Expanding the State: Rural Mobilization and Bureaucratic Presence in Peru

Abstract

Governments in the Global South have historically lacked bureaucratic capacity in peripheral areas, making it difficult for them to maintain order and implement their preferred policies. This paper explores when central states invest in establishing bureaucratic presence. We argue that twentieth-century peasant mobilization provided an important impetus for incumbents to create bureaucratic agencies. Using a novel dataset of bureaucratic presence and municipal-level data on peasant collective action in mid-twentieth-century Peru, we show that the Peruvian president responded to peasant mobilization by investing in new bureaucratic offices that could respond to peasant demands. We further demonstrate that these investments in bureaucratic offices endured, enabling the implementation of pro-peasant policies and reducing violence during the Shining Path insurgency of the late 20th century. This suggests a surprising reversal: early peasant mobilization is associated with a long-term reduction in later rural unrest.

Keywords

Latin American politics bureaucracies state building social movements politics of growth/development rural mobilization

Central states in the Global South have historically been characterized by an uneven territorial reach and persistent inability to implement their desired policies in peripheral areas.¹ Many of these issues arise from a lack of bureaucratic institutions and agents to reliably enforce central state directives (Soifer, 2015). This paper investigates when governments overcome these challenges and establish a bureaucratic presence in the periphery.

We posit that peasant mobilization often provided a powerful impetus for incumbents to invest in bureaucratic agencies. The 20th century was characterized by extensive rural unrest from China and Algeria to Vietnam and Mexico. A central demand of many of these movements involved extensive structural policy changes, including land reform and an end to labor exploitation. Incumbents could rarely depend solely on traditional strategies of addressing social mobilization: a one-time distribution of targeted benefits (“fire-fighting”) failed to fully address the broader structural concerns of peasant movements, and violent repression risked backfiring (Finkel, 2015; Moore, 2000). Bureaucratic presence, on the other hand, could complement the weaknesses of these other strategies. It was non-violent—and thus less likely to provoke backlash. Furthermore, by establishing a permanent office in the periphery, the government signaled a more enduring commitment to address peasants’ structural demands. Through land reform agencies and peasant response centers, the government would not only provide a one-time payout to peasants but would enact a broader and more transformative set of programs. Yet, these agencies were also costly to establish, such that incumbents could not deploy them everywhere.

We argue that incumbents targeted bureaucratic presence to areas with a recent history of localized peasant mobilization. These were areas where future mobilization was most likely to occur (past mobilization predicts future mobilization, absent government action) and from which a broader rural rebellion might draw support. As such, the downside risk of failing to respond to peasant demands was high. Furthermore, these areas were ones in which peasant demands were clearly structural; addressing peasant grievances thus required more than short-term fire-fighting. As we argue, this strategy often worked: bureaucratic presence proved an effective way of channeling localized peasant demands into institutions and preventing the emergence of widespread rural unrest and rebellion.

The above argument is most likely to hold in cases where incumbents had a financial windfall (from foreign entities or resource shocks) that could offset the high cost of expanding bureaucratic presence and that was under the near-full control of the executive.² This scope condition was especially likely to arise in an understudied context for bureaucratic expansion: 1960s Latin America. The successful Cuban revolution of 1959 provided non-Marxist incumbents with a powerful demonstration of what might occur if rural mobilization were not contained (Weyland, 2019).³ The demonstration effect of the Cuban revolution also increased foreign assistance for non-Marxist incumbents to address rural unrest. Most notably, the US-backed Alliance for Progress (AFP) sought to curb Marxist uprisings in Latin America by providing financial support to national executives to foster rural development and conduct land reforms, both of which entailed establishing bureaucratic agencies in the periphery.⁴ Absent such funding, incumbents lacked the own-source revenue to fund expensive bureaucratic investments.⁵

We evaluate our argument using the empirical case of 1960s Peru, which was characterized by extensive but generally localized peasant mobilization. During this period, rural agricultural workers invaded large estates (haciendas), engaged in locally organized protests, and participated in strikes. These mobilizational events were frequent and impactful but generally contained to the estates or localities where they occurred. Peru’s President Fernando Belaúnde, who had been popularly elected, invested in two programs designed to respond to peasants’ primary concerns for land reform and rural development: Popular Cooperation (CP) and an agrarian reform agency (SIPA).⁶ Our argument suggests that Belaúnde should have established offices of CP and SIPA (i.e. bureaucratic presence) in areas with a recent history of peasant mobilization.

To test this prediction, we use a research design that exploits spatial and temporal variation in peasant mobilization during Belaúnde’s first term in office (1963–1968). We use a municipal-level measure of peasant mobilization, which includes all instances of peasant-led protest, targeted violence, strikes, and land invasions.⁷ Our measures of bureaucratic presence include the location of CP and SIPA offices, which were each staffed by central government-appointed bureaucrats. Using a difference-in-differences approach, we show that mobilization is a very strong and significant predictor of the location of CP and SIPA offices. Using archival records produced by the Belaúnde administration, we further demonstrate that Belaúnde and the members of his government were primarily concerned about peasant collective action and that these dynamics shaped the decision to allocate bureaucratic agencies.

We further show that these investments in bureaucratic presence reflected a sustained commitment to addressing peasant concerns. Future governments maintained and utilized the offices established by Belaúnde to implement rural development programs and land reform.⁸ Ultimately, these agencies appear to have been effective in achieving their desired aim: municipalities with bureaucratic presence had lower levels of rural violence during Peru’s prolonged civil conflict of the late 20th century. As such, we document a somewhat surprising reversal in peasant mobilization. Initially high levels of rural mobilization are associated with lower levels of rural violence in subsequent decades. We attribute this to investments in bureaucratic presence that occurred in the interim.⁹

The findings in this paper build upon a growing literature on the historical determinants of bureaucratic capacity (Ansell & Lindvall, 2020; Herbst, 2014; Saylor, 2014; Soifer, 2015; Vogler, 2024).¹⁰ Much of this work focuses on Latin America and argues that investments in state capacity arose from negotiations between economic and political elites, particularly in the 19th century (Garfias, 2018; Kurtz, 2013; Mazzuca, 2021; Sánchez Talanquer, 2017; Saylor, 2014; Soifer, 2015).¹¹ We show, however, that institutional trajectories of state strength in the 20th century were also shaped by grassroots mobilization that inspired investments in bureaucratic presence. Ultimately, the establishment of these bureaucratic offices, as we demonstrate, had important implications for state capacity, shaping the ability of Latin American governments to implement policies and maintain order in the periphery.¹²

The argument and evidence further consider the enduring effects of understudied but consequential instances of localized peasant collective action. A rich literature exists on the consequences of large-scale, twentieth-century peasant movements and rebellions (Migdal, 1974; Paige, 1978; Skocpol, 1979), but there exists much more limited research on geographically constrained instances of peasant mobilization, which were common during the same historical period. As we demonstrate, these localized collective action events proved transformative in shaping bureaucratic capacity and limiting future rural uprisings.

Our paper also offers insight into the perhaps surprising decision of governments to invest in bureaucratic capacity. In many cases, repression or a one-time targeted distribution of resources may be preferable to address popular sector mobilization. These strategies require shorter time horizons for an incumbent, may offer more certain results, and cannot be harnessed by the incumbent’s opponents in the future. Yet, these strategies cannot be easily deployed in all cases. Democratically elected civilian leaders may be hesitant to employ large-scale repression for sincere or strategic reasons.¹³ Likewise, mobilized groups may not be appeased by short-term, targeted resource distribution. Peasants may view these “fire-fighting” as a non-credible response to their more structural or revolutionary demands (Acemoglu & Robinson, 2006, 26). In these cases, our argument suggests that incumbents should be more likely to invest in bureaucratic presence, an intermediate strategy between targeted goods transfers and more transformative, national-level changes to social policy, such as land reform.

Finally, this article contributes to the literature on contentious politics. Prevailing theories disagree over whether government responses to popular mobilization through concessions increase or reduce the likelihood of protest escalation.¹⁴ While some scholars argue that concessions generally discourage further violence (Dugan & Chenoweth, 2012), others argue that concessions can exacerbate social unrest (Leuschner & Hellmeier, 2023; Rasler, 1996). Yet, in considering whether concessions may mobilize or demobilize aggrieved societal groups, it is useful to consider that concessions can take various forms (Dugan & Chenoweth, 2012). We find evidence to support the claim that costlier concessions—through investments in bureaucratic presence—appear more likely to prevent the escalation of societal mobilization to violent conflict.

The paper proceeds by outlining how peasant mobilization sparks investments in bureaucratic presence. Our argument details the intuition and assumptions behind our central prediction: incumbents are most likely to invest in establishing bureaucratic offices in areas where peasants have a demonstrated ability to mobilize collectively. We then proceed to the analysis of our central case, 1960s Peru under Fernando Belaúnde. We use archival data to demonstrate that Belaúnde established bureaucratic presence based on the prevalence of peasant mobilization. We conclude with a discussion of the theoretical and practical implications of these investments in bureaucratic capacity.

Explaining government Responses to Peasant Collective Action

Societal mobilization offers a way for otherwise neglected groups to influence government decisions. The threat—or reality—of mobilization escalating to an uncontrollable level can provide a substantial inducement for incumbents to respond to the demands of these populations. If societal mobilization becomes sufficiently widespread or intense, it may evolve into a violent revolt against the national government. Alternatively, collective action by marginalized groups can lead political elites and other citizens to question the incumbent’s competence. Incumbents who are perceived as failing to address a crisis are more likely to be either deposed through a coup or replaced in an election. It is, therefore, in the incumbent’s interest to address instances of collective mobilization that could feasibly escalate to violent unrest.

In this paper, we are particularly interested in explaining national incumbents' responses to localized instances of twentieth-century peasant mobilization, which we define as geographically contained collective action by small-scale agricultural workers. These forms of mobilization can be disruptive to local-level economic elites, but the greatest threat they pose is in a yet unrealized future when they might coalesce into a larger-scale peasant movement that could displace the incumbent and other entrenched elites. As such, this categorization of localized collective action excludes the large-scale instances of peasant unrest and rebellion whose causes and consequences have been well theorized and explored in the existing literature (Moore, 1993; Paige, 1978; Skocpol, 1979). In some cases, localized peasant mobilization targeted landowners who controlled and exploited rural workers’ land and labor. Peasants employed a variety of strategies, including strikes, land invasions, lawsuits to denounce mistreatment, and even violence against landlords and their families. Other types of peasant mobilization targeted the state, including protests and movements that demanded land reform and more favorable government policies (e.g., price supports). Peasants’ demands were, thus, primarily structural, reflecting a centuries-long struggle to regain control over their land and labor.

We begin with the assumption that incumbent strategies to address localized peasant mobilization are motivated—first and foremost—by a desire to maintain power.¹⁵ Localized peasant mobilization can threaten the incumbent’s long-term hold on power in two key ways. First, if not contained, it could escalate into broader and more disruptive collective action through either the emergence of violent, large-scale rebellion or the formation of a political party or movement that challenges the incumbent at the ballot box. Second, peasant mobilization may create sufficiently high social and economic instability that—if sustained—could induce entrenched elites to overthrow the incumbent.¹⁶ As such, containing and addressing peasant collective action, even when localized, often constitutes a primary priority for incumbents.¹⁷

Incumbents can draw on a portfolio of non-exclusive strategies to address mobilization, each of which implies distinct costs and benefits. The first strategy is repression, or the violent suppression of peasant movements. A second strategy, fire-fighting, involves the targeted and short-term distribution of goods directly to mobilized groups.¹⁸ A final strategy—and the central focus of this paper—involves the establishment of bureaucratic agencies in the periphery. We define this strategy as the creation of physical government outposts designed to distribute goods and respond to popular demands.

Bureaucratic agencies are costly to establish—making this strategy a surprising one. This approach requires an investment in physical infrastructure, the maintenance of a team of appointed bureaucrats, and the creation of a system of oversight to ensure compliance with national-level directives. Furthermore, these offices can provide resources (i.e., networks and knowledge) that an incumbent’s political challengers can exploit in the future. Finally, while bureaucratic offices can generally be established relatively quickly—particularly given our scope condition of external funding assistance—they can take longer than fire-fighting to yield dividends.¹⁹

When, then, do governments make costly investments in bureaucratic presence? We argue that incumbents will invest in bureaucratic offices in areas where localized peasant mobilization has already occurred. If a large-scale peasant movement is to emerge, existing mobilized groups are likely to provide the kindling. Incumbents have imperfect information about where mobilization will occur in the future and thus rely on past mobilization to predict what might lie ahead.²⁰ Establishing bureaucratic offices provides certain benefits over fire-fighting in addressing peasant demands. The creation of a physical office represents a long-term commitment to government responsiveness to rural needs, signaling that the incumbent plans to address not only peasant groups’ present demands but also their concerns that may arise in the future. These offices are particularly well suited to processing the structural demands that animate most peasant movements.²¹ Furthermore, bureaucratic offices go beyond the distributive functions of fire-fighting, serving as centers of information gathering and—potentially—policy innovation. Absent a reliable agent in the periphery to gather information, the state may have only a noisy signal of the concessions that would lead to peasant demobilization.²² Bureaucratic presence likewise offers a benefit over repression in that it can be used in contexts where repression has been shown to backfire: cases where there is electoral competition (e.g., Davenport, 2007) or where previous efforts at repression did not quell dissent (Moore, 2000).

Thus, we can think of bureaucratic agencies as the product of an implicit negotiation between peasant groups and the state. These offices allow governments to more reliably identify the key concerns of mobilized groups in the periphery and provide a credible commitment to respond to them. Bureaucratic presence, once established, may lead peasants to refrain from participating in future rural unrest.²³ Rather than engaging in contentious mobilization, peasants can channel their demands through established government institutions.²⁴ Physical offices also create a space for knowledge accumulation and bureaucratic embeddedness (Hassan, 2020). Over time, this may allow the central government—through greater responsiveness, monitoring, and cooptation—to prevent the escalation of mobilization into unrest. As such, the returns to bureaucratic presence grow with time and especially iterated interactions between peasants and reliable central government agents tasked with addressing rural workers’ demands. This encourages their preservation by future administrations, even those not aligned with the incumbent who initially established them.

Ultimately, these dynamics suggest a perhaps unexpected reversal in rural patterns of mobilization. In the first stage, peasant mobilization leads an incumbent to invest in creating bureaucratic presence. The establishment of these offices, through either cooptation or responsiveness, channels rural workers’ demands into political institutions, reducing the likelihood they will support later contentious—or even violent—mobilization. This theoretical prediction suggests that early patterns of peasant mobilization may, in some cases, reduce the likelihood of later violent rural unrest.

The preceding argument focuses on incumbent preferences. Yet, it is likewise important to consider whether and when incumbents have the capacity to act on these preferences. In many cases, bureaucratic agencies may be too costly for incumbents to construct. This is especially true where incumbents have limited own-source revenue or face legislative opposition that is unwilling to support resource allocation to the incumbent’s desired projects. In these cases, foreign assistance, particularly that which is earmarked for rural development or resource windfalls under the exclusive discretionary control of the incumbent, can provide the financial capacity to produce bureaucratic agencies.²⁵ The existence of such external financing thus serves as a key background condition for our argument.

In this section, we have developed an argument to explain the divergence in an understudied outcome of contentious politics: enduring investments in bureaucratic presence. We have argued that incumbents were more likely to build bureaucratic presence in areas characterized by peasant mobilization (Figure 1). Over the long term, we expect these investments to have channeled peasant demands into institutions and to have reduced the likelihood of rural violence. In the next section, we present the context in which we test the hypothesized relationship between rural mobilization and bureaucratic presence.

Figure 1.

Argument: Peasant mobilization and bureaucratic agencies.

Peasant Mobilization in Mid-Twentieth Century Peru

The early 1960s constituted a critical period in Peruvian history. Extensive but localized mobilization by mostly disenfranchised peasants posed a growing and clear threat to the national government. Inspired by the Bolivian Revolution of 1952 and the Cuban Revolution of 1959, peasant demands for land reform in rural Peru exploded beginning in 1956 (Cant, 2012, 26; Thorp & Paredes, 2010, 128).²⁶ Rural land invasions accelerated, and insurgent groups emerged to channel these demands into armed conflict. This mobilization touched most parts of Peru but was especially acute in rural areas of central and southern Peru (Figure 2).

Figure 2.

Prevalence of peasant mobilization (number of collective action events) in Peru, 1956–1964. Note: Data taken from Guzmán and Vargas (1981). Collective action events include rural protests, strikes, grassroots unionization efforts, land invasions, and violence.

Most of these instances of collective action occurred on haciendas (large estates), which had rapidly expanded at the turn of the 20th century, largely at the expense of longstanding Indigenous communities. The encroachment of hacienda owners on traditionally native lands forced Indigenous peasants into exploitative debt peonage arrangements to maintain their access to land. Subsequent collective mobilization by peasants thus arose primarily on the haciendas and coalesced around structural demands for better working conditions and, ultimately, a return of community land. Hacienda owners feared such mobilization and deliberately limited their workers' exposure to outside influences (McClintock, 1981, 76). Access to schools and literacy, a key condition to vote and participate in life outside the hacienda, was especially restricted; Cotler (1970) observes that hacienda peasants in one region of Cusco demonstrated much lower rates of literacy than a nearby Indigenous community. Lacking the ability to vote, hacienda peasants voiced their displeasure with the rural status quo in contentious ways, through land invasions, strikes, and even violence.²⁷ The fear of escalation had its clearest expression in the 1962 presidential elections, when all candidates, civilian and military, committed to some form of agrarian reform to appease the movements (Cant, 2021, 27).

Mobilization spiked in 1963, reaching 164 instances of peasant collective action, more than three times the number in 1961 and almost twice as many as 1962 (Appendix Figure A2). This increase in peasant mobilization occurred during the first term in office of President Fernando Belaúnde (1963–1968), who assumed the presidency through a mostly free and fair election. As Crabtree (1992) observes, “The Belaúnde government witnessed one of the most extensive waves of peasant protest [in the 20th] century, with land seizure in the sierra [highlands] on a large scale” (13).

Belaúnde faced robust political opposition that exacerbated the threat posed by this localized mobilization. He and his center-left political party, Acción Popular (AP), had narrowly defeated the center-right APRA candidate Víctor Raúl Haya de la Torre to win the presidency: 33 percent to 32.2 percent. Congress, however, was firmly in the hands of APRA, which joined a right-wing party, UNO, to oppose the government.²⁸ The APRA-UNO coalition controlled 63 percent of the seats in Congress and 65 percent of the seats in the Senate. This legislative opposition actively blocked Belaúnde’s attempts to respond to the demands of mobilized peasants, which included a land reform program (Collier & Collier, 2002, 703; Thorp & Paredes, 2010, 128).²⁹

Government officials were particularly concerned about the opposition parties, APRA and UNO, exploiting rural mobilization and landowner dissatisfaction to discredit the AP government. In a 1963 cabinet meeting, Belaúnde claimed that: “This situation [of rural mobilization] is developing with the support of APRA and General Odria [leader of UNO] and the Communist legislator Ledesma…and undoubtedly, the oligarchy is also cooperating” (Belaúnde Terry, 1964, 78–79, author translation). Julio Trelles, President of the Council of Ministers, claimed that opposition parties were to blame: “[T]he instigators are elements linked to the extreme right and extreme left. The newspapers, institutions and entities that defend the Peruvian right…are instigating the squatters and are discrediting the government” (Belaúnde Terry, 1964, 43–44, author translation). This political opposition was creating an image that the “government was weak and that the situation was out of control completely” (Belaúnde Terry, 1964, 44–45, author translation). If left unaddressed, Trelles argued, “the consequences would be obvious: unrest, violence, a state of seige, chaos, and the long-expected revolutionary coup” (Belaúnde Terry, 1964, 44, author translation).

Thus, when Belaúnde assumed the presidency in 1963, he faced expanding peasant mobilization in the periphery, which posed a growing threat to his administration. Our argument suggests that he should have responded to this challenge with increased investments in bureaucratic agencies. In the next section, we describe the empirical strategy we use to test this hypothesis.

Research Design and Data

A quantitative analysis of our theoretical predictions requires us to address a key empirical challenge. Governments may target bureaucratic infrastructure to certain areas based on a host of factors, some of which are likely correlated with levels of peasant mobilization. This challenge arises from the fact that peasant mobilization is not plausibly randomly assigned by nature or otherwise. To address concerns of confounding, we rely on a difference-in-differences design. The model we estimate is chosen based on the plausibility of the identifying assumptions and is specified to carefully address potential sources of confounding.³⁰

The primary causal quantity of interest for this study comprises the effect of peasant mobilization on the allocation of bureaucratic offices in a given municipality, i. This is captured by the parameter, β₁, in equation (1). Each of our analyses endeavors to recover an unbiased estimate of this parameter.

\begin{matrix} {Bureaucratic presence}_{i} = β_{0} + β_{1} {Mobilization}_{i} + ϵ_{i} . \end{matrix}

(1)

To measure our outcome, we focus on indicators of bureaucratic presence at the subnational level (Table 1).³¹ We analyze variation in the presence of two agencies managed by the central state, which were established to address peasant demands: the agricultural research and extension services agency, SIPA (Servicio de Investigación y Promoción Agraria) and a rural development program, CP (Cooperación Popular, Popular Cooperation).

Table 1.

Measures of Variables Used in Analysis.

	Variables	Description	Coding
Outcomes	SIPA office	Agrarian reform Agency (1962/1963 1964–1967)	1 if agency constructed During period 0 otherwise
Outcomes	CP office	Cooperación popular Office (1967)	1 if agency is present 0 otherwise
Treatment	Peasant mobilization	Peasant mobilization in 1963 or 1964	1 if municipality Experienced mobilization 0 otherwise
Controls	Altitude	Altitude of district Capital (1k meters)	Continuous
	Enrolled students	Primary school Students (per cap. 1961)	Continuous
	Teachers	Teachers (per cap. 1961)	Continuous
	Schools	Primary schools (per cap. 1961)	Continuous
	AP mayor	Municipality has An AP mayor (1963)	1 if municipality has AP Mayor, 0 otherwise
	Road access	Municipality is Accessible by road (Feb. 1964)	1 if municipality has Access, 0 otherwise
	Total population	Municipal population (1961)	Continuous
	Rural population	Municipal population in rural areas (1961)	Continuous
	Hacienda size	Population per Hacienda (1940)	Continuous
	Haciendas	Number of haciendas (Logged, 1940)	Continuous
	State presence	State officials in municipality (1961)	1 if municipality has Official, 0 otherwise
	Region	Region in Peru	1 if municipality is in (North/South/Central) Region, 0 otherwise

SIPA was a branch of the Ministry of Agriculture established in 1960 to conduct agricultural research and provide extension services to farmers. The agency was tasked with surveying land quality and productivity, establishing experimental stations to improve seed quality and livestock breeding, and providing agronomic services and education to local farmers. To carry out those tasks, SIPA maintained a network of agricultural engineers trained domestically and abroad.³² Crucially for our argument, SIPA was under the direct control of the executive. As Lincoln (1978) writes, “Decision-making in [SIPA]… was highly centralized with zonal offices and departmental offices subject to national office approval for all actions and programs” (102).

SIPA differed from a pure fire-fighting strategy in three key ways. First, while SIPA offered targeted material assistance (e.g., seeds, fertilizers, machinery), it did much more than this (Inter-American Committee for Agricultural Development, 1968, 300–301).³³ The agency’s yearly reports show that the bulk of its activities were focused on conducting studies of production profiles, developing and distributing high-yield seeds to small-scale farmers, vaccinating cattle, and carrying out demonstrations of good cultivation techniques. SIPA extension agencies also produced reports on local conditions and activities for the national government, thus serving an important information-gathering role (Inter-American Committee for Agricultural Development, 1968, 293). Second, unlike short-term fire-fighting strategies, SIPA offices endured for years. All the offices created in 1964—except for eight—remained in place until at least 1967. Third, SIPA was not politicized to the degree that fire-fighting often is.³⁴ Most employees of the extension agencies were not partisan appointees but technocrats. As of 1964, 66 percent of all the agency’s bureaucrats were agronomists and agricultural technicians (Ministerio de Agricultura, 1964, 2–3).³⁵ The Alliance for Progress financing likely contributed to the agency’s capacity and autonomy, as international actors could deter politicians from using the institution for patronage.³⁶

Belaúnde also embraced a second program, Cooperación Popular (CP), which he established to respond to peasant demands. This program “would have rural men and women participate directly in Peru’s reformation, combining their labor with the state’s material resources to build local infrastructure” (Heilman, 2010, 123). Projects included roads, bridges, health clinics, and reservoirs. Many of the projects completed under Cooperación Popular involved education and roads. By the end of the second full year of the program, 2,600 km of roads had been constructed, and 500 schools had been built (Llosa Larrabure, 1966, 234).³⁷ The CP program had enormous reach; between 1963 and 1967, it “completed over 2,000 projects, [had] begun 4,000 others, and [had] requests to assist 11,000 more” (United States Congress Senate Committee on Foreign Relations, 1967, 31).

Yet, CP did not simply distribute benefits to peasants. It also had a more complex set of information-gathering, information-providing, and capacity-developing functions. Its stated goals were to revive traditional forms of self-help and communal labor, to train local leaders, to teach communities to innovate, to mobilize the population to participate in the economy, and to increase awareness among current and future elites of the structural problems facing rural areas (Llosa Larrabure, 1966, 223). Peasants often “mobilised [sic] themselves in…support” of CP and actively demanded CP projects (Llosa Larrabure, 1966, 230–231).

Like SIPA, Cooperación Popular program was centrally administered and housed within the Ministry of Development and Public Works; rather than channeling resources to local governments, a new bureaucratic structure was created with a “Central Office in [the capital of] Lima, four Regional Centers, ten Major Centers, and fifty Basic Centers” (Lyon, 1968, 67).³⁸ Young bureaucrats, who were assigned to staff the offices, implemented directives that came directly from Lima (Lyon, 1968, 67). This was necessary because local elected officials could not always be trusted to implement the program. As Bourque and Palmer (2015) argue, “[T]he opposition perceived Cooperación Popular as a political instrument of AP and did what they could to thwart it” (200).³⁹

Our primary outcome employs measures of SIPA and CP offices. For SIPA, we construct an indicator for whether a municipality received a SIPA office between 1964 and 1967—a range that corresponds to Belaúnde’s first and last full years in office before being deposed in a 1968 military coup. For CP, we use a binary indicator that denotes whether a municipality had a CP office in 1967. We then create an index of bureaucratic presence by summing the 1967 CP measure and the post-1963 SIPA measure. We also construct a pre-1964 version of this index.⁴⁰

Our treatment variable documents municipal-level variation in peasant mobilization in 1963 and 1964.⁴¹ These two years—which were also the first two of Belaúnde’s presidential administration—corresponded to the largest wave of peasant mobilization to that point in Peruvian history and were arguably the most significant in the 20th century (Crabtree, 1992, 13). Around 7 percent of Peruvian municipalities experienced peasant collective action during these years (Appendix Figure A2), and the vast majority of this mobilization involved land invasions on large estates (Appendix Figure A3).⁴² In the appendix, however, we demonstrate that our results are robust to defining treatment using movements that occurred between 1956 and 1964—the years for which relatively comprehensive data exists.⁴³

Our primary research strategy is a two-period difference-in-differences approach (Equation 2). We regress the pre-post (before/after 1963) difference in the values of our bureaucratic presence index on a binary measure of whether a municipality, i, experienced a peasant movement in either 1963 or 1964.⁴⁴ The coefficient on this latter term constitutes the primary quantity of interest to our study and yields our difference-in-differences estimate (β₁ in equation 2). To improve precision, we include province-level fixed effects (α_j) in our preferred specification. We also include an indicator for pre-1963 mobilization and an interaction between this variable and our treatment.

\begin{matrix} {Bureaucratic_post 1963}_{i} - {Bureaucratic_pre 1963}_{i} = β_{1} {Mobilization 1963 / 1964}_{i} + α_{j} + ϵ_{i} \end{matrix}

(2)

The central identifying assumption behind the difference-in-differences design is parallel trends: observed pre-1963 differences between treated and control municipalities (i.e., those with and without mobilization in 1963–1964) should have remained the same after 1964 had peasant mobilization not occurred. A potential threat to inference would arise if trends in the presence of bureaucratic offices across our treated and control groups had already begun to diverge before 1964. As a first test of this assumption, we can show that there was remarkable stability in the aggregate number of SIPA offices in the two years before 1964 but a rapid expansion in subsequent years. This data provides suggestive evidence that the divergence in bureaucratic presence we observe post-1964 had not already begun in the years immediately preceding our treatment (Figure 3).⁴⁵

Figure 3.

Trends in SIPA and CP office creation (1962–1967). Sources: Ministerio de Agricultura (1962, 1963, 1964, 1965, 1966); Belaúnde Terry (1967).

We also incorporate a more formal test of parallel trends in Figure 4. We plot the probability of having a SIPA office for every year between 1962 and 1967. We break our treated group (having a movement in 1963 or 1964) into two categories: places that had no movement prior to 1963 and places that had a movement prior to 1963. We compare these to places that had no movement at all during this period. We find that the trends prior to Belaúnde taking office are parallel but diverge notably after 1963. For clarity of presentation, we omit confidence intervals from this plot, but we include these in the Appendix (Figure A5).

Figure 4.

Trends in SIPA offices by treatment status. Note: The lines indicate places that had a peasant movement only in 1963/1964, that had a movement prior to 1963, and those that did not have a movement. Sources: Ministerio de Agricultura (1962, 1963, 1964, 1965, 1966); Guzmán and Vargas (1981).

A further assumption is that there are no spillovers between the treated and control groups. A possible concern, for example, is that the Peruvian state was weak and could only allocate bureaucratic offices in a finite number of municipalities. Given such a budget constraint, any allocation of a bureaucratic office to a treated municipality would preclude the state from creating one in a control municipality. Yet, the rapid expansion of bureaucratic offices denoted in Figure 3 suggests that this does not appear to have been the case; the Peruvian government seemed willing to bear the cost of sustaining and expanding bureaucratic presence as long as there was a powerful enough incentive to do so, such as that provided by peasant mobilization.⁴⁶

The design also assumes strict exogeneity. This assumption would not be met if peasants had waited to mobilize until Belaúnde was elected, expecting that collective action under his administration would be more likely to generate concessions than under the prior government. This seems highly unlikely. Elections in 1962 yielded a civilian candidate from APRA, but these results were promptly rejected by the military, which took power through a coup. Documents produced at the time suggested that there was widespread skepticism that the military would comply with its promises to hold fair elections in 1963 (United States Congress, 1963, 16). It is, therefore, unlikely that peasants would have expected or predicted that a civilian government, much less Belaúnde, would take power in 1963.

A final assumption is common support, which we analyze in Figure A4. This assumption requires sufficient overlap in the pre-treatment distribution of covariates between the treated and control groups. If there is not overlap, the two groups are not comparable, and the estimated effects may be biased. To test common support, we generate a probability of treatment assignment conditional on our covariates listed in Table 1. We then plot the distribution of these propensity scores by treatment status (having a movement in 1963 or 1964 vs. not). We find that 90 percent of our observations fall in the range of common support. In the Appendix, we show that our results remain strong and significant even when subsetting only to the observations within this range (Table A6).

In our analyses, we include a comprehensive set of control variables to enhance the precision of our estimates and to account for potential alternative explanations (Table 1). One concern is that Belaúnde was simply ideological and targeted his programs to the areas of most need: rural highland communities where historical land loss and peasant labor exploitation were most severe. These areas may have also been more likely to produce peasant mobilization. To address this issue, we include measures of municipal altitude, the presence of large estates (haciendas), and the percentage of the municipal population living in rural areas.

Partisanship could also explain both municipal levels of peasant mobilization and the receipt of bureaucratic agencies. Belaúnde may have targeted his investments in bureaucratic offices to places that supported him. Likewise, given his support among rural peasants, it may have been in pro-Belaúnde municipalities that peasant movements were most likely to emerge. Contrariwise, Belaúnde may have invested in bureaucratic offices in areas of political opposition as a way of maintaining control; peasant movements—if directed at Belaúnde—may have been more likely to arise in these areas. To address these concerns, we incorporate a control for electoral support for Belaúnde, measuring whether a member of his AP party was elected mayor in 1963.⁴⁷

A further concern is that both movements and bureaucratic offices are shaped by relative isolation. Municipalities that were better connected to urban areas through roads and communication systems may have been more exposed to unions and other organizations that often organized land invasions, strikes, and protests. Likewise, less isolated municipalities may have been more legible to the Peruvian central state, facilitating the construction of bureaucratic offices. To address this, we include a control variable for whether a municipality had road access in early 1964. While this is technically post-treatment, roads that were completed by February 1964 were started—and generally finished—well before 1963.⁴⁸

We might also expect that movements are more likely to emerge—or be documented—in more populated areas, which are also the places where bureaucratic agencies may be most prevalent. To address this concern, we include controls for a municipality’s population as documented in the 1961 census. We also address the possibility that historic bureaucratic capacity may have reduced both peasant movements and investments in future bureaucratic offices by controlling for the number of government employees in a given municipality in 1961.

Finally, we include controls for education specifically enrollment and the number of schools and teachers in 1962. This addresses several concerns. First, if education is a tool of indoctrination (Paglayan, 2022), then places with more access to schooling may be more subservient and demobilized. They may also be less likely to make demands on the state for land reform or other peasant-related concessions. Second, education is a potential source of upward mobility and human capital development. Those who had schooling may have been less likely to become peasants, reducing the prevalence of peasant mobilization as well as the demands for—and the usefulness of—the agricultural and rural development projects achieved through Belaúnde’s SIPA and CP programs.

We have—to our knowledge of the historical record—complete data on our independent and dependent variables. Yet, there is some missingness in the controls. For almost all covariates, missingness affects well under 10 percent of observations, and for these, we impute missing values using the unconditional mean of each variable (Lin et al., 2016).⁴⁹ In addition to our battery of controls described above, we also include province-level fixed effects to address time-invariant confounders. The province is the second-level tier of subnational administration in Peru—below only the department. Historically, this was the lowest consequential level of subnational administration, with provincial-level sub-prefects having extensive authority over taxation and public goods provision. Nearly all of our control variables were recorded between 1961 and 1963, reducing the risk of post-treatment bias.⁵⁰

Results

Our argument predicts that Belaúnde should have targeted his investments in bureaucratic infrastructure to areas where there was relatively more peasant mobilization. Our difference-in-differences strategy provides strong evidence in support of this hypothesis. Table 2 examines the effect of mobilization on changes in municipal-level bureaucratic presence after 1963. We find that mobilization in 1963/1964 increases the values of the index by between 0.17 and 0.24 units. This is a change that is large in magnitude (between 0.55 and 0.78 sds), significant, and robust to the inclusion of provincial-level fixed effects and our battery of controls. These results hold even when we control for movements that occurred prior to 1963—also a significant, though somewhat weaker, predictor of bureaucratic offices. The interaction term offers some evidence—though inconclusive—that the effects are strongest in places where there is new peasant mobilization in 1963 and 1964.⁵¹

Table 2.

Difference in Differences Estimation: Effect of Peasant Mobilization (1963–1964) on Changes to Bureaucratic Presence.

	Dependent variable
	Differenced index (pre/post 1963)
	(1)	(2)	(3)	(4)
1963/1964 movement	0.214** (0.053)	0.235** (0.076)	0.174** (0.050)	0.195** (0.069)
Pre-1963 movement		0.178** (0.047)		0.132** (0.045)
1963/1964 movement x Pre-1963 movement		−0.196† (0.115)		−0.139 (0.105)
Intercept	0.070** (0.007)	0.059** (0.007)
Num.Obs	1581	1581	1581	1581
R2	0.031	0.049	0.210	0.218
R2 Adj	0.030	0.047	0.124	0.131
RMSE	0.31	0.30	0.28	0.27
Fixed effects	No	No	Yes	Yes
Controls	No	No	Yes	Yes

Note. † p $<$ 0.1, * p $<$ 0.05, ** p $<$ 0.01. Estimates obtained from a regression of the dependent variable (an index that sums a dummy indicator of the municipal-level presence of a CP office in 1967 and the differenced SIPA offices pre- and post-1963) on a binary indicator of peasant mobilization in a given municipality in either 1963 or 1964. Column 2 includes all covariates in Table 1, while columns 3 and 4 contain provincial-level fixed effects and all covariates in Table 1, except the regional indicators, which are perfectly collinear with the fixed effects.

Sources: Belaúnde Terry (1967); Ministerio de Agricultura (1962, 1963, 1964, 1965, 1966); Comisión Ejecutiva Interministerial de Cooperación Popular (1964); Guzmán and Vargas (1981); Peru Dirección Nacional de Estadística y Censos (1968); Jurado Nacional de Elecciones (2017).

The findings also hold when we examine the individual components of the index (Table 3). A peasant movement in 1963 or 1964 corresponds to a 7–10 percentage point increase (0.3 − 0.4 sds) in the likelihood a municipality receives a SIPA office after 1963 (p < 0.05).⁵² Peasant mobilization is likewise associated with a 15–17 percentage point increase (0.8 − 0.95 sds) in the probability of having a CP office in 1967 (p < 0.01). Once again, earlier mobilization (i.e., pre-1963) is likewise important but does not substantially alter the effect of peasant movements in 1963 or 1964.⁵³

Table 3.

Relationship Between Peasant Mobilization (1963–1964) and Bureaucratic Presence Index Components.

	Dependent variable: Offices
	SIPA (pre/post 1963)		CP (1967)
	(1)	(2)	(3)	(4)
1963/1964 movement	0.071* (0.035)	0.096* (0.046)	0.172** (0.039)	0.153** (0.050)
Pre-1963 movement		0.085* (0.035)		0.091** (0.033)
1963/1964 movement x Pre-1963 movement		−0.119 (0.076)		−0.041 (0.079)
Num.Obs	1581	1581	1581	1581
R2	0.165	0.172	0.125	0.139
R2 Adj	0.082	0.087	0.038	0.051
RMSE	0.22	0.22	0.17	0.17
Fixed effects	Yes	Yes	Yes	Yes

Note:† p $<$ 0.1, * p $<$ 0.05, ** p $<$ 0.01. Estimates obtained from a regression of the dependent variables (the differenced SIPA office measure and the CP office measure) on a binary indicator of peasant mobilization in a given municipality in either 1963 or 1964, a binary indicator of having a movement between 1956 and 1962, and an interaction of the two. Provincial fixed effects are included. The first two columns present difference-in-differences estimation.

Sources: Belaúnde Terry (1967); Ministerio de Agricultura (1962, 1963, 1964, 1965, 1966); Comisión Ejecutiva Interministerial de Cooperación Popular (1964); Guzmán and Vargas (1981); Peru Dirección Nacional de Estadística y Censos (1968); Jurado Nacional de Elecciones (2017).

We further demonstrate the robustness of our results to an alternative coding of our treatment variable: whether a peasant movement occurred in 1964. Excluding movements that occurred in 1963 allows us to examine only those mobilizations that definitively occurred after Belaúnde took office in July 1963. Furthermore, this analysis ensures that all covariates are pre-treatment. The results only strengthen in magnitude through this redefinition of the treatment (Table A3).

Beyond our quantitative analysis, the archival record produced by the Belaúnde administration demonstrates that peasant mobilization was a primary concern for the president. Minutes from cabinet meetings held in 1963 and 1964 suggest that Belaúnde viewed mobilized peasant groups as a key political threat that—if left unaddressed—could destabilize his government:

[L]and invasions of haciendas have been taking place by members of different communities…The tense situation generated by the continuous invasions has worsened and is taking on proportions that, with time, can escape the control of the government if an immediate action plan isn’t agreed upon…It is mandatory for the government to confront this urgent situation. (Belaúnde Terry, 1964, 43–44, author translation)

A chief fear within Belaúnde’s administration was that landowners would mobilize against the government if peasant mobilization endangered their land and labor supply. Weil (1972) observes,

[T]he principal association of landowner groups ran full-page advertisements in the newspapers almost daily and sponsored radio and television messages attacking the government’s failure to curb the rural agitation and protect property rights…They also worked through the UNO-APRA coalition in Congress to press for government action to restore order (Weil, 1972, 204).

Importantly—and consistent with the assumptions of our argument—Belaúnde did not address peasant mobilization through repression (Handelman, 2014, 122). Instead, Belaúnde sought to invest in new bureaucratic agencies to facilitate his goals of preventing the spread and escalation of peasant mobilization. The Belaúnde government’s 1963 “Action Plan,” for example, advocated, “send[ing] multiple commissions to different parts of the country where [land] invasions have happened or are likely to occur due to existing tensions” (Belaúnde Terry, 1964, 45, author translation). These commissions would be comprised of civilian bureaucrats, who were “representatives of government ministries and institutions” and who would be in charge of “establishing themselves in the place” and “making possible the regular channeling of the comuneros’ [peasants’] demands” (Belaúnde Terry, 1964, 45–46, author translation). A core part of this Plan entailed establishing bureaucratic agencies, like SIPA and CP, in the areas where mobilization had occurred to respond to the peasants’ demands. His administration stipulated that this work should occur “immediately” (Belaúnde Terry, 1964, 47–50).⁵⁴ Through these agencies, Belaúnde hoped to preserve his hold on power in the face of a growing rural threat.

This section has shown that peasant mobilization was an especially salient concern during Belaúnde’s first years in office. To address this issue, the Peruvian national government invested in the establishment of bureaucratic offices in the peripheral areas where there was a recent history of peasant mobilization. In the next section, we analyze the persistence of this bureaucratic infrastructure and its longer-term effects.

The Legacies of Bureaucratic Presence

A central characteristic of investments in bureaucratic agencies is their tendency to endure. Creating government offices can be expensive—both in absolute terms and the opportunity cost of foregoing other policy priorities. Even in areas where local officials are politically aligned with the national executive, state-appointed bureaucrats provide something of an insurance policy; if local political conditions change, the incumbent retains a reliable agent in that area.⁵⁵ As such, once agencies are established, they are unlikely to be actively disbanded by future incumbents. Furthermore, the returns to state presence increase over time. While bureaucrats may initially have limited knowledge about the contexts in which they work, their personal knowledge of an area increases over time.⁵⁶

Consistent with these observations, Belaúnde’s investments in bureaucratic offices did, in fact, endure. The military government of General Juan Velasco (1968–1975), who had deposed Belaúnde in a military coup, launched a radical land reform program that expanded on the prior work of Belaúnde’s government. The speed with which Velasco implemented his land reform demonstrated his reliance on existing bureaucratic infrastructure. He “grafted land reform implementation on top of existing Agrarian Zones and their operational offices,” which had been established by SIPA (Albertus, 2021, 229). As Table A9 in the Appendix shows, there exists a strong positive relationship between the existence of a SIPA office established during Belaúnde’s government and land redistributed under the later military government. While peasant mobilization may not have led to the adoption of land reform (Albertus, 2015), it appears to have created the bureaucratic infrastructure required for its implementation.

Cooperación Popular also endured, serving as the basis for the influential National System of Support for Social Mobilization (SINAMOS) implemented under Velasco (Jaquette, 2015, 410–411). As Cant (2021) argues, SINAMOS was built using existing state infrastructure: “The organization was formed from eight existing institutions and retained large numbers of civil servants from the previous regime” (217). Because the infrastructure of CP had largely remained in place even during the military government—though under a different name—Cooperación Popular later reemerged under the second Belaúnde administration (1980–1985). Observing data on CP offices in the 1980s, we find that location of CP offices in the 1960s strongly predicts the location of CP offices two decades later (Table A8). It is little surprise that subsequent governments saw a need to maintain these response centers. Llosa Larrabure (1966) wrote that if CP offices failed to function as they were supposed to or if the budget were cut, peasants would resent government failures to meet their demands, and “the consequences would be serious” (235). Perhaps the greatest symbol of the resilience of CP emerges from the fact that APRA, the chief rival of AP, maintained the program when its candidate, Alán Garcia, assumed the presidency in 1985. Garcia not only preserved CP agencies but used them to administer his most sweeping social program: Programa de Apoyo de Ingreso Temporal (Temporary Income Support Program, PAIT).⁵⁷

Importantly, as our argument predicts, these enduring investments in bureaucratic presence may be associated with a reversal of peasant mobilization. By channeling peasant discontent and demands into bureaucratic institutions, governments can prevent the escalation of protests, invasions, and strikes into violent rebellion. Specifically, peasants might have fewer incentives to support insurgent groups, which are more likely to emerge in contexts of unsatisfied demands by the rural poor (Albertus, 2020; Albertus & Kaplan, 2013). As such, places of high initial mobilization may be characterized by a long-term decline in violent forms of mobilization.

To demonstrate this, we examine the most severe period of civil conflict in Peru’s history: the struggle between the Peruvian government and the Shining Path rebel group. The Shining Path was a Marxist rebel movement, which was concentrated in the rural Andean highlands. While not a peasant-led movement, the Shining Path looked to recruit peasant support (McClintock, 1984, 48; Tapia, 1997, 137).⁵⁸ The Peruvian government sought to win the loyalty of peasants, particularly through bureaucratic agencies. Soifer and Vieira (2019) write that the government expected that “state agencies would fund economic development to win over peasants” in the civil conflict (117).

Using data compiled by De la Calle (2017), we code areas of Shining Path control based on the incidence of electoral boycotts by insurgent forces. Places where elections are never held or are annulled—at least once between 1980 and 1995 are coded as “1.” Our analyses support the posited reversal in trends of rural mobilization. Early peasant mobilization is correlated with a long-term reduction in later unrest, as measured by Shining Path control (Table 4). Specifically, municipalities with mobilization in 1963 or 1964 experienced an 11 percentage point reduction in the likelihood of rebel control between 1980 and 1995 (

\approx 0.2

SDs). Strikingly and consistent with our argument, we find that investments in bureaucratic agencies under Belaúnde are associated with a similar reduction in later rebel control (columns 3–6). While these analyses cannot be assigned a causal interpretation, they provide suggestive evidence of our argument of a reversal in mobilization attributable to the cooptation or responsiveness facilitated by bureaucratic agencies. Consistent with the work of Albertus (2020), we find that some of this effect may operate through land reform agencies. However, the evidence for CP offices suggests an even broader effect of bureaucratic presence—beyond land reform—in preventing the escalation of peasant mobilization into violent rural unrest.

Table 4.

Relationship Between Mobilization, Bureaucratic Agencies, and Later Rural Unrest.

	Dependent variable
	Rebel control (binary, 1980–1995)
	(1)	(2)	(3)	(4)	(5)	(6)
1963/1964 movement	−0.113** (0.042)	−0.116** (0.036)
SIPA (1964–1967)			−0.181* (0.031)	−0.054* (0.026)
CP office (1967)					−0.131* (0.058)	−0.122* (0.046)
Intercept	0.335** (0.012)		0.347** (0.013)		0.331** (0.012)
Num.Obs	1547	1547	1547	1547	1547	1547
R2	0.004	0.550	0.015	0.547	0.002	0.548
R2 Adj	0.003	0.504	0.014	0.501	0.002	0.502
RMSE	0.47	0.31	0.47	0.32	0.47	0.32
Fixed effects	No	Yes	No	Yes	No	Yes

Note. † p $<$ 0.1, * p $<$ 0.05, ** p $<$ 0.01. Estimates obtained from a regression of the dependent variable (rebel control, as measured by electoral boycotts) on a binary indicator of peasant mobilization in a given municipality in either 1963 or 1964, a binary indicator of a SIPA office created anytime between 1964 and 1967, and a binary indicator of a CP office in 1967. Columns 2, 4, and 6 contain provincial-level fixed effects.

Sources: Belaúnde Terry (1967); Ministerio de Agricultura (1964, 1965, 1966); Guzmán and Vargas (1981); De la Calle (2017).

A potential challenge to our interpretation of these results would arise if peasant mobilization simply led to greater resistance to the Shining Path. The same groups that mobilized to resist the state may have also mobilized to resist the violent recruitment efforts of the Shining Path movement. If this were true, we would expect peasant mobilization to be associated with a higher instance of contested Shining Path control (areas where the insurgency tried but failed to monopolize its control). We find no evidence that this is the case. Using again the data from De la Calle (2017), we show null or negative effects of 1960s-era peasant movements on the likelihood that a municipality experienced contested control during the civil conflict (Table A10).

Conclusion

This paper has developed and tested an argument about the near- and long-term effects of twentieth-century peasant mobilization on state presence. We have demonstrated that Peruvian president Belaúnde was most likely to establish bureaucratic agencies in areas where there was a recent history of peasant mobilization. We have further shown that these investments endured beyond Belaúnde’s term in office and seem to have achieved their goal of preventing the escalation of mobilization into rural unrest. We have shown that municipalities with bureaucratic offices in the 1960s had less rural unrest in future decades.

While bureaucratic presence increased the Peruvian state’s capacity to implement policy in a defined set of rural areas, we should not assume that the Peruvian state became “strong” as a result of these changes. The path-dependent forces that generated persistent weakness in Peru, as detailed in Kurtz (2013), Schenoni (2024), and Soifer (2015) continued. Yet, the changes we document did generate new patterns of heterogeneity within Peru, as state capacity and—especially—rural development increased in some areas but not in others. The presence of bureaucratic offices, thus, did not generate a new national-level critical juncture, but our evidence suggests it did lead to meaningful divergences in development trajectories at a more local level.

These effects go well beyond violent rebellion, as bureaucratic presence may also shape prospects for democracy. With respect to the former, state bureaucratic agencies may co-opt and demobilize disenfranchised groups, like peasants in 1960s Peru, who could otherwise advocate for democracy. On the other hand, these same agencies might promote democracy. By creating institutional opportunities for demand-making within regimes with widespread disenfranchisement, functional bureaucratic agencies can increase popular expectations of the government, thus sparking demands for the creation of more such spaces (e.g., the franchise).⁵⁹

The establishment of bureaucratic offices also has key implications for redistribution. To implement large-scale, redistributive programs, such as land reform, the central government requires not only physical outposts but also reliable agents who can commit to following state directives that are likely to be opposed by entrenched, local elites. Without bureaucratic presence in the periphery, efforts at ambitious and transformative social programs are unlikely to be successful. We have shown that the offices established under Belaúnde, for example, were central not only for his administration to respond to peasant demands and maintain order but also for future administrations—which were less institutionally constrained—to implement land reform and social assistance programs.

Looking beyond Peru, we believe that our argument is most likely to hold when several factors jointly obtain: there is relatively localized peasant mobilization that threatens to escalate into coordinated rebellion; institutional constraints (e.g., electoral competition) prevent a sole reliance on widespread repression; state capacity is uneven; and substantial external funding assistance is available to the incumbent. The first condition creates the incentive to respond to rural unrest, while the second, third, and fourth dictate the strategies incumbents can use to respond to this unrest, leaving bureaucratic agencies as a costly but viable way to prevent the escalation of peasant mobilization. While these scope conditions may seem restrictive, they hold in many Latin American cases in the mid-twentieth century, including Bolivia, Ecuador, Colombia, Venezuela, and Chile.⁶⁰

The enduring effects of twentieth-century peasant mobilization have been well-studied by generations of scholars (Moore, 2000; Paige, 1978; Skocpol, 1979). Yet, the enduring effects of more localized instances of rural mobilization have generally escaped scholarly attention. In this paper, we have demonstrated that the legacies of these instances of collective action were sometimes profound. Despite lacking the right to vote, peasants mobilized to advocate for meaningful concessions and encouraged governments to embark on costly and consequential expansions of state presence. While bureaucratic capacity created opportunities for cooptation, it also generated new space for rural responsiveness within regimes that otherwise provided limited opportunities for peasants’ demands to be heard.

Supplemental Material

Supplemental Material - Expanding the State: Rural Mobilization and Bureaucratic Presence in Peru

Supplemental Material for Expanding the State: Rural Mobilization and Bureaucratic Presence in Peru by Christopher L. Carter and Madai Urteaga Quispe in Comparative Political Studies

Footnotes

Acknowledgements

We are grateful to Natália Bueno, Anna Callis, Eduardo Dargent, Alisha Holland, Tanu Kumar, Steve Levitsky, Jorge Mangonnet, Paula Muñoz, Maritza Paredes, Brian Palmer-Rubin, Ben Smith, Hillel Soifer, and Yuhua Wang for their helpful feedback. We also thank participants at the 2022 Annual Meeting of the Red para el Estudio de Economía Política de América Latina (REPAL), Universidad del Rosario, the 2023 Annual Meeting of the American Political Science Association, and the 2023 Annual Meeting of the Midwest Political Science Association.

Author Contributions

Both authors contributed equally to this project

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs

Christopher L. Carter

Madai Urteaga Quispe

Data Availability Statement

Replication materials can be found in the Harvard Dataverse (Carter & Urteaga Quispe, 2025).

Supplemental Material

Supplemental material for this article is available online.

Notes

Author Biographies

Christopher L. Carter is Assistant Professor of Politics and John L. Nau III Assistant Professor of the History and Principles of Democracy at the University of Virginia. His research examines the historical evolution of Indigenous-state relations in Latin America—from independence to the present.

Madai Urteaga Quispe holds a PhD in Political Science from Harvard University. Beginning in August 2025, she will be a postdoctoral fellow at the Kellogg Institute for International Studies at the University of Notre Dame. Her research examines the historical political economy of agricultural policies in Latin America.

References

Acemoglu

Robinson

J. A.

(2006). Economic backwardness in political perspective. American Political Science Review, 100(1), 115–131. https://doi.org/10.1017/s0003055406062046

Albertus

(2015). Autocracy and redistribution: The politics of land reform. Cambridge University Press.

Albertus

(2020). Land reform and civil conflict: Theory and evidence from Peru. American Journal of Political Science, 64(2), 256–274. https://doi.org/10.1111/ajps.12466

Albertus

(2021). Property without rights: Origins and consequences of the property rights gap. Cambridge University Press.

Albertus

Kaplan

(2013). Land reform as a counterinsurgency policy: Evidence from Colombia. Journal of Conflict Resolution, 57(2), 198–231. https://doi.org/10.1177/0022002712446130

Ansell

B. W.

Lindvall

(2020). Inward conquest: The political origins of modern public services. Cambridge University Press.

Bates

(1981). Markets and states in tropical Africa: The political basis of economic policy. University of California Press.

Belaúnde Terry

(1964). Actas del Consejo de Ministros del Gobierno Constitucional del Presidente Arquitecto Fernando Belaunde Terry: 28 julio 1963-21 julio 1964 (volume 1). Presidencia de la República del Perú.

Belaúnde Terry

(1967). El Peru construye: mensaje presentado al Congreso Nacional por el presidente constitucional de la república, arquitecto Fernando Belaunde Terry, el 28 de julio de 1967. Editorial Minerva.

10.

Bourque

S. C.

Palmer

D. S.

(2015). Transforming the rural sector: Government policy and peasant response. In Lowenthal

A. F.

(Ed.), The Peruvian experiment (pp. 179–219). Princeton University Press.

11.

Brierley

Lowande

Potter

R. A.

Toral

(2023). Bureaucratic politics: Blind spots and opportunities in political science. Annual Review of Political Science, 26(1), 271–290. https://doi.org/10.1146/annurev-polisci-061621-084933

12.

Cant

(2012). Land for those who work it’: A visual analysis of Agrarian reform posters in Velasco’s Peru. Journal of Latin American Studies, 44(1), 1–37. https://doi.org/10.1017/s0022216x11001106

13.

Cant

(2021). Land without masters: Agrarian reform and political change under Peru’s military government. University of Texas Press.

14.

Carter

C. L.

Urteaga Quispe

(2025). Replication data for: Expanding the state: Rural mobilization and bureaucratic presence in Peru. https://doi.org/10.7910/DVN/ABIUIL

15.

Centeno

M. A.

(1997). Blood and debt: War and taxation in nineteenth-century Latin America. American Journal of Sociology, 102(6), 1565–1605. https://doi.org/10.1086/231127

16.

Centeno

M. A.

Kohli

Yashar

D. J.

(2017). Unpacking States in the developing world: Capacity, performance, and politics. In Kohli

Yashar

D. J.

Centeno

M. A.

(Eds.), States in the developing world (pp. 1–32). Cambridge University Press.

17.

Collier

R. B.

Collier

(2002). Shaping the political arena: Critical junctures, the labor movement, and regime dynamics in Latin America. University of Notre Dame Press.

18.

Comisión Ejecutiva Interministerial de Cooperación Popular . (1964). Datos distritales para fines de desarrollo. Perú.

19.

Cotler

(1970). Traditional haciendas and communities in a context of political mobilization in Peru. In Stavenhagen

(Ed.), Agrarian problems and peasant movements in Latin America (pp. 533–558): Doubleday Anchor.

20.

Crabtree

(1992). Peru under García: An opportunity lost. Pitt Latin American series. University of Pittsburgh Press.

21.

Crabtree

(2016). Peru under Garcia: An opportunity lost. Springer.

22.

Culbertson

(1963). Statement of Robert Culbertson, director, US operations Mission, Lima, Peru. In United States aid operations in Latin America: Hearings before a subcommittee of the committee on government operations house of representatives, eighty-seventh congress, first session, November 27, 30, December 1, 4, 8, 9, and 11, 1961 (pp. 218–229). United States Government Printing Office.

23.

Dahlström

Lapuente

(2022). Comparative bureaucratic politics. Annual Review of Political Science, 25(1), 43–63.

24.

Dargent

(2018). The Rise and Fall of the Instituto Nacional de Planificación in Peru, 1962–1992: Exploring the Limits of State Capacity Building in Weak States. In Centeno

M. A.

Ferraro

A. E.

(Eds.), State and nation making in Latin America and Spain: The rise and fall of the developmental state (pp. 134–156): Cambridge University Press.

25.

Davenport

(2007). State repression and political order. Annual Review of Political Science, 10(1), 1–23. https://doi.org/10.1146/annurev.polisci.10.101405.143216

26.

De la Calle

(2017). Compliance vs. constraints: A theory of rebel targeting in civil war. Journal of Peace Research, 54(3), 427–441. Publisher: Sage Publications Ltd. https://doi.org/10.1177/0022343316686823

27.

Doner

R. F.

Ritchie

B. K.

Slater

(2005). Systemic vulnerability and the origins of developmental states: Northeast and Southeast Asia in comparative perspective. International Organization, 59(2), 327–361. https://doi.org/10.1017/s0020818305050113

28.

Dugan

Chenoweth

(2012). Moving beyond deterrence: The effectiveness of raising the expected utility of abstaining from terrorism in Israel. American Sociological Review, 77(4), 597–624. https://doi.org/10.1177/0003122412450573

29.

Finkel

(2015). The Phoenix effect of state repression: Jewish resistance during the holocaust. American Political Science Review, 109(2), 339–353. https://doi.org/10.1017/s000305541500009x

30.

Garfias

(2018). Elite competition and state capacity development: Theory and evidence from post-revolutionary Mexico. American Political Science Review, 112(2), 339–357. https://doi.org/10.1017/s0003055417000715

31.

Gingerich

D. W.

(2013). Political institutions and party-directed corruption in South America: Stealing for the team. Cambridge University Press.

32.

Graham

C. L.

(2010). Safety nets, politics, and the poor: Transitions to market economies: Brookings Institution Press.

33.

Grossman

Slough

(2022). Government responsiveness in developing countries. Annual Review of Political Science, 25(1), 131–153. https://doi.org/10.1146/annurev-polisci-051120-112501

34.

Guzmán

Vargas

(1981). Cronología de los movimientos campesinos, 1956-1964. Investigación, Documentación, Educación, Asesoría, Servicios.

35.

Handelman

(2014). Struggle in the andes: Peasant political mobilization in Peru. University of Texas Press.

36.

Hanson

J. K.

Sigman

(2021). Leviathan’s latent dimensions: Measuring state capacity for comparative political research. The Journal of Politics, 83(4), 1495–1510. Publisher: The University of Chicago Press. https://doi.org/10.1086/715066

37.

Hassan

(2020). Regime threats and state solutions: Bureaucratic loyalty and embeddedness in Kenya. Cambridge University Press.

38.

Heilman

(2010). Before the shining path: Politics in rural Ayacucho, 1895-1980. Stanford University Press.

39.

Herbst

(2014). States and power in Africa: Comparative lessons in authority and control. Princeton University Press.

40.

Inter-American Committee for Agricultural Development . (1968). Perú: estudio de educación, investigación, y extensión agricolas, 1967. In Union Panamericana, Secretaría General de la Organízación de los Estados Americanos.

41.

Jaquette

J. S.

(2015). Belaunde and velasco: On the limits of ideological politics. In Lowenthal

A. F.

(Ed.), The Peruvian experiment: Continuity and change under military rule (pp. 402–437). Princeton University Press.

42.

Jurado Nacional de Elecciones (2017). Infogob: Resultados de elecciones municipales 1963.

43.

Kurtz

M. J.

(2013). Latin American state building in comparative perspective: Social foundations of institutional order. Cambridge University Press.

44.

Leuschner

Hellmeier

(2023). State concessions and protest mobilization in authoritarian regimes. Comparative Political Studies, 57(1), 3–31. https://doi.org/10.1177/00104140231169022

45.

Lin

Green

D. P.

Coppock

(2016). Standard operating procedures for Don Green’s lab at Columbia. https://alexandercoppock.com/Green-Lab-SOP/Green_Lab_SOP.pdf

46.

Lincoln

J. M. K.

(1978). Land to the peasants: The Peruvian military in action-An agrarian reform policy study. Dissertation: The Ohio State University.

47.

Llosa Larrabure

(1966). Cooperación popular: A new approach to community development in Peru. International Labour Review, 94(3), 1.

48.

Lyon

P. J.

(1968). Cooperative labor and community development in Southern Peru (39). Papers - Kroeber Anthropological Society.

49.

Mann

(1984). The autonomous power of the state: Its origins, mechanisms and results. European Journal of Sociology, 25(2), 185–213. https://doi.org/10.1017/s0003975600004239

50.

Mazzuca

(2021). Latecomer state formation: Political geography and capacity failure in Latin America. Yale University Press.

51.

McAlexander

R. J.

Ricart-Huguet

(2022). State disengagement: Evidence from French West Africa. International Studies Quarterly, 66(1), 1–13. https://doi.org/10.1093/isq/sqab040

52.

McClintock

(1981). Peasant cooperatives and political change in Peru. Princeton University Press.

53.

McClintock

(1984). Why peasants rebel: The case of Peru’s Sendero luminoso. World Politics, 37(1), 48–84. https://doi.org/10.2307/2010306

54.

Migdal

J. S.

(1974). Peasants, politics and revolution: Pressures toward political and social change in the third world. Princeton University Press.

55.

Ministerio de Agricultura . (1965). Memoria Anual del SIPA y Balance General 1965. Oficina Nacional de Reforma y Promoción Agraria.

56.

Ministerio de Agricultura . (1962). Memoria del Servicio de Investigación y Promoción Agraria Año 1962. Oficina Nacional de Reforma y Promoción Agraria.

57.

Ministerio de Agricultura . (1963). Memoria del Servicio de Investigación y Promoción Agraria Año 1963. Oficina Nacional de Reforma y Promoción Agraria.

58.

Ministerio de Agricultura . (1964). Memoria Anual del SIPA y Balance General 1964. Oficina Nacional de Reforma y Promoción Agraria.

59.

Ministerio de Agricultura . (1966). Memoria Anual del SIPA y Balance General 1966. Oficina Nacional de Reforma y Promoción Agraria.

60.

Moore

(1993). Social origins of dictatorship and democracy: Lord and peasant in the making of the modern world. Beacon Press.

61.

Moore

W. H.

(2000). The repression of dissent: A substitution model of government coercion. Journal of Conflict Resolution, 44(1), 107–127. https://doi.org/10.1177/0022002700044001006

62.

Niedzwiecki

(2014). The effect of unions and organized civil Society on social policy: Pension and health reforms in Argentina and Brazil, 1988–2008. Latin American Politics and Society, 56(4), 22–48. https://doi.org/10.1111/j.1548-2456.2014.00247.x

63.

O’Donnell

G. A.

(1996). Illusions about consolidation. Journal of Democracy, 7(2), 34–51. https://doi.org/10.1353/jod.1996.0034

64.

Paglayan

A. S.

(2022). Education or indoctrination? The violent origins of public school systems in an era of state-building. American Political Science Review, 116(4), 1242–1257. https://doi.org/10.1017/s0003055422000247

65.

Paige

J. M.

(1978). Agrarian revolution. Free Press.

66.

Paulston

R. G.

(1971). Education and community development in Peru: Problems at the cultural interface. Council on Anthropology and Education Newsletter, 2(2), 1–8. https://doi.org/10.1525/aeq.1971.2.2.01x

67.

Peru Dirección Nacional de Estadística y Censos . (1968). Censos nacionales de población, vivienda y agropecuario, 1961. Gobierno de Peru.

68.

Purcell

(2019). The peace corps in South America: Volunteers and the global war on poverty in the 1960s. Springer Nature.

69.

Rasler

(1996). Concessions, repression, and political protest in the Iranian revolution. American Sociological Review, 61(1), 132–152. https://doi.org/10.2307/2096410

70.

Sánchez Talanquer

(2017). States divided: History, conflict, and state formation in Mexico and Colombia Ph.D. Thesis. Cornell University.

71.

Saylor

(2014). State building in boom times: Commodities and coalitions in Latin America and Africa. Oxford University Press.

72.

Schenoni

Luis L.

(2024). Bringing war back in: Victory, defeat and the State in nineteenth-century Latin America. Cambridge University Press.

73.

Skocpol

(1979). States and social revolutions: A comparative analysis of France, Russia, and China. Cambridge University Press.

74.

Soifer

H. D.

(2015). State building in Latin America. Cambridge University Press.

75.

Soifer

H. D.

Vieira

(2019). The internal armed conflict and state capacity: Institutional reforms and the effective exercise of authority. In Soifer

Vergara

(Eds.), Politics after violence: Legacies of the shining path conflict in Peru (pp. 109–131): University of Texas Press.

76.

Stokes

S. C.

Dunning

Nazareno

Brusco

(2013). Brokers, voters, and clientelism: The puzzle of distributive politics. Cambridge University Press.

77.

Tapia

(1997). Las fuerzas armadas y Sendero Luminoso: Dos estrategias y un final. In Serie Ideología y política (1. ed). Instituto de Estudios Peruanos.

78.

Thorp

Paredes

(2010). Ethnicity and the persistence of inequality: The case of Peru. In Conflict, inequality and ethnicity: Palgrave Macmillan.

79.

Tilly

(1990). Coercion, capital, and European states, AD 990-1990. Studies in social discontinuity. B. Blackwell, Cambridge, Mass.

80.

United States Congress . (1963). Reports and documents: 88th congress, first session (1): US Government.

81.

United States Congress Senate Committee on Foreign Relations . (1967). Latin American summit Conference: Hearings, Ninetieth Congress, First Session …march 17 and 21, 1967. U.S. Government Printing Office.

82.

United States Congress Senate Committee on Foreign Relations Subcommittee on American Republics Affairs . (1968). Survey of the alliance for progress: Hearings, Ninetieth Congress, second session. U.S. Government Printing Office.

83.

Vogler

(2024). Bureaucracies in historical political economy. In The Oxford handbook of historical political economy (pp. 373–399). Oxford University Press.

84.

Weil

T. E.

(1972). Area handbook for Peru: U.S. Government Printing Office.

85.

Weyland

Kurt

(2019). Revolution and reaction: The diffussion of authoritarianism in Latin America. Cambridge University Press.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.83 MB