eta ha a ee mee age Cette Ae aT aa106 Latin American Research Review Soto, Hemando de 1986 ‘Thompson, Edward P- (0963) 1988 La formation Valenzuela, Samuel J, Eugenio Tironiy Timothy R. Scully, comps. 2006 ‘Vez, Claudio 1984 Verret Michel, yJ.Creusen 1995 La culture ouorére. Paris: UHarmattan. Elo ender La revi ioral. Lima El Barranco. dela tase ourte anglaise. Paris: ations de Maison des Sciences ‘de Hemme lelebn prs Famili, mderaancionyBenesay en Che, Santiago: Aguilas “The New Ward ofthe Gothic Fox: Culture and Economy in English and Spanish America, Berkeley: University of Califocnia Press. 1889 neg is: Isttut National de a Recherche tral te Etude detinographie ouoviére. Pais inst ‘Kgronomigue Editions de FEcee des Hastes Brads en Sciences Sociales. EXPLAINING PATTERNS OF REDISTRIBUTION UNDER AUTOCRACY The Case of Peru's Revolution from Above Michael Albertus University of Chicago “Aboteect. Wi Denefts and we loses during redistribution under dictatorship? This article argues that expropriating powerful preexisting economic elites can serve to dem astra a dictator or just’ loaly to thir launching organication hile destroying lite rivals eut of government that could potentially threaten the dictators surcioat Expropriation also provides resources for buying the support of key nonelte groups that could otherwise organize destabilizing resistance. An analysis ofthe unserse of Aifteen thousand land exproprations under military rule in Peru from 1968 to 1980 demonctrates the plausiity uf this argument as a ease of redistributive military rile that destroyed traditional elites and empowered the militery. Land tons redistributed fo “imddle-class” rural labore nso hod the grentot expat ho rgunce rete Te, sistance if they were excluded from the reform. Ths fring directly challenges a core as sumption of social conflict theory: that nondemoacratc leaders will act as faithful agents af econo eles. A discussion of ther modernizing militaries and data on large-scale exproprations of land, natural resources, and banks across Latin America fromt 1935 to 2008 suggests that the theory generalizes beyond Per How do nondemocratic regimes that choose to expropriate decide who to farget, and why do they redistribute confiscated assote tn certain groupe while neglecting others? Case-study accounts and formal theoretical work have iden- tied several plausible hypotheses to explain patterns of redistribution under Aictatorship, including ideological orientation (Verdery 1991) and the ideological bilization of subaltern sectors against entrenched elites (Laclau 1977; Canowan 1, “father of the poor” strategies used to cultivate popular support for a re- (Levine 1998; Turits 2002), and staving off the threat of revolution (Acemagi Robinson 2001, 2006). These explanations have shed important light on how * the composition ofa regime's inner circle or the presence of external thronte ‘condition regime behavior. Nonclemocratic leaders’ responses to the dual challenges posed from within Wutside the regime, however, are often intertwined. Expropriating power- Preexisting elites can simultaneously demonstrate a dictator’ loyalty to his coalition while also providing resources to reduce the threat from below108 Latin American Research Review and buy the support of key groups that could organize resistance to the autocrat’ rule, Viewed from this perspective redistribution under autocracy arises from in- traelite splits and competition among factions, and it serves two complementary functions in consolidating a dictator's rule. ‘This article develops a theory to explain patterns of redistribution under au- tocracy, outlines the theoretical bases for alternative explanations of autocratic redistribution, and then empirically examines support for these explanation: the historically prominent case of redistributive military rule in Peru. From 1968. to 1980, Generals Juan Velasco Alvarado and Francisco Morales Bermiidez and their coalitions implemented a set of radical reforms known as the “revolution from above.” Amid this wide-ranging experiment with “state capitalism,” one of the most significant projects was a land reform program that destroyed landhold- ing elites in favor of rural laborers. Roughly fifteen thousand properties were ex- propriated and redistributed, constituting 45 percent ofall agricultural land. Ale though the reform left out key sectors of the rural poor, many peasants benefited ‘materially, and it drastically changed land tenure relations in Pera, According to prominent Peru scholar Enrique Mayer (2009, 3), “It was the first government ever to execute significant income distribution in a society of great inequalities. It come pleted the abolition ofall forms of servitude m rural estates, a momentous shift in the history of the Andes, akin to the abolition of slavery in the Americas.” The ‘urban middle class also gained from pro-worker industrial laws, exproptiations in the urban and natural resource sectors; and employment, pension, and health benefits driven by the expansion of the state What explains how Peru's military regime targeted its land redistribution? Us 1g uriginal data on the universe of land expropriation along with data on land tenure and landholdings, I demonstrate that the regime targeted the largest, most influential Iauntorrnecs for expropriation, redistributing to peasant worker be leaving out the poorest rural inhabitants—landless workers and indigenous com> munities. Furthermore, expropriation and redicteibution was not focused on the areas that had previously formed the greatest threat of revolution. ‘The findings are indicative of a pragmatic military regime that pursued it ‘own autonomous interests while undercutting its rivals and solidifying its sup port base. This poses a diract challenge to influential recent scholarship such a ‘Acemoglu and Robinson (2006) and Boix (2003) that assumes that nondemocrai Tadore will act ae faithful agonts af altos Those authare 3 elites and their political or military allies will enter a coalition to repress th ‘masses and will jointly choose policies such as taxes and transfers, thus elim ing the possibility of an autocratic regime acting contrary to elite interests. B (003, 714-219) briefly discusses redistributive left-wing dictatorship but argu that these regimes are rare, typically arise when the poor take power t revolution, and are ineffective at implementing redistributive reforms. Fot ‘moglu and Robinson (2001, 939n2), “dictatorships that are against the inter the richer segments of society... fll outside the scope of our model.” ‘This article demonstrates that nondemocratic leaders can and frequently destroy the power of economic elites and operate in power according to auton j sme that eran ‘tical arguunent advanced here also has democracy. In addition to implementing 1966), « sedis ibutive dictator can make EXPLAINING PATTERNS OF REDISTRIBUTION UNDER AUTOCRACY 109 mous interests. The Peruvian militar elites who had acted to regulate them to overthrow threatening der hg he ile ype ec tes nad SRrole pes nt inet amet aie ned hr ore slants and dee ema ses mins ate ge mn Sie sean ote Eon etyg nd ci Ce Ws ining tnd ence nue es ue cto amo teen erm eda ean a a st en Tindal ppdstey cae ee ae end se pvr tras eno pele ference aa a, cine as Bi ee ‘The divergence between preexisting powerful elites in Peru and the coalition that brought Velasco to power is not uncomm not uncommon in autocratic rule. Nondemur cratic leaders and powerful elites eas tl Norllinger 1977; Trimberger 1978). : Beyond enhancing the understanding of autocratic redistribution, the theo- Observable implications forthe study of policies that condition the likelihood of certain and democrat sabi sachs wesenig eed ee tions and expanding the middle class (eg, Ansell and Samuels 2010: Mons Powerful elites more wary of autocratic i and therefore more likely to support democracy, particularly ithe ean de re anstely influence policy (Albertus and Menaldo 2014. The potential fora atic leader to favor a support coalition distinct from preexisting elites coup it they imperfectly control a dictators ind dictator cycling. It can also lead elites to ‘mildly redistributive democracy over the prospect of tent and possible reformist autocratic rule. Democratic ca whee elltes ate powertul (eg, Colombia and Ven. 41985) are less surprising under this account than current ior or potential countercoups as ppport a more predictable, milly + Bular leader replacem Asition in unoqual otat lain 1958, Braz talure anticipates.“ao Latin American Resch Review EXPLAINING PATTERNS OF REDISTRIBUTION UNDER AUTOCRACY 141 EXPROPRIATION AND REDISTRIBUTION UNDER AUTOCRACY ‘The theory advanced here therefore sheds ight on both the targets and the beneficiaries f expropriation, When a distotor’s LO diverges om Ge PE deen tors have an incentive to target powerful preexisting elites to reduce then cise ‘capacity to threaten the autocrat’ rule. Destroying te PE also demonstrates toc 4 ‘Recent lierature on the political economy of autocratic rule suggests hat the ‘most serious threat faced by dictators or juntas emanates from within their sup- port coalition (Dueno de Mesquila et al. 2003, Svolik 2012)! To survive in office, therefore a dictator must consolidate his authority by cultivating the favor ofthis expropriation ofthe Poe ‘group. Yet how does the dictator's launching organization (LO) of individuals ‘who helped him grab power learn to trust him, especially if the dictator took power by overthrowing and thus betraying the previous leader? Expropriation is one powerful policy that dictators can avail to reduce their political inaceurity. Thia oteategy io particularly useful given common problems of incomplete information at the outset of a dictator's rule: the costs of honoring promises made during coup planning, private appeals by economic elites to de liver the dictator rents in return for respecting the status quo atthe expense of his LO, and the dictator's risk acceptance for generating benefits for the LO in ways that raise popular resistance and the likelihood of a countercoup. intraelite splits and competition among factions, and il van elon ‘ ions, and it caut elongate a dictator's tenure by (1) consolidating his coalition and (2) eliminating external threats, Expropriation When the PE and LO Overlap When the support base of the regime is drawn lgely from preexisting ei in contrast expropratng the PE is likely Eeproprioice ahi undercut the ditator's only supporters and amen serainly walkin ere Dictators with an LO drawn fom the PE—such ne Pinochet in Chg ete in Nicaragua, and the most recent military regimes in Argentine ae ta Expropriation and Coalition Building amid Splits between the Hkely to serve as agents of the PE and forgo expropriation Lacnching Organcaton and Pers es Expropriation of preesstng elites (PE the individuals privileged under the] Alleratce Explanations of Expropriation under Dictatorship revious regime, can help a dictator preserve his ability to act independently in The fire hyeliinatng rivals with tgstanding power. The 10 alka benef from this policy when distinct from the PE. By expropriating the PE, the dicta reveals that he intends to remain loyal to his LO. He not only forgoes the ls of rents and political support from the PE but also accepts the risk of being. with no support if the LO turns its back on him. The degree to which the Bena rm xpopition hens an mesg fc of he sega ee the PE. Simply expropriating nonelite groups or buying off the PE is ulti [The theoretical argument advanced here also contrasts with explanations of seleefeating a this ead the LO to fer thatthe dictator plans to maintain the Wg wise to al sats wth eatin a eg a Rac opin! LO mares rte rn fom the rendering gsr ogi asus that dao at se oe ee, Tl less accountable to the LO. This would incentivize the LO to withdraw suppor Itc nat densa psa tie! wo from the dictator, thus destabilizing his rule. sea Binson 2001, 2006, Box 2003). Redistribution under dictatorship therelone Expropriation can also serve the complementary function of providing Grcurs only when a high revolutionary threat pushes elites to accede some ne, sources to reduce pressure from below and win the support of key nonelite group Atcibton tard eens ease Ee that have the potential to organize resistance to the dictator if their interests an Seal ah ipa er a ssmton dew vere neglected, Dittors who fae potential opposition or requis cooperation fa ship between « dictator and existing elites, focusing instead on intel important groups in society to sustain their rule will be more likely to yield pol dynamics. Some point to ideology as a key determinant of pro-poor re- concessions to these groups (Gandhi and Przeworski 2006). In developing stat ributive policies under dictatorship, Verdery (1991), for example, highlights ‘where land is key component of wealth for both rich and poor, converting tribution as a central legitimating principe in the socialist economics of ra laborers tosmallholders through land redistribution can turn otherwise gg} Hemet Soviet Union and Eastern Europe. The Soviets during this period of Pence ipasegnis tntu/a bonsetvalive foere supposing the sinals gun ace ay in competition with US capitalism, substantially influenced theories and threaten property ownership (Ansell and Samuels 2010; Huntington 1968). ftevies of development in develyping Latin American countries ana beyond. OF an Laclau (1977) and Canovan (1981) associate the redistributive ‘emphasis ideawcitional populist policies, often implemented under dictatorshio wate awinat 2 Jn contrast t existing predictions that a dictator will engage in regressive re- distrButon when ine horizons are shart eg, Olson 195), theca Trent here predicts that dictators will target powerful preexisting cites etter piu the Poor when empowered by an LO drawn from outside the ranks of elites Furthermore, because doing so helps a dictator demonstrate nmmitmont to hey eee hd teretore consolidate his rule, redistribution should occur cary‘American Resear Rev! EXPLAINING PATTERNS OF REDISTRIBUTION UNDER AUTOCRACY 133 sz Latin American Research Review ification for providing material benefits to the former. Others argue that Velasco's Launching Organization lt reararertadion instrumental reasons: to court the lower teatro tthe poor pursued fe ineirumente meses ons to In October 1968, General Velasco seized power from Belainde with the help of eoker to lene than sek ones given diminishing marginal wily of income 4 military launching organization ina coup that would begin the “Dover's Frcnac locomonte in the ciel lterature the poorest individuals chou riod of military rulelastng until 1960, Vlasco relies onacorccoalition the -Farahe iynamic documented in the clientes iterature, be the chief benefactors. Although these theories both shed light on the beneticia- quake Group") with whom he had plotted the coup to construct and guide major poorest segments of society, lices. The four key colonels in this group were Leonidas Rodrigues Figuena, ‘ening ranged eterno: Fes Pr mn pap ae hg they have less specific predictions forthe tar Scurrah 1880, 43) Civilians at times were drawn closely to this nner evele, bot typically for advising on specific issues. The broader set of politica lito sence er snes LAD Wer PERL tung Velasco government inched army, ai fore and navy offices changed : rule in Peru serves to illustrat in with implementing the regime's polices. Key members included José Graleen ening cerns ty ene pe By ee eas cenilly thrwtening peering eles aver of ts launching organization Angel de ia For, jorge Barandiarén, Litis Barandiaén, Pedro Richter lave fom, Pet te distributed te ized eves vo Key cectos ofthe population, Perl taleén, Enrique Valdez Angulo, and Lats Vargas Caballero (Cleaves and'seancgh Een ed Sane notte, eg frre be pois in ah “Soe ewan ge Pac Kenia ins seagicattare Bat there was longstanding presure for land reform. The 191 nitted civilians were brought into this group, including Carlos Delzado, Augusto cae aetuonartes te severe Mequaliy Sule The layet 1 peal Zea, and Glens Pat val 1, whereas 83 percent of farmers held All of these key supporters were expected to secure the support and coop- Seneca nvehecarerion eeecnngoniyepacertt oalprvatclnd) cation of eter enciiy Seen he e saeE nd oop properties ofthese wat were archaic in many regions, the most ‘exchange for their positions in Velasco's government. Their views were diverse on Ewing a ham ms get] tnt pti tino eri Ter eon eee cetera plots for family consump: loose, shifting “tendencies”: bourgeois liberals, progressives, and “La Mision,” tame user’ land in xchange fo labor onthe hacienda, tpl uring || Plip (979) cegerzes then a valence ia an scan and harvesting ines when campesinos most needed t fend co ther Wf vatves, While MeClinixck (983 28) eephreeesee tendencies and ee es ee ee ret fectosorsandhocrer cold tot tones es iat ee wre nde and vont pluls, Furthermore, these coloneo were n¢ mben il Politically at times to achieve his preferred outcomes. Velasco's key supporters is diseretion. Peru lagged its peers on & nu Sy Supp a ene ee incrooe etn Gle depleted nara 7 gelato ge Barthquake Group) against the landed ‘social and economic dimensions in the’ i lite and in favor of agrarian reform. They were also united again: t elites in the ynal Peruvians attributed this to the lopsided lite of ag rhey ited against elites is ene Pen eo ee a a Industry, nance, and ex; ‘port sectors and favored early redistributive initiatives "The mantle of agrarian reform was first taken up in the 1930s by the polit inthese areas, Be re gree rae apart ting landowning elit logy, aie arene mee ing organization did not have an explict, cohesive ‘most offcers did nonetheless support agrarian reform on the basis of Principle, Why wore Velasco's officers predipusetl toward reform (ehullp 1¥/8)? form until Fernando Belatinde was elected president in 1963. The military Most key officers came from Peru's provinces, born to largely impoverished fami- 1m until Fernando Belatinde was elect ; owe {and because oftheir anticipated tes (Kruit 1994 46. This was true of the entre Earthquake Group. Rodrigues cee nn enn eae If Reser ot fe tie ane cep Rogue inh es ch during his Gasnpalgn, and abinnigh ani ina small town in remote Moquegua; and Hoyos was from Cojaunarca, Se eae agli Casein in or lanaiauding inieess ed joining the army as a volunteer soldier. Military education shifted while thess a Ears tacesnty aioe coal Ofer climbed the military ranks particularly in the Cones for Acero Nahe ee TaD, Dasani tote wes frvatated by erles: Land invasions ol Studies (CAEM) and the army’s Intelligence School. Both CAEM and mili. rah 1980). i es in the Andes from 1963 to 1965, and loans from forei intelligence introduced changes toward more merit based prousutivit atl CESIT seein zonomfc ont ererated publi mall toward the en aed the military tl in economic and sol his term. initiallyx14 Latin American Research Review EXTLAINING FATTERNS OF REDISTRIBUTION UNDER AUTOCRACY 145 ‘was not lost on entrenched elites. Belatindle's Prime Minister Pedro Beltrén, the to be expropriated. Compensation wi president ofthe powerful landowners’ National Agrarian Society, ordered CAEM declared by the lneesoenpesation was based on the to cut nonmutary matters trom sts curriculum (Kruijt 1994, 39). This angered military officials, who interpreted Beltrén’s actions as an encroachiment similar to the elite’ tight monitoring of its budget and to their past use of press ownership and export sector dominance to caole military rulers into complying with their esired policies Gilbert 1977) Threats to political stability in Peru in the late 1950s and 1960s brought the military in direct contact with Peru backward agrarian structure andthe landed elites that dominated it, deepening its sense that landed elites must be eliminated quad ilosesentuvent of manipulation at Ue atid of these same elites, Hugo Blancos, farmers’ union movement in the La Convencién and Lares valleys in the late 1950s resulted in land invasions that ullimately requises anu ier vention, The highly unequal distribution of holdings, archaic land tenure relations, and ap palling conditions ofthe poor convinced many officers that Peru was overdue for agrarian reform. In anticipation of land reform under Belasinde in 163, peesants again launched large-seale land invasions in the Andes in 1960-1964. Yet landed lites relentlessly opposed change, dramatically watering down Belainde’s 1964 land roform in Congroae. A rural gucrrilla movement in the Andes in 1965 agate required a military response, and Belatinde's reform failed to respond effectively ‘The Morales Coup and Subsequent Redistribution to rural demands. These events helped solidify the idea of officers currounding ‘Velasco regarding reform. Property value previous ner for tx purposes un Well eo marketed mally pad in longterm goverment bone tht by st very high nftion athe endothe reimbursement was pr next to worthless agains aay Unlike land reform efforts under previous governments, Law altered lan tenure relationships and property ;Todeepon thera) property ownership. To deepen the plaza sxpportorelon beneficiaries ndhamness Nsmuprerotaeee ie al lasco created the Confederacién Nacional Agraria (CNA), an agricultusslecer: organization hat beneficiaries Were push tin inonderioseicd he gnncn progress Yet opposition tothe CNA geew from groups that id enn sgssian reform. Some ofthese individuals joined the Conese tee Peruana (CCP), which times grew fser than the governmnntonense tee Fusthemore, Velasco implemented several polices in the renee ee that diminished rural siportand eae lea tothe eter an Expropriation and Redistribution under Velasco The military regime under Velasco forged a more interventionist, statist eco nomic policy ae itil “ctato capitalism” The military quickly coised the Inter national Petroleum Company's Talara installations upon taking power, It subse= quently expropriated foreign mining companies and privately owned Peruvi Motales eventually veered ri companies deemed to be in its national interests, including banks, utilities, ish later changes, the support coal enterprises, and major newspapers, The regime created state enterprises with: elites (Kruijt 1994). The new nopoly privileges that hobbled private businesses in the export sectors of c sugar, minerals, coca, and petroleum marketing Gaulniers 1988). Furthermore, il created manufacturing laws (eg, the Industrial Community Law) that speci worker participation in profit distributions. worker shareholding. and participa tion in company management in all industries. One of Velasco most prominent initiatives among these, the agrarian form Decree Law 17716 of 1968, was aimed squarely at preexisting elites. The stipulated that all landholdings larger than 150 hectares on the coast and ight in the face of economic turmoil. Despite these ition sil didnot overlap substantially with landed (Kru ‘minister of agriculture after Gallegos, General Litis Abul Isez uly 175July 197 ax a longtime miltry sen wie detgns estarian seform “irreversible.” Morales therefore never reversed the ref sm {icClintock 1981), but he did tape it intensity substantially in his lntn yee Lanne Autwolilated™ te revolution, leading the CNA and CCP tojoinin two ‘and increasing coalitional than 15 to 55 hectares in the Sierra (depending on the locaton) were subject overlap with rem: a ‘ e asing coalitiona 5 ining urban economic elites who had forged tacit allianes wi expropriation without exception? Those in violation of labor laws were subj Revly poworfalcobinet meuters suchas Pad and Cece hee ih | to expropriation regardless of property size, and capital assets on expropr os. These policies 200mm) amas saa 9° gl? 28 0 ‘guerrilla accounts and the secondary literature (Befar 1969; Ministerio de Guerra 1966; Masterson 1991), [also constructed two measures of land invasions to test the robustness of the results for guerrilla violence. The firsts an ordinal measure of land invasion intensity based on accounts in the literature, and the second a. dichotomous measure (Befar 1969; Handelman 1975), Both yielded results similar to those for guerrilla violence. “The final alcernative explanation ofthe Feruvian lad reforut hulls tat rede tribution was aimed at undermining support for the politcal party APRA. To| ‘whether the reginwe way tying to undercut a political foe, I measure APRA s port using its vote share in the 1963 presidential election, the last national elect Defore tle military coup" Beyond the alternative hypotheses and consistent with the main hypot advanced earlier, another major factor that might have impacted the pattern redistribution was latent pressure from below by rural workers with a capacity ‘organize if excluded from the reform. Lereate a proxy for the influence of pressure by value added agriculture per agricultural worker. The size ofthe. cultural cector relative to the number of economically active workers in agri ture captures land pressure, which should be lower when the amount of land the value of agriculture are high relative tothe size ofthe agricultural labor Because peasant pressure and organization can provide problematic to nondemocratic rule in many ways other than outright revolution (gst road blockades), this measure captures pressure differently from the gue violence varishle Yearly agrienltiral pravtctinn data are measired in. con 1979 nuevos soles, with data from the National Statistics Institute (Instituto cional de Fstadistica ¢ Informatica, or INED, The number of economically ri ‘workers in agriculture, also from INEEL is measured in 1965. When this mea cal logic of the Peruvian land reform and its material inclided with the rural poor presence. i should tap a greater presence of com haciendas, which formed the basis of peasant unions that became incre strong and active throughout the 1960s (Handelman 1975), peace 4 measur of waitin rom census ats) fo proxy forte ier areas ately alleen are HS ODE era mall econome ol and where meen elites therefore challenge or displace landowning baridres iagiaa ti 5 interests (Hunan The value ofagricultaral production m peluning ager seer inte ibftenarenson for reform and ne Baten aren nd the presence and pre oflanded ces ep he income comes rom valuable agin (ccna eet ). Agricultural production data come from INE ome ay also affect redistribution, An under- Context of a traditional land tenure system A hie savy ex sory of heavy expropriation may reduce future redistribution or alterna- ly signal administrative infrastructure necessa ications with panel correct ea ors ae eassquares (LS) ean wth ponte anarsees eres OS Pavel het an earcomienerane {Lana unaer meseregumesquatinea for reform. 9 Three acldtional measures yielded similar results APRA vote share in the 1962 pres tion in the 1966 municipal elections and the change in APRA vote share fom 198 01968. 2 wo adits varias hatha seme tan ro measurable te on exept. inna es Al nang pep contre ena124 Latin American Research Review to address serial correlation arising from temporal dependence between yearly observations." All models include year dummies to control for contemporane- ‘ous shocks and exogenous trends in land reform, of which figure 2 is indicative, Because the dependent variable of land expropriations is right skewed, I log this, variable to normalize its distribution. I first estimate regression models to test cach explanation of land reform individually, followed by a set of encompassing ‘models that jointly test these theories to determine which has the most support Model 1 of table 2 includes the rural poor along with the controls In contrast to the ideology or “father of the poor” hypotheses, the coefficient for rural poor is, negative and insignificant. Expropriation was not simply targeted at areas where poverty was higher. 1nis 1s consistent with much of the literature, noting that ‘some of the poorest segments of the population—many in indigenous communis ties, those working minfuncios, and most of the landiless—did not benefit from the reform (Mayer 2009; McClintock 1983), nor did the military want to cultivate active popular support among the very poor (Pease Garcia 1977), Furtheriwore, there is little evidence that Velasco had a radical political perspective prior to becoming presideut that stjght have motivated him to redisttibute to the poor for ideological reasons. Typical of Peru's military, he had an explicitly anticommuni strat (MeCliuWch 1961, 32-54). The divesse aud alftnng tcudenies i top sl tary circles were fr from constituting ideologies (Pease Garcia 1977). Could the model I results be the artifact of a relative lack of available land cexpropriate in highland areas where the poorest communities were concentra as the prominent economist José Maria Caballero hao suggested (e, Cabal and Alvarez 1980)? Several pieces of evidence suggest this is not the case. Fi agricultural value per worker is strongly negatively associated with expropri tion. Net of accounting for the presence of the rural poor, departments where Heultural value por worlr was higher —a proxy for lowar land proesure by workers with a capacity to organize—witnessed fewer seized properties. Sec the results hold including regional fixed effects for the Const, Andes, and Soi regions, indicating that land was expropriated in more unequal depart with rural populations that had a greater capacity to organize urn within the geographical region. Finally, more land was distributed in the highlands than the coast, and the “intensity” of highlands reform relative to productive land: lower, indicating that the regime could have gone even farther in the highla ishanization is positively linked with expropriation in model 1 and i per capita is negative but insignificant. Prior expropriation is positive, as is. cultural production. The latter indicates that regions with more productive! led agribusiness—such as the profitable sugar agro-industry—were targeted! the government. ‘Model 2 tests the revolution hypothesis b rilla violence. Its coefficient is positive and borderline significant but loses si cance in models 6-8, suggesting thatthe military governments seemingly rroducing a variable for 11, Because land expropriation and ineguaity (or lt presence) may be susceptible to Inatored common shocks rend, aso eatimated models using Driscoll Kraay standard Table 2 Determinants of lana expropriation Peru, 1968-1960 Model 2 el Motels Model Model5 Modi Medel? Noaald Model 1 Agricultial value per worker Urbanization Agricultural production ‘come per capita ‘riot exp-opri sural poce oar (01) uerzilla violence asa PRA vot 1963, (0753) 6a9ge (1265) é S = z egresslon; par corrected Propiatons. All moles are OLS, isthe og nunberof x ime dummies te not sown ‘The dependent variable led) x z126 Latin American Research Review EXPLAINING PATTERNS OF REDISTRIBUTION UNDER AUTOCRACY 127 redistribute property justo rdees peasant untest or in aras where land was [funding more than quit reel Pope ee jcCinock (98% 289 agues, hepa | democratization. than quintupled (Krujt 1994) This legacy extended beyond Sere reetin a not consistent with natrowly countering "communist Could the model 45 fn ver cere octal peace” While junior officers ighting guerils inthe 19 | by Velasco? ‘Although de eon itined by an “anioligarchy” ideology held or fostering Socal Ey rnesng te appalling coniions of the por eg, | mel 4-8 sasucal esl, wo lets of evidence cant some dosh on a Feeney reece counternsurency campaign temiitry didnot | le top military cies soght to imi See ri 196 fer one wih en of ol pensar rte ana gern [consent clogs bt rater hing tod SS dpitation, since they were affected last by the agrarian reform. the results forthe landholding Gini tng fensdencies (Pease Garcia 1977). Second, ation nce they wer et iniheuy regimes and reformitinately com | fatal thesample tothe post 075 period whes onic ermal ict onal rial in fa a Calombias pacar ling, spite th fact hat ew sche would cab ne ee aa tributed to Shining Paths Wane fan ls) subsequent Kuclle ity ME Wlenes Ties nx imply Rover that eas about eees nee a cer vnantrated in regions like Ayacucho and Apurfmac with less comprehen- | As lscussed earlier, there were ah ieee ee mmpatane ee oa een poorer, indigenous peasant cominunites that were | outof principle. west of reasons many officers supported reform rei cut of te reform (Hunefeldt 1997). In contrast othe substantial number of Columns 6-8 of table 2 present several lout he fom Gms ang aon nargaizedcommuntia |] aula ray ans tiene of mates sana nce ees Regie by and refers ery few members fhe costal cooperative mar] reoflanded lies remains rey etm epitome elected by land rel Peasant faroreoganizedcommaniiee remains dighiScot, ands magitade sabe ndieging tatenores ede ee in Puno ‘and the rondas in Cajamarca) also tended to resist rather than join redistribution were greater where there ws se dnatcating that expropriation eral ining Becictanco. Cucrrilla violeme love aera ae potential for organized rut ode! 3 inchudes a measure of APRAS 1963 vote share. The coefficients neg Bs Soc inte support Goth equity na ae hypotheses sin el nr meret AA i er fms nl eens ce pene ean pote a tveandinsgn a i While any intallangeacaleexpropiatins more heavy rgeed whee there were more lange Lndigtice corners ce cana northern cowstal APRA strongholds el aldo nor) further exp ring the main hypottani, ne Nowe ge landholders strongly oon deren the central coast and southern highlands, where APRA hed lit Sn argo soe Thare ate also case ike the adjacent departments of Cajamarca Isis iarytiy en stead erste tnoooeer aes : otandard deviation below its mean to one stan. Piura: Cajamarca had more than twice the APRA support but a landholding dard deviation above whil ir ae cane ee ne hans ure, ent onpertnend lve expropriation, ieee rignet Decelicpreed are eyremlons ‘Model 4 introduces a measure for landholding inequality to test the main’ lxpropration over the twelve-year reform mould ge abou ete ce ree as ged tates whore equally es hight || santsiated 15001 heroes eFand Thesatne cer clac rece End the landed oligarchy strongest to reduce their influence over the milit Simla. An increase in lafundics from is minima ne cate eee 2 ae a clynigntcant Might inequality bea pos) aes mean reoultsin an sited 7s pecantincresesntoene eee Inequaly eyne and iy Ser ance edge mle sal peyns or 0 ver hewole en rete ee tary to acto an agent of elite to undercut this threat by redistebuting from 8 form period for the average department. sar dis aan or weaker lites to poorer pessans? Model 5 eects this poss id eee ey menured asthe al muni gf SS Tels of Expropriation a ee nie is postvely and signal asocated with gel Bs test the robustneos ofthe rents t | cual expropriation. Redistribution was aimed squarely at the most privileg that follow the able 2 specifintian bat urea nega brea oe | landed lite, Indeed, the miltarys policies surprised and angered the elit oc Wit the dependent variable as the numb af porertce eone meat Velasco himeclf acknowledged i Ben department yoae 6 netic Ban rum of properties expropriated ina “Te more effectively target elites while reducing their collective resistance CP ate cee cna tive binomial estimator models cazeo of land ex regione adopted = policy of stealth Tt created landholding ceilings tat bes ard esos, specifications include year dummies and robust radualy stricter overtime, winnowing elites by seting those under the Gonsisten withthe table grnatcinetthove ahrwe ther. then later redefining the rues (McClintock 1 mary was ideconcaly or ee 2 ite suppor fo hypotheses Thivaltimately destroyed the landed elite and their long-standing abil call or strategically motivated to help the poor ‘A long period of military rule ensued that elites I oxic ot eto oe ei mnlomized, doubled in 528 in ee ea mnie tenn <8, Rand ‘manipulate the militarMode. Model5 Model __Model7 0000, hse Constans ae Medel 4 Model 3 Model 2 Model 1 1963 le 3. Robesiness ests ofdterrinants of land expropriation in Pew, 1969-1980 fundies Jessions roburtstanda-derron are inparent he mumber of expropiations. All mal aze negative binomial eas fnre dummies arenotshown. Fieulturdl value per worker some percapita ios expropriation serilla violence and inequality Gini) ‘ear dummies Ibservations departments ral pot PRA vot EXTLAINING PATTERNS OF REDISTRIBUTION UNDER AUTOCRACY 129 above all else, And although the literature provides some support for the idea that counterinsurgency radicalized officers in the 196Ne, there ie weak: tatictical cevidence that they focused the reform in areas of guerrilla violence that posed 4 greater revolutionary threat. The only time the APRA variable gains signifi cance, in model 3, ithasa negative sign. In contrast elite presence is positively and strongly associated with land expropriation in models 4-8, whether measured asa landholding Gini coefficient or by the presence of latifundios. As in table 2, greater potential for pressure by organized rural peasants as proxied by agricul. tural value per worker is linked to greater expropriation, and the magnitude ofits coefficient is consistently higher in models that include the rural poor measiiro Prior expropriation is now positive and significant. The substantive effect of in- equality in model 8 is significant and similar to that in table 2 [REDISTRIBUTIVE DICTATORSHIP: BEYOND THE PERUVIAN CASE Per is far from unique in its history of redistributive dictatorship. Consider table 4, which lists all nondemocratic Latin American leaders that implemented large-scale seizures of land, firms operating in the natural resource sector, and "commercial banks trom 1935 to 2008, Twelve of eighteen Latin American coun- ties experienced at least one episode of large-scale expropriation under autoc- racy during this period. Furthermore, the median tentire of the table 4 leaders was 5 years, compared to 1.5 years for all autocratic leaders in Latin America exganization of incoming leaders (Albertus, forthcoming; Albertus and Menalao 2012) That this divergence is not uncommon in autocracy is supported by Hun- ington’ (1960, 209) class treatment of praetorianism, which holds that militaries are frequently key forces for progressive change in the shift from oligarchy or tra- ditional monarchy to middle-class empowerment: “In these early stages of polit- al modernization, the military officers play a highly modernizing and progres- sive role. They challenge the oligarchy, aul they promote social and economic teform.” Middle-class military groups pushed ruling generals and juntas they empowered to implement radical programs of suval reform atthe expense ofthe oligarchy in Chile and Brazil in the 192s, and in Bolivia, Venezuela, El Salvador, Panama, the Dominican Republic, and Dewouh suing aud after World Wat U The phenomenon of politically autonomous militaties that introduce large- Scale changes attacking preexisting clites and use seized assets to build political port among previously excluded groups is not limited to Latin America (Al- tus frtheaming; Finer 1988), Modernizing, redistributive miliary takevvers urred in Syria in 1949, Egypt in 1952, Iraq in 1958, Pakistan and Burma in 1956, land in 1982, andl Turkey in the 1920s (Huntington 1968, 203-221). A similar moccurred under a host of populist dictators in West Africa following inde- nce (Bienen 1985) and sindar lilo Misiam in Ethiopia. In these aru inaniy «ases, politically autonomous militaries were a prominent feature rather ‘an anomaly ofthe political landscape Bisthermaen ar Ancomntntnd130. Latin American Kesearch Review Table £ Cases of large-seule expropriation under autocracy in Latin America, 1935-2008 Yeartook Typeoflarge-scale _ Lengthaf Country Leader power expropriation _ tenure in Bolivia ‘Toro 1936) R 1 Paz Estenssoro 1952 R 4 Siles Zuaz0_ 1956 b 4 Paz Estenssoro 1960 iD 4 Ovando Candia 1969 LR 1 Torres 1970 CR 0 Banzer wri L 7 Garefa Meza Tejada 1980 r 1 Torrelio Villa 1981 L 0 Brazil Vargas 1930 B 5 Medici 1969 L 5 Geisel 1974 it 5 chite Pinochet! wr L a Cuba Castro 1959 LEB 2 Dominican Republic Balaguer 1961 io 0 Ecuador Velasco Ibarra 1968, R 4 Rodriguez Lara wR, R 4 Yoveda #urbano 1% K 3 El Salvador Duarte 1980 LB 2 Guatemala Ubico 1821 L B Castillo Armas! 1954 io 3 Mexico Cardenas 1934 LR 6 Avila Camacho 1940 L 5 Lépez Mateos 1958, L 6 Diaz Ordaz 1964 LR 6 Echeverria 1970 iD 6 Lépes Portillo 1976 3 6 Nicaragua Ortega 1973 LB W Panama Torrjos 1968 i B Pera Velasco 1968; ERB 7 Morales Bermidez 1975 L Note: Talo nclades all cases of majo expropriation under dictatorship foo 1835 to 208 with regime coded by Chelbub, Gandhi and Veeeland (09) Types of large-scale expropriation areas follows: t= expropriation exceeding 3 percent fcltivable land in a given year. R= natural resource expropriation form ofl mineral or gas firms. B= expropriation of foreign o¢ domestic ims in the banking sector segime Europe after World War Il, China, the Kuomintang in Taiwan, Cuba, and under the Partido Revohicionario Institucional, large-scale redistribution occur under single-party rule when the launching organization ofthe party verges from preexisting elites, even when the military is subordinate to the sgime. These autocratic regimes all built new political coalitions with their {tnbutive policies. ine theory advanced here to explain patterns of redistn under dictatorship can therefore apply to a range of eases beyond Peru, APLAINING PATTERNS OF REDISTRIBUTION UNDER AUTOCRACY 151 ‘coNcLusiON ‘Ths article advances a th leary autocracy. I argue that demonstrate a dictators loyalty cite rivals out of government th, to explain patterns of redistribution under ating powerful preexisting elites can serve to {o his launching organization while destroying SORE eanpunclctn clones apace ta natch oe peration and representatives of democratic nsttatine a st i poverof Rev oligarchy wat sigan mista sting ee 2 Nee 196 we ely * at ther expense. The tiban so bei rowers Pay La Rese ee 2010, Although democracy in Por boi down under Fujin, MClinoc (993350 argued are Secs elven fom Peru previous democrat: beaktowne ne a col of socal enon, o pola terse tat ld a ru te democrat proces” aly ht he mst vente eect 1 Peruvian history happened under miltay ale ae Pee {Say elie to support tation oa or predicate niin cera, ocracy over the prospest nf farther cfontit autora nae eee ally 198 Furthermore thas helped deter elites from moon ee country witha long history ofelie-deten poled rea oglu, Daron, and James Robinson “A Theory of Political Transitions” American Economie Revicw 9 Economic Origins of Dictatorship a See Bins of Dictatorship and Democracy. Cambridge: Cambridge University eae oming Anocracy and Redistribution Pridge Unteeedy a Relistibution: The Polite of Land Reform, New York: Cam— 432 Latin American Research Review cuteness soa a ee ee eater pocaad mea “Ansel, Ben, and David Samuels a0 siaatality and Democratization.” Compartie Poli Studies 4 (12): 1543-1578 Bear He acs. nates Ceri Experience New York Monthly Review Press. Bere, NEMO Inequality, and Ilsion in the Choice for Far Bections” Comartie CS ae Soe HE ey Reine in WA Amel oe on Set 0 oe wet Rearinti Coie Cant ve Sn nee ues coe re eee arc crs ase option, New York: Harcourt Brace Jovanovich “nye nr aa gi om oe” ee Militry Coneroment in Boru Uhaca, NY: Cornell Un wane conan Ca a pe iy To iineliing Salecrafz Democracy and Nel if on “Political Crisis and Military Populism in Peru. 1 Pes Tnattate de Bot aia, 1909-1979. U ine, Samuelsen on Horseback: Te Roe ofthe Miltary in Poles. 2nd ea, Boulder, CO: ree i Jenifer an Adam Paes a. Gand ee a aon and Rebellion under Dicatorahip” noni and sores ee he ” lissertation, Geet Pen Gusaehy and the Old Regime in Peru” PRD disertat Ones andelman, Howard a Handelman Howard sine Psant Pal Mobizaton ner Austin: Univer Testes auf Chrstne 1 deere tere Ha Se pagcape aid Cheging Pain oer, om isin Te uma Labyre aS mai 188194" The Fe roduc nd yy wel Cameron an lp Maser 10710 U ‘She Park: Ponnsvlvania State University EXPLAINING PATTERNS OF REDISTRIBUTION UNDER AUTOCRACY 133 Huntington, Samuel 1968, Poti Order in Changing Sct, New Have, CT Yale University Press i. Die 4 "Revolution by Dere: Peru, 1968-1975, Amsterdam: Thela Publishers. Laclau, Ernesto 1977 Plc nl euagy in Marat Theory: Capita, bases, Populism. London: Lapp, Nancy D. ” vee . “een 2008 Landing Vote: Reesntatin and Land Retr in Lain Arca. New York: Palgrave Levine, Robert M. Lh ler Bor Vrs and i Er Cambridge, Cate Unveraty Pres. 197 “Authortarianism, Corporatism and Mobilization in Per.” In The New Corporat: ‘sme Socal and Potial Structure nthe Prin World, edited by Frederick Pike and ‘Thomas Stritch, 52-84, Notre Dame IN: University of Notre Dame Pres. ‘Masterson, Daniel M. 1991 Miltarism and Polit in Latin Americ: Per fom Séncez Cero fo Sendeo Lominso [New York: Greenwood Pres ie Mayer Encique 2009 Ugly Stores ofthe Peruoian Agrarian Reform. Duthaim, NC: Duke University Press. ‘McClintock, Cynthia a ™ ‘yi 1881 ls Conprotoes and Pita Change Pe, Princeton, NJ: Prieto University Ks sini etn Gan mete het Ea peel ad Ss en nt Rn Ca ter Peete ve HRC SE Dena Dee oo Ce a a ea Ieee ac oan etn et snes Moore, Barrington, Je ” TS El tint ema rn Pe ‘Modern Worl. Boston: Beacon, maa ee nee ri iy Cop Gimeno Ci FP Hal Tay Donny al Dome A Re Sen Rk (367-576. “ " ae in 0h 1 lion eM ney end ep Pt ‘se oo Bune Cece noua aco El ppt cere 95-195 ina ot wa meee WT sco fic rin Rk 8681976 tener en BD Reng ey a 2% tn eure lier, Alfred H. 1985 Pubic Enerpries in Po: Publi Ser Growth and Reform, Boulder CO: Westview Linda 185 Between Reform and Revolution: Politica! Struggles in the Peruvian Andes, 1969-1991 nity, CA Stanford University Pres. ‘The State and Society: Peru in Comparative Perspective. Princeton, Nj: Princeton Uni- The Sta nd Sc Pr in Comp spec fon, Nj Princeton Unt434 Latin American Research Review valily Milan WE "013 The Politics of Authoritarian Rule. Cambridge: Cambridge University Press. ‘Thiesenhusen, William C, ed. "989 Searching for Aguiar fori Latin Americ, Boston: Unwin Hyman. ‘Triemberger Ellen Kay 1978 Revolution from Above: Military Bureaucrats and Developmen in Japan, Turkey, Egypt, and Peru. New Brunswick, Nf Transaction Books “Turts, Richard Lee "2003 Foundations of Despotom: Peasants, the Trujillo Regime. an Modernity ix Dominican History Stanford, CA: Stanford University Press. Verdery, Katherine BOL National IAology wnder Soils: deity ava Clan Puts in Caansenr’s Romi Berkeley: University of California Press. Zibiat,Maniel "2009 "Shaping Democratic Practice and the Causes of Electoral Fraud." American Political ‘Science Review 108 (1-21. tos juridico la sociedad liza este pi laborales y cexistencias Las migrac