foreign policy objectivesby destabilizing friendly regimes. As indicated above,a close foreign policy relationship with the UnitedStates may help promote peace and stability; it may beassociated with unrest and instability; and, of course, itmay have little or no effect at all. I rst describe belowwhy states that maintain close ties with the UnitedStates might experience more negative consequences,before describing the rationale behind the opposingarguments.

U.S. Foreign Policy and Regime Instability.

Why would a cooperative relationship with the

United States precipitate political and societal instabilityin the host nation? There may be numerous, specicU.S. policies the local populace nds objectionable, but


16/73

6

here I focus on three broader sets of concerns that mightengender opposition to the United States and ultimately

its allies in the host regime. First, the U.S. relationshipwith the friendly or client regime may undermine thepopular legitimacy and sovereignty of the governmentor interfere with local political processes. Given thatthe United States is pursuing its own foreign policyinterests and seeking to sway regimes to adopt morefavorable domestic and international policies, we mustassume that (a) the local regime has not adopted policiessufciently close to the U.S.-preferred position alreadybecause its own political interests and key constituentsdictate otherwise; and (b) the United States is seeking toconvince the regime to adopt policies that run contraryto its perceived interests. Ultimately, no matter howmuch the United States may seek to hide or disguiseits efforts, many local ofcials and ordinary citizens

are likely to resent U.S. actions as intrusive and beoffended by the perceived subversion of their nationalinterests. The more extensive the relationship betweenthe United States and the local regime, the greater theprobability that such opposition will grow.

Second, political ties with the United States oftenimpact local economic conditions. Whether it iseconomic ties per se the United States is seeking toadvance through opening markets, providing economicassistance, or promoting U.S. multinational corporation(MNC) interests, or it is the economic spillover effectsfrom a U.S. military presence, local market conditionsare bound to be inuenced by the actions of the worldslargest economy within the local borders. For example,U.S. economic assistance that fosters free market reforms

may, in turn, lead to reduced government subsidies tosome economic sectors; spending reductions in socialwelfare programs; and the advancement of the interests


17/73

7

of U.S. MNCs and local economic elites at the expenseof small businesses and workers. A large U.S. military

and/or political presence may lead to the growthof industries that cater more to the demands of U.S.Government personnel than to the local population,and may distort wages in some sectors of the economy.Even if the aggregate impact of all these U.S. actionsis positive and leads to growth in the local economy,there will still likely be resentment at U.S. involvementin local affairs and the creation of winners and losersin the local economy that will breed further oppositionto the U.S. role.

Third, the local population may also be opposed tothe broader U.S. foreign policy goals with which U.S.ofcials are seeking acquiescence or cooperation. Somedegree of opposition to U.S. interests will form eitherbecause governments and their citizens oppose them

because they areAmerican interests and/or because theyare the interests of a hegemon. Hegemonic interests,regardless of their substance, will always be met withsome resistance from some quarters as the hegemonschallengers seek to assert their own interests. Tensionsand jockeying for power in global affairs are a nearconstant feature of international relations throughouthistory. Specic U.S. interests will also provokeantagonism as the populations of other states takeexception to the ends or the means of U.S. foreignpolicy, and to their regimes degree of identicationwith such interests. And even though regimes mayseek to distance themselves from the particular U.S.actions that arouse such opposition, nonetheless, thedeeper their overall ties with the United States, the

greater the likelihood that U.S. unpopularity will leadto local unrest.

Therefore, given these several reasons why citizensand even some regime leaders would oppose U.S.


18/73

8

foreign policies, U.S. inuence, and a U.S. presence intheir nations, the rst, general hypothesis I will test is

that the greater the degree of local U.S. involvement,the greater the likelihood of domestic unrest. Theextent and nature of the local U.S. presence will actas a lightning rod for opposition to U.S. policies moregenerally and as a source of contention locally. It is bothemblematic and symbolic of the relationship betweenthe United States and the host government that cangalvanize opposition to the (perceived) negativeconsequences of U.S. inuence, and can serve to inspirethose opposed to U.S. policies more generally. Andwhile greater involvement often also provides greaterlocal benets as U.S. aid and military resources owinto a country, the political and economic negativeexternalities of extensive involvement increase as well.I do not attempt to assess whether, on balance, the U.S.

presence provides greater benets than costs. Rather, Iassume that there are direct and indirect relationshipsbetween the size and nature of the U.S. presence andthe likelihood of local opposition, and ultimatelyunrest. Further, I must acknowledge that the level ofopposition to U.S. foreign policy in any given nationwill not remain constant over time, but will uctuatedepending on the salience and visibility of U.S. actions.Thus, the causal pathways between the U.S. presenceand local, domestic unrest are several and complex.

U.S. Foreign Policy and Regime Stability.

On the other hand, U.S. foreign policy means andends are intended and designed to promote positive

relations and maintain stability in those nations withwhom the United States seeks to foster amicable andcooperative relationships. A strong U.S. presence can


19/73

9

promote multiple positive conditions. First, to theextent that a U.S. presence promotes both internal and

external security for a nation, it provides the protectionand stability a state needs to develop economically andpolitically. U.S. friendship can deter interstate rivalsfrom overtly aggressive behavior and can dissuadeinternal, political rivals from sowing unrest. Second,to the extent a U.S. military presence or U.S. militaryaid alleviates the need for a government to expendresources on its own security, a regime is betterable to utilize freed up resources on economic andsocial development that should further the nationsprosperity. Third, a U.S. military presence and militaryaid can stimulate the local economy and provide jobsfor many nationals who are involved in businesses thatcontract with and supply the U.S. military, and canopen avenues of opportunity for citizens to take part

in educational, economic, and military interactionswith the United States. Therefore, the second, generalhypothesis I test is that the greater the degree of localU.S. involvement, the lesser the likelihood of domesticunrest. I describe below several hypotheses basedon these potential relationships and note where bothpositive and negative effects may result from U.S.foreign policy actions.

HYPOTHESES

U.S. Troop Presence.

The Potential Negative Impact. The U.S. Governmentmaintains a military presence in most nations of the

world. The troops that are stationed on foreign soilrange in responsibilities from small contingents ofMarines that provide diplomatic protection, to themany thousands of U.S. forces stationed in allied states


20/73

10

including Europe, Japan, and South Korea. While theirmissions may vary and change over time, the impact

of the presence of such forces can go far beyond theiractual purposes. However, the U.S. military presencewill likely serve as a lightning rod when the U.S.Government takes unpopular foreign policy positionsand actions. Opposition to U.S. hegemony will likelybe greater the more visible the manifestation of suchinuence in the presence of U.S. military personnel. Asthe foremost symbol of U.S. inuence, a U.S. militarypresence will likely serve as both an inspiration andtarget for local opposition to the United States. Imeasure a U.S. military presence by the number of U.S.military personnel listed as stationed in all nations ofthe world.6 I save for later a lengthy discussion of thedependent variables.

Hypothesis 1a: The greater the number of U.S. forcesstationed in a foreign nation, the greater the level ofdomestic unrest, terrorism and war in that nation.

The Potential Positive Impact. Contrarily, a U.S.troop presence can potentially have many positiveinuences on local political conditions. Some of thelarger deployments can have substantial benecialeffects on the economy by generating growth in localindustries and services to support the U.S. personnel.U.S. personnel make many positive contributions to thesociety and the government through training programs,community involvement, and other endeavors.While not generally charged with inuencing theinternal affairs of these states, the military presence

nonetheless may have important secondary effects onthe likelihood of stability and peace in foreign nations.Local citizens, who live near, work at, or simply follow


21/73

11

developments at American overseas military bases,become cognizant of a variety of American cultural and

political customs. Long-term exposure to these ideasand practices may help inculcate democratic valuesin the populace and lead to calls for positive politicalchange. The U.S. military also sponsors a number ofprograms and classes for foreign military personnel toimpress upon them the importance of civilian controlof the armed forces and other democratic and humanrights values. In order to ensure continued Americanaccess to foreign bases, the U.S. Government alsohas an interest in stable and legitimate governmentsin these nations and may push some regimes towarddemocratization rather than risk less predictable formsof political change (e.g., the Philippines, 1986).

Hypothesis 1b: The greater the number of U.S. forces

stationed in a foreign nation, the lesser the level ofdomestic unrest, terrorism, and war in that nation.

U.S. Military Aid.

The Potential Negative Impact. The provision of U.S.military assistance no doubt wins the United Stateskey friends in foreign regimes and militaries, butmay also provoke anger and resentment on the partof many outside the government who disapprove ofits uses. U.S. military aid may be used to help buttressunpopular repressive regimes; it may free up funds thatregimes can then use to support private military forces,and it may be used to support unpopular wars or otherprograms, such as drug eradication. Ultimately, many

in the population will likely see little direct benetfrom U.S. military assistance and believe that U.S. aiddollars would be better invested in social programs


22/73

12

and other initiatives designed to help the people. Andto the extent that the population perceives the military

assistance program as furthering U.S. dominance,whether locally or globally, its unpopularity mayprovoke dissention and unrest. Scholars have soughtto determine if linkages occur between the decisionto provide foreign aid and the level of such aid, andimprovements in a nations human rights practices anddemocratization. Regan nds in his study of U.S. aid onhuman rights repression in 32 developing nations that. . . U.S. economic aid has had little or no impact on thehuman rights practices of the recipient governments.7Similarly, in a study of the impact of U.S. foreignassistance on democratization, Knack nds that Theevidence presented here does suggest that either thefavorable impacts of aid on democratization are minor;or they are roughly balanced by other democracy-

undermining effects of aid dependence.8

I measureU.S. military assistance using annual data in constantU.S. dollars from the U.S. Agency for InternationalDevelopment (USAID).9

Hypothesis 2a: The greater the level of the U.S.military assistance spending in a foreign nation, thegreater the level of domestic unrest, terrorism andwar in that nation.

The Potential Positive Impact. While most U.S. foreignassistance is given for political, economic, and securityinterests, it has also been used to advance and rewarddemocracy, human rights protections, and other suchgoals that should enhance stability and peace in foreign

nations. Insofar as military aid exposes foreign citizensto U.S. political values and helps to create a civilsociety, the underpinnings of stability and democracy


23/73

13

are encouraged. The net impact of aid in generalaccording to Carothers is . . . usually modestly positive,

sometimes negligible, and occasionally negative.10

Hypothesis 2b: The greater the level of the U.S.military assistance spending in a foreign nation, thelower the level of domestic unrest, terrorism and warin that nation.

Militarized Disputes.

The Potential Negative Impact. Nothing symbolizesU.S. hegemony more than its use of military force toinuence international politics and to effect politicalchange in other nations. While some regimes andgroups may support U.S. military action in somecrises, often the use of force is viewed as evidence of

U.S. heavy-handedness in global affairs. Many regimesand individuals will likely view U.S. military actionsas protective of U.S. national interests rather than localinterests, and believe the United States cares little forthe value of civilian lives in those nations it enters.Thus, even though the U.S. military may be dispatchedto provide order and stability in foreign nations, itmay also precipitate more violence and unrest. Severalscholars are skeptical of the utility of U.S. attempts toenforce its values and practices on other nations, andargue that military force is far too blunt an instrumentwith which to export values that take time, commitmentand resources to grow.11 It is certainly possible as well,however, that there is reciprocal causation occurringbetween the use of military force by the United States

and foreign unrest, for such military operations maybe authorized in response to violence and instability inforeign nations. In order to account for such reciprocal


24/73

14

causation, I lag this variable 1 year. I measure thisvariable as the number of militarized interstate disputes

the United States was involved in with each nation ofthe world.12

Hypothesis 3a: The greater the number of U.S.militarized disputes involving a foreign nation inthe previous year, the greater the level of domesticunrest, terrorism and war in that nation.

The Potential Positive Impact. Throughout U.S.history, the military has been employed on behalfof friendly relations and liberalist ideals, such as thepromotion of democracy and human rights. Presidentsdispatched the armed forces into Central America andthe Caribbean in the early part of the 20th century,as well as the 1980s, to ensure the peace and oversee

elections. More recently, the U.S. military has played amajor role in the democratization process in Haiti andBosnia after civil strife and war tore apart those nations.Indeed, one of the ve major objectives of U.S. militarystrategy in theAnnual Defense Report 2000 is fosteringan international environment in which Democraticnorms and respect for human rights are widelyaccepted.13 Scholars have discovered, however, thatwhile the utility of military force depends on a deepercommitment among U.S. policymakers to regimestability, democratization, and the promotion of humanrights, military operations do inuence the likelihoodof democratic transitions.14 After spending substantialsums of money and incurring a great many politicalcosts in major military deployments, policymakers will

seek to help build friendly and peaceful regimes. Andof all the tools in the U.S. foreign policy arsenal, noneprovides the degree of direct inuence that military


25/73

15

force does. Operations may be designed to facilitatestability, human rights, and democratic transitions

(e.g., Haiti), or compel them (e.g., Germany and Japanafter World War II).

Hypothesis 3b: The greater the number of U.S.militarized disputes involving a foreign nation in theprevious year, the lesser the level of domestic unrest,terrorism and war in that nation.

U.S. Foreign Policy Similarity.

The Potential Negative Impact. Taking a stand infavor of U.S. foreign policy positions is often a riskyundertaking as U.S. actions often have a way ofalienating some people and regimes that do not sharethe U.S. world view. Democratically-elected leaders

in particular may sometimes pay a high price for theirsupport of unpopular U.S. policies as we have recentlyseen with regard to the war in Iraq. To the extent thatthe publicly expressed positions of governments alignthese regimes with U.S. foreign policies, we wouldexpect that opposition to close identication with U.S.interests would increase, and such regimes would bemore susceptible to domestic unrest. To measure thecorrespondence between a nations foreign policypositions and those of the United States, I utilize anindicator of UN General Assembly voting similaritydeveloped by Eric Gartzke, which ranges from -1(representing nations whose voting similarity is leastlike the United States) and 1 (representing nationswhose voting similarity is most like the United

States).15


26/73

16

Hypothesis 4a: The more similar the voting behaviorof a nation in the UN General Assembly to that of

the United States, the greater the level of domesticunrest, terrorism and war in that nation.

The Potential Positive Impact. Contrarily, regimesthat closely align with U.S. foreign policy interests mayenjoy U.S. support and favoritism on a wide varietyof issues. U.S. inuence on international political andeconomic organizations is substantial, and the UnitedStates can use its inuence to help ensure outcomesfavorable to friendly states. Close alignment with U.S.foreign policy interests may also result in a regimereceiving more tangible rewards, such as governmentcontracts, foreign assistance, and other types of largessthat can have a positive effect on regime stability.

Hypothesis 4b: The more similar the voting behaviorof a nation in the UN General Assembly to that of theUnited States, the lesser the level of domestic unrest,terrorism and war in that nation.

Control Variables.

To ensure a properly estimated model, I alsoinclude several control variables that have generallybeen found to exercise a substantial impact on thedependent variablesunrest, terrorism, and war.First, the level of democracy in a society should havea negative impact on these events. More democraticnations provide for outlets for citizens grievancesagainst their governments, such as legal protests,

freedom of speech, freedom of the press, and thepower to change ofce holders through regular andfree elections. These opportunities should diminish


27/73

17

the attractiveness of more violent forms of protestand should increase regime stability and decrease the

incidence of terrorism and war. I use several measuresof the level of democratization from the Polity IVdata base.16 I rst include a measure of the numberof constraints the executive branch of a governmentmust labor under. Previous research has found thatthe more constraints the executive must contendwith, the more difcult it is for the executive to takestrong measures to counter unrest and violence.17I use the Polity IV measure ExConst that is coded1 for unlimited authority; 3 for slight to moderatelimitations on executive authority; 5 for substantiallimitations on executive authority; 7 for executiveparity or subordination; and values 2, 4, and 6as intermediate categories. I also include a measureof the degree of political competitiveness in a regime.

In contrast to the previous variable, we would expectthat, as political competition increases, there is lessneed for individuals to resort to violent methods ofexpressing their dissatisfaction with government.I also use the Polity IV variable measuring the levelof political competition, PolComp, but create apolitical competition variable that is coded 1 forthose nations that receive the highest PolComp score,and 0 otherwise. Political Competition refers to theextent to which alternative preferences for policy andleadership can be pursued in the political arena18 andwhether, Participation is regulated to the extent thatthere are binding rules on when, whether, and howpolitical preferences are expressed.19 Finally, I use ameasure of regime durability. We would expect that

regimes that have been in existence for longer periodsof time would have better experience and institutionscapable of handling dissent and unrest, or channeling


28/73

18

it in appropriate directions. This variable also comesfrom Polity IV and is described in extensive detail in

their codebook.20

It is simply a measure of the numberof years a regime has been in existence.

Second, states with larger populations have beenfound to be more likely to experience various typesof unrest, particularly terrorism.21 The greater the sizeof the population, the more likely it is that there aregroups of people within the society whose grievanceshave not been satisfactorily dealt with. Thus, stateswith large populations should show more signs ofdomestic unrest, experience more terrorism, and bemore likely to be involved in wars. I measure totalpopulation using data from the World DevelopmentIndicators CD from the World Bank. Third, stateswith greater levels of economic development shouldbe more stable. An economy that effectively produces

wealth, and in which there is a reasonable distributionof such rewards, should lessen many citizens potentialfor unrest by providing employment and more tangiblerewards, and by promoting an economically moreprosperous future. Combined, these effects shouldsubstantially detract from the attractiveness of violentaction. I measure economic development using percapita gross domestic product (GDP) from the WorldDevelopment Indicators CD from the World Bank.Fourth, I control for state power. States with a greatershare of power internationally are likely to make moreattractive targets for terrorists because of their greaterability to inuence international politics.22 I measurestate power using the Correlates of War CompositeIndicator of National Capability.23 The Correlates

of War Composite Indicator of National Capabilityencompasses total population, urban population, ironand steel production, and energy consumption (as


29/73

19

indicators of economic size comparable across time),number of military personnel, and defense expenditures,

and is measured as each states percentage of the worldpower total across all these factors. Finally, I control forregional effects by including binary variables for theWestern hemisphere (Latin America, the Caribbean,and Canada); Sub-Saharan Africa; Europe; and Asia.The Middle East is used as the reference category.

Hypothesis 5: The greater the number of executiveconstraints in a regime, the greater the level ofdomestic unrest, terrorism and war in that nation.

Hypothesis 6: The greater the level of politicalcompetition in a regime, the lesser the level ofdomestic unrest, terrorism and war in that nation.

Hypothesis 7: The greater the number of years aregime has been in existence, the lesser the level ofdomestic unrest, terrorism and war in that nation.

Hypothesis 8: The greater a states population, thegreater the level of domestic unrest, terrorism andwar in that nation.

Hypothesis 9: The greater a states level of economicdevelopment, the lesser the level of domestic unrest,terrorism and war in that nation.

Hypothesis 10: The greater a states power, the greaterthe level of domestic unrest, terrorism and war in thatnation.


30/73

20

Measuring Domestic Unrest, Terrorism, and War.

I analyze the impact of the variables described aboveon three separate sets of indicators. First, I look at severalmeasures of domestic unrest: riots, government crises,assassinations, and anti-government demonstrations.These data are measured on an annual basis for allcountries of the world and are from the Banks Cross-National Time Series Data Archive.24 Second, I analyzedata on terrorism from the Oklahoma City NationalMemorial Institute for the Prevention of Terrorism(MIPT) Terrorism Knowledge Database.25 Thesedata include all acts of terrorism, both domestic andtransnational, that occur within a country.26 Finally,I analyze civil and international wars as denedand measured in the Correlates of War Intrastate

and Interstate data bases.27

Civil wars are dened asdisputes where military action was involved, at least1,000 battle deaths resulted during the civil war, thenational government at the time was actively involved,and there was effective resistance (as measured by theratio of fatalities of the weaker to the stronger forces).28International wars are dened as disputes betweentwo or more members of the international state systemin which there are at least 1,000 battle deaths. Alldependent variables are measured on an annual basisfor all nations of the world for which there are data.All independent variables are also measured on anannual basis, but are lagged 1 year to account for thelikelihood that the impact of such variables will not beinstantaneous, but will take some amount of time to

manifest.


31/73

21

METHODOLOGY

To address and answer the questions and hypothe-ses posed above, I will use several statistical techniquesto analyze these data. For the assessment of the factorsthat predict regime instability and terrorism, I usea statistical technique known as negative binomialregression. Because the data on regime instability andterrorism are counts of events (riots, government crises,anti-government protests, assassinations, and acts ofterrorism), use of ordinary least squares regressionis inappropriate as its results would be inefcient,inconsistent and biased.29 I use robust standard errors toaddress the likelihood of heteroskedasticity among theerror terms as has been done in many other studies.30

Analyzing Acts of Terrorism.

The results of the model estimating the impact ofthe variables described above on acts of terrorism arepresented in Table 1. I will address the impact of theindependent variables rst and then discuss the overallt of the model, its ability to predict acts of terrorism,and analyze which nations are most at risk for suchactions. Most of the coefcients for the independentvariables are statistically signicant at the .05 level.I note rst that all indicators of U.S. foreign policyrelationsthe size of a U.S. military presence, if any;the amount of military aid provided to a regime; andinvolvement in a militarized dispute with the UnitedStates, are positive. Close ties with the United States, asevidenced by these indicators, are correlated with risk

for increasing numbers of acts of terrorism, with oneexception. I stress again that we cannot assume thatthere is a causal relationship at work here, given thelimitations of all statistical models. We can, however,


32/73

22

make qualied inferences regarding the nature of theserelationships. I use the incidence rate ratio to interpret

the statistical effect of the variables. This measure tellsus the impact of a variable on the number of terroristactions holding other variables constant at their meanvalue. Thus, the incidence ratio for the variable U.S.Military Presence is approximately 1.0, which tellsus that for every unit increase in the number of U.S.troops deployed in a nation, all other things beingequal, the model predicts that the incidence rate ofterrorist actions increases by a factor of .000002 percent,which is to say, very little. However, if the metric of aU.S. military presence we use is 1,000 troops insteadof 1, the predicted impact is .002 percent. Therefore,we would conclude that while there is a statisticallysignicant relationship between the size of a U.S.military presence and acts of terrorism, the substantive

impact is quite small.

Table 1. Predicting Terrorism Across States,1953-2003.a


33/73

23

We nd a similar relationship existing between

U.S. military aid to regimes and the expected levelof terrorist activity. Here the incidence rate ratio is1.000594. For every $1 million increase in U.S. militaryassistance ceteris paribus (other things being equal), thepredicted level of terrorist actions in a nation increasesby a factor of .000594 percent. Here we see that it takes asubstantial, but not altogether rare, increase in the levelof military aid to truly impact the number of terroristattacks. When aid increases by $100 million, the modelpredicts a signicant likelihood of an upsurge interrorist activityterrorism would increase by a factorof 5.94 percent at this level of military assistance.

There is a slightly different story, however, whenwe examine the impact of prior U.S. involvement inmilitarized disputes with a nation. Here we see that

with each additional militarized dispute the modelpredicts an increase in the frequency of terrorist actionsby a factor of 1.49, or approximately 49 percent. Giventhat there have been several instances in which theUnited States has used force within or against a nationin a given year, the potential for a signicant andsubstantial rise in the level of terrorist activity is quitereal. As hypothesized above, these types of militarizeddisputes may iname relations between the UnitedStates and the regime in question, or catalyze terroristgroups into attacking the regime (if the regime isperceived as being too close to the United States, orU.S.-related individuals, businesses, installations, ororganizations in that nation). Thus, there are clearpolicy implications here. Subsequent to U.S. military

actions occurring within or against other nations, basedon the models estimates, there would seem to be asubstantial likelihood of increasing terrorist actions inthose states.


34/73

24

Interestingly, the model shows that states whoseforeign policies are similar to those of the United States,

as reected in the closeness of their voting patternsin the UN, are less likely to experience terrorism. Asvoting similarity increases on this -1 to 1 scale bya factor of 1, the incidence of terrorism declines by afactor of 1 percent (1the incidence rate ratio of .99),ceteris paribus. Since a change of one full point on thisscale is quite rare, (except perhaps in cases where a newregime comes to power that is diametrically opposed to[or supportive of] the United States where the previousregime exhibited just the opposite behavior), this impactfactor is not necessarily meaningful. If, however, therewere a change on the order of a .1 increase, ceteris

paribus, we should expect to nd a corresponding 0.1percent decrease in terrorist activity.

Together, these rst results would seem to suggest

that while a close relationship with more tangibleaspects of U.S. foreign policya military presence,military aid relationship, and involvement in U.S.militarized actionstends to be associated with anincrease in the risk of terrorism, a similarity in foreignpolicy orientation toward the world per se does not seemto enhance the probability of terrorist activity within astate. Terrorists may well be inspired and galvanizedinto action by the more manifest military policies theUnited States engages in rather than some of the moreabstract and removed policies at issue at the UN. Imust caution again, however, that these estimates donot prove that a states close foreign policy relationshipwith the United States directly causes terrorism. Allthat we may reasonably conclude is that those states

that do maintain such ties with the United States areat greater risk for increasing levels of terrorist activity.The long-term and structural factors that give rise to


35/73

25

the conditions that make terrorism more likely mayinvolve the nature of the regimes ties to the United

States, but they may also reect the regimes domesticand international policies that are correlated with bothcloseness to the United States and terrorist activity.

As Li nds, the impact of democracy on predictedlevels of terrorism depends on the extent to whichthere are constraints placed on the executive, thecompetitiveness of the electoral process, and regimedurability.31 I had argued, as have other researchers, thatdemocratic societies provide for more opportunities forindividuals to present their grievances in a nonviolentmanner and so ought to experience fewer incidentsof terrorism.32 Thus, we nd that there is a negativerelationship between electoral competitiveness andterrorism, albeit one that is statistically insignicant.On the other hand, executives that face a great many

constraints on their power, as in more democraticsocieties, are handicapped in their ability to preventterrorism. Totalitarian or authoritarian regimes, bycontrast, through deeper and broader governmentmonitoring of individual behaviors, are able to preventsuch individuals from coalescing into effective actiongroups. For example, one of the few nations for whichthere is no record of any terrorism in the period understudy is North Korea, one of the most repressivepolice states in the world. Thus, we see that the moreconstraints placed on the executive branch, the morelikely that nation is to experience terrorist actions. Aone-unit increase in the measure of executive constraintsis associated with a 1.16 factor increase in terroristactivity, all other things being equal. Regimes that have

existed for longer periods of time are statistically lesslikely to experience terrorist attacks. For every year aregime has been in existence, terrorist attacks decrease


36/73

26

by a factor of 2 percent, which is not large, but stillindicates that longer-lasting regimes tend to be more

stable and capable of addressing grievances that mightlead to terrorism, or addressing through their criminaljustice systems the potential for terrorist activity.

Neither a strong economy (as evidenced by percapita gross domestic product [GDP]) nor a largepopulation appear to exercise any kind of meaningfulimpact on the incidence of terrorism in a given country.The coefcients for both variables are statisticallyinsignicant. Some of their impact, however, may beabsorbed by the state power variable. The incidencerate ratio for this variable is statistically signicant andpositive, which indicates that more powerful states aremore likely to experience terrorist violence. As I arguedabove, such states are attractive targets for terroristsbecause their power and inuence in international

relations are not only more likely to incur the wrathof terrorists and like-minded individuals that viewthemselves and others as oppressed by such powers,but also because attacks against these nations are alsomore likely to generate tremendous publicity, therebyfurthering the terrorists cause(s).33 The incidencerate ratio indicates that for every one unit increasein a states share of global GDP, terrorism increasesby a factor of 28 percent. I stress again, however,that it is not possible to claim that there is a causalrelationship between state power and terrorism thatcan be identied from this analysis. As well, most ofthese more powerful nations tend to be Western states,against whom much terrorism is directed because ofcultural and other differences. State power certainly

plays a role as an underlying factor that makesconditions ripe for the emergence of terrorist activity.Yet, just as the collision of two weather fronts does not


37/73

27

automatically produce storms, but creates a conduciveset of environmental conditions, so, too, does state

power create an underlying and facilitating conditionthat makes terrorism more likely. Thus, even though wecannot then predict precisely when and where terroristactions will occur in powerful states, we know at leastto look for such events to occur where the conditionsare most favorable, and not in other states where suchconditions are absent.

The coefcients for the regional variables are allstatistically signicant and negative, which is exactlyas I expected. Because the Middle East, which tends toexperience more than its share of terrorists incidents,is the reference category, the other regions of theworld look more peaceful in comparison. In particular,those nations in the Western hemisphere, (with theexception of the United States, which is not included

in the analysis), Asia and Sub-Saharan Africa are farless likely to experience acts of terrorism.

The next stage in the analysis is to evaluate thepredictions of the model to determine which nationsare predicted to be most likely to experience acts ofterrorism. Using the models parameters, I am able togenerate an annual count of the predicted number ofacts of terrorism, which can be compared against theactual number of such incidents. The vast majorityof the nations of the world are not predicted toexperience more than one act of terrorism per year, soI focus instead on those nations that are predicted toexperience three or more such actions in a given year.Those nations are listed in Table 2 in descending orderof frequency. Note that the number associated with

each nation is the number of years each nation is predictedto experience at least three acts of terrorism. According tothe models estimates of the number of years a nation is


38/73

28

predicted to experience three or more acts of terrorism,four nations in particular stand out. Japan is predicted

to experience multiple acts of terrorism in 33 of theyears under study; Israel is predicted to experiencemultiple acts of terrorism in 29 of the years understudy; the Peoples Republic of China is predicted toexperience multiple acts of terrorism in 26 of the yearsunder study; and Turkey is predicted to experiencesuch violence in 22 of the years under study. There arethen several nations the model predicts to experiencefewer, but still multiple years in which there are at leastthree acts of terrorism: Egypt (in 13 of the years understudy), Germany (in 12 of the years under study), andIran (in 9 of the years under study). Then there areseveral nations that the model predicts to experience5 or fewer years in which there were at least three actsof terrorism including: Sudan, South Korea, France,

Kuwait, and India. As can be seen when comparing theactual number of terrorist incidents with the predictednumber, the model tends to overestimate the numberof such attacks. Given that a nation ts a prole of thetype of nation most likely to experience multiple actsof terrorism (because many of the characteristics thatpredict such actions tend not to change dramaticallyfrom year to year), such nations are predicted to be atrisk for multiple acts of terrorism in many years. Thus,the model tends to err on the side of over-predictingterrorist actions.


39/73

29

Table 2. Terrorism Across Time in Specic States.

Acts of terrorism may often be the nal culmination

of domestic unrest, protest, and violence. Whilecertainly not all countries that are characterized by suchconditions can be expected to subsequently experienceterrorist violence, these problems may well help set thestage. Therefore, I next analyze the incidence of fourtypes of domestic unrest to determine which nationsare most susceptible to this kind of instability andthe impact exercised by U.S. foreign policy actions onthese indicators. The types of unrest I examine are:riots, government crises, anti-government protests,and assassinations. I assess the impact of each ofthe independent variables across all four of thesedependent variables. I again use the negative binomialPoisson model to estimate these models.


40/73

30

Analyzing Riots, Government Crises, Assassinations,and Anti-Government Demonstrations.

I turn rst to examine the effect of the indicatorsof foreign policy relationships with the UnitedStates. The presence of U.S. military forces stationedin foreign states has a mixed effect across the fourmodels. The incidence rate ratio for this variable isstatistically signicant and positive in the estimates ofthe number of government crises and assassinations,but statistically insignicant in the models of riots andanti-government demonstrations. Interestingly, theeffect of the U.S. military presence is greater on the twoindicators of what we might consider to be elite unrestand instability. Government crises may have deep andwide societal implications, but they often most directlyinvolve regime leaders. Assassinations may involve

government, militia, or other disaffected groups, butthe targets are generally chosen for their high proleelite status. Taking this logic a step further, but asalways bearing in mind the difculties associated withmaking direct, causal inferences, it might be that aU.S. military presence may lead to political difcultiesfor those in authority, but it may not always inspireprotesters in the streets. The substantive impact, aswe saw in the model of terrorist actions, however, isstill rather slight even when the incidence rate ratiois statistically signicant. A rather sizeable increasein the number of U.S. troops stationed in a foreigncountry is required before the predicted incidence ofthese instability indicators will increase.


41/73

31

Table 3. Predicting Riots Across States, 1953-2003a

Table 4. Predicting Government Crises AcrossStates, 1953-2003.a


42/73

32

Table 5. Predicting Assassinations Across States,1953-2003.a

Table 6. Predicting Anti-GovernmentDemonstrations Across States, 1953-2003.a


43/73

33

The incidence rate ratio for the variable measuringthe degree of a military aid relationship between a

foreign regime and the United States is statisticallysignicant and positive in all but one of the models.Greater levels of U.S. military assistance provided to aregime are associated with a greater incidence of riots,assassinations, and anti-government demonstrations,ceteris paribus. The impact of this variable is alsosubstantial across all three of these models. A $100million increase in U.S. military assistance is associatedwith a factor rate increase of 8 percent in the number ofriots; a factor rate increase of approximately 5 percentin the number of assassinations; and a factor rateincrease of 7 percent in the number of anti-governmentdemonstrations.

Some words of caution are in order, however, forunderstanding the nature of the relationship between

U.S. troop levels, military assistance, and regimeinstability. Those regimes to which the United Statesprovides signicant levels of military assistance,as well as those nations wherein the United Statesstations a sizeable number of military personnel, aremostly states with whom the United States sharesongoing, close ties. Thus, aid levels and the numberof military personnel stationed in these countries arenot likely to uctuate widely from one year to the next.Thus, while aid and troops are not constants, there arerelatively long-lasting features of the U.S. relationshipwith certain regimes. As such, their impact may beregistered in a more subtle and fundamental way thatthe statistical analysis cannot always detect. Thus,while we do nd evidence of a positive correlation in

several cases, we must be mindful of the difculties ofuntangling the true causal nature of the relationship. Ireturn to this point in the conclusion.


44/73

34

When the United States is involved in a militarizeddispute with a nation the year prior, the incidence

of domestic unrest in that nation does not appear tobe substantially affected. None of the indicators ofdomestic unrest were statistically related to theseincidents. American involvement in these disputes cancertainly affect political developments in the targetstates, but this inuence does not appear to extend tothe level of (in)stability in a society.

Contrarily, we see evidence of a strong relationshipbetween the extent to which a government votes ina fashion similar to the United States in the GeneralAssembly of the United Nations and the frequencyof riots, government crises, assassinations, and anti-government demonstrations, although the incidencerate ratio slightly misses statistical signicance in thelast type of unrest. In fact, a 10 percent rise in vote

similarity between a foreign regime and the UnitedStates increases the incident rate ratio by 5 percentin the estimates of riots; 9 percent in the estimates ofgovernment crises; and 3 percent in the estimates ofassassinations. While we cannot rule out the possibilitythat it might be the United States whose voting patternsare most closely resembling those of other states, giventhe relative consistency of U.S. foreign policy overtime, its regime stability, and its enduring hegemonicinterests, it is much more likely the case that otherstates whose foreign policies can and do shift morefrequently based on international conditions, andwhose regimes and constitutions are more at risk ofupheaval, change their foreign policies to move closeror farther away from U.S. foreign policies. Thus, the

United States may well nd itself in a conundrumbecause of these trends. On the one hand, it would liketo encourage states to shift their policies more to its


45/73

35

liking, yet at the same time the United States certainlydoes not want to be forced to address the domestic

unrest that might occur when a regime substantiallyand dramatically changes its foreign policies to alignthem with those of the United States. Many sectors inthese societies may have strong anti-American viewsthat are enamed by close ties with the United States,which then manifest themselves in domestic unrest.Indeed, even those nations that host U.S. militaryforces and receive substantial amounts of U.S. militaryaid often take positions contrary to American interestsin order to demonstrate their independence from theUnited States. For example, in the Middle East it maybe difcult enough for a nation to host U.S. forceswithout incurring the wrath of the anti-Americanists.If such a nation, or any nation in the region, took theU.S. position on an issue involving Israel in the UN, the

likelihood of unrest would likely increase dramatically.Symbolic politics, of just the sort that occur in the UN,can often generate as much instability and violence asmore tangible elements of politics in certain parts ofthe world on certain issues.

As we would expect, regime type plays a criticalrole in predicting a states propensity to suffer domesticunrest. In all models, the extent to which the executiveis constrained in the exercise of power is positivelyrelated to riots, government crises, assassinations,and anti-government demonstrations. The incidencerate ratio for this variable is statistically signicantand powerful in all four models. The frequency ofriots increases by a factor of 12 percent with everyunit increase in executive constraints; the incidence of

government crises increases by a factor of 18 percent;the frequency of assassinations increases by a factorof 21 percent; and the incidence of anti-governmentdemonstrations rises by a factor of 12 percent for every


46/73

36

unit increase in executive constraints. Strong executivesare able to clamp down on dissent or the potential for

dissent.On the positive side for democratic states, we

see that the greater the competitiveness of elections,the lower the level of all forms of domestic unrest.I nd that the frequency of riots decreases by afactor of 61 percent, with a unit increase in electoralcompetitiveness; the incidence of government crisesdecreases by a factor of 36 percent; the frequency ofassassinations decreases by a factor of 71 percent; andthe incidence of anti-government demonstrations fallsby a factor of 74 percent in states where there is fullelectoral competitiveness. Hence, while the constraintsdemocratic regimes typically require executives tooperate under are associated with increases in domesticunrest, the competitiveness of their electoral systems

has the opposite impact. It is also important to note thatthe impact of political competitiveness on reducinginstability is greater than the impact of executiveconstraints in discouraging instability. Presumablyfree and fair elections allow citizens to express theirvoices and opposition in such a way that energies thatmight have been directed toward more violent formsof protest are channeled into peaceful and healthydemocratic practices and discourse.

Of course, elections can also be a mixed blessing.Elections held too early in a critical period of transition,such as in the aftermath of war, may harden politicaland sectarian cleavages and precipitate more violence,especially from the losing side(s). Thus, it is alsoimportant to consider the longevity of the regime. All

things being equal, I argue that more established andlonger-lasting regimes exhibit greater effectiveness ataddressing citizens aspirations, thus allowing themto survive longer. The results tend to bear out this


47/73

37

supposition. The longer a regime has been in existence,the less likely it is to experience domestic unrest, ceteris

paribus. The incidence rate ratio for regime durability isnegative and statistically signicant in two of the mod-els. The frequency of government crises and assassina-tions decreases by a factor of 1 percent in each model.The incidence rate ratio is statistically insignicant inthe riots and government demonstrations models.

The impact of population size on the indicators ofcivil unrest is mixed. I nd that the larger the populationsize, the more infrequent are riots and anti-governmentdemonstrations, while the incidence rate ratio isstatistically insignicant in the cases of assassinationsand government crises. Yet, the effect of increasingpopulations is so minuscule that its substantive impactbarely registers. Large states may be advantaged tosome very small degree perhaps because collective

action problems may be more difcult to resolve withlarger numbers of potential rioters. On the other hand,one would assume there would be more opportunitiesfor civil unrest in large states simply because thelikelihood that all citizens would not engage in suchbehavior would diminish with size.

Per capita GDP is negatively related to the frequencyof riots, government crises, assassinations, and anti-government demonstrations, although the incidencerate ratio for this variable in the government crisismodel is not statistically signicant. As the economicwell-being of a nation becomes healthier and stronger,there is less reason to engage in such forms of domesticunresta prosperous citizenry is a content citizenryby and large (although we must also recognize that

in many of the wealthiest countries in Europe andelsewhere, there will be those who, because of theirsocietys prosperity, have the time to engage in the


48/73

38

sorts of protests we normally see when the G-8, theWorld Bank, or the International Monetary Fund have

a gathering). Additionally, when a nations economy isrunning smoothly and people are working hard, thereis much less opportunity to engage in violent formsof protest. Contrarily, those nations in which largenumbers of citizens are unemployed, especially youngmen, there is ample opportunity for such protests, andultimately violence.

Finally, I note the impact of the regional variableson domestic unrest. There is a positive relationshipbetween the Western hemisphere dummy variableand the frequency of all forms of domestic unrest,although the incidence rate ratio is statisticallyinsignicant in the government crises model. Europeis statistically more likely to experience riots andanti-government demonstrations. Asia is unlikely to

experience government crises or assassinations, but isstatistically more likely to experience anti-governmentdemonstrations. Africa is unlikely to experiencegovernment crises or assassinations. Thus, if onewere searching for a stable part of the world that ismost unlikely to experience these forms of unrest andviolence, the nations of Sub-Saharan Africa would seemto be least inclined toward these particular measuresof domestic unrest, while Latin America appears to bethe most prone to various forms of domestic instability.Does this mean that there is little to fear in Africa andmuch to fear in Latin America, the Caribbean, andCanada? Not exactly, it may be that the democraticfreedoms that are enjoyed by citizens residing in theWestern hemisphere protect them from terrorism, as

we saw above, but also enable the masses and elitesto express their grievances in other inappropriate, butless feared ways.


49/73

39

The nal step in the analysis of domestic unrestinvolves further investigation of the predictions of

the model. From each of the four sets of estimates, Igenerate the predicted number of events at issue andlook to see which nations are estimated to suffer fromthe greatest number of these incidents in the overallperiod 1998-2003. I focus on the period from 1998through 2003 (the last year for which we have completedata on all independent variables) since it is the mostrecent time period. Beginning with the predictions forthe number of riots in Table 7, I nd that the modelpredicts the following nations are most likely to besusceptible to multiple riots in multiple years in thisperiod: South Korea, China, India, and Brazil. Chinaespecially is fairly consistently predicted to be the siteof numerous riots. Brazil, China, and India in particularare large, populous nations with diverse groups of

citizens and their economies are rapidly improving,but much of their wealth has yet to lter down to theurban and rural poor. The enormous concentrations ofpoor people in the large cities of Brazil and India inparticular would seem to pose a number of potentialpolitical problems for these regimes. Both states havesignicantly weaker executives than China, whichincreases their predicted likelihood of experiencingriots. The nal state to address, South Korea, has beenthe scene of several riots and mass demonstrationsthroughout its history. It also has received a signicantamount of U.S. military assistance over the years, andof course, hosts a sizeable contingent of U.S. militarypersonnel. It too, continues to be at substantial riskfor riots in the coming years. I note, however, that

there is a signicant divergence between the numberof predicted riots and the number of actual riots inthese nations. The model tends to over-predict riots


50/73

40

in all cases. Among the states predicted to experiencethe greatest number of riots, only Brazil in 2 years and

China in 1 year actually experienced any such violence.I caution, however, that we are most interested in theimpact of the particular coefcients rather than theoverall t of the model, given that we cannot hope toinclude all the relevant factors that would explain riotsin every nation of the world for over 30 years.

Table 7. Predicted Number of Riots in MostRiot-Prone States.

Because government crises in any given nation arerelatively infrequent events, I calculated the averagenumber of predicted government crises across the1998-2003 period and rank-ordered the nations by theirpropensity to experience such events in Table 8. SouthKorea also has the distinction of having the highestnumber of predicted government crises in the 1998-2003 period. Again, because these events are quite rare,their predicted numbers for any nation never exceed a

fraction. Thus, and for example, the predicted numberof government crises for South Korea is .577 over theperiod 1998-2003. Behind South Korea, we nd thefollowing nations in order of their predicted number of


51/73

41

crises: India, Moldova, Bulgaria, Estonia, DominicanRepublic, Poland, Nicaragua, Slovakia, and Romania.

A comparison of the predicted number of governmentcrises and the actual number of such events revealsthat the models ability to predict such events acrossall nations over time has improved relative to the othermodels. Several of the nations that are at higher riskfor such events have experienced government crises,including Israel, Haiti, Romania, and India.

Table 8. Predicted Number of Government Crisesin Most Crisis-Prone States.

Using the same technique described at the beginningof this paragraph, I now examine those nations themodel predicts to be most at risk for assassinations.The nations that are predicted to be most susceptibleto this form of political instability are: South Korea,

Brazil, China, Haiti, Nicaragua, Mexico, the DominicanRepublic, Ecuador, Paraguay, and Guatemala. Thus,eight of the top ten nations most likely to experience


52/73

42

assassinations are in the Western Hemisphere, whichmirrors the positive coefcient for the Western

Hemisphere variable in this model. Indeed, as onemoves further down the list, one nds even more U.S.neighbors at risk for this type of violence. Interestingly,the one nation we do not nd in this predicted riskcategory is Panama, which had long been host to a largeU.S. military presence. The model correctly predictsthat many of the countries mentioned above are at agreater risk for experiencing assassinations, such asHaiti, Guatemala, Ecuador, and Mexico. However, aswe have seen in other such comparisons, the modeltends to over-predict these events for China and SouthKorea.

Table 9. Predicted Number of Assassinationsin Most Assassination-Prone States.


53/73

43

The group of countries most at risk for experiencinganti-government demonstrations (Table 10) largely

mirrors those most likely to suffer riots. China, India,South Korea, Brazil, and also Japan are predicted to bemost at risk for these types of events. The model doesquite well at accurately predicting anti-governmentdemonstrations in India, Brazil, and China, and not aswell in Japan and South Korea. Indeed, the latter twostates are, as we have seen before, home to sizeablenumbers of U.S. forces that tend to place them atgreater risk for these types of events. Yet, despite theirenhanced risk, both states have not been experiencingthese kinds of events in recent years although theyhave in the period prior to 1998-2003.

Table 10. Predicted Number of Anti-GovernmentDemonstrations in Most Anti-Government

Demonstration-Prone States.


54/73

44

Analyzing Civil Wars.

I turn next to an examination of the determinants ofcivil or intrastate wars and the impact of U.S. foreignpolicy actions on these conicts. The dependent variableis measured 1 for every year a civil or intrastate war, asmeasured by the Correlates of War project, is occurringin a given country. In this and the next analysis, I makeuse of the statistical technique known as probit, whichis specically designed to model binary dependentvariables. I use robust standard errors to control for theeffects of heteroskedasticity, or unequal variance acrossnations. The results indicate again that the greaterthe size of U.S. military presence in a nation and thegreater the amount of U.S. military aid, the more likelya nation is to experience a civil war in the followingyear. As we have seen in earlier estimates, however,

the impact is rather slight. I utilize the marginal effectsof the independent variables to interpret their impacts.The marginal effect is the increase in probability ofobserving the event of interest (a civil war) given a unitincrease in the independent variable while holdingall other variables constant at their mean value. Anincrease of 10,000 troops stationed in a foreign countryis associated with only a .07 increase in the predictedprobability of a nation experiencing a civil war, while a$100 million increase in military assistance only raisesthe probability of war by .003 percent. Involvementin a militarized dispute with the United States in theprevious year, however, tends to increase the predictedprobability of a civil war by 4 percent. While this effectis not enormous, we must remember that civil wars

are still comparatively rare across the world, and soeven an increase of 4 percent can have an importantimpact on the likelihood of such conict. A similarforeign policy outlook with the United States does not


55/73

45

appear to play a role in predicting intrastate war. Thecoefcient for this variable is statistically insignicant.

Table 11. Predicting Civil Wars Across States,1953-2003.a

We also see that an important role in predicting civilwar occurrence is played by regime characteristics.Those states in which there are more constraintsplaced on the exercise of power by the executive arestatistically more likely to lapse into civil war. Withevery unit increase in such constraints, the predictedlikelihood of intrastate war occurring in a nation risesby 1 percent. A stronger executive branch should be

better able to act more quickly and decisively to stopdomestic unrest from spiraling into open warfare,or failing that, to take actions to stop such warsfrom lasting for long periods of time. We also see


56/73

46

that political systems that permit greater electoralcompetition are less likely to experience civil wars. In

political systems where there are full and free electoralcompetitiveness, the predicted likelihood of intrastatewar occurring in a nation diminishes by 7 percent,which is quite substantial. Regime durability is notrelated to the likelihood of civil war. Thus, these resultsspeak rather clearly: Regimes with strong executivesand strong electoral competition are best poised toprovide the opportunity for vigorous leadership, buttheir legitimate and open avenues for healthy politicalcompetition provide effective outlets for grievances.

States with less developed economies, and largestates, are at greater risk for experiencing civil wars.The smaller a states per capita gross domestic product,the greater the predicted probability of civil war.For every $1000 decrease in per capita GDP, there is

a corresponding .2 percent increase in the likelihoodof civil war involvement. For every increase of onemillion people in a nation, the probability of civil waroccurrence increases by .02 percent. I stress again thatwhile these numbers are not large, their cumulativeimpact on the likelihood of the occurrence of such arare event can be important. I also note that a statesshare of international power is negatively associatedwith civil war. Simply put, more powerful states are notsusceptible to the same sorts of threats to their powerand status as are other states. For every percentagepoint increase in international power, the predictedlikelihood of civil war occurrence declines by 2 percent.Lastly, I note that all of the regions listed, includingAfrica, are less likely to experience civil wars. Rather, it

is the Middle East which forms the reference categoryfor a region that is the area most likely to experiencethis form of warfare, as it is the case with so many otherindicators of instability and violence.


57/73

47

Table 12. Predicted Probability of Civil WarOccurrence By Year.

Table 13. Predicted Probability of Civil WarOccurrence from 1998-2003.

I next describe which countries appear to be mostsusceptible to civil war violence. I rst calculated thepredicted probability that a nation would experiencea civil war and selected out just those states whoseprobability was 25 percent or greater in any given yearsince 1998. I then examined their average probabilitiesof experiencing civil wars across the years 1998-2003. Ifocus on the period from 1998 through 2003 (the last yearfor which we have complete data on all independentvariables) since it was the most recent period. The

states that exhibited the greatest predicted likelihoodof such violence were Turkey (an average of 22 percentacross 1998-2003 and reaching a high of 29 percent in


58/73

48

1999); Egypt (an average of 26 percent and reachinga high of 32.8 percent in 2000); Israel (an average of

20 percent across 1998-2003 and reaching a high of33 percent in 1999); China (an average of 25.7 percentacross 1998-2003 and reaching a high of 33 percent in2000) and India (an average of 52 percent across theentire period and reaching a high of 52.5 percent in2000). I would note that both Egypt and Israel receivesubstantial amounts of U.S. military assistance, whichmay help explain why their predicted probabilitiesare fairly high. As well, based on their governmentpolicies, the number of dissident and terrorist groupswith grievances against them, in addition to the usualarray of conict-producing politics in the Middle East,we should expect that these states would be primecandidates for intrastate war.

Of those nations that the model predicted would

experience civil wars in this time frame, it correctlypredicts that in 1998, 1999, and 2000 India wouldbe involved in such conict. The model incorrectlypredicted that Turkey would not experience suchviolence in 1999, when it did. The model registers anincreased likelihood of civil war for China, Israel andEgypt, but the predicted probability does not exceed.50 and so the model does not generate a predictionof civil war violence. Overall, the model accuratelypredicts 91.6 percent of the cases correctly for a slightimprovement over predicting the modal category ofno civil war (91.4 percent) in all cases.

8/6/2019 U.S. Fore

u.s. foreign policy and regime instability

Documents