After a decade and a half of counterinsurgency operations in Iraq and Afghanistan, U.S. policymakers want to change their approach to COIN by providing aid and advice to local governments rather than directly intervening with U.S. forces. Both this strategy and U.S. counterinsurgency doctrine in general, however, do not acknowledge the difficulty of convincing clients to follow U.S. COIN prescriptions. The historical record suggests that, despite a shared aim of defeating an insurgency, the United States and its local partners have had significantly different goals, priorities, and interests with respect to the conduct of their counterinsurgency campaigns. Consequently, a key focus of attention in any future counterinsurgency assistance effort should be on shaping the client state's strategy and behavior. Although it is tempting to think that providing significant amounts of aid will generate the leverage necessary to affect a client's behavior and policies, the U.S. experience in assisting the government of El Salvador in that country's twelve-year civil war demonstrates that influence is more likely to flow from tight conditions on aid than from boundless generosity.


After nearly fifteen years of war in Afghanistan and Iraq, the United States is seeking to reorient its approach to counterinsurgency (COIN). Rather than intervening with U.S. forces, the United States wants to focus on supporting local governments' counterinsurgency efforts with offers of aid and advice.1 Without U.S. troops on the ground capable of independent action, success will depend on the policies and choices of the client government. Herein lies the problem. This approach, and U.S. thinking on COIN in general, does not acknowledge the difficulty of convincing local governments to follow U.S. counterinsurgency prescriptions. Despite a shared aim of defeating an insurgency, the historical record suggests that the interests and priorities of the United States and a local partner can diverge significantly. Maintaining power is frequently a competing priority for the incumbent regime, making many of the standard reform prescriptions—streamlining the military chain of command, ending patronage politics, engaging in economic reform, or embracing disaffected minority groups—as threatening to the besieged government as the insurgency itself.2 Although analysts in the United States have spent much time and effort in recent years trying to divine the key to success in counterinsurgency, they have not given the same level attention to understanding when a local government will follow U.S. guidance.

The Barack Obama administration's 2015 decision to expand the training of Iraqi forces fighting the Islamic State gives added salience to the question of how to shape the counterinsurgency policies of a local government that the United States is supporting.3 This problem, however, has plagued U.S. counterinsurgency assistance efforts for more than five decades. Writing before the “Americanization” of the Vietnam War, the senior adviser to the South Vietnamese I Corps warned in his end-of-tour report that “the development of techniques and means to increase U.S. leverage in Vietnam is the single most important problem facing us there and it will be a fundamental problem in any future counterinsurgency effort in which we become involved.”4 Since then, numerous critics have noted that despite providing partner governments with huge amounts of money and material to support their counterinsurgency operations, the United States has displayed an inability to convince them to follow its counterinsurgency doctrine or address what it considers to be the political and economic “root causes” of their insurgencies.5

The failure to recognize the problem of divergent interests between patron and client stems, in part, from the general assumption in the counterinsurgency literature that counterinsurgent forces are unitary actors. Consequently, scholars have not integrated issues of alliance behavior into their studies of the dynamics of revolutionary and counterrevolutionary war.6 Focusing on indirect interventions to support local governments with aid and advice, this article examines a patron's ability to shape a client state's strategy and behavior in counterinsurgency.

The first section of this article uses agency theory to explore the potentially divergent preferences of patrons and their local clients in a counterinsurgency support effort. The second section examines the relative leverage generated by two distinct aid-giving strategies: inducement, which assumes that the unilateral provision of assistance to a client, coupled with strong public statements of support, will be reciprocated by compliance with a patron's preferred policies; and conditionality, which tries to shape the client's behavior by making delivery of assistance contingent on a client's prior implementation of a patron's preferred policies. The third section analyzes the relative utility of these two strategies in a case study of U.S. support to the government of El Salvador during that country's twelve-year civil war (1979–92), which is based on extensive archival research, including recently declassified materials. The fourth section evaluates sixteen discrete influence episodes during the civil war and concludes that the client government complied with U.S. preferences when the United States attached conditions, but not when it provided inducements. It also examines the challenge of making credible threats to withhold aid from clients being supported in COIN. The conclusion summarizes the findings and offers some recommendations for effectively employing conditionality in future counterinsurgency assistance efforts.7

The Trouble with Allies in Counterinsurgency

Much of the United States' experience with counterinsurgency has involved assisting another government in combating an insurgency. Yet, the challenges entailed in working with a partner nation are not widely recognized in the counterinsurgency literature. Classical counterinsurgency theorists, such as the influential British counterinsurgency expert Robert Thompson, have implicitly assumed that the preferences and goals of the host-nation government and its supporting partner will be closely aligned.8 This belief is reflected in the 2006 U.S. Army/Marine Corps counterinsurgency field manual, FM 3–24, which shaped the thinking of U.S. political and military leaders on civil wars for nearly a decade.9 The manual asserts that “the primary objective of any COIN operation is to foster development of effective governance by a legitimate government.”10 Toward that end, U.S. forces were enjoined to build trust with host-nation authorities and work closely with them to enhance their legitimacy by undertaking reform and responding to popular grievances. No suggestion is made that these goals might not be in the interest of the ruling government.11

In recent experience, however, the United States' local partners in Iraq and Afghanistan have often sought to subvert U.S. counterinsurgency efforts. An inability to restrain Iraqi Prime Minister Nouri al-Maliki's sectarian agenda prevented the military gains from the 2007 surge from being translated into positive political outcomes and laid the foundation for the rise of the Islamic State.12 In Afghanistan, President Hamid Karzai's use of patronage politics was seen by outsiders as a form of corruption, which undercut public support for the very government that U.S. and NATO forces were trying to assist. The inability of the United States to shape the Afghan government's behavior led one European diplomat to marvel that “never in history has any superpower spent so much money, sent so many troops to a country, and had so little influence over what its president says and does.”13 FM 3–24's assumption of a unanimity of interests between patron and client was clearly misplaced.

This misalignment of counterinsurgency priorities is a variant of the classic principal-agent problem that occurs when one party delegates responsibility for carrying out a task to another in an environment of asymmetric information.14 In such relationships, the interests of the principal are not completely aligned with those of the agent. For example, employers generally want their employees to work hard all day long; employees, on the other hand, may prefer to get paid for shirking their responsibilities, particularly if they know their employer may not catch them in the act. This is not to suggest that the employees (agents) are lazy, evil, or stupid, merely that they are strategic actors, seeking to achieve their own goals within the restrictions put upon them by their employer (the principal).15

A primary challenge in assisting another nation in counterinsurgency is that the patron and the client have independent goals and priorities. Broadly speaking, both the United States and the local government may want to see an insurgency defeated. Retaining power is a competing priority for the client government, however, which puts a premium on continuing the domestic social and economic arrangements that benefit its core supporters, even if these same measures are driving support for the insurgency. A client state may accept external assistance to pursue its interests, but this is more frequently the result of necessity than choice.16 The client will try to manipulate the dynamics of the relationship with its patron to maximize the amount of political, economic, or military assistance it receives, while seeking to maintain its autonomy. As a result, a key source of frustration for American policymakers when assisting a state's counterinsurgency efforts is the inability to get the client to comply with U.S. counterinsurgency prescriptions.17 In reality, the local government is an independent agent and the power supporting it has, at best, only indirect control over its client's economic, political, and military policies.

As Thomas Grant notes, “It is rare to have a morally splendid ally in counterinsurgency work, simply because morally pristine, administratively effective governments do not provide the inspiration or excuse for a guerrilla war.”18 Thus, assisting counterinsurgency suffers from what is known in principal-agent parlance as “adverse selection,” because the only governments needing external assistance to combat domestic political opponents are almost by definition flawed in some key respects—be they incompetent, fraudulent, abusive, or all of the above. The same governmental shortcomings that facilitate the emergence of an insurgency also undercut the effectiveness of the counterinsurgent response.

The local government is unlikely, however, to be eager to address these problems of its own volition. As Bruce Bueno de Mesquita and Alastair Smith have argued, the behavior of political leaders is based primarily on their desire for political survival: “Decisions are not taken to improve the welfare of the people unless coincidentally this simultaneously aids survival.”19 The priority for a besieged government is often to bolster its position within its society by ensuring that the leadership of the state's security forces is too divided to mount a coup, using economic patronage to co-opt rival elites who could pose a threat, and restricting political decisionmaking to trusted loyalists—all of which are likely to diminish counterinsurgency effectiveness. Consequently, Daniel Byman warns that “U.S. COIN doctrine, no matter how well thought out, cannot succeed without the appropriate political and other reforms from the host nation, but these regimes are likely to subvert the reforms that threaten the existing power structure.”20

This leads to the second principal-agent problem that a patron state faces, what economists have termed “moral hazard.” In this situation, possessing insurance against risk inadvertently leads a party to act less carefully than it otherwise would have because it does not bear the full consequences of its actions.21 By providing assistance, the patron state can alter the burden of risk in a way that might unintentionally change the behavior of the local government. If a regime believes that an external power is committed to its survival, it can ignore the potential risks of its own actions (or lack thereof), confident that the patron will protect it from harm if the situation deteriorates too far. Absent the ability to compel a client to address the various shortcomings that facilitated the outbreak of an insurgency, some critics contend that external assistance can sap the local government's motivation to defeat the insurgents on its own or undertake the reforms that such aid was intended to encourage. Consequently, external support can render a supported government less stable than it might otherwise have been.22 These different priorities can produce a situation in which the intervening power supports a regime that refuses to implement its counterinsurgency prescriptions and may even seek to free ride, forcing the patron to bear the costs of fighting the insurgents.

Shirking by an agent can never be completely prevented; however, agency theory identifies several tools for mitigating principal-agent problems. The preferred means is to carefully screen agents and select only those whose preferences closely align with those of the principal.23 A patron assisting a client state in counterinsurgency, however, rarely has the option of selecting an optimal agent beforehand. Therefore, the next-best solution is to shape the client's behavior through the offer of inducements or the threat of sanctions. This topic is taken up in the subsequent section.

Patron-Client Influence Strategies

The literature on international patron-client relationships focuses on the influence dynamics between patron and client, yet it has not previously been brought to bear on the challenge of shaping a smaller partner's behavior in counterinsurgency. This body of scholarship, which reached its apogee in the last decade of the Cold War, emphasizes the reciprocal nature of the relationship between a great power and a client state. Although their relationship many be not formally codified as an alliance, the patron and the client each has something the other desires, often economic, military, or diplomatic aid on the one hand and loyalty and compliance on the other.24 When it comes to the question of influencing a client, and given that both sides value the consensual partnership, some scholars argue that coercion will not play a role in bilateral relations, nor will a patron exploit its superior strength to command obedience from its client.25 Instead, the patron achieves compliance by providing its client with inducements to change its behavior.26

A second group of scholars of international patron-client relations accepts the basic premise that bilateral relations are consensual and mutually valued. They note, however, that as utility maximizers clients do not readily reciprocate their patron's aid. Rather, they “generally seek maximum support in exchange for minimal concessions.”27 Consequently, within the context of an overall benign relationship between patron and client, shaping a client's behavior requires more coercive tools.28

The remainder of this section explores in greater detail the logic of the use of inducements and conditions on aid to influence a client state.


Inducement strategies focus on a patron's use of aid to encourage its client to undertake specific actions. Patronage can include intangible benefits, such as security guarantees and diplomatic backing, as well as tangible ones such as military hardware and economic aid. As Celia Reynolds and Wilfred Wan describe, “Senders use positive inducements as persuasive measures to cajole the recipient into changing its behavior.”29 Given the cooperative relationship between the two parties, unilateral grants of aid are expected to generate positive reciprocity and cooperative behavior from the recipient.30 Moreover, because such incentives can mitigate the client's cost of undertaking a policy change desired by the patron, proponents believe that these incentives will increase the chances of compliance.31 Befitting a strategy that seeks to gain influence through the unilateral provision of aid, inducement strategies favor public and unambiguous commitments to client states to gain their trust and bolster their confidence.

In the context of patron-client relations, inducements have a number of characteristics to recommend them. Unlike sanctions, they do not require multilateral cooperation to be effective; inducements can succeed unilaterally.32 Moreover, from a reputational standpoint, a target state may find it easier to accede to an incentive than to acquiesce to a threat.33 Finally, unlike more coercive approaches, inducements help to maintain friendly relations between the states involved, which is important in the context of patron-client relations.34

Past scholarship on the use of this approach by the United States finds that inducements were generally met with reciprocity from the other party and that positive sanctions have a better track record in influencing a target state's behavior than punitive pressures.35 Among U.S. policymakers, William Mott reports that faith in the efficacy of inducements has traditionally been bolstered by the twin beliefs that the client state shares the United States' priorities and that the provision of aid will result in “a powerful ability to influence the actions and policies of U.S. recipients.”36 U.S. officials believed that, in a number of counterinsurgency efforts, making unilateral grants of assistance to a local government would create the trust necessary to encourage the desired political and economic reforms.37


In contrast to an inducement strategy, which focuses solely on the provision of rewards to shape a client's behavior, conditionality includes a combination of rewards and threats to suspend or withhold assistance in the absence of client compliance.38 The latter measure is a necessary result of the potentially diminishing utility of foreign aid in generating influence.39

David Baldwin reports that the suspension of aid is as successful in influencing state behavior, if not more so, than are other means, including the use of military force.40 Conditionality may be particularly suited for dealing with recalcitrant clients, because economic coercion has been found to be more effective against friendly states than it has against adversaries.41 The success rate of World Bank and IMF conditionality to engender economic reform in target states is mixed; however, the fact that their conditions are typically ex ante—local governments receive loans and aid after promising to reform but before any reforms have been implemented—may explain why they have been less effective at generating compliance.42

As an influence strategy, conditionality is not a tool that the United States has readily turned to when managing client states. On the face of it, sanctioning or coercing a partner government appears to run contrary to the point of assisting it in counterinsurgency. Coercion is for adversaries, not allies.43 Moreover, even when a patron wishes to shape a client's behavior, several scholars report that patrons are remarkably hesitant to exercise their potential leverage over smaller allies to actively influence their policy choices because they are often unsure of the degree of influence they actually have over their clients.44

U.S. Influence in the Salvadoran Civil War

This section tests the competing logics of inducement and conditionality against evidence from the U.S. effort to support the Salvadoran government against the insurgents of the Farabundo Martí National Liberation Front (FMLN) from 1979 to 1992. This is one of the longest and most expansive cases of the United States providing aid and support to a local counterinsurgency effort short of the full-scale commitment of combat troops. Moreover, El Salvador has been identified by both scholars and the U.S. military as a template for future counterinsurgency assistance missions.45

The El Salvador case is ideal for testing both the inducement and conditionality strategies because the United States pursued both at different times during the civil war with varying degrees of success in influencing the behavior of the local government.46 The case contains sixteen discrete influence episodes, each beginning with the Salvadoran government or military actively opposing U.S. entreaties for reform, reorganization, or policy change, thus indicating that compliance was not its preferred course of action. Congruence, a within-case tool of causal inference, is employed to assess the relative effectiveness of inducement and conditionality in influencing the degree to which the Salvadoran government implemented specific political, military, or economic reform measures favored by the United States as part of El Salvador's counterinsurgency strategy.47 The Salvadoran government undertook important reforms and policy changes when the United States attached strict conditions to its delivery of aid, but not when inducements were given. To avoid spurious correlation, the study is also alert to any external shocks that could explain the client's changed behavior. Although this study employs multiple observations, any theoretical proposition derived from or tested against a single case can run afoul of omitted variable bias or interaction effects. At a minimum, this analysis can provide circumstantial evidence of the responsiveness of a client regime to a patron's preferences under alternate influence strategies in one of the most significant U.S. counterinsurgency assistance efforts on record.


Challenged by both Marxist guerrillas and reactionary death squads, during the 1980s the Salvadoran government suffered from corruption, widespread human rights abuses by state security forces, and a lack of democratic legitimacy. Hoping to prevent the government from falling to either the extreme left or the far right, the United States sought to strengthen the military and encourage democratization and socioeconomic reforms that could bolster the regime's legitimacy and reduce popular support for revolution. Twelve years of U.S. aid and advice allowed the Salvadoran government, which had been on the brink of collapse in late 1979, to withstand the insurgency and eventually conclude an externally brokered peace agreement. This outcome was achieved, however, at the cost of $6 billion in U.S. economic and military aid and some 75,000 Salvadoran lives.

The Salvadoran economy of the late 1970s was built around the export of labor-intensive crops such as coffee, cotton, and sugar—an undertaking dominated by a landed oligarchy. As the smallest state in Latin America, El Salvador had the region's highest population density, and agriculture could not provide sufficient income for the country's growing rural population. Consequently, only Haiti and Guatemala had worse living standards in the region for the poorest of society.48 With the support of the staunchly anticommunist upper and upper-middle classes, a series of military-led governments had dominated the country's political affairs since the 1930s.49 Paramilitary police maintained order in the countryside, employing physical intimidation tactics and a grassroots network of informants to identify and eliminate potential subversives.

The country's most important institution was the Armed Forces of El Salvador (ESAF), whose 500-man officer corps provided the leadership for both the army and the state's security forces, the latter composed of three paramilitary police agencies: the National Police, the National Guard, and the Treasury Police. In theory, control of the armed forces was highly centralized in the office of the defense minister; in practice, however, the defense minister's authority depended on the concurrence of the heads of the security forces and the quasi-autonomous military commanders of El Salvador's fourteen departments (provinces).50

A global plunge in the price of coffee in 1978 economically destabilized large portions of the countryside. In the wake of the July 1979 revolution in neighboring Nicaragua, demands for change in El Salvador among labor unions, peasant groups, students, and segments of the Catholic Church grew stronger and more radical. This dynamic political environment saw the emergence of no fewer than five revolutionary groups, each with its own guerrilla faction as well an overt political wing capable of marshaling tens of thousands of demonstrators.51 To stave off rebellion, a movement of reform-minded junior officers overthrew the government in October 1979, establishing a civilian-led junta that included members of centrist and left-wing opposition parties. The Junta Revolucionaria de Gobierno promised to equitably distribute national wealth, to rein in the security forces, and to hold “genuinely free elections” in 1982 for a constituent assembly, which would write a new constitution for the country.52 To improve El Salvador's human rights environment, the junta purged 80 officers and 1,400 enlisted men accused of human rights abuses.53

Although the coup had been virtually unopposed, only 20 percent of the officer corps strongly supported radical change. The vast majority begrudgingly supported the minimum reforms necessary to prevent a revolution, and the reformers were quickly outmaneuvered in the struggle for control of the armed forces.54 In the ensuing months, as demonstrations and antigovernment protests continued, repression against trade unionists, students, members of peasant federations, and other dissidents escalated and state-linked political murders spiked.55 By January 1980, all of the civilian members of the junta had resigned, blaming the defense minister, Col. José Guillermo García, for obstructing their reform agenda and for the ongoing violence by the security forces.56

While the junta imploded, the far left consolidated. In December 1979, the Cuban government brought together the leaders of the five Salvadoran revolutionary groups to form the Farabundo Martí National Liberation Front. Collectively, the FMLN possessed 4,000 guerrilla fighters backed by 5,000 part-time militiamen and a political wing that could put 200,000 demonstrators on the streets.57 It received weapons, military training, and assistance from neighboring Nicaragua and Cuba, as well as from the Soviet Union, Bulgaria, East Germany, and Vietnam.58

conditional aid preserves the junta. The Carter administration's aim in El Salvador was to prevent either the far left or the extreme right from seizing power; support the junta's economic and political reforms, which could reduce popular support for a revolution as well as weaken the power of the oligarchy; and reduce state-linked violence by bringing the security forces under the junta's control. To preserve a centrist government, the administration brokered a deal with the military to allow José Napoléon Duarte's Christian Democrat Party (PDC) to rejoin the junta. This was hardly an easy bargain to make because widespread electoral fraud engineered by the military denied Duarte victory in the 1972 presidential election, and the two sides had been bitter enemies for more than two decades. Nevertheless, Carter's envoy, Deputy Assistant Secretary of State James Cheek, informed all parties that U.S. aid was contingent on the PDC joining the military as a partner in the government.59 As Cheek later recounted, U.S. leverage over the ESAF derived almost solely from “conditional promises of future assistance.”60

The coalition, U.S. observers described, was a highly tenuous “marriage of convenience” that was “driven by a sense of mutual need but with important differences of orientation.”61 The military wanted to focus on restoring law and order, whereas the Christian Democrats pushed economic and social reforms as the answer to El Salvador's problems.62 Both groups viewed the other's approach as ill conceived. As the PDC's price for joining the government, the military agreed to nationalize the banking sector; control the export of coffee, cotton, and sugar; expropriate 2 million acres of farmland for purchase by sharecroppers and agricultural co-ops; and end repression against the population.63 To support these reforms, which would greatly diminish the power of the country's landed oligarchy, the U.S. government pledged $5.7 million in nonlethal military aid and $50 million in economic assistance.64

Although Carter administration officials believed that a violent challenge by the far left posed the greatest threat to the junta's survival, the FMLN was not the junta's only enemy.65 El Salvador's large landowners and certain segments of the business community violently opposed the junta's reforms, particularly the redistribution of highly productive agricultural land to sharecroppers and co-ops.66 Taking a page from its communist foes, the far right organized its own overt political and clandestine military wings. Their public face was Robert D'Aubuisson, a charismatic military intelligence officer who had been purged by the junta.67 Leading “an anti-Communist crusade” on national television, he denounced Christian Democrats and reform-minded military officers as subversives.68 Many of those whom D'Aubuisson named were subsequently murdered by the far right's covert military wing, the so-called death squads. The majority of these right-wing paramilitaries comprised active-duty or retired members of the police and military intelligence, functioning, according to the Central Intelligence Agency, “with or without the knowledge of immediate superiors.”69 In 1980, political killings would claim more than 8,000 lives, 90 percent of which were attributed to the security forces and the death squads.70 Among the victims were several reform-minded military officers; the country's archbishop, Oscar Romero; the attorney general; and hundreds of PDC activists.

By early 1980, rumors had begun to circulate that Defense Minister García was planning to seize control of the junta to avert a coup by the far right.71 In response, the Carter administration warned the ESAF high command that it would withdraw its recent promise of $50 million in economic aid and $5.7 million in nonlethal military assistance if the government was overthrown. D'Aubuisson urged the military not to be intimidated by such threats; the need for U.S. assistance, however, proved compelling, as Defense Minister García publicly pledged support for the junta and its reform program.72

the helicopter deal and leverage lost. With a coup temporarily averted, the Carter administration turned its attention to El Salvador's security forces. A debate over the best way to restrain their repressive measures was initiated by a Salvadoran request to lease six surplus Vietnam-era transport helicopters—to compensate for the army's limited ground mobility—and to deploy American technicians to maintain them.73 Despite their immediate utility, the U.S. ambassador to El Salvador, Robert White, advocated conditionality, arguing that the helicopters should not be given “until the worst of the right-wing violence is brought to an end.” Otherwise, such a visible U.S. commitment would bolster the military at the expense of the civilians in the junta.74 While sharing White's goals, Deputy Assistant Secretary Cheek argued for an inducement strategy: the United States should quickly dispatch the helicopters and technicians to signal the credibility of its commitment, which was necessary if the United States hoped to separate the ESAF from its “traditional patrons on the far right.”75 Contrary to White, Cheek argued that a clear signal of U.S. support would strengthen reformers relative to the reactionaries, whereas a failure to provide aid would be interpreted as a lack of enthusiasm for the junta.76

White's arguments for conditionality ultimately prevailed. The United States offered to lease at no cost six helicopters, which the ESAF was “anxious” to receive “to the point of desperation.” Delivery was conditioned on the Salvadoran government completing the following five steps within sixty days: (1) issue a directive denouncing the “indiscriminate violence and human rights violations” occurring in the country, including the repudiation of the abduction, torture, and execution of suspected subversives; (2) improve command and control of counterinsurgency operations to reduce abuses; (3) replace senior officers and military units in areas where significant violence against civilians had occurred; (4) demonstrate a commitment to suppress the influence of the far right, particularly within the military; and (5) commit to defend the judiciary from intimidation and violence.77

Civilian and military members of the junta found it “galling” that “reforms were being enforced by another country.” Rather than explicit conditions, they suggested an informal understanding that the above terms would be implemented if the United States supplied the helicopters and technicians.78 Despite warnings that delaying delivery of the helicopters endangered the upcoming coffee harvest, and therefore the country's fragile economy, the United States refused to act until it had received the junta's written reply. In late September 1980, the Salvadoran government officially accepted the U.S. terms.79

In the ensuing months, the ESAF made some progress in meeting the U.S. conditions, officially denouncing human rights violations, preparing a code of conduct for the armed forces condemning the abuse of civilians, and strengthening the military chain of command to improve accountability. Conditionality had compelled Salvadoran military leaders to accept an agreement they did not like; the credibility of the threat to withhold the helicopters, however, vanished with Carter's crushing loss to Ronald Reagan in November 1980. To many in El Salvador, Carter's defeat signaled the end of conditions on U.S. military aid and the start of a new policy of unconditional support for pro-American regimes.80

Consequently, the Salvadoran government took no firm action against officers implicated in human rights abuses or involved in the death squads.81 With killings by security forces on the rise, Ambassador White advised the State Department at the end of the sixty-day monitoring period that “there is no way that any objective observer could state that the Government has complied with the five steps we proposed.”82

aid suspension and junta reorganization. During its last months in office, the Carter administration attempted to regain leverage over the Salvadoran government, but its efforts were derailed by the murders of six U.S. citizens by members of the Salvadoran security forces.83 In the first instance, four American missionaries were raped and killed; in the second, two American land-reform experts were gunned down in a hotel alongside the head of the Salvadoran land reform agency.84

The killings provoked immediate outrage in the United States and raised questions about the degree of influence Washington wielded over its erstwhile client. Although not part of the counterinsurgency effort per se, the murders became a litmus test for U.S. observers as to whether the Salvadoran military could ever be convinced to conduct a clean counterinsurgency war. Carter suspended all aid and dispatched a high-level delegation to San Salvador to undertake a final attempt to compel the junta to bring the military under civilian control, remove hard-liners from positions of authority, and control the death squads.85 U.S. embassy analysts estimated that without the embargoed aid, the government's military operations—particularly its efforts to protect the coffee harvest from guerrilla sabotage—would be adversely affected “in a matter of weeks,” and that the Salvadoran economy could suffer a total collapse within a month.86

A major weakness of the junta was that its civilian members had little control over the armed forces. To strengthen the position of the Christian Democrats vis-à-vis the military, the Carter administration turned to conditionality, announcing it would resume aid if there was a “significant restructuring of the Government and shifts in military personnel.”87 Privately, the leader of the Christian Democrats, Napoléon Duarte, was informed that if he could reach an agreement with the military that improved civilian control over the security forces, reduced indiscriminate killings, and removed several “hard-liners” from command positions, the United States would release $95 million in economic aid, start delivery of the helicopters, and deploy up to fifty U.S. military trainers to assist the armed forces.88

After several days of intense negotiations, a deal was brokered that made Duarte president of a reorganized junta and committed the military to remove from office the vice defense minister, the head of the Treasury Police, and ten middle-ranking officers associated with extremist violence.89 The Salvadoran government also pledged to hold free elections in 1982 and urged the FMLN to peacefully join the political process. As a reward for the partial implementation of the desired reforms and the promise to purge hard-line officers, the United States immediately released $67 million in economic aid.90 Former Venezuelan Foreign Minister Arístides Calvani, who was an interlocutor between the military and the PDC during the negotiations, reported that U.S. conditions on aid “had been an important element in the shaping of the new agreement.”91

the 1981 final offensive and the end to conditionality. While U.S. economic assistance to El Salvador was fast-tracked, nonlethal military aid was discretely made available on a “phased incremental basis”: tranches would be delivered based on tangible progress in the investigation of the murders of the four missionaries, removal of senior officers implicated in significant human rights abuses, and a reduction in death squad violence.92 Duarte acknowledged that U.S. conditions on military aid provided him with leverage over the armed forces, but he urged the Carter administration to wave them in light of growing evidence that the FMLN was preparing for a “final offensive” to seize power.93 An interagency split developed within the administration, with Undersecretary of Defense Robert Komer arguing that tying military aid to human rights issues was akin to “fiddling while Rome burns.” Forestalling an insurgent victory in El Salvador had to take priority, Komer insisted; once peace was restored, “we can later use aid as leverage to enhance human rights.”94 Secretary of State Edmund Muskie and Ambassador White took the opposite view, counseling President Carter that releasing military aid would send “the wrong signal” to the Salvadoran military, which had already “misinterpreted” the resumption of economic assistance as a sign that the United States was not serious about pressing it to purge human rights abusers from its ranks.95

The debate was overtaken by events when, on January 10, 1981, several thousand guerrilla fighters launched simultaneous attacks on forty-three locations across El Salvador. The FMLN initially achieved impressive success, but its call for “all the people to rise up as one” went unheeded.96 In nine days of hard fighting, which nearly exhausted the Salvadoran government's stock of munitions and consumables, the ESAF forced the insurgents to surrender most of their gains.97 Marred by poor coordination and insufficient urban organization, the operation proved a major failure for the insurgents.98

In the wake of the offensive, the Carter administration lifted its conditions on military aid. As one of his final acts in office, President Carter used the “unforeseen emergency” provisions of the Foreign Assistance Act to dispatch six helicopters; twenty-eight military trainers; and $5.9 million in military hardware—including new weapons and millions of rounds of ammunition—to El Salvador without congressional approval.99 By the time the military supplies began arriving in country, however, the ESAF had regained the initiative over the insurgents, as Defense Minister García proudly proclaimed, “without one bullet” from the United States.100

Carter attempted to justify the emergency aid release by citing Salvadoran compliance with U.S. conditions; Secretary of State Muskie and Ambassador White, however, disagreed that tangible progress had been made.101 By unilaterally abrogating its conditions on military aid, the Carter administration confirmed the Salvadoran high command's long-held belief that the United States would never risk a break with El Salvador.


With the ascent of Ronald Reagan to the presidency, observers in San Salvador and Washington braced themselves for a major reorientation of U.S. policy.102 Yet, in many respects there was significant continuity, as the new administration shared its predecessor's belief that “the alternative to the existing junta today is extremism on either side.”103 Although some of the junta's economic reforms, such as the nationalization of the banking sector and the collectivization of agriculture, clashed with Reagan's free-market economic principles, the administration believed that such measures were succeeding in the political battle against the insurgency.104 Rather than embracing the Salvadoran right, within weeks of taking office Reagan approved a proposal to provide covert financial support to the Christian Democrats and the remaining reform-minded officers in the junta.105 “Prompt, free, and open” elections were the administration's solution for broadening the Salvadoran government's base of support and enticing both the left and the right to compete for power politically instead of through force.106 As a result of these continuities, the head of Latin American policy planning in Carter's State Department, Luigi Einaudi, described the Reagan approach as “Carter plus”—the “plus” being enhanced military aid.107

In February 1981, the United States granted El Salvador $25 million in military aid—more than it gave to the rest of Latin America combined—as well as $74.4 million in economic assistance.108 To give the junta “breathing space” in its battle with the FMLN, Reagan followed Carter's lead in invoking an “unforeseen emergency” to rush modern rifles, machine guns, mortars, medical equipment, four more helicopters, and an additional twenty-six military trainers to El Salvador without congressional approval.109 Failure to support El Salvador, Reagan feared, would not only put Costa Rica, Honduras, and Panama at risk, but it would harm the credibility of the United States, weakening U.S. alliances worldwide.110

inducements without results. The Reagan administration's primary point of departure from its predecessor was the embrace of an inducement strategy. Unconditioned aid and unambiguous statements of support for El Salvador were advised in the belief that dispensing U.S. largesse provided leverage over the local government, particularly the military, that could be used to influence their behavior.111 In contrast, Reagan officials believed that the Carter administration had demonstrated that conditions on aid were “counterproductive,” leading to a “waning” of U.S. influence over client states.112 The government in San Salvador could not count on the United States, Secretary of State Al Haig, argued, because “assistance was promised one day and turned off the next.”113 Such coercive measures should be used to influence hostile states, not friendly ones.114 The administration would continue to try to convince Salvadoran leaders to reduce human rights abuses by the security forces and sustain the junta's reforms, but it would do so using inducements and friendly persuasion rather than conditions on aid.115

The Reagan administration's new approach produced immediate results—none positive—as the ESAF reneged on its promise of reform made in December 1980. The murder investigations stalled; the pace of extrajudicial killings spiked; hard-line military officers retained their posts; and land reform was soon “in serious danger of failure.”116 With Robert D'Aubuisson publicly alleging that there was “high-level Washington support” for a military coup, Secretary of State Haig firmly informed the Salvadorans that the United States would terminate all military aid if the government was overthrown.117 Again, the potential loss of military aid spurred Defense Minister García into action. Publicly warning supporters of the far right in the armed forces that a coup would “sink” El Salvador, he insisted that “without foreign support, especially that of the United States, there could be no survival.”118 Similarly, the missionary murder case moved forward, not because of inducements, but because of threats to withhold aid. In late April 1981, Haig called Duarte to warn him that growing impatience in the White House and in Congress over the progress of the case “threatens our ability to continue to assist you.”119 Three days later, the six guardsmen suspected of the murders were arrested.

congress pushes conditions. Following their failed “final offensive,” the insurgents retreated to their rural strongholds in the north and east of the country and regrouped for a protracted fight. For the next several years, 12,000–14,000 FMLN guerrillas fought a quasi-conventional war against a Salvadoran military that grew from 20,000 members in 1980 to 54,000 by the second half of the decade. The poorly trained Salvadoran army was organized for a conventional war with neighboring Honduras. To defend against the insurgents, the majority of the force was deployed in small units to guard critical economic infrastructure. Infrequent offensive operations consisted of ponderous, short-duration sweeps of FMLN territory by large units that the guerrillas easily avoided.120 Instead, local civilians bore the brunt of the armed forces' brutality under the logic that, in the words of one army commander, “civilians who don't want to cooperate [with the insurgents] leave the area and those who remain are collaborating.”121 In December 1981, elements of the American-trained Atlacatl rapid reaction battalion rampaged through the village of El Mozote, in Morazán department, torturing and killing 767 men, women, and children over the course of three days. It would be the single largest massacre of the war.

Combat between the army and the insurgents was initiated largely by the FMLN, which inflicted terrible losses on the ESAF. In 1981, government troops suffered an astounding 3,827 casualties, or 19.1 percent of the force.122 Without even rudimentary medical evacuation capabilities, one-third of soldiers wounded in the field died from their injuries.123 The ESAF high command soon abandoned offensive action for fear its forces were on the verge of collapse.124 In September 1981, senior officers at U.S. Southern Command (SOUTHCOM) warned that the Salvadoran military was losing the war, a view echoed by the State Department and the National Security Council.125 This gloomy assessment was punctuated by an FMLN raid on the Ilopango Air Base, in the capital of San Salvador, that destroyed or damaged 70 percent of the country's military aircraft.126 With an estimated 13,000 political murders in 1981, mostly attributed to the security forces and the death squads, and human rights groups denouncing “a systemic and brutal policy of government-sponsored intimidation and repression,” the Salvadoran government was increasingly isolated internationally.127 Western European countries suspended assistance, and some even extended diplomatic recognition to the FMLN.128

The situation in El Salvador emerged as a subject of contention on Capitol Hill. In December 1981, congressional Democrats passed an amendment to the foreign aid bill requiring President Reagan to certify every six months that the Salvadoran government was (1) making a concerted effort to comply with international human rights standards; (2) achieving substantial control over the armed forces to prevent the torture or murder of civilians; (3) making continued progress in implementing economic, political, and land reforms; (4) promising to hold elections; and (5) attempting to bring the murderers of the American missionaries and land-reform experts to trial.129 If the president could not make this certification, Congress would cut off all military assistance to El Salvador. This certification requirement attempted to push the United States back toward a conditionality strategy, which the administration opposed because, in the words of United Nations Ambassador Jeane Kirkpatrick, it was “undermining the confidence of vulnerable allies.”130 The all-or-nothing nature of the terms of certification left little room to calibrate aid to the level of Salvadoran compliance. Consequently, on three occasions in 1982 and 1983, the Reagan administration complied with the letter but not the spirit of the law by certifying progress in order to keep aid flowing, despite limited or even nonexistent evidence of the consolidation of civilian authority, improvements in military discipline, or a reduction in human rights abuses.131

military reforms languish, political reforms advance. In the early 1980s, the ESAF lacked trained officers, standardized equipment, enough soldiers to sustain offensive operations, and an adequate intelligence capability for irregular war.132 U.S. military observers judged that the force was “unprepared, strategically, tactically, organizationally, or equipment-wise to confront a credible guerrilla force or insurgency.”133 As a result, the 55-man U.S. military assistance group in El Salvador had two key goals: (1) to expand and modernize the ESAF so that it could take on the FMLN, and (2) to alter the ESAF's operational and tactical approach to counterinsurgency. Conditionality was not employed to pressure the Salvadoran military to heed U.S. advice.

Over the course of the civil war, with U.S. assistance, the army nearly tripled to 56,000 men, who were issued modern weapons and equipment; several new heliborne rapid reaction battalions were created to attack large FMLN formations; and the Salvadoran air force was modernized.134 Providing a military with advanced weapons and recruiting new enlisted personnel is significantly easier, however, than it is to change its mode of operations. The United States called for the Salvadoran army to abandon its defensive posture and undertake aggressive, small-unit operations, carried out day and night, to seize the initiative from the FMLN. Rather than chasing the guerrillas all over the country, the ESAF was advised to concentrate its efforts on the country's main population centers and economic heartland. It was envisioned that the ESAF would systemically clear the insurgent presence from the most important sections of the country before moving on to the peripheral areas dominated by the guerrillas. In a classic “clear, hold, build” model, civic action and development in these cleared zones would be undertaken to win popular support.135

A lack of effective leadership and a continued tolerance of institutional violence, however, stymied efforts to implement these measures. Despite the serious military challenge posed by the FMLN, many officers approached the conflict as a “nine-to-five war,” returning to San Salvador to see their families on weekends while their troops remained in the field.136 U.S. entreaties to promote capable and aggressive officers to leadership roles went unheeded, as ESAF leaders prioritized advancing the careers of classmates and cronies rather than enhancing military effectiveness.137 With the exception of the rapid reaction battalions, which embraced small-unit tactics and achieved battlefield results disproportionate to their size, Salvadoran officers selectively adhered to U.S. guidance. Thus, the army retained its preference for ineffectual large-unit sweeps by road-bound forces and confined its operations to the daylight hours to avoid ambushes, surrendering the night to the insurgents. But even if the army had embraced aggressive small-unit operations against the guerrillas, efforts to win popular support would have been stymied by the ESAF high command's inability or unwillingness to acknowledge the problem that “extreme rightist terrorism and institutional violence” posed for an effective counterinsurgency strategy.138 No amount of civic action or development projects would rally civilians to the government's cause when the armed forces were the primary source of human rights abuses.

While attempts to transform the Salvadoran military foundered, political reform, which was one of the congressional certification requirements, showed greater promise. Elections for a constituent assembly that would draft a constitution were scheduled for March 28, 1982. The Reagan administration saw the electoral exercise as a key means of building support for the Salvadoran government, while disarming its domestic opponents by enticing the far right into the political process and demonstrating the limited political support for the insurgents.

The Christian Democrats and the former party of the oligarchy, the PCN, were joined in the contest by the new Nationalist Republican Alliance (ARENA), headed by Robert D'Aubuisson. The Salvadoran electoral commission also recognized several parties allied with the FMLN as legitimate political entities and offered to facilitate their campaigns from exile via television if security concerns made physically returning to El Salvador too dangerous.139 The insurgents rejected the offer, however, insisting that they should be brought into a power-sharing government with authority proportional to their relative military capability.140

In a contest that has been judged to be “relatively free, fair and competitive,” 85 percent of El Salvador's eligible voters went to the polls in the face of FMLN threats to kill anyone seeking to participate in the election.141 The Christian Democrats finished on top; ARENA finished second; and the PCN came in third. The country's archbishop declared the turnout a clear “vote in favor of peace, democracy, and justice,” and called upon the FMLN to “accept the judgment of the people and lay down their arms.”142

In Washington, elation quickly turned to alarm when it became clear that ARENA and the PCN had won enough seats between them to form a right-wing government that sidelined the Christian Democrats and named D'Aubuisson as provisional president. The United States promised to cut aid to a D'Aubuisson-led government, but its past inducement policies reduced the credibility of the threat.143 As a senior ARENA official told a reporter, “The United States has never cut off aid anywhere for very long or even entirely. Reagan will never let the Communists win here. It's just a complete bluff.”144

In the face of this intransigence, the Reagan administration warned the Salvadoran military and political leaders that continued U.S. support required (1) a national unity government that included the Christian Democrats; (2) sustained political and economic reforms, including land reform and progress in controlling violence; and (3) presidential elections.145 Alarmed by the U.S. threat, the military announced that exclusion of the Christian Democrats was unacceptable and compelled the formation of a provisional government led by Álvaro Alfredo Magaña Borja, a nonpolitical banker with close ties to the military and the imprimatur of the U.S. embassy.146

Following the election, the Reagan administration turned its attention to sustaining the land reform program. The agrarian reform program initiated by the junta in 1980 had three parts. In phase 1, the country's largest landholdings, those in excess of 500 hectares (2 square miles), accounting for approximately 15 percent of farmland, had been seized and given to peasant cooperatives. Phase 2 affected medium-sized estates from 100 hectares to 500 hectares, but its implementation had been delayed. The final component was the “land-to-the-tiller” program, which allowed sharecroppers to purchase up to 7 hectares of farmland that they currently leased from their landlords. With El Salvador's economy stagnating, the right-wing parties dominating the constituent assembly attempted to boost the production of export crops by halting the land-to-the-tiller program.147 This violation of aid conditions drew a swift reaction, as the Senate Foreign Relations Committee voted to cut scheduled military aid by 60 percent if the Salvadoran government failed to implement the program.148 Although the Salvadoran officer corps did not have particularly strong views on the importance of land reform, in the face of this threat both Defense Minister García and the army chief of staff publically endorsed the program, dispatching soldiers to return thousands of illegally evicted peasants to their farms.149 In moments of crisis, conditions on American military aid and clear threats to suspend it repeatedly proved to be the key to mobilizing the military to support U.S.-backed reforms.150

conditionality and the “new contract”. From the U.S. standpoint, the 1982 elections had generated positive political momentum in El Salvador; however, the death squads were undermining efforts to defeat the insurgency and build a democracy. In a direct challenge to the Salvadoran and U.S. governments, right-wing paramilitaries unleashed a wave of violence against union organizers, peasant groups, and Christian Democrats. The Reagan administration hoped to co-opt the leaders of the far left into the political process, but that would never happen so long as their lives were constantly at risk.

On the battlefield, the FMLN concentrated its forces into large formations that sought to confront and defeat the ESAF militarily in a series of quasi-conventional offensives. Between June 1982 and June 1983, the army suffered a crippling 6,815 casualties (21 percent of the force), which was double the number of losses from the previous twelve months.151 “There were days at the end of 1983,” U.S. Ambassador Thomas Pickering recalled, “when we wondered whether we would make it through the next two or three months.”152 In Hugh Byrne's judgment, “According to all indicators—ESAF casualties, arms taken, prisoners captured, terrain controlled, major towns and army positions taken, infrastructure damaged or destroyed, level of enemy morale—the guerrillas were winning the war.”153

In the United States, public support for the administration's El Salvador policy was anemic, and congressional opposition was growing. With progress on the investigation into the missionary murders “marginal at best,” Congress conditioned 30 percent of the $64.8 million in military aid appropriated for fiscal year 1984 on a trial and verdict in the case.154 Given the deteriorating military situation in El Salvador and the strong opposition to its policies in the United States, the Reagan administration needed a new approach.

An interagency policy review in July 1983 concluded that U.S. efforts in El Salvador were being hindered by “critical issues of troop motivation, treatment of civilians and military discipline.”155 In San Salvador the U.S. ambassador, Dean Hinton, insisted that the time had come to force the Salvadorans to follow American guidance: “The more the [Salvadoran government] needs our assistance, the more leverage we have. What we need to do is use it in concrete cases…. [Otherwise,] we will have strengthened the position of those Salvadorans, probably a majority, who believe and assert that no matter what they do or do not do, the U.S. will support and protect them against a Communist takeover.”156 In theory, congressional certification might provide such leverage; however, the administration's decision to issue certifications irrespective of Salvadoran compliance was subverting the process. Consequently, the State Department argued Salvadoran leaders have “not been motivated to take the minimal actions required to help us sustain our support.”157 Embracing the need for real conditionality, the interagency task force recommended forging “a new and reliable contract” that explicitly identified what the Salvadorans must do to win—including the “elimination of military participation in death squads”—and what assistance the United States would provide in return.158

In December 1983, Vice President George H.W. Bush was secretly dispatched to San Salvador to negotiate the new contract. He carried with him a letter from Reagan stressing that it was more important than ever to show dramatic progress in areas such as elections, land reform, and human rights.159 Forcefully denouncing the death squads as “right-wing fanatics” and “cowardly terrorists,” Bush warned Salvadoran leaders that “every murderous act they commit poisons the well of friendship between our two countries.”160 In a series of private meetings with President Magaña, the minister of defense, and senior military commanders, Bush laid out the Reagan administration's terms: (1) arrest former National Guard Capt. Eduardo Avila for the murders of the American land-reform experts and agree to a timeline for the prosecution of that case, as well as for the murders of several other U.S. citizens; (2) explicitly condemn death squad violence, send a list of three military officers and three civilians with known links to death squads into exile, and implement due process procedures for suspects detained by the security forces; (3) publicly commit to support the March 1984 presidential elections; and (4) complete constitutional action on land reform that would protect the agricultural co-operatives and the land-to-the-tiller program while moving forward on the expropriation of smaller landholdings.161 Emphasizing that these issues “cannot be set aside since our support hinges directly on all of them,” Bush announced a one-month deadline for their execution, so that there would be tangible evidence of progress when the U.S. Congress returned from its Christmas recess.162 “Without actions in these areas, there is no point in trying to obtain additional funds for El Salvador,” he warned, “and to be honest we will not even make the effort because it would be fruitless.”163

The reward for compliance was to be sufficient U.S. military aid to field forty-two additional army battalions, as well as attack aircraft, helicopters, and an enhancement of El Salvador's airborne medical evacuation capability, the latter of which was critical for boosting troop morale.164 Behind the scenes, President Magaña had advised Bush to set tight deadlines for compliance because “our military will agree to anything if there is no time deadline connected with it.”165 Indeed, the Salvadoran leader suggested that Congress's threat to cut aid was “very important in persuading the commanders to act positively and act soon.”166

Within the military, several senior officers denounced the Reagan administration's “arm-twisting tactics” at a time when “Salvadoran soldiers are shedding their blood in the fight against Communism so that U.S. soldiers will not have to do the same.”167 Ultimately, however, the threat of an aid suspension and the promise of sufficient support to defeat the insurgents was too much for the high command to ignore. As the defense minister bluntly admitted, “We know that improving our image is worth millions of dollars of aid for the country.”168

Days after Bush's visit, thirty-one senior officers signed a proclamation supporting a crackdown on the death squads. The three military officers on Bush's death squad list were removed by the ESAF high command. Several of the named civilians also lost their jobs in government; however, the latter could not legally be forced into exile.169 In accordance with the U.S. conditions, the Salvadoran government introduced new due process requirements for suspects detained by the security forces; the ESAF pledged to defend the 1984 presidential election; and the army pressured the Constituent Assembly into extending the land-to-the-tiller program. The army also saw that Captain Avila was arrested for his role in the murders of the American land-reform experts, and prevented his uncle, the president of the Salvadoran Supreme Court, from interfering in the case.170

Initial assessments by Central Intelligence Agency analysts suggested that the Salvadoran government was taking only symbolic steps against the rightwing paramilitaries; however, death squad activity declined significantly in the wake of the U.S. ultimatum and remained low for the next several years.171 By the first quarter of 1984, the U.S. embassy was reporting that it was unaware of any assassinations of political activists, nor had any known death squads claimed responsibility for any killings since Bush's visit.172 As the high command continued to back reform, the gap between the military and the far right appeared to be growing.173 The improvements were incremental, and Salvadoran compliance was grudging and subject to backsliding; but compared with past efforts, the Reagan administration's conditionality approach had achieved significant success.

electing duarte. In 1984 Napoleon Duarte defeated Roberto D'Aubuisson to become president in the first truly democratic election in Salvadoran history.174 Marshaling support from a coalition of labor unions and peasant groups, and bolstered by U.S. financing, Duarte promised to accelerate land reform, negotiate with the FMLN, and allow nonviolent protests.175

In his first weeks in office, Duarte disbanded several units in the Treasury Police and the National Police known to house death squads. He also replaced the three security force commanders with respected officers favored by the U.S. embassy.176 Several senior army officers and more than 100 members of the public security forces suspected of human rights abuses were also removed from their positions or dismissed.177 Although these reforms did not eradicate the death squads, extrajudicial killings plunged 80 percent, from an average of 105 per month in 1983 to 18 per month in 1984 and fewer than a dozen per month in 1985.178 Duarte's ability to reshape the military in the face of an officer corps seeking to protect its institutional prerogatives above all else was unprecedented, and resulted directly from the ESAF's recognition that access to continued U.S. aid required winning congressional approval.179

Within weeks of Duarte's election, the suspects accused of murdering the American missionaries were prosecuted. Five guardsmen were convicted and sentenced to thirty-five years in prison. With its pressure tactics having achieved the desired result, the United States released the $19.4 million in military aid tied to the outcome.180

Although the insurgents had pushed the Salvadoran military to the brink of collapse in late 1983, an expanding army, supported by U.S.-supplied aircraft, was able to blunt the FMLN's quasi-conventional large-unit operations and force the insurgents to reconsider their military strategy.181 By mid-1984, U.S. efforts were finally registering visible success: El Salvador had its first democratically elected president; the far right had been co-opted into the political process but kept out of power; the insurgency's military offensives had been contained; extrajudicial killings had fallen significantly; and the murderers of the American missionaries had been brought to justice. Yet, this very success would undermine the means by which it had been achieved.


Following Duarte's presidential win, Congress and the Reagan administration relaxed their use of conditions on aid to El Salvador.182 Some proponents continued to insist that conditionality was still required “to spur reforms that must occur if a military victory by the left is to be avoided,” but these voices were a distinct minority.183 As the assistant secretary of state for the Western Hemisphere testified to Congress, although conditionality “served a useful purpose, the need has passed.”184 In a bipartisan effort to bolster Duarte, Congress granted El Salvador a military aid package of $196 million for 1984, nearly two-and-a-half times the $81.3 million approved in 1983, and authorized an additional $123 million for the following year.185 With the exception of a symbolic $5 million linked to the arrest and prosecution of a second officer involved in the murders of the land reform experts—and the proviso that aid would be suspended in the event of a coup—U.S. military assistance was no longer conditioned on reforms in El Salvador.

The impact of this policy change was seen almost instantly, as the ARENA/PCN–dominated National Assembly repealed the land-to-the-tiller law over Duarte's objection.186 Despite winning a majority in the 1985 assembly elections, without conditions on U.S. aid supporting their reform proposals, the Christian Democrats could not exert much influence over their political opponents. With military aid freely flowing, the ESAF blocked civilian investigations into human rights abuses by officers. The end of conditionality also removed the pressure that had kept death squad activity in check, and by 1988 a modest increase in extrajudicial killings was raising concerns within the U.S. embassy.187

Under Duarte, the one area where the United States continued to employ conditionality was economics. Real gross domestic product in El Salvador had declined 23 percent since 1979, unemployment was 30 percent, and inflation hovered around 22 percent. With the costs of the war consuming half of the national budget, the government was running an annual deficit—financed by the United States—of $100 million.188 To reduce government spending, control inflation, and decrease the economy's dependence on commodity exports, the United States Agency for International Development began to attach conditions to its economic aid. In November 1984, it required Duarte to devalue the colon in order to gain $65 million in economic assistance.189 When that failed to turn around the ailing economy, fourteen months later the Agency again pressured Duarte into devaluing the currency and adopting austerity measures—such as capping wages and raising gas prices by 50 percent.190 Despite these measures' significant unpopularity, the Salvadoran government had little choice but to comply with the United States' conditions. As the country's planning minister bluntly stated, “Without U.S. aid, we would be absolutely broke and inflation would be totally out of control.”191

resisting counterinsurgency. As the ESAF grew in strength, the FMLN ceased its use of large formations and moved to a more traditional guerrilla war strategy. The insurgents' use of ambushes and land mines inflicted a steady toll on government forces, but the ESAF remained unwilling to adopt a small-unit counterinsurgency strategy.192 Instead, military commanders continued to favor a conventional approach: employing battalion-sized units in operations backed by close air support and heavy artillery to chase down small bands of guerrillas.193 Unwillingness to follow U.S. advice was not an issue of capacity. As a frustrated American military trainer noted, “The Salvadoran Army has been thoroughly trained in U.S. counterinsurgency tactics and they can do them well—the problem is getting them to actually use these tactics.”194 Salvadoran officers believed that confronting and defeating the guerrillas could resolve the conflict faster than U.S. counterinsurgency strategies could; yet ironically, doing so was returning the war to a stalemate.195 Nevertheless, the United States did not condition military aid on the ESAF's adoption of its tactical advice. Instead senior officials told their Salvadoran counterparts that “our partnership does not require you to adopt any particular doctrine,” provided you “respect the welfare of civilians and [fight] clearly in support of democracy.”196

On the other side, a steady stream of defections saw the FMLN shrink from a high of 12,000 guerrilla fighters in 1983 to 6,000 by 1987. Although the insurgents could deny the government control over one-third of the country, they could not exploit the growing opposition to Duarte's government and U.S.-imposed economic austerity measures to expand their support base.197 Three successful elections had convinced large majorities that the ballot box could bring a degree of change. Moreover, the guerrillas' economic sabotage, as well as their widespread use of land mines and a campaign of urban terrorism, alienated major segments of the population. The military's unwillingness to adapt to counterinsurgency and the insurgents' limited appeal stalemated the conflict: neither side could gain the upper hand, nor was either at risk of defeat. One Salvadoran pithily summarized the plight of his fellow peasants, “The army comes and goes. The guerrillas come and go. We hide under our beds.”198

the 1989 “final offensive”. For Duarte and his Christian Democrats, the military stalemate only deepened their problems: their political ineffectiveness and the unpopularity of U.S.-imposed economic austerity measures, which saw the economy grow by 1 percent between 1984 and 1989, led their approval ratings to plummet. In the 1989 elections to succeed Duarte, ARENA put forth Alfredo Cristiani, a Georgetown-educated coffee grower with no ties to the death squads. Over the course of the decade, urban entrepreneurs, rather than traditional landlords, came to dominate the party, and its politics moderated. Running on a platform promising economic recovery and openness to negotiations with the FMLN, Cristiani swept thirteen of the country's fourteen departments.199 In office, he attacked corruption and instituted economic reforms to jump-start the ailing economy.200 Duarte had signed on to Costa Rican President Oscar Arias's Central American Peace Accord in 1987, which sought to resolve the various civil wars in El Salvador, Guatemala, and Nicaragua. Cristiani was, however, better positioned to gain the military's support for dialogue with the insurgents, and he called for negotiations two days after winning office.

Cristiani's political overtures were conducted from a position of perceived strength, but in fact, the FMLN was consolidating its forces for one final military thrust. On November 11, 1989—two days after the Berlin Wall came down—the FMLN launched its largest offensive of the war. In a coup de main, 2,000 insurgents captured parts of San Salvador, ranging from the poorest barrios to upper-class neighborhoods, while unsuccessfully attempting to assassinate President Cristiani and the military high command.201 The intensity of the surprise assault sent shockwaves around the country. Nevertheless, as in 1981, the Salvadoran people did not respond to the call for a mass uprising. In a three-week campaign, the military fought block by block to recapture the city, at times employing helicopter gunships and fighter aircraft to dislodge the insurgents from their strongholds.202 During the fighting, the security forces detained union members, opposition leaders, and left-wing clergy. Despite declaring a state of emergency, however, the government did not initiate a systematic campaign of violence against suspected subversives.203 William Stanley has attributed this marked restraint to the ESAF's fear of alienating the United States. The military had only a one-month supply of fuel and ammunition, while operations relied heavily on helicopters that required maintenance and spare parts that only the United States could provide.204

The insurgent assault shocked Washington, where it was widely compared to the 1968 Tet Offensive in Vietnam—itself a considered a “dismal military failure and a brilliant political success.”205 The FMLN spent nearly two years preparing for the operation and suffered significant casualties as a result. Nevertheless, in the aftermath of the 1989 offensive, the Salvadoran and U.S. governments radically reevaluated the insurgents' military capabilities.206 Within weeks, the SOUTHCOM commander informed Congress that the Salvadoran government could not defeat the guerrillas.207

the jesuit murders and the return of conditionality. Despite its relative restraint during the fighting, the ESAF managed to snatch political defeat from the jaws of military victory. Five days into the offensive, elements of the Atlacatl rapid-reaction battalion raided the Jesuit-run Central American University. Acting on the instructions of Col. Guillermo Benavides—the commander of a special security zone in the capital—members of the unit executed the university rector, five priests, and two bystanders. The Jesuits were prominent scholars and a key conduit for the government's dialogue with the far left. Many in the ESAF, however, viewed them as the intellectual architects of the revolution.208

The murders shattered the bipartisan consensus in Washington. Even though the FMLN offensive was still under way, some members of Congress called for the immediate suspension of aid, while other legislators warned that next year's appropriations would be cut unless the perpetrators were quickly arrested.209 To underscore this position, the U.S. ambassador and the SOUTHCOM commander bluntly warned Salvadoran military leaders that “if by the end of January 1990 the [government and military] have not done everything humanly possible to find the guilty parties, all security assistance to El Salvador could be halted by Congress as its first order of business in the new year.”210

The threat led to rapid results. In early January 1990, President Cristiani announced that nine suspects, including Colonel Benavides and two lieutenants, had been arrested for the murders.211 Yet, the following month, while congratulating Cristiani on progress in the investigation and emphasizing the need to move it forward, now-President George H.W. Bush sent the Salvadoran government a conflicting signal by promising to increase El Salvador's $131 million economic aid package by $50 million.212

In the months after President Bush's announcement, the Jesuit case stagnated. Despite U.S. urging for “a thorough and timely investigation,” the leadership of the Salvadoran military stonewalled, in the apparent belief that interest in the case would eventually disappear.213 As evidence accumulated that senior officers had attempted to cover up the crime, however, Bush administration officials who had previously been “vehemently opposed” to conditionality began to recognize that “now we have to play hardball … we're now basically supporting the move to condition or cut military aid.”214

To demonstrate displeasure with the lack of progress in the case, the United States withheld $19.65 million in military assistance.215 The aid would be released, President Cristiani was informed, only if the military leadership compelled potential witnesses in the ESAF to cooperate with the investigating judge.216 The impact of these conditions on military aid was undercut by the fact that the embargoed items were chosen on the basis that “withholding these materials will not greatly affect the ESAF's basic ability to fight the war, but will cause inconvenience.”217 Unsurprisingly, Salvadoran interlocutors informed the U.S. defense attaché that senior officers did not feel much pressure to comply with the U.S. demands.218

The United States further undermined its position when it released $250,000 worth of spare parts for Salvadoran helicopters in late September 1990. The aid was notionally granted in return for “positive, though insufficient, progress” in the Jesuit case, but the parts in question were identified as “critical components” for the Salvadoran air force to remain ready against a guerrilla threat.219 President Cristiani and Salvadoran military leaders were informed that, despite the release of the spare parts, military assistance remained conditioned on the ESAF's cooperation with the Jesuit murder case. Within weeks, however, military commanders were pressing the U.S. embassy for the release of additional “vital” aid to maintain the army's armored personnel carriers without having met U.S. conditions.220 In San Salvador, the frustrated U.S. ambassador saw this as a clear indication of the hollowness with which the Salvadorans perceived the American threat.221

conditions, negotiations, and prosecution of murder suspects. In the aftermath of the 1989 “final offensive,” the political positions of the Salvadoran government and the FMLN underwent a transformation. The scope of the offensive appeared to dash any hopes of an imminent government victory. Moreover, the FMLN's acquisition of advanced surface-to-air missiles hobbled the Salvadoran air force which, in turn, constrained the aggressiveness of the army's operations.222 At the same time, the insurgents recognized that they were unlikely to trigger a popular insurrection in El Salvador.223 Within weeks of the 1989 final offensive, both the Salvadoran government and the FMLN had contacted the UN to help with mediation. This began a two-year period in which the two sides fought while they negotiated. To exert pressure on the Salvadoran government to make a deal, Congress returned to conditionality. It cut military aid for El Salvador by 50 percent and warned that it would withhold the remaining balance if the government failed to negotiate in good faith with the FMLN for a permanent settlement, if it did not thoroughly investigate the Jesuit killings, and if it employed large-scale violence against civilians.224 As an incentive for the FMLN to reach a compromise as well, Congress included a provision restoring the suspended military aid if the insurgents failed to negotiate with the government in good faith or if they received significant military assistance from abroad. The sharp reduction in military assistance was a shock for those in El Salvador who believed that the United States would sustain them indefinitely.225 After initially signaling that he would restore the full amount of military aid, President Bush, who was interested in winding down U.S. involvement in El Salvador, embraced conditionality as a means to move the peace process and the Jesuit murder investigation forward.226

In response to the pressure from Washington, the Salvadoran government conducted a series of intense negotiations with the FMLN during the fall and winter of 1991. On January 16, 1992, the two sides signed a UN-brokered peace accord, bringing an end to a war that had taken 75,000 lives and exhausted the country. In the final settlement, the size of the army was reduced by 50 percent, and civilian supremacy over the armed forces was finally established. A new civilian police force replaced the security forces. The FMLN abandoned its demand for a power-sharing agreement, instead becoming a political party that would compete for power legally.

Meanwhile, Colonel Benavides was found guilty of ordering the murders of the Jesuits and, along with an accomplice, was sentenced to thirty years in prison. Three more officers were convicted for their roles in attempting to cover up the crime. The case demonstrated how far El Salvador had come in the past decade. Not only was this the first time a senior officer had been tried for human rights abuses, but as William LeoGrande noted: “For the first time in modern Salvadoran history, two officers were convicted for the politically motivated murder of civilians.”227 This success, however, would at best be incomplete: the investigation did not extend to the senior officers who ordered the killings.228

The Dynamics of Patron-Client Relations in the Salvador Civil War

At the outset of U.S. involvement, El Salvador was governed by a weak civilmilitary junta that possessed only notional control over elements of the security forces and the far right, who pursued their own dirty war against suspected leftists and government reformers. Repression, as well as economic and social grievances, drove support for a guerrilla movement that became the most potent in Latin America. In response, the United States tried to encourage a broad counterinsurgency approach that focused on building the Salvadoran government's legitimacy by ameliorating grievances and advancing democratization, while developing a military capable of confronting the insurgents. Although the United States and its Salvadoran partners had a shared aim of preventing the FMLN from taking power, the measures that the United States deemed necessary to achieve these aims frequently met with resistance from elements of the Salvadoran military or government. Nevertheless, the United States succeeded in convincing the Salvadoran government to implement some political and economic reforms; however, it achieved less influence over the military's counterinsurgency tactics, and its impact on the military's human rights record was mixed. What accounts for this variance in outcomes?

In the course of its twelve-year involvement in El Salvador, the United States employed two strategies for using aid to influence the local government: inducement and conditionality (see table 1). On balance, the El Salvador case provides more evidence to support the effectiveness of the conditionality strategy than it does for the inducement approach. The use of conditions on U.S. aid was associated with Salvadoran compliance, even in areas where reform or policy change had previously been resisted. The Carter administration successfully employed conditionality to broker the Christian Democrats' entry into the junta and stave off a right-wing coup attempt in 1980. The Reagan administration's threat to suspend military aid spurred the ESAF to defend the land-to-the-tiller program and compel the formation of a coalition government including the Christian Democrats. Vice President Bush's “new contract” in 1983 tied aid to suppressing right-wing paramilitary groups, which led to a sharp and lasting decline in death squad murders. Both congress and the Reagan administration made assistance contingent on democratization. Even scholars critical of U.S. involvement in El Salvador acknowledge that “had the United States not insisted on elections as a condition for aid, El Salvador's transition to elected government and the first steps towards transforming the social elite into a political class would have come later, if at all.”229 Similarly, unpopular economic austerity measures, negotiations with the FMLN, and the prosecution of the Jesuit case were all advanced by conditions on aid.

Table 1.

Summary of Sixteen Patron-Client Influence Episodes in El Salvador's Civil War

DateU.S. GoalStrategyCompliance
1980 Christian Democrats to join junta conditionality high 
1980 prevent military coup conditionality high 
1980 improve in human rights in return for helicopters conditionality lowa 
1980 restructure junta under President Duarte and purge of human rights abusers conditionality high 
1980 control human rights abuses and death squads inducement low 
1981–83 sustain junta reforms and reduce human rights abuses inducement low 
1981–91 adapt the ESAF for counterinsurgency by embracing small unit tactics inducement low 
1981 prevent right-wing coup and move missionary murder case forward conditionality high 
1981–84 improve human rights standards, ensure elections, defend land reform, and investigate murders of missionaries and land-reform experts conditionality lowc 
1982 include Christian Democrats in a coalition government conditionality high 
1982 defend the “land-to-the-tiller” program conditionality high 
1983 Bush's “new contract”: condemn death squad violence, support elections, defend land reform, pursue murder investigations conditionality high 
1984–85 implement economic austerity measures conditionality high 
1984–86 prosecute additional suspects in land-reform murders conditionality lowd 
1990 promote ESAF cooperation with Jesuit investigation conditionality lowe 
1990–91 prosecute Jesuit murders and negotiate with the FMLN conditionality high 
DateU.S. GoalStrategyCompliance
1980 Christian Democrats to join junta conditionality high 
1980 prevent military coup conditionality high 
1980 improve in human rights in return for helicopters conditionality lowa 
1980 restructure junta under President Duarte and purge of human rights abusers conditionality high 
1980 control human rights abuses and death squads inducement low 
1981–83 sustain junta reforms and reduce human rights abuses inducement low 
1981–91 adapt the ESAF for counterinsurgency by embracing small unit tactics inducement low 
1981 prevent right-wing coup and move missionary murder case forward conditionality high 
1981–84 improve human rights standards, ensure elections, defend land reform, and investigate murders of missionaries and land-reform experts conditionality lowc 
1982 include Christian Democrats in a coalition government conditionality high 
1982 defend the “land-to-the-tiller” program conditionality high 
1983 Bush's “new contract”: condemn death squad violence, support elections, defend land reform, pursue murder investigations conditionality high 
1984–85 implement economic austerity measures conditionality high 
1984–86 prosecute additional suspects in land-reform murders conditionality lowd 
1990 promote ESAF cooperation with Jesuit investigation conditionality lowe 
1990–91 prosecute Jesuit murders and negotiate with the FMLN conditionality high 

Partial compliance derailed by Ronald Reagan's election


Conditions attached, but U.S. aid rushed in the wake of 1981 final offensive before Salvadoran government compliance


Congressional certification requirements subverted by the Reagan administration


Only $5 million in military aid out of $100 million was conditioned on resolving the case


Conditions attached to equipment specifically chosen so that “withholding these materials will not greatly affect the ESAF's basic ability to fight the war”

In contrast, grants of U.S. assistance were not met with reciprocal compliance by the Salvadorans. After Carter rushed military aid to El Salvador in 1981, the ESAF reneged on reform pledges while death squad violence increased. Following Duarte's win in 1984, U.S. military aid was no longer conditioned on the ESAF's human rights practices, and by 1988, extrajudicial killings had made a modest resurgence.230 Conditionality was also absent in areas where the United States had the least influence: the military aspects of the counterinsurgency campaign. Although at times military aid was conditioned on the ESAF's treatment of civilians or its support for reform measures, it was never contingent on changes to its counterinsurgency tactics. Even when it was clear that the army's failure to embrace small-unit operations had produced a military stalemate, the United States did not attempt to force a change.231 The fact that the United States achieved more success in influencing the Salvadoran government in the political and economic spheres than in the military sphere is not surprising given that conditionality was employed in the former, but not in the latter.

The use of conditionality did not result in total compliance with U.S. preferences on every occasion. In some instances, other factors undercut the credibility of U.S. conditions. The 1980 helicopter deal, for example, generated partial compliance by the ESAF, but the credibility of the threat to withhold aid was dissipated by Carter's defeat in the 1980 election before all the terms of the deal were fulfilled. At other times, conditionality's credibility was affected by the relative importance of the aid involved. The Bush administration efforts to use conditionality to pressure the Salvadoran military into cooperating with the investigation of the Jesuit murders was undercut by the fact that the suspended aid was specifically chosen so that it would not adversely affect ongoing military operations.

All things considered, the evidence reviewed here suggests that a patron's influence over a client state flows from the use of tight conditions on aid rather than on the granting of inducements. As the chief of staff of the Salvadoran army, Gen. Adolfo Blandón, acknowledged: “I'll be frank, though some don't want to admit it: The conditions the U.S. placed on us helped.”232

A key challenge in employing conditionality when assisting counterinsurgency, however, is making a credible threat to withhold aid if the client does not comply. The very act of helping a local government signals that its survival is important and links the patron's reputation to the client's success. Given that the patron has an interest in the client's long-term stability, some scholars suggest that the client may not believe that aid will really be withheld for noncompliance.233 Rather, the client may conclude that it will receive aid irrespective of what it does, because the patron's desire to defeat the insurgency outweighs whatever reform or policy change the patron is trying to compel the local government to make.

The fact that withholding assistance will impede the broader goal of arresting an insurgency is no guarantee that the patron will not punish noncompliance. In a number of international interactions, ranging from economic sanctions to nuclear deterrence, states attempt to coerce each other by making credible threats to take actions that, if implemented, could inflict serious harm on the implementing party.234 Unlike nuclear threats, however, the withholding of aid pending client compliance does not pose an existential threat to the client's survival. Except in the most extreme cases, it is unlikely that the local government will immediately collapse from a single instance of aid suspension.235 Even in situations where the threat of a complete cut-off of aid is not believable, the patron can still calibrate conditions to make them more credible. Breaking an aid package into smaller tranches, for example, each clearly linked to a specific action, offers patrons a way to make believable threats to withhold aid from their clients without leaving them completely helpless.236 Thus, there is good reason to believe that a client cannot be totally confident that a patron will not withhold aid simply because it has a long-term interest in the client's success.237

Nevertheless, it is worth asking if a client state could be so important to U.S. national security that it would not be possible to make such threats credible? It is difficult to predict where the United States will intervene to assist counterinsurgency in the future, but from a historical standpoint, the countries that have received aid and support in counterinsurgency from the United States, short of the intervention of ground troops, were overwhelmingly small, resource-poor states whose material capabilities (as measured by the Composite Index of National Capability) were a bare fraction of those of their patron (see table 2). These were not countries that were too important to fail; the loss of any one of these minor client states would have had a minimal impact on the standing or alliances of a superpower such as the United States.238 Of course, irrespective of their validity, geopolitical considerations, reputational concerns, or the fear of falling dominos could lead U.S. policymakers to conclude that a state is vital to U.S. security even if its material endowments are minimal.239 This was certainly true in El Salvador, where successive administrations feared the country would be the first of a string of falling dominos set off by the Nicaraguan revolution and that a failure to arrest the insurgency would undermine the confidence of U.S. allies worldwide.240 This belief that El Salvador was critical to U.S. foreign policy in the region and the reputation of the country did not prevent Carter, Reagan or Bush from bringing coercive pressure to bear on their clients in San Salvador via conditions on aid, even on occasions when the local government was clearly losing to the insurgents and appeared to be on the brink of collapse. At times, it may take more effort to influence a strategically important client government via conditionality, but the client's importance should not be a barrier to the use of this tool.

Table 2.

Completed Counterinsurgency Campaigns Supported by U.S. Indirect Intervention, 1944–2010

CountryInitial Year% USA CINCastResource Endowment
Greece 1947 1.07% poor 
Philippines 1946 0.95% poor 
Laos 1959 0.19% poor 
Guatemala 1960 0.21% poor 
Cambodia 1967 0.51% poor 
Morocco 1975 1.88% poor 
Nicaragua 1978 0.21% poor 
El Salvador 1979 0.35% poor 
Croatia 1992 0.85% poor 
CountryInitial Year% USA CINCastResource Endowment
Greece 1947 1.07% poor 
Philippines 1946 0.95% poor 
Laos 1959 0.19% poor 
Guatemala 1960 0.21% poor 
Cambodia 1967 0.51% poor 
Morocco 1975 1.88% poor 
Nicaragua 1978 0.21% poor 
El Salvador 1979 0.35% poor 
Croatia 1992 0.85% poor 

CINC stands for Composite Index of National Capability

SOURCE: The list of externally supported insurgencies comes from Christopher Paul et al., Paths to Victory: Lessons from Modern Insurgencies (Santa Monica, Calif.: RAND Corporation, 2013), pp. 77–78. The Composite Index of National Capability comes from the National Material Capabilities Dataset version 4.0, as discussed in David J. Singer, “Reconstructing the Correlates of War Dataset on Material Capabilities of States, 1816–1985,” International Interactions, Vol. 14(1987), pp. 115–132. Resource endowment measures the amount of oil and/or diamonds in a country and comes from Desha Girod, “Foreign Aid and Post-Conflict Reconstruction,” Ph.D. dissertation, Stanford University, 2008, p. 39.


In contemporary and historical interventions to assist counterinsurgency, the United States has found that its local partners were significant obstacles to success. A shared interest in defeating an insurgency does not guarantee that a patron and its client government will prefer the same policies for carrying out that task. In theory, U.S. policymakers should assist only those countries with which their interests and priorities clearly align. In practice, however, the problem of adverse selection means that the types of governments most in need of counterinsurgency assistance are largely those that are least likely to heed U.S. guidance. Consequently, policymakers planning to assist counterinsurgency will be best served by expecting a contested relationship with the client from the outset.

The divergent preferences of local allies (or powerful elements within their regimes) can lead them to resist U.S. recommendations for responding to internal violence and instead pursue policies or behaviors that are antithetical to U.S. counterinsurgency doctrine. Under such circumstances, the patron's advice and material support will not necessarily enhance the client's counterinsurgency prowess. Therefore, a key focus of attention should be on bringing the client's behavior into line with the requirements of the patron.

The case of U.S. aid to El Salvador indicates that imposing conditions on aid, rather than granting inducements, is more closely associated with client compliance with the patron's counterinsurgency preferences. Although further empirical testing is necessary to validate the results, the lesson that conditionality is more effective at generating influence is an important one because many of the same contentious patron-client dynamics found in El Salvador plague the United States' relations with its counterinsurgency partners today.

This is not to suggest that El Salvador should necessarily be seen as a model counterinsurgency assistance effort simply because the FMLN was prevented from overthrowing the Salvadoran government. This outcome came at the cost of $6 billion in U.S. military and economic assistance over twelve years and the loss of tens of thousands of Salvadoran lives, and required significant intrusion into the political and economic policies of the local government. Even then, the ability of the United States to affect change in vital areas such as the conduct of the armed forces or the country's economic structures had limits. There was some improvement in the military's respect for human rights and a moderate degree of agrarian reform, but abuses and inequities persisted throughout the war.241 Merely possessing the ability to influence a client's counterinsurgency choices is no guarantee that an assistance effort will succeed or that the costs involved will be proportional to the objectives sought. Conditions work better than inducements, but their use alone cannot guarantee strategic success.

Five policy findings can be extrapolated from this study. First, the need to be able to credibly threaten to withhold aid in the absence of compliance suggests that use of a conditionality strategy favors looser and more ambiguous commitments to the client in an effort to create some doubt about the lengths to which the patron will go to aid it. In other words, if the level of influence that a patron has over its client varies inversely with the perceived level of commitment to that client, the patron will want to maintain a degree of uncertainty about its future intentions.242

Second, employing conditionality to press a local government for changes to its counterinsurgency operations is a controversial practice, because the reflexive response of many officials would be to prop up a failing partner. If shaping the counterinsurgency behavior of a client state is a priority, however, pressing it to make changes when it is weak is the way to succeed. The more directly the insurgents threaten the ruling elites in the local government, the greater the client's incentive to cooperate with the patron.243 Conversely, the client has little reason to undertake reform when the situation is stable and it is assured of support. When the Salvadoran economy was on the brink of collapse and the insurgents were defeating the ESAF in the field, successive U.S. administrations took advantage of the opportunity to use conditionality to force reform on the Salvadoran government.

Third, in a conditionality-based influence strategy, the conditions and their desired policy outcomes should be as unambiguous as possible, with easily measureable metrics and clear deadlines for compliance. Vague requirements that tie aid to demonstrations of “effort” or “progress” introduce a high degree of subjectivity into the process.

Fourth, even after an initial reform or policy change occurs, it may be necessary to maintain pressure on the client regime to ensure that it does not seek to backslide. The protracted effort required to reduce the number of extrajudicial killings in El Salvador, and their modest resurgence in the late 1980s after pressure was removed, suggests that the patron may have to continue conditions on aid for a significant period of time to ensure that reforms remain in place long enough to have their desired effect.

Fifth, if the patron grants one type of aid unconditionally, it may undermine the credibility of attempts to attach conditions to aid in other realms. A client government whose survival is deemed important enough to merit unconditional military assistance, for example, may have reason to question a patron's threats to withhold economic aid in the face of noncompliance with a proposed reform. If conditions on aid are to be credible, they must be applied in all instances.

Counterinsurgency is a challenging undertaking, the difficulties of which are only magnified when one state attempts to assist another state in its counterinsurgency effort. Policymakers facing that challenge must recognize that the prevalent assumption in both counterinsurgency scholarship and U.S. doctrine—that patron-client preferences will closely align—is wrong. Although it is tempting to think that significant amounts of assistance will easily shape a client's behavior and policies, influence is more likely to flow from tight conditions on aid than from boundless generosity. These conditions must be carefully structured so that the requirements are measurable and achievable, and that the aid the client desires most—in all likelihood military aid—is offered as the reward. To enhance its leverage, a patron may need to press its client government hard to make reforms, even when it is at its weakest, rather than take immediate measures to strengthen it against the insurgents. These recommendations to condition aid to a friendly government, to bargain hard with it, and to exploit its vulnerability may run counter to the instincts of many policymakers. But sometimes being a good ally means being a stern friend.


For their helpful comments and advice on prior drafts, the author thanks Janine Davidson, Todd Greentree, Yuen Foong Khong, Catherine Lang, Nina Silove, Hew Strachan, and the anonymous reviewers. This research was made possible by the generous support of the Smith Richardson Foundation's World Politics and Statecraft Fellowship and the John Fell Fund of Oxford University Press.



U.S. Department of Defense, Sustaining U.S. Global Leadership: Priorities for 21st Century Defense (Washington, D.C.: U.S. Department of Defense, 2012), p. 6; and Andrew F. Krepinevich, “Overhauling the Army for the Age of Irregular Warfare,” Wall Street Journal, February 18, 2016.


For an example of such prescriptions, see David Kilcullen, Counterinsurgency (Oxford: Oxford University Press, 2010), p. 160.


Michael R. Gordon and Julie Hirschfeld Davis, “In Shift, U.S. Will Send 450 Advisers to Help Iraq Fight ISIS,” New York Times, June 10, 2015.


Bryce Denno, “Debriefing of Senior and Designated Key Officers Returning from Field Assignments,” September 6, 1963, Senior Officer Debriefing Reports: Vietnam War, 1962–1972, Military History Institute, Carlisle, Pennsylvania (MHI), p. 7.


Important works in this vein include Douglas S. Blaufarb, The Counterinsurgency Era: U.S. Doctrine and Performance, 1950 to the Present (New York: Free Press, 1977); D. Michael Shafer, Deadly Paradigms: The Failure of U.S. Counterinsurgency Policy (Princeton, N.J.: Princeton University Press, 1988); and Benjamin C. Schwarz, American Counterinsurgency Doctrine and El Salvador: The Frustrations of Reform and the Illusions of Nation Building (Santa Monica, Calif.: RAND Corporation, 1991).


Daniel L. Byman, Going to War with the Allies You Have: Allies, Counterinsurgency, and the War on Terrorism (Carlisle, Pa.: Strategic Studies Institute, 2006), p. 3.


The topic under examination is the relative utility of inducement and conditionality to influence the behavior of a client state in counterinsurgency. Why a patron chooses one particular influence approach over another and under which conditions these approaches would be more or less effective are extremely important questions for future research, but answering them is beyond the scope of the present research design.


See, for example, Robert Thompson's comments on the subject in Thompson, “Civic Action in Low-Intensity Warfare,” in Proceedings of the Low Intensity Warfare Conference (Washington, D.C.: U.S. Department of Defense, 1986), p. 74.


Department of the Army, Counterinsurgency, Field Manual (FM) 3–24 (Washington, D.C.: U.S. Department of the Army, 2006).


Ibid., p. 37. The 2014 edition of FM 3–24 recognizes that a local government's interests may not always coincide with those of the United States. It does not, however, examine in great detail the difficulty of influencing a local government to adopt U.S. counterinsurgency principles. See Department of the Army, Insurgencies and Countering Insurgencies, FM 3–24 (Washington, D.C.: U.S. Department of the Army, 2014), pp. 1–8, 10–15.


Stephen Biddle, “The New U.S. Army/Marine Corps Counterinsurgency Field Manual as Political Science and Political Praxis,” Perspectives on Politics, Vol. 6, No. 2 (June 2008), p. 348.


Peter Beinart, “The Surge Fallacy,” Atlantic, September 2015, http://www.theatlantic.com/magazine/archive/2015/09/the-surge-fallacy/399344/.


Rod Nordland, Melissa J. Rubin, and Matthew Rosenberg, “Gulf Widens between U.S. and a More Volatile Karzai,” New York Times, March 17, 2012.


Agency theory does not require a formal delegation to apply. See Ray Rees, “The Theory of Principal and Agent: Part 1,” Bulletin of Economic Research, Vol. 37, No. 1 (January 1985), p. 3.


Jean-Jacques Laffont and David Martimort, The Theory of Incentives: The Principal-Agent Model (Princeton, N.J.: Princeton University Press, 2002), p. 3.


Michael Handel, Weak States in the International System (London: Frank Cass, 1990), p. 121.


Schwarz, American Counterinsurgency Doctrine and El Salvador, p. 77.


Thomas A. Grant, “Government, Politics, and Low-Intensity Conflict,” in Edwin G. Corr and Stephen Sloan, eds., Low-Intensity Conflict: Old Threats in a New World (Boulder, Colo.: Westview, 1992), p. 261.


Bruce Bueno de Mesquita and Alastair Smith, “Foreign Aid and Policy Concessions,” Journal of Conflict Resolution, Vol. 51, No. 2 (April 2007), p. 254.


Daniel L. Byman, “Friends Like These: Counterinsurgency and the War on Terrorism,” International Security, Vol. 31, No. 2 (Fall 2006), p. 82.


Yehuda Kotowitz, “Moral Hazard,” in John Eatwell, Murray Milgate, and Peter Newman, eds., The New Palgrave: A Dictionary of Economics (New York: W.W. Norton, 1987), pp. 207–213.


William E. Odom, On Internal War: American and Soviet Approaches to Third World Clients and Insurgents (Durham, N.C.: Duke University Press, 1992), pp. 9, 103–104, 213–215; and James E. Cross, Conflict in the Shadows: The Nature and Politics of Guerrilla War (New York: Doubleday, 1963), p. 140.


D. Roderick Kiewiet and Matthew D. McCubbins, The Logic of Delegation: Congressional Parties and the Appropriation Process (Chicago: University of Chicago Press, 1991), p. 48.


Christopher P. Carney, “International Patron-Client Relationships: A Conceptual Framework,” Studies in Comparative International Development, Vol. 24, No. 2 (June 1989), pp. 44–45; and S.N. Eisenstadt and Louis Roniger, “Patron-Client Relations as a Model of Structuring Social Exchange,” Comparative Studies in Society and History, Vol. 22, No. 1 (January 1980), pp. 49–50.


Jacob Bercovitch, “Superpowers and Client States: Analyzing Relations and Patterns of Influence,” in Moshe Efrat and Jacob Bercovitch, ed., Superpowers and Client States in the Middle East: The Imbalance of Influence (London: Routledge, 1991), p. 15; and Christopher C. Shoemaker and John Spanier, Patron-Client State Relationships: Multilateral Crises in the Nuclear Age (New York: Praeger, 1984), p. 20.


Bercovitch, “Superpowers and Client States,” p. 16; and Joseph Helman, “The Politics of Patron-Client State Relationships,” Ph.D. dissertation, George Washington University, 2002, p. 316.


John D. Ciorciari, “A Chinese Model for Patron-Client Relations? The Sino-Cambodian Partnership,” International Relations of the Asia-Pacific, Vol. 15, No. 2 (2015), p. 248.


Abraham Ben-Zvi, The United States and Israel: The Limits of the Special Relationship (New York: Columbia University Press, 1994), p. 2; and Carney, “International Patron-Client Relationships,” p. 44.


Celia L. Reynolds and Wilfred T. Wan, “Empirical Trends in Sanctions and Positive Inducements in Nonproliferation,” in Etel Solingen, ed., Sanctions, Statecraft, and Nuclear Proliferation (Cambridge: Cambridge University Press, 2012), p. 58.


Tit-for-tat strategies produce optimum outcomes in an iterative prisoners' dilemma game. See Robert Axelrod, The Evolution of Cooperation (New York: Basic Books, 1984).


William J. Long, Economic Incentives and Bilateral Cooperation (Ann Arbor: University of Michigan Press, 1996), p. 11.


David Cortright, “Incentives and Cooperation in International Affairs,” in Cortright, ed., The Price of Peace: Incentives and International Conflict Prevention (Lanham, Md.: Rowman and Littlefield, 1997), p. 10.


Thomas C. Schelling, Arms and Influence (New Haven, Conn.: Yale University Press, 1966), pp. 70–78.


Klaus Knorr, The Power of Nations: The Political Economy of International Relations (New York: Basic Books, 1975), pp. 166–206; and David A. Baldwin, “The Power of Positive Sanctions,” World Politics, Vol. 24, No. 1 (October 1971), p. 32.


Lloyd Jensen, “Negotiating Strategic Arms Control, 1969–1979,” Journal of Conflict Resolution, Vol. 28, No. 3 (September 1984), pp. 273–303; Martin Patchen, Resolving Disputes between Nations: Coercion or Conciliation? (Durham, N.C.: Duke University Press, 1988), pp. 262–263; and Miroslav Nincic, “Getting What You Want: Positive Inducements in International Relations,” International Security, Vol. 35, No. 1 (Summer 2010), p. 150.


William H. Mott IV, United States Military Assistance: An Empirical Perspective (Westport, Conn.: Greenwood, 2002), pp. 66, 307.


This was most notable in Vietnam. See, for example, Memo, Lansdale to O'Donnell, September 20, 1960, Foreign Relations of the United States (FRUs), 1958–1960, Vol. 1: Vietnam (Washington, D.C.: Government Printing Office [GPO], 1986), p. 580; Telegram 70, Saigon to State, July 14, 1961, FRUS, 1961–63, Vol. 1: Vietnam, 1961 (Washington, D.C.: GPO, 1988), p. 218; and Memorandum, Rostow to Kennedy, August 4, 1961, box 194, NSF: Vietnam, John F. Kennedy Presidential Library, Boston, Massachusetts.


The imposition of such ex-post conditions differs from the approach taken by institutions such as the World Bank, which typically imposes conditions ex ante.


George F. Kennan, Realities of American Foreign Policy (New York: W.W. Norton, 1966), p. 55.


Baldwin, Economic Statecraft, pp. 318–319.


Daniel W. Drezner, The Sanctions Paradox: Economic Statecraft and International Relations (Cambridge: Cambridge University Press, 1999), pp. 27–35.


Paul Collier et al., “Redesigning Conditionality,” World Development, Vol. 25, No. 9 (1997), pp. 1399–1407; David Dollar and Jakob Svensson, “What Explains the Success or Failure of Structural Adjustment Programmes?” Economic Journal, October 2000, pp. 894–917; and Nicolas Van de Walle, African Economies and the Politics of Permanent Crisis, 1979–1999 (New York: Cambridge University Press, 2001), p. 214.


For a discussion about why policymakers would hold this belief, see Drezner, The Sanctions Paradox, p. 4.


Christian Catrina, Arms Transfers and Dependence (New York: Taylor and Francis, 1988), p. 356; and Mott, United States Military Assistance, p. 10.


Department of the Army, Counterinsurgency, FM 3–24 (2006), p. 201; Anthony James Joes, America and Guerrilla Warfare (Lexington: University Press of Kentucky, 2000), p. 277; and Mark Moyar, A Question of Command: Counterinsurgency from the Civil War to Iraq (New Haven, Conn.: Yale University Press, 2009), p. 187.


Alexander L. George and Andrew Bennett, Case Studies and Theory Development in the Social Sciences (London: MIT Press, 2005), pp. 83–85.


John Gerring, “What Is a Case Study and What Is It Good For?” American Political Science Review, Vol. 98, No. 2 (May 2004), p. 348.


Yvon Grenier, The Emergence of Insurgency in El Salvador: Ideology and Political Will (Pittsburgh: University of Pittsburgh Press, 1999), pp. 12–13.


This is the major theme of William Stanley, The Protection Racket State: Elite Politics, Military Extortion, and Civil War in El Salvador (Philadelphia: Temple University Press, 1996).


Fred Woerner, “Report of the El Salvador Military Strategy Assistance Team,” November 16, 1981, accession no. EL00383, Digital National Security Archive, George Washington University (DNSA), p. 47, http://nsarchive.chadwyck.com; and John D. Waghelstein, interview, January 31, 1985, Senior Officers Oral History Program, MHI, p. 37.


These groups were the Central American Workers' Revolutionary Party, the People's Revolutionary Army, the Farabundo Martí Popular Liberation Forces, the Armed Forces of National Resistance, and the Communist Party of El Salvador's Armed Forces of Liberation.


“Proclamation of the Armed Forces of the Republic of El Salvador,” October 15, 1979, reprinted in Stanley, The Protection Racket State, pp. 267–269.


This represented approximately 16 percent of the officer corps and 10 percent of the enlisted ranks. See Telegram 7163, San Salvador to State, October 16, 1980, State Department Freedom of Information Act Virtual Reading Room (DOS/FOIA), https://foia.state.gov/search/search.aspx; and Letter, Office of Central American Affairs to Harrison, November 14, 1980, DOS/FOIA.


Brian J. Bosch, The Salvadoran Officer Corps and the Final Offensive of 1981 (London: McFarland, 1999), p. 43; and CIA, “El Salvador: The Right Wing,” April 17, 1981, El Salvador Human Rights Collection, Hispanic Reading Room, Library of Congress (ESHRC).


Todd Greentree, Crossroads of Intervention: Insurgency and Counterinsurgency Lessons from Central America (Westport, Conn.: Praeger, 2008), pp. 78–79.


Bosch, The Salvadoran Officer Corps and the Final Offensive of 1981, p. 34; and Tommie Sue Montgomery, Revolution in El Salvador: From Civil Strife to Civil Peace (Boulder, Colo.: Westview, 1994), p. 128.


Montgomery, Revolution in El Salvador, p. 116.


Hugh Byrne, El Salvador's Civil War: A Study of Revolution (Boulder, Colo.: Lynne Rienner, 1996), pp. 77, 151.


Stanley, The Protection Racket State, p. 183.


Ibid., p. 296 n. 6.


Telegram 529, San Salvador to State, January 24, 1980, accession no. EL00652, DNSA; and Fred Woerner, interview, November 7, 1987, Oral History of the Conflict in El Salvador, Vol. 2, MHI, p. 5.


Telegram 312, San Salvador to State, January 11, 1980, DOS/FOIA; and Telegram 523, San Salvador to State, February 6, 1980, DOS/FOIA.


Bureau of Public Affairs, “El Salvador: U.S. Policy,” March 1980, DOS/FOIA.


Telegram 7724, San Salvador to State, November 4, 1980, DOS/FOIA.


The far right is described as a “latent threat.” See Memorandum, Vance to Carter, December 26, 1979, accession no. EL00648, DNSA.


Byrne, El Salvador's Civil War, pp. 43–44.


Telegram 96, San Salvador to State, January 6, 1981, accession no. ES01131, DNSA; and “El Salvador: The Role of Roberto D'Aubuisson,” March 4, 1981, document no. 0000654925, Central Intelligence Agency FOIA Electronic Reading Room (CERR), http://www.foia.cia.gov.


Telegram 1257, San Salvador to State, February 20, 1980, DOS/FOIA.


Telegram 8084, San Salvador to State, November 19, 1980, DOS/FOIA; and “El Salvador: Controlling Rightwing Terrorism,” January 5, 1985, document no. 0000046974, CERR, p. 7.


William M. LeoGrande, Our Own Backyard: The United States in Central America, 1977–1992 (Chapel Hill: University of North Carolina Press, 1998), p. 50.


Memorandum, Situation Room to Brzezinski, January 22, 1980, document no. CK2349696662, Declassified Documents Reference System (DDRS), http://gdc.gale.com/products/declassified-documents-reference-system; and Telegram 1336, San Salvador to State, February 22, 1980, DOS/FOIA.


Telegram 1358, San Salvador to State, February 23, 1980, DOS/FOIA; Telegram 1886, San Salvador to State, March 13, 1980, DOS/FOIA; and Telegram 2611, San Salvador to State, April 12, 1980, DOS/FOIA.


Telegram 1099, San Salvador to State, February 17, 1980, DOS/FOIA; Telegram 1116, San Salvador to State, February 19, 1980, DOS/FOIA; and Telegram 50819, State to San Salvador, February 26, 1980, accession no. ES00463, DNSA.


Telegram 1886, San Salvador to State; Memorandum, Bushell to SecState, March 13, 1980, DOS/FOIA; and Telegram 2038, San Salvador to State, March 19, 1980, DOS/FOIA.


Memorandum, Bushell to SecState; and Alan Riding, “U.S. Aid to Salvador Army,” New York Times, February 23, 1980.


Memorandum, Bushell to SecState.


Memorandum to Carter, “Helicopters for El Salvador,” June 18, 1980, DOS/FOIA; and Telegram 210613, State to San Salvador, August 9, 1980, DOS/FOIA.


Telegram 5810, San Salvador to State, August 23, 1980, DOS/FOIA.


Telegram 6729, San Salvador to State, September 29, 1980, DOS/FOIA.


Telegram 8332, San Salvador to State, November 30, 1980, DOS/FOIA; and CIA Cable, San Salvador to Langley, December 1, 1980, accession no. EL00042, DNSA.


Telegram 8281, San Salvador to State, November 27, 1980, DOS/FOIA.


Telegram 8117, San Salvador to State, November 20, 1980, DOS/FOIA; and Telegram 8281, San Salvador to State.


Memorandum, Derian to Deputy Secretary of State, December 4, 1980, DOS/FOIA.


Greentree, Crossroads of Intervention, pp. 83–84.


“Evening Reading: El Salvador,” December 5, 1980, accession no. ES00930, DNSA; and Memorandum, Bushnell to Deputy Secretary, December 11, 1980, DOS/FOIA.


Memorandum, Bushnell to Deputy Secretary.


“December 12 Statement on the Special Presidential Mission to El Salvador—December 6–9, 1980,” December 12, 1980, accession no. ES00983, DNSA.


Telegram 329627, State to San Salvador, December 13, 1980, accession no. ES00985, DNSA.


Telegram 8717, San Salvador to State, December 13, 1980, accession no. ES00988, DNSA.


Telegram 335609, State to All ARA Posts, December 20, 1980, accession no. ES01033, DNSA.


Telegram 11023, Caracas to State, December 14, 1980, DOS/FOIA.


Telegram 333735, Sate to San Salvador, December 18, 1980, accession no. ES01018, DNSA; and Memorandum, Muskie to Carter, “Security Assistance to El Salvador,” January 8, 1981, accession no. ES01153, DNSA.


Telegram 8963, San Salvador to State, December 23, 1980, DOS/FOIA; and Telegram 9059, San Salvador to State, December 31, 1980, DOS/FOIA.


From 1967–1968 Komer had been the head of rural pacification programs in South Vietnam. See Memorandum, Under Secretary for Policy to Secretary of Defense, January 8, 1981, accession no. EL00380, DNSA.


“Evening Reading: Military Assistance to El Salvador,” December 29, 1980, DOS/FOIA.


Bosch, The Salvadoran Officer Corps and the Final Offensive of 1981, p. 85.


Ibid., p. 90; and Telegram 380, San Salvador to State, January 16, 1981, accession no. EL00698, DNSA.


Bosch, The Salvadoran Officer Corps and the Final Offensive of 1981, p. 105.


Telegram 12962, State to San Salvador, January 17, 1981, accession no. ES01223, DNSA.


Garcia quoted in Bosch, The Salvadoran Officer Corps and the Final Offensive of 1981, p. 97. See also Telegram 410, San Salvador to State, January 16, 1981, accession no. EL00700, DNSA.


Telegram 9213, State to San Salvador, January 14, 1981, accession no. ES01182, DNSA; Memorandum, Muskie to Carter; and Telegram 380, San Salvador to State.


The distinction between Reagan's rhetoric on Central America and his actions is discussed in Russell Crandall, The Salvador Option: The United States in El Salvador, 1977–1992 (Cambridge: Cambridge University Press, 2016), pp. 201–223.


National Security Advisor Richard Allen quoted in Christopher Dickey, “Envoy Assails Reagan Aides on El Salvador,” Washington Post, December 10, 1980. Reagan described the situation as a “three-way civil war” between “a moderate government, a right-wing faction and a left-wing faction.” Bob Levin, “El Salvador's Short Fuse,” Newsweek, January 19, 1981, p. 42.


Telegram 26592, State to San Salvador, February 2, 1981, box 29, Executive Secretariat, NSC Country File, Ronald W. Reagan Presidential Library (RRL), Simi Valley, California; and Juan de Onis, “U.S. Firmly Endorses Salvador Reforms,” New York Times, February 19, 1981. For Reagan's views on the link between social inequity and revolution, see Douglas Brinkley, ed., The Reagan Diaries (New York: HarperCollins, 2009), p. 24.


This was part of a broader $19.5 million Central Intelligence Agency program that also sought to interdict arms supplies for the FMLN from Nicaragua and Honduras. See Bob Woodward, Veil: The Secret Wars of the CIA, 1981–1987 (New York: Simon and Schuster, 1987), p. 91; and LeoGrande, Our Own Backyard, p. 89.


U.S. Department of State, American Foreign Policy: Current Documents, 1981 (Washington, D.C.: GPO, 1982), pp. 1326–1330.


Quoted in LeoGrande, Our Own Backyard, p. 94.


LeoGrande, Our Own Backyard, p. 89.


State Department official quoted in Loren Jenkins, “Arms Put U.S. Stamp on Salvadoran War,” Washington Post, May 7, 1981.


Francis X. Clines, “U.S. May Increase Salvador Advisers,” New York Times, March 5, 1983; and U.S. Department of State, American Foreign Policy: Current Documents, 1983 (Washington, D.C.: GPO, 1984), p. 1320.


Memorandum, Shultz to Reagan, “Managing Our Central America Strategy,” May 25, 1983, RRL; “Summary NSC Paper on El Salvador,” attached to Memorandum, Richard Allen, “Paper for NSC Meeting on February 11, 1981,” February 10, 1981, box 1, Executive Secretariat, NSC Meeting Files, RRL; and Telegram 204069, Abrams to SecState, July 23, 1982, DOS/FOIA. Continued U.S. assistance was seen as a key factor restraining the ESAF from adopting the same kind of “self-defeating brutality” tactics employed by the Guatemalan government in its own civil war. See State Department Briefing Paper, “Central America,” June 26, 1982, accession no. ES03190, DNSA.


Memorandum, Shultz to Reagan; and Bernard Weinraub, “U.S. Envoy to Salvador Ordered to Stop Criticizing Rights Abuses,” New York Times, November 10, 1982.


Memorandum, Haig to Reagan, January 26, 1981, box 1, Executive Secretariat, NSC Meeting Files, RRL.


For evidence that senior U.S. officials believed in using conditionality against hostile states, see the discussion at the National Security Council meeting on February 11, 1981, in Jason Saltoun-Ebin, ed., The Reagan Files: Inside the National Security Council, Vol. 2 (Santa Barbara, Calif.: Seabec, 2014), pp. 6–9. For a more general discussion of the value of positively engaging with friendly Latin American nations in need of reform rather than sanctioning them, see the National Security Council meeting on February 6, 1981, in ibid., pp. 2–6.


Crandall, The Salvador Option, pp. 219–220.


Telegram 85909, State to San Salvador, April 4, 1981, accession no. ES01530, NSA; and Al Kamen, “Salvador Conflict Centers on Army's Role,” Washington Post, April 21, 1981.


Telegram 54855, State to San Salvador, March 6, 1981, box 29, Executive Secretariat, NSC Country File, RRL; and Juan de Onis, “Haig Opposes Coup by Salvador's Right,” New York Times, March 5, 1981.


Telegram 2671, San Salvador to State, April 8, 1981, accession no. ES01549, DNSA.


Telegram 112041, State to San Salvador, April 30, 1981, accession no. ES01637, DNSA; and Telegram 3279, San Salvador to State, April 30, 1981, accession no. ES01635, DNSA.


Greentree, Crossroads of Intervention, p. 100.


Montgomery, Revolution in El Salvador, p. 152.


Figure includes total killed, wounded, and missing in action. See Edward Cody, “Army's Toll Doubles in El Salvador,” Washington Post, August 12, 1983.


The comparable figure for the United States in Vietnam was 10 percent.


Loren Jenkins, “Silent Stares of Death Routine in El Salvador,” Washington Post, May 3, 1981.


LeoGrande, Our Own Backyard, p. 622 n. 49; Memorandum, Allen to Reagan, “Secretary Haig's Recommendations on El Salvador,” [n.d.], box 30, Executive Secretariat, NSC Country File, RRL; and Memorandum, Schweitzer and Fontaine to Allen, “Situation in El Salvador,” August 19, 1981, box 30, Executive Secretariat, NSC Country File, RRL.


Crandall, The Salvador Option, p. 364.


LeoGrande, Our Own Backyard, pp. 153–154.


José Napoleón Duarte with Diana Page, Duarte: My Story (New York: G.P. Putnam's Sons, 1986), p. 170.


Sections 728(b), (d), and (e) of the International Security and Development Cooperation Act of 1981, P.L. 97–113, December 29, 1981.


Crandall, The Salvador Option, p. 260.


Cynthia J. Arnson, Crossroads: Congress, the President, and Central America, 1976–1992 (State College: Penn State University Press, 1993), pp. 84–91, 100–105, 119.


Telegram 803, San Salvador to State, November 18, 1981, accession no. ES02222, DNSA.


Woerner, interview, p. 37.


Greentree, Crossroads of Intervention, pp. 140–141.


Byrne, El Salvador's Civil War, pp. 78–79.


Moyar, A Question of Command, p. 174.


Waghelstein, interview.


LeoGrande, Our Own Backyard, pp. 137–138.


Montgomery, Revolution in El Salvador, p. 158; and Duarte with Page, Duarte, p. 178.


Greentree, Crossroads of Intervention, p. 139. The idea that “an armed minority” could “shoot its way into power” was anathema to senior U.S. officials. See Memorandum, Bremer to Clark, March 7, 1983, box 30, Executive Secretariat, NSC Country File, RRL.


Enrique A. Baloyra, “The Salvadoran Elections of 1982–1991,” Studies in Comparative International Development, Vol. 28, No 3 (Fall 1993), p. 24.


Christopher Dickey, “Salvadoran Prelate Warns of Return to Violence,” Washington Post, April 5, 1982.


Telegram 84176, State to San Salvador, April 6, 1982, box 29, Executive Secretariat, NSC, Country File, RRL.


Joanne Omang, “As Salvadoran Politics Boil, U.S Envoy Shifts Attention,” Washington Post, April 24, 1982.


A personal representative of Secretary Haig hand carried letters to President Duarte, General Garcia, and the heads of the major political parties explicitly specifying the conditions required for economic and military aid to continue. See Telegram 104143, State to San Salvador, April 18, 1982, DOS/FOIA.


Telegram 3497, San Salvador to State, April 24, 1982, accession no. ES02959, DNSA; and Jose Guillermo Garcia, interview, July 2, 1987, Oral History of the Conflict in El Salvador, Vol. 4, MHI, pp. 49–52.


Intelligence Community Assessment, “El Salvador: Performance on Certification Issues,” July 27, 1982, ESHRC.


Bernard Weinraub, “Proposal for Aid to Salvador Cut by Senate Panel,” New York Times, May 27, 1982.


Christopher Dickey, “El Salvador Gives Farmers Land to Show Progress in Reforms,” Washington Post, June 5, 1982; Shirley Christian, “Salvadorans Battle Erosion of Land Reform,” Miami Herald, June 13, 1982; and Intelligence Community Assessment, “El Salvador: Performance on Certification Issues.”


LeoGrande, Our Own Backyard, p. 170.


Cody, “Army's Toll Doubles in El Salvador.”


Thomas Pickering, interview, August, 28, 1987, Oral History of the Conflict in El Salvador, Vol. 1, MHI, p. 4.


Byrne, El Salvador's Civil War, p. 104.


CIA, “El Salvador: Performance on Certification Issues,” July 13, 1983, accession no. EL00100, DNSA; and Arnson, Crossroads, p. 140.


Philip Taubman, “U.S. Said to Weigh 40% Increase in Military Funds for Latin Allies,” New York Times, July 17, 1983.


Telegram 1543, San Salvador to State, February 23, 1983, document no. CK2349619526, DDRS.


Taubman, “U.S. Said to Weigh 40% Increase in Military Funds for Latin Allies.”




Letter, Reagan to Magaña, December 10, 1983, document no. CK2349554603, DDRS.


Bureau of Public Affairs, U.S. Condemns Salvadoran Death Squads, Current Policy No. 533 (Washington, D.C.: U.S. Department of State, December 11, 1983).


Telegram 11567, San Salvador to State, December 14, 1983, accession no. EL00815, DNSA.








Telegram 11456, San Salvador to State, December 11, 1983, accession no. EL00813, DNSA.




CIA, “Military Commanders' Resentment and Opposition to U.S. Government Pressure,” January 25, 1984, document no. 0000049079, CERR.


Cynthia McClintock, Revolutionary Movements in Latin America: El Salvador's FMLN and Peru's Shining Path (Washington, D.C.: United States Institute of Peace Press, 1998), p. 151.


Telegram 11882, San Salvador to State, December 22, 1983, accession no. EL00818, DNSA; “El Salvador: Crackdown on Death Squads,” December 23, 1983, document no. 0000049367, CERR; and CIA Cable, “Actions Taken by the Minister of Defense against Salvadoran Officials Allegedly Involved in Right-Wing Death Squad Activities,” December 23, 1983, document no. 0000049076, CERR.


CIA, “Recommendations by Salvadoran Army Political Commission on Points Raised during the Visit of the United States Vice President,” December 14, 1983, accession no. EL00113, DNSA; Telegram 11707, San Salvador to State, December 17, 1983, accession no. EL00817, DNSA; Memorandum for the Record, “Briefing of the House Permanent Select Committee on Intelligence Staff on the Central American Finding,” January 26, 1984, document no. 0001309773, CERR; and Robert McCartney, “U.S. Lauds Drive to Halt Death Squads,” Washington Post, January 8, 1984.


CIA, “El Salvador: Dealing with Death Squads,” January 20, 1984, accession no. EL00117, DNSA, p. 1; and “El Salvador: Controlling Rightwing Terrorism,” p. iii. The human rights group Tutela Legal reported 1,259 death squad killings in 1983 and 224 such murders in 1984. See Congressional Research Service (CRS), “El Salvador, 1979–1989: A Briefing Book on U.S. Aid and the Situation in El Salvador” (Washington, D.C.: Library of Congress, April 28, 1989), p. 88.


Telegram 2447, San Salvador to State, March 8, 1984, DOS/FOIA.


Ibid., p. 4.


Per the Polity IV dataset, the election of 1984 was the first time that El Salvador passed the threshold of democracy. See “El Salvador,” Polity IV Project: Political Regime Characteristics and Transitions, 1800–2008 (Vienna, Va.: Center for Systemic Peace, 2008), http://www.systemicpeace.org/polity/sal2.htm.


Fearing that D'Aubuisson would block U.S.-backed reforms—leading Congress to cut aid to El Salvador—President Reagan authorized the CIA to give $1.4 million to the Christian Democrats and PCN. See Robert McCartney, “U.S. Seen Assisting Duarte in Sunday's Salvadoran Vote,” Washington Post, May 4, 1984.


James LeMoyne, “A Salvadoran Police Chief Vows to End Rights Abuses,” New York Times, July 1, 1984; Lydia Chavez, “Salvador to Restructure Security Forces,” New York Times, June 14, 1984; and CIA, “El Salvador: Duarte's First 100 Days,” September 14, 1984, ESHRC, p. 1.


CIA, “El Salvador: Duarte's First 100 Days,” p. 1; and “El Salvador: Controlling Rightwing Terrorism,” p. 10.


CRS, “El Salvador, 1979–1989,” p. 88.


Terry Lynn Karl, “After La Palma: Christian Democracy, U.S. Policy, and the Prospects for Democratization in El Salvador,” World Policy Journal, Winter 1985, pp. 318–319.


LeoGrande, Our Own Backyard, p. 257.


Telegram 2104, San Salvador to State, February 19, 1985, accession number EL00866, DNSA; A.J. Bacevich et al., American Military Policy in Small Wars: The Case of El Salvador (Boston: Potomac, 1988), pp. 32–33; Byrne, El Salvador's Civil War, p. 88; and LeoGrande, Our Own Backyard, pp. 267–268.


LeoGrande, Our Own Backyard, pp. 256–259.


Congressman Gerry Studds quoted in “House Takes Up First of Three Salvador Military Aid Plans,” Associated Press, May 10, 1984.


LeoGrande, Our Own Backyard, p. 273.


Ibid., pp. 257–258; and “Congress Gives Reagan Aid for El Salvador,” in CQ Almanac 1984 (Washington, D.C.: Congressional Quarterly, 1985), http://library.cqpress.com/cqalmanac/cqal84-1152061.


James LeMoyne, “Salvadoran Right Blocks Land Plan,” New York Times, June 30, 1984.


Telegram 8711, San Salvador to State, June 29, 1988, accession no. EL00960, DNSA.


Montgomery, Revolution in El Salvador, p. 190.


James LeMoyne, “Duarte's Power Seems Eroded as Voting Nears,” New York Times, February 17, 1985.


Majorie Miller, “Thousands of Salvadorans March to Protest Duarte's Economic Austerity Plan,” Los Angeles Times, February 22, 1986; and Stanley, The Protection Racket State, p. 237.


Peter Ford, “Civil War Undermines Efforts to Boost Salvador's Ailing Economy,” Christian Science Monitor, May 21, 1986, http://www.csmonitor.com/1986/0521/osalv2.html.


CIA, “The Salvadoran Military: A Mixed Performance,” June 1984, document no. 0000761618, CERR, p. 9.


Bacevich et al., American Military Policy in Small Wars, p. 37.


Quoted in Michael Childress, The Effectiveness of U.S. Training Efforts in Internal Defense and Development: The Cases of El Salvador and Honduras (Santa Monica, Calif.: RAND Corporation, 1995), p. 31.


Ibid., p. 28.


Memo, Levitsky to Lord, “Vice President Quayle's Meeting with Salvadoran Armed Forces Officers,” February 1, 1989, accession no. EL00990, DNSA.


LeoGrande, Our Own Backyard, p. 277; and Brands, Latin America's Cold War, p. 205.


Quoted in Dan Williams, “Signs Positive but Reality Is Still Painful in Salvador,” Los Angeles Times, May 30, 1985.


As a sign of El Salvador's political liberalization, a party affiliated with the FMLN participated in the elections, winning 3.8 percent of the vote.


LeoGrande, Our Own Backyard, pp. 264, 565–568.


Intelligence Information Report [redacted] to DIA, “FMLN Final Offensive 1989—A Wrap Up,” December 22, 1989, document no. CK2349582498, DDRS.


Schwarz, American Counterinsurgency Doctrine and El Salvador, p. 21.


Stanley, The Protection Racket State, p. 249.


Ibid., pp. 249–250.


Chris Norton, “After Salvador's Rebel Offensive,” Christian Science Monitor, December 8, 1989, http://www.csmonitor.com/1989/1208/oguer.html. The assessment of the Tet Offensive comes from James J. Wirtz, The Tet Offensive: Intelligence Failure in War (Cornell, N.Y.: Cornell University Press, 1994), p. 17.


Greentree, Crossroads of Intervention, p. 151; and Byrne, El Salvador's Civil War, p. 173.


Michael Gordon, “General Says Salvador Can't Defeat Guerrillas,” New York Times, February 9, 1990.


Telegram 2104; and Telegram 6383, San Salvador to State, May 22, 1985, accession no. EL00875, DNSA. For ESAF views of the Jesuits of Central American University, see Greentree, Crossroads of Intervention, p. 153. For a similar perspective from U.S. military trainers, see Arnson, Crossroads, p. 247.


Robert Pear, “House Rejects Curb on Salvador Aid,” New York Times, November 21, 1989.


Telegram 15668, San Salvador to State, December 5, 1989, accession no. EL01061, DNSA.


“Cristiani Names Colonel, Eight Others as Presumed Culprits in Massacre,” Associated Press, January 13, 1990. The effects of U.S. pressure on Cristiani and the pace of the investigation are discussed in CIA message, [redacted], “New Developments in the Investigation into the Jesuit Murders,” January 12, 1990, ESHRC.


Maureen Dowd, “Bush Seeks a Rise in Aid to Salvador,” New York Times, February 2, 1990.


Telegram 2787, San Salvador to State, February 27, 1990, accession number EL01116, DNSA; Memorandum, Byron to Assistant Secretary of Defense for International Security Affairs, “El Salvador Update,” August 15, 1990, accession number EL00529, DNSA; and Memorandum, Byron to Assistant Secretary of Defense for International Security Affairs, “El Salvador Security Assistance,” August 16, 1990, accession number EL00530, DNSA.


Phil Bronstein, “U.S. Officials Reverse View on Salvador Army,” San Francisco Examiner, May 6, 1990.


Memorandum, Byron to Assistant Secretary of Defense for International Security Affairs, “El Salvador Security Assistance.”


Telegram 275329, State to San Salvador, August 17, 1990, accession no. EL01173, DNSA; and Telegram 11127, San Salvador to State, August 18, 1990, accession no. EL01176, DNSA.


Telegram 10914, San Salvador to State, August 14, 1990, accession no. EL01170, DNSA.


Telegram 13617, DAO San Salvador to DIA, October 5, 1990, accession no. EL00541, DNSA.


Telegram [excised], State to San Salvador, “$250,000 to Be Released of Remaining FY90 Military Aid,” September 1990, accession no. EL01184, DNSA; Telegram 331044, State to San Salvador, September 29, 1990, accession no. EL01214, DNSA; and Telegram 13730, San Salvador to State, October 18, 1990, EL01208, DNSA.


Telegram 13730, San Salvador to State; Telegram 12054, San Salvador to State, October 7, 1990, accession no. EL01203, DNSA; and Telegram [illegible], San Salvador to State, October 15, 1990, accession no. EL01206, DNSA.


Letter, CDR USMILGRP El Salvador to CINCSOUTH, February 10, 1991, ESHRC.


JTIC Advisory 15–90, “The Gremlin Comes to El Salvador,” November 29, 1990, accession no. EL00563, DNSA.


After the fact, some guerrilla commanders suggested that the real purpose of the 1989 offensive was to strengthen their negotiating position vis-à-vis the government.


Section 531(d) of the Foreign Operations, Export Financing, and Related Programs Appropriations Act, Public Law 101–513, November 5, 1990.


LeoGrande, Our Own Backyard, p. 575.




Ibid., p. 576.


The UN Commission on the Truth for El Salvador subsequently alleged that the kill order originated with Col. René Emilio Ponce and was given in the presence of several other senior officers. See UN Security Council, From Madness to Hope: The 12-year War in El Salvador—Report of the Commission on the Truth for El Salvador, S/25500, 1993, pp. 38–47.


Stanley, The Protection Racket State, p. 254.


The killings, however, occurred on a vastly smaller scale than in the early 1980s. See Telegram 8711, San Salvador to State.


Memo, Levitsky to Lord.


Mark Hatfield, Jim Leach, and George Miller, Bankrolling Failure: United States Policy in El Salvador and the Urgent Need for Reform (Washington, D.C.: Arms Control and Foreign Policy Caucus, November 1987), p. 26.


Glenn Snyder, Alliance Politics (Ithaca, N.Y.: Cornell University Press, 1997), pp. 168–170; Stephen M. Walt, The Origins of Alliances (Ithaca, N.Y.: Cornell University Press, 1987), pp. 43–44; Shoemaker and Spanier, Patron-Client State Relationships, pp. 20–21; and Shafer, Deadly Paradigms, p. 120.


James D. Fearon, “Rationalist Explanations for War,” International Organization, Vol. 49, No. 3 (Summer 1995), pp. 379–414; and Paul K. Huth, “Deterrence and International Conflict,” Annual Review of Political Science, Vol. 2 (1999), pp. 25–48.


If that appears legitimately imminent, the patron can always reverse its policy.


Thomas C. Schelling, The Strategy of Conflict (Cambridge, Mass.: Harvard University Press, 1960), p. 41.


As an independent actor, the local government will simultaneously take actions to avoid the patron's pressure and to secure a stronger commitment. Such measures can range from raising the specter of military defeat or internal collapse should the patron not provide sufficient aid, to cultivating alternative sources of assistance or even threatening to defect to a rival patron. For the logic of the threat of collapse, see Schelling, The Strategy of Conflict, p. 37. The effectiveness of threats to defect to a rival depends on the prevailing geopolitical circumstances. It would have been hard, for example, for a client government fighting against a communist-linked insurgency during the Cold War to credibly threaten to defect to the Soviet or Chinese camp if the United States withheld aid.


Kenneth N. Waltz, Theory of International Politics (London: Addison-Wesley, 1979), p. 169.


Historically, great powers have repeatedly overestimated the threat that disorder in peripheral states poses to their core interests. See Jack Snyder, “Imperial Myths and Threat Inflation,” in A. Trevor Thrall and Jane K. Cramer, eds., Politics of Fear: Threat Inflation since 9/11 (New York: Routledge, 2009), p. 41. On the exaggerated fear of falling dominos during the Cold War, see Jerome Slater, “Dominos in Central America: Will They Fall? Does It Matter?” International Security, Vol. 12, No. 2 (Fall 1987), pp. 105–134.


LeoGrande, Our Own Backyard, p. 33; and Saltoun-Ebin, The Reagan Files, Vol. 2, pp. 6, 9.


An estimated 25 percent of the rural population (87,547 small-farm families) benefitted from the transfer of 22 percent of the country's farmland. See CRS, “El Salvador, 1979–1989,” p. 10.


Robert Jervis, The Logic of Images in International Relations (Princeton, N.J.: Princeton University Press, 1970), pp. 87–88.


Steven R. David, Choosing Sides: Alignment and Realignment in the Third World (Baltimore, Md.: Johns Hopkins University Press, 1991), p. 6.