We employ three metrics to gauge enforcement. First, we mea-sure deforestation in yellow and red areas, where the NFPR forbidsit. Second, we measure overall deforestation rates since the enact-ment of the NFPR. This metric allows us to capture non-enforcement of the NFPR in provinces that misclassified forest-lands into lower conservation categories as well as to measurethe deforestation that took place—and sometimes accelerated—be-tween the enactment of the NFPR and the OTBNs, when clearingswere forbidden. Sanctions constitute another potential metric forenforcement.

Given the lack of cross-provincial and cross-temporal data on sanctions, we believe deforestation rates in for-bidden areas are a good proxy of enforcement as sanctions aremeant to discourage clearings in those areas. Finally, we assesswhether provinces re-categorized individual farms.

The NFPR does not specify what sustainable management entails,and provinces show important variation in both the activities thatare allowed and in how these activities are regulated. OTBNs allowfor forest grazing and controlled timber extraction, both of whichare vulnerable to ‘‘concealed clearings.”

The Argentine Chaco is one of the world’s deforestation hot-spots. In order to assess the implementation of the NFPR, we focuson the four core provinces that comprise the majority—63percent—of the forest and that have at least 50 percent of theirprovincial area covered by forest. Since the enactment of the NFPR,these provinces have concentrated 75 percent of all deforested areain Argentina (AGN: 2017, 114)

In order to show how the mechanisms identified in our frame-work shape governors’ choices, we also present two in-depth casestudies of the NFPR’s implementation, focusing on Chaco and San-tiago del Estero

Wefocus on the core Chaco provinces: Chaco, Formosa, Salta, and San-tiago del Estero, which exhibit significant variation in deforestationlevels but are similar in terms of economic development, statecapacity, partisan alignment, and electoral competition

Governors in these cases are more likely to invest incapacity-building to enforce the law, as the conservationist coali-tion is more or less actively involved in denouncing illegal forestclearings and producers are not sufficiently powerful to undermineenforcement.

Conservationist Coalition

Conservationist coalitions are organized societal and economicinterests opposing the expansion of the agricultural frontier overforest areas. 16 These actors may advance environmental principlesor may act on self-interest. In the core Chaco provinces, conserva-tionists are primarily concerned about the negative effects of soyexpansion on their own economic activity and livelihoods.

We assume largelandowners have greater capacity to influence policymaking inprovinces where they control larger shares of provincial productiveland, which increases both their structural and instrumentalpower.

Large producers are here understood as landowners and inves-tors seeking to exploit production and real estate opportunitieswith parcels of at least 2500 ha. 10 Motivated by spiking prices dur-ing the commodity boom, large producers drove soybean cultivationinto previously unexploited areas. The expansion of agriculturefavored real estate speculation, incentivizing forest clearings as landprices climbed.11 Producers were lured by the relatively lower landprices in the Chaco provinces, the availability of large farms—whichallowed them to maximize profit margins—and the absence of actualrestrictions on forest clearings prior to the sanctioning of the NFPR

The NFPR requires provinces to enact implementation regula-tions, or a Territorial Classification of Native Forests (OTBN), com-prising land use regulations and a zoning map.

In linewith studies of decentralized forest management, we assume thatsubnational executives have no inherent preferences regarding for-est governance (Andersson et al., 2006) but rather prioritize subna-tional political and economic interests over environmentalprotection (Koontz, 2002).

We integrate both producers’ and conservationists’ interests inour theoretical framework with the aim of accounting for subna-tional variation in forest protection in the Chaco and contributingto the broader literature on enforcement in weakly institutional-ized contexts

we pay special attentionto two factors that have received less attention in the recent liter-atures on enforcement of forest protection regulations in theArgentine Chaco and on weak institutions more generally: eco-nomic interests and land structure.

Scholars further note that the more leeway subnational govern-ments have to define the content of laws and rights, the higher thechances are that local politics will shape these institutions produc-ing subnational inequality in enforcement

This article is organized as follows. In the next section we dis-cuss the relevant literature on forest protection and subnationalpolicy implementation in weak institutional contexts. Section 3presents our explanatory framework for the implementation offorest protection legislation as conflict avoidance and the theoret-ical expectations that we derive from this framework. Next, wedescribe our research design, methods, and data sources. Section 5then presents and measures our dependent variable, subnationalvariation in the implementation of the NFPR in the Chaco pro-vinces. The final sections present our empirical evidence. Weemploy process tracing based on multiple data sources, includingfieldwork research in the four Chaco provinces, and quantitativeanalysis of an original department-level dataset to assess our argu-ment and alternative explanations. Two in-depth case studies ofNFPR implementation in Chaco and Santiago show how the mech-anisms identified in our explanatory framework account for gover-nors’ implementation choices.

n doing so,we dialogue with studies that view implementation as a domain inwhich policy content is also shaped (e.g., Krott et al., 2014),6 andwe add to the literature on enforcement by showing other sourcesof institutional weakness beyond politicians’ decisions (not) to sanc-tion non-compliance (see Holland, 2016)

We findthat when provincial conservationist groups are strong—pressingthe state and engaging in policymaking—governors design andenforce strict regulations; this is the case with the province ofChaco. However, if both conservationist groups and large produc-ers are powerful, governors design strict regulations to satisfy con-servationists, creating a veneer of commitment withenvironmental goals, but enforcement of these regulations is lowas governors attempt to respond to the demands of large produc-ers.

While large producers pressured for permissive reg-ulations and nonenforcement of the NFPR, conservationistsadvocated strict implementation of the NFPR in order to containdeforestation and protect their economic activity and livelihoodsfrom agricultural expansion

In explaining cross-provincial variation in forest protection, weargue that the combined pressure from large producers drivingcultivation into forest areas and from local groups challengingthe expansion of the agricultural frontier—here called conserva-tionist coalitions—shapes governors’ implementation choices.Governors engage in a strategy of conflict avoidance, granting con-cessions to competing interests in the design and enforcement ofthe law in order to prevent mounting discontent from destabilizingtheir governments.

The coreChaco provinces—where the lion’s share of the forest is locatedand which account for most of the deforestation in Argentina—de-signed significantly different regulations and have displayedremarkable variation in deforestation rates since the NFPR’sapproval.

In 2007, duringthe heyday of the commodity boom, the Argentine Congress sanc-tioned a national forest protection regime (NFPR) in response toenvironmental activists seeking to curtail deforestation.

While environmental laws are the initiative of national author-ities, their implementation generally falls under the responsibilityof subnational governments. 3 Comparative scholarship on LatinAmerica suggests that implementation of national laws correlateswith subnational state capacity (Amengual, 2016), the politicalalignment between national and subnational authorities(Niedzwiecki, 2018), and the electoral incentives of subnationalpoliticians (Holland, 2016; Smulovitz, 2015).

Given the prevailing view that institutions areweakly enforced in Latin America (Levitsky & Murillo, 2009, 2013),a key question concerns whether and under what conditions forestprotection regulations are implemented at times of extraordinarilyhigh commodity prices

ranching

The commodity boom of the 2000s intensified environmentaldegradation and conflict over natural resources across the develop-ing world

we argue that subnational variation in the implementation of forestprotection legislation is driven by governors’ attempts to avoid conflict produced by agricultural expan-sion. Through process tracing, we show how governors’ implementation decisions—regarding both thedesign and enforcement of provincial regulations—sought to mitigate pressures from large producersopposed to clearing restrictions and from various groups contesting agricultural expansion.

what factors drive subnational authorities to implement for-est protection regulations in active agricultural frontiers?

Beginning in the late 1960s, Rumbaugh notably advanced the study of com-parative primate learning and intelligence by developing the transfer indexand applying it to statistically significant small samples of apes, monkeys,and lemurs, initially at the San Diego Zoo and then at the Yerkes NationalPrimate Research Center, which had complemented their large colony ofchimpanzees with gorillas, orangutans, and gibbons (Figure 10.2C).146 Be-fore Rumbaugh’s ambitious project, only a smattering of psychological testshad been conducted on great apes other than chimpanzees, and comparativepsychologists had neglected siamang and most other hylobatid species

Chimpanzees are moderately sexually dimorphic in cranial capacity.

Much of the increase in human brain size is achieved postnatally. A new-born monkey has 60 percent and a newborn chimpanzee has 46 percent ofits adult brain mass. But a human neonate has only 25 percent of its adultbrain mass.

To a large extent, brain size is determined by overall body size in primates.7Apes follow the general primate pattern, but humans have brains that are aboutthree times larger than one would expect for primates of our size

Because humans view apes as mentally limited, some currentcaptive environments may appear idyllic while offering only anillusion of appropriate care derived from a simplistic view ofwhat apes are rather than what they might be. This perceptionof apes determines their handling, which determines theirmental development, which perpetuates the prevailing percep-tion. Only breaking this cycle will allow the current perceptionof apes to change

Nonetheless, some African apes, both wild and captive, provide tantalizinghints that altruism, sympathy (feeling for another being and wishing toincrease its welfare), empathy (Chapter 10), consolation, reconciliation,and cooperation might occur in beings with brains one-third the mean sizeof ours.9

There are clear adaptive advantages to individuals in societies that havecommon values, share resources gathered or produced according to individ-ual skills, protect one another, and can rely on others to assist in child careand tuition instead of each individual having to provide entirely for herself,himself, and dependent progeny. At some point, communication system(s)became vital for coordinating subsistence activities and resolving disputeswithin and between group

To engage in joint attention,• To share a common code,• To recognize the objects and phenomena to which a signaler refers, or tocalibrate social activity based on referent-focal joint attention,• To surmise one another’s intentions,• To surmise a speaker’s thoughts and level of knowledge and beliefs aboutthe topic,• To assess the validity and importance of messages exchanged, and• To decide whether and how to respond to them.

My bias at this point is that emergent group-living hominids had or rela-tively quickly developed a fairly rich communication system comprising amixture of vocal, facial, and bodily gestures that facilitated their social, mat-ing, and subsistence systems. Eventually, facial gestures, body language, andmusical features were subordinated to speech.69 Nonetheless, they continuedto have information content that astute recipients of speech could detect andby which they might assess the affective state, sincerity, and truthfulness ofthe sender. Music and dance continued to evolve as somewhat autonomousmusical and bodily kinesthetic intelligences sensu Gardner that, like speech,sign language, and culture, were greatly facilitated by imperfectly under-stood symbolic processes

Finally, as the language of a hominid community grew more analytic,strong selective pressure for rapid analytic learning by children and sharingtruthful information among close relatives resulted in language that wasmostly composed of anatomic meaning units (morphemes or words). Fitchfurther speculated that the analytic urge was genetically fi xed in the finalstage of human language evolution

First, phonology developed via complex vocal learning during an initialperiod of songlike communication that lacked propositional meaning, per-haps driven by sexual selection or kin selection or a combination of bothfactors.

Advocates for a lexical protolanguage posited that the hominid communi-cation system began with a large learned lexicon of meaningful words butno complex syntax, which was a culminant development in the evolution ofhuman language.

Each language and musical tradition is culturally transmitted to children,who must learn its distinctive phonological system as a whole. Like language,music is a culturally transmitted human universal that entails a generative sys-tem constituted by tonality, timbre, and rhythm, composed of hierarchicallycombined small sets of notes or tonal nuclei to form syllabic sequences i

Manual grasping, which characterizes all Primates.2. A mirror system for grasping shared with Macaca.3. Simple imitation shared with Pan troglodytes but not Macaca.4. Complex imitation that is beyond the capacity of Pan troglodytes.5. Innovation of protosign, which underpins an open repertoire.6. Innovation of protospeech (vocal control) by extending volitional neocor-tical motor control of the hands and face to the neighboring motor andpremotor cortex for musculature of the tongue and larynx.7. Human language

musical protolanguage proposedthat speech began as complex learned vocalizations that were more like songthan speech, with semantics developing later

Those favoring lexical protolanguage proposed that lan-guage started with isolated meaningful spoken words

Advocates of gestural protolanguage suggested thathominid language began in a manual modality and that syntax and seman-tics preceded speech.

The various theories on hominid language ori-gins fall into three basic categories—gestural protolanguage, lexical protol-anguage, and musical protolanguage—none of which in isolation suffices asa phylogenic precursor of fluent syntactic recursive vocal and signed proposi-tional human language.

Although the subject returned to scientific dis-course in the mid-1970s, it is still impossible to devise a scenario, let alone aformal model, for the origin and evolution of the modern human communi-cation system upon which linguists, psychologists, and evolutionary anthro-pologists would agree.

In brief, it is difficult to imagine hominid symbolic language developingde novo without reliance on iconic and indexical signs. The grand mystery isto determine how and when iconic and indexical signs were combined intosyntactic patterns to which arbitrary signs—that is, symbols— could beincorporated to create ever more complex communicative forms and cus-tomary expressions

In primates and other animals, learning sets are based on formative spansof co-occurrence of events leading to similar indexical cause-effect relation-ships based on event-memory. Classic conditioning of stimulus-response as-sociations in laboratory animals establishes memory-based indexical regu-larities: for example, pressing levers or touch screens or traversing labyrinthsresults in food rewards

n both humans and great apes, social communication is an intrinsicallycreative process that unfolds as partners continuously adjust their behaviorsto one another: coregulation. That is to say, our social communication ismore than a linear transfer of information from a sender to a receiver whodecodes the signal for its information content.15 Nevertheless, while humanand nonhuman communication systems share much of the same semioticproperties— of indexical and iconic types—research on the evolution ofhuman language, which is uniquely symboli

Effective communication between individuals fundamentally requires a meansof production by senders and comprehension by receivers.

Deacon and Donald are correct in seeing symbol use as themost fundamental factor in language evolution

After the wave of market-oriented reform of the 1990s, the pendulum swung back in the2000s in most countries to greater state intervention. For some, this trend signaled aworrisome revival of state capitalism, but in most instances states did not reverse the major liberalizing reforms of the 1990s but rather devised new policies to promote development in more open economies.

Resource dependence is not newin Latin America, but it did have dramatic effects in the 2000s by generating current account and fiscal surpluses (not seen in decades) and pressuring currencies to overvalue(with consequent Dutch disease effects such as specialization in natural resource exportsand deindustrialization). Commodity price surges also boosted revenues, investment, andvaluation of dominant, hierarchical business groups in major commodity sectors (groupslike Vale, EBX, and Votorantim in Brazil; Agnelli, Luksic, and Matte in Chile; and Cemexand Grupo Mexico in Mexico) that displaced slower-growing manufacturing firms.

For example, neoliberal reforms of the 1990s attracted a new wave of MNC investment to Latin America. Privatization in particular attracted large investments, and shifted the composition of the largest firms by reducingthe state share and increasing the MNC share

Last, some world leading manufacturing firms in HMEs(some of the best-known cases are Embraer [aircraft, Brazil] and Techint [steel tubes, Argentina]), have managed to create pockets of lasting investment in skills and well-mediated employment relations and consequently look more like CME firms

Dispersed ownership in LMEs shifts decisional authority tomanagers, but also subjects them to short-term monitoring and performance pressures.Owners have greater control in the other non-LME varieties, where investors tend to bemore “patient.”

My more deductive point of departure is that contemporary capitalist systems—definedby the predominance of mostly free markets and private property—accommodate a limited number of alternative mechanisms for allocating resources, especially the gains frominvestment, production, and exchange. These mechanisms are markets, negotiation, trust,and hierarchy, and correspond in “varieties of capitalism” terms to, respectively, liberalmarket economies, coordinated market economies, network market economies (NMEs),and hierarchical market economies (HMEs)

liberal market economies (LMEs) and coordinated market economies(CMEs).

ames Mahoney (2010), for a recent example, concludes that more liberal versus more mercantile patterns of incorporation into theSpanish Empire explain long-term differences in economic and social development. Others emphasize the long-term effects of mining enclaves versus diversified agriculture onclass structure and prospects for development (Cardoso and Faletto 1979). More recently,patterns of engagement with the latest wave of globalization also set countries on different trajectories, as for example Mexico and other countries of Central America and theCaribbean that integrated deeply through global production networks into the U.S. economy and expanded low-skill employment in manufacturing (Stallings and Peres 2000)

comparison with previous cases of successful commodity-led development(for example, Finland or Australia) can pinpoint factors and policies (such as heavy investment in education) underlying those successes

(1) an initial spread of "light industries,"producing for a protected home market; (2) a politically weakdomestic bourgeoisie, which unlike their nineteenth-centurycounterparts, were not impelled to lead the search for new mar-kets or technologies; (3) the emergence of"populist" distributiveand political pressures relatively early in this process of industri-alization; and (4) the reliance on imported technology and capi-tal goods, financed through external credits or commodity ex-ports. My focus here has been on the transitional problems ofcount

velopment of a vertically integrated industrial sector was at mostpartial--often set aside in favor of other, more "conventional"objectives, such as facilitating the importation of supplies neces-sary for the expansion of existing public and private enterprises.It should also be clear that the heavy industrial expansion whichoccurred in Argentina and Brazil did not, in any absolute sense,imply dramatic "breakthroughs" in reducing the vulnerability ofthese countries to the external economic environment

B-A rule may be instrumental for convert-ing vicious developmental circles into "virtuous" ones. Twomajor developmental projects-arguably the newest and mostfundamental of the options discussed above-have been em-phasized in this connection. O'Donnell, as noted, stresses the po-tential role of B-A regimes in cutting through the conflicts asso-ciated with deepening.74

The problem of controlling inflationary pressures-what Al-bert Hirschman once called a form of peaceful civil war-62wasboth a symptom and cause of these changing circumstances. Nopopularly elected government (and only a few authoritarian re-gimes) could impose the sacrifices necessary to contain thesepressures.

The immediate crises which triggered B-A coups can betraced to these bottlenecks. The connections, however, are com-plex and indirect; for such crises were neither the unique prod-ucts of this phase of industrialization nor did they follow in-･ｶｩｴｾ｢ｬｹ＠in its ｷ ｡ ｾ ･ Ｎ ＠ Indeed, Mexico and Brazil in the 195o'ssurvived the declme of "easy" lSI opportunities relatively une-ｶ･ｮｴｦｾｬｬｹＺ＠without serious problems of "stagflation

In short, despite compa-rable levels and patterns of industrialization, populism operatedwithin these countries with different degrees of strength andsuccess, in the context of both open and closed political regime

irst, common external stimuli seemed toweigh heavily in each case-prompting existing manufacturingfirms, import merchants,

quite distinct from the religious, inward-looking, anti-capitalistand traditionalistic themes which prevailed in Spain and Por-tugal until the 196o's. Instead, cooperation with internationalbusiness, a fuller integration into the world economy, and astrictly secular willingness to adopt the prevailing tenets ofinternational economic orthodoxy, all formed a different, but noless ideologically bounded, set of intellectual parameters withinwhich the technocrats could then "pragmatically" pursue there-quirements of stabilization and expansion.

!he use rf. ｴ ｨ ｾ ＠ ｳ ｴ ｡ ｾ ＠ apparatus to depoliticize organized labor.Despite the vanatlon m the bases of their class support, eachgovernment focused on organized labor as a primary target ofrepression, employing various mixtures of police surveillance,arrests, and cooptive "corporatist" labor laws to penetrate andcontrol the union movement. In each case, strikes were strictlyprohibited and the unions purged of all leaders viewed asthreatening by ruling officials

A sharp restriction rf "public contestation. "31 Each governmentchanged the prevailing rules of the game in ways that were quiteclearly authoritarian. Government officials did not tolerate seri-ous public challenge of their incumbency or basic policies; elec-toral competition and civil liberties were severely curtailed

Bordaberry's civilian administration was formedprimarily by rancher-politicians; conservative technocrats, as agroup, played no significant role

Another possible difference between Brazil and Argentina isthe relative emphasis which the latter placed on "deepening," al-though here the evidence is less clear. In fact, both the Argen-tines and the Brazilians frequently chose to supply capital andequipment needs through imports, rather than through the en-couragement of local production, relying on export promotionand external credits to furnish the necessary foreign exchange.

1) the expansion and diversification ｯ ｾ ＠ ･ ｸ ｰ ｯ ｲ ｾ ｳ Ｌ ＠( 2 ) development of the economic infrastructure (particularly mrailroad, electricity and port facilities, whe:e the ｧ ｯ ｾ ･ ｲ ｮ ｭ ･ ｮ ｴ ＠promised to rationalize and streamline ｰ ｵ ｢ ｾ ｉ ｣ ＠ enterpns.es); and(g) the ｰ ｲ ｯ ｾ ｯ ｴ ｩ ｯ ｾ ＠ of ｾ ｡ ｳ ｳ ｩ ｶ ･ ＠ ｰ ｲ ｩ ｶ ｡ ｴ ･ Ｍ ｳ ･ ｾ ｴ ｯ ｲ Ｎ ｭ ｶ ･ ｳ ｴ ｭ ｾ ｮ ｴ ＠ ｉ ｾ ＠ heavyindustry, pnmanly with the aid of multmational capital

Congress and issued the famous Fifth In-stitutional Act, which granted the president unrestricted powersto protect "national security." The ensuing wave of repres-sion-arrests, dismissals, police surveillance, and strict censor-ship-continued unabated during the first three years of theMedici administration (1g6g-1974), which became particularlynoted for its use of torture in dealing with actual or suspectedopponents. In

usterity measures had pro-､ｵ｣･､ｾ＠precipitous (and perhaps unnecessary) ､ ･ ｣ ｬ ｩ ｮ ｾ ＠ in ｾ ｯ ｲ ｫ ﾭing-class living standards, with only a ｭ ｯ ､ ･ ｲ ｡ ｾ ･ ＠ ｲ ｾ ､ ｵ ｾ ｴ ｩ ｯ ｮ ＠ m. therate of inflation itself. Although external pubhc md did flow mtoBrazil during this period, private investors and lenders wereslow to respond to the liberal incentives offered by ｴ ｾ ･ ＠ ｃ ｡ ｳ ｴ ｾ ｬ ｯ ＠government, and the ｲ ･ ｾ ･ ｳ ｳ ｩ ｯ ｮ ＠ whic.h had gripped the ｾ ｮ ､ ｵ ｳ ｴ ｮ ｡ ｬ ＠southeast showed few signs of abatmg.

Brazil. The central theme that dominates much of the histori-cal analysis of this case-the most enduring and "successful" ofthe new B-A regimes-is the tentativeness and uncertainty withwhich the new military government groped toward thepolitical-ec??omic ｾ ｯ ｲ ｭ ｵ ｬ ｡ ＠ for which it is now so famous.

The gov-ernments with which we are concerned display a rhetoric and in-tent which is modernizing and impersonal. Their "arbitrary"rule over workers, politicians, students, and intellectuals is ac-companied by attempts to establish pragmatic and predictablerelations with the private entrepreneurial sector, particularlyinternational business, and to rationalize and advance the econ-omy as a whole. It is this combination of repression and "ra-tionality" which warrants empirical exploration and theoreticalexplanation.

Nevertheless, the South American governments referredto above have employed public instruments of surveillance, re-pression, and torture with unusual ruthlessness and efficiency

Why are so many of the industrially advanced South Americancountries dominated by governments which express strongcommitments to economic development and simultaneouslytrample so thoroughly on human rights?

ntegrative Party System (Mexico and Venezuela). These cases had a stablecentrist majority bloc in the electoral arena, and the labor movement was or-ganizationally tied to the political center and thus linked to the governing co·alition. These regimes generally preempted or defeated leftist and oppositionmovements, contained social conflict and polarization, and were stable andhegemoni

Electoral Stability and Social Conflict (Uruguay and Colombia). These re·gimes had a stable centrist majority bloc in the electoral arena, but unionswere not organizationally linked to it. In Uruguay the unions were consistentlyoriented to parties of the left and hence generally played an oppositional role,and in Colombia they were increasingly oriented in a similar way. The resu

The combination of new political hopes on the left and new political fearsin other parts of the political spectrum set the stage for a major polarizationwithin the region. Amid these hopes and fears, political dynamics revolvedin part around the "objective" potential for radicalization in each country,but also around the "perception of threat" (O'Donnell 1975) on the part ofthe military and other more conservative sectors within each country.

We argue that the varying scope of thisopposition and crisis in each country can be explained in part by .character-istics of the party system and its political or hegemonic resources. Somecountries experienced severe polarization, whereas in others the polarizationwas more mild and to one degree or another was effectively contained byestablished political actors. In this part of the analysis we explore both theeconomic challenges reflected in the politics of stabilization policy and thepolitical challenges that derived from the emergence of new oppositionmovements in the party arena and in labor and peasant organizations

For the cases of stateincorporation, it begins with the restoration of a competitive regime withina year of the end of the incorporation period

Labor Populism. Peru (1939-48) and Argentina (1943-55) experienced ac-tive electoral 'mobilization of labor support and a major effort to link unions toa party or political movement, but the incorporation project did not encompassa peasantry. 3 Because the more extensive mobilization of this type remainedrestricted to labor in the modern sector, we refer to it as labor populism.

Electoral Mobilization by Traditional Party. Colombia (1930-45) and Uru-guay (190;3-16) experienced active electoral mobilization of labor support, butthe effort to link unions to the party was either limited or nonexistent, and theincorporation project did not encompass the peasantry.

ontrol and Support. Was the major goal of the political leaders whoinitiated incorporation primarily to control the working class, with at mostmarginal concern with mobilizing its support, or was the mobilization ofsupport part of a political strategy to gain and maintain power of at leastequal importance?

Our basic thesis is that the incorporation periods were a crucial transition,in the course of which the eight countries followed different strategies ofcontrol and mobilization of the popular sectors. These differences had a long-term impact on the evolution of national politics. We do not intend to sug-gest that once the initial incorporation period had occurred, the patterns es·tablished remained unchanged. Quite the contrary, these periods set into mo-tion a complex sequence of reactions and counterreactions, and the legacy ofincorporation is to be found in the working out of this sequence

THE PERIOD of initial incorporation of the labor movement is defined as thefirst sustained and at least partially successful attempt by the state to legiti-mate and shape an institutionalized labor movement.

Yet in all eight cases a well-defined change in governmentserves as a benchmark in the inauguration of this period of reform and of thetransformation of the oligarchic state: in Chile the election of Arturo Ales-sandri in 1920; in Brazil the Revolution of 1930 and the assumption of powerby Getulio Vargas; in Uruguay the election of Jose Batlle y Ordonez in 1903;in Colombia the beginning of the Liberal period and the election of EnriqueOlaya Herrera in 1930; in Venezuela the death of Juan Vicente Gomez in1935; in Mexico the 1910 Revolution and the fall of Porfirio Diaz in 1911; inPeru the second presidency of Augusto Leguia, beginning in 1919; and in Ar-gentina the beginning of the Radical period with the election of Hipolito Yri-goyen in 1916.

In countries where worker protest in the modern sectorhad been accompanied by peasant rebellion and protest and/or by the wide-spread displacement of peasants from their land in the traditional rural sec-tor, issues of peasant incorporation, land reform, and in some cases of thereform of policy toward Indians was also raised

They debatedthe appropriate role of the newly emerging working classes within the eco-nomic and political system and the problem of mitigating the exploitativeconditions of work that appeared to encourage this new social protest.

Other interna-tional actors played a role as well, such as the international communistmovement, whose evolving policy had a major impact on the coalitional po-sition not only of national communist parties but also of national labormovements, thereby strongly influencing domestic coalitional patterns. Bothworld wars had major ramifications in Latin America

TABLE 0.

labor populism

The core similarity in each pair derives fromthe analysis of the incorporation periods, presented in Chapter 5. The casesof state incorporation, where the state sought primarily to impose new meth-ods of control, are Brazil (1930--45) and Chile (1920-31). Among the cases ofparty incorporation, where the concern with control was accompanied by amajor effort at support mobilization, we distinguish three subtypes. First, inColombia (1930--45) and Uruguay (1903-16), the mobilization of workers wascarried out by traditional parties as an aspect of electoral competition withinan established two-party system. Since these parties were founded in the19th century and had strong ties to the economic elite, not surprisingly thistype involved the most limited mobilization of the working class, being re-stricted largely to electoral mobilization. We refer to this category as elec-toral mobilization by a traditional party

The literature ~n the emergence of labormovements in Latin America1 exhibits a broad consensus on the socioeco-nomic conditions of special rekvance to creating this opportunity: (1) urban-commercial development, (2) industrial growth, (3) the emergence of isolatedenclaves of export production, (4) conditions of labor shortage (or labor sur-plus), and (5) European immigration.

These international influences contributed to a common rhythm of labormovement development shared by a number of countries. Within that frame-work, major variations among countries in the scope of worker organizingand protest can best be understood in light of internal economi

The terms on which the labor movement was initially incorporated dif-fered greatly within Latin America. In some countries the policies of the in-corporation period aimed primarily at establishing new mechanisms of statecontrol. In other cases the concern with control was combined with a majoreffort to cultivate labor support, encompassing a central role of a politicalparty-or a political movement that later became a party-and sometimesproducing dramatic episodes of worker mobilization.

The argument is not that the socioeconomic context of politics is unim-portant. Rather it is that the political arena is not simply fluid, constantlyresponding to socioeconomic change. Instead, because of an autonomous po-litical logic and vested interests, it may be resistant to such change over sig-nificant periods of time. Socioeconomic change is important to political out-comes, but the political arena may to some degree follow its own pattern andpace of change, that at times takes a highly discontinuous form.

On the basis of thesedimensions, four broad types of outcomes are identified: integrative partysystems, multiparty polarizing systems, systems characterized by electoralstability and social conflict, and stalemated party systems

wo sequences of change may initially be identified. In cases of state in-corporation, the incorporation project was principally concerned with statecontrol of the labor movement and was implemented under an authoritarianregime. Correspondingly, the initial regime breakdown brought with it a pro-cess of democratization. In the cases of party incorporation, the incorpora-tion period promoted progressive social policies and the political mobiliza-tion of the working class, and the regime under which incorporationoccurred was in most cases more democratic and competitive.

In the first decades of the 20th century, the relationship between the stateand the labor movement changed fundamentally. Prior to that time, statepolicy commonly involved extensive repression of working class organiza-tion and protest, repression that on many occasions resulted in the death ofdozens or even hundreds of workers.

we do not intend to suggest that workers and laborleaders did not themselves play a major role in constituting labor move-ments. Their role has been amply documented, 2 and at various points it playsan important part in the present analysis. 3 However, our primary attentioncenters at a different level: the repercussions for the larger evolution of na-tional politics of alternative state strategies for dealing with the labor move-ment.

e goal of this book is to provide a framework for this comparisonand to offer a methodological and analytic basis for assessing the causal im-pact of the incorporation periods on the national political regime

f one connects this insight with the fact that Latin American countriesare some of the most unequal societies in the world, and that in spite of socialrevolutions, land reform, the construction of partial welfare states, redis-tributive transfers to the poor regions, and (sometimes) economic growththe differences between rich and poor in the Latin American federationsremain abysmal, one cannot help but consider the possibility that the solu-tions each country has found to the dilemma of centralized fiscal federalbargains have played a key role in the reproduction of this inequality

The nationalizationof party systems provides precisely that kind of insulation because localpoliticians can attenuate local risks through a hedging effect generated bynational electoral processes. In my view, this process precedes the concen-tration of tax authority in federal hands and lends credibility to the transferbargain through which local elites are assured resources

A related issue is that too much attention in the scholarly literature isdevoted to the question of the size of government.10 Fukuyama (2004) hasnoted that competence in collecting taxes should be kept distinct from dis-cussions concerning the size of the state.

Comparisons between Africa and Latin America would be very usefulfor understanding the nature of centralized political authority. Although inLatin America population densities are more similar to those of Europethan of Africa, national leaders in the vast federal countries did find“African” problems of how to extend their power beyond the limits of thecapital city and its surrounding area

More generally, my study of fiscal centralization suggests that, in prin-ciple, there is no direct link between modernization and fiscal centraliza-tion;

To the extent that transfer systems involveda credible commitment from the federal government, local elites were will-ing to cooperate with the federal government rather than compete againstit: Their political survival was enhanced, rather than threatened, by the sta-bility brought about by the federal government.

Neitherthe military in Brazil nor the PRI in Mexico eliminated elections at that locallevel because those contests conveyed important information about citizens’dissatisfaction with the regime, allowed local elites to compete within clearlyestablished institutional channels, and selected out the leaders with whomthe center would bargain to construct a national coalition

stability was based not on force but on the distribution of funds to theregions.

My findings suggest that in Latin America both democratic and author-itarian regimes constructed ruling coalitions by using similar devices toappease regional interests, an insight originally formulated by Ames (1987).Both the Mexican dictablanda emerging from the Revolution and the Brazil-ian democradura of the military governments between 1964 and 1988 wereremarkably stable political regimes.

heformula involved a large degree of centralization. In this sense, it is notobvious that Latin America is a model of successful federalism. But on thegrounds of stability, the Latin American solution for state-building was verysuccessful.

On one hand, state politicianscan exercise threats to the extent that they believe their careers and politicalfuture depend on their defense of local interests. Off-the-path threats bylocal politicians are binding when politicians are more locally oriented butmay become noncredible when politicians are nationally oriented. Hence,one can think of outcomes of centralization as being divided accordingto whether local politicians are nationally or locally oriented. This is thepolitical dimension of the centralized fiscal bargain

he implication of these predictions is therefore that the system ofrevenue-sharing (and in general any system of resource allocation wherestates can threaten to go on their own) will not be particularly redistribu-tive because states will only accept a share that is highly correlated with theireconomic capacity.

Third, the more“productive” a state is in collecting federal taxes and the more important itsshare of total federal revenue, the more resources it can bargain for

Revenue centralization opens the gate to regional redistribution. When thefederal level collects the most important sources of revenue, future increasesin tax collection can be distributed among the partners to the fiscal federalbargain in many ways. All actors in the federal bargain benefit from addi-tional resources that are brought in by centralized taxation, but they maydisagree on how the extra revenue should be distributed

is to link the political fates of state or provincial politicianswith those of the federal level of government. This can be achieved throughnational political parties.

The game of commitment over time thus highlights the role of insti-tutions as credibility-enhancing mechanisms that might solve problems oftemporal inconsistency. Institutional rules governing the size of federaltransfers have often been placed in constitutions, or federal governmentshave opted for some third-party enforcement of the fiscal pact, placing theobligation to pay revenue shares on an independent central bank. Anotherpossibility is that a malapportioned senate with significant budget authoritycan grant veto power to minority political actors who can find assurancein that decision body to the effect of generating compliance on the part ofthe federal government. These institutional devices can reduce the amountof resources the national government needs to transfer in order to make afederal bargain palatable to the states or provinces.

Hence, two insights emerge: namely, that states canacquiesce to a unitary system, provided that they are not willing to riskconflict; and that national governments can impose a unitary system if theadvantage of the federal organization is not too large

This suggests that the emergence of centralized federal bargains isattributable to the preservation of state and provincial threats of open con-flict and the expansion, through tax technologies, of potential revenue gen-erated through centralized income or value-added federal tax arrangements.When states can no longer threaten conflict, a unitary imposition would bepreferred by the federal government. The centralized fiscal federal bargainis contingent not only on how much revenue it can generate but on thepower of states or provinces

hat is, the national government pays transfers toavert conflicts. Feasible transfers are thus bound by the value of conflictand, implicitly, the risk of civil war.

owever, federalism only makes sense if there is some-thing to gain out of a federal arrangement,

Accepting the federal fiscal compromise means obtaining a transfer fromthe federal government but giving up the capacity to tax at will.17 Acceptingthe unitary model means losing independent local power (while perhapsretaining a dependent administrative position) and forfeiting the tax base.Rejecting the federal bargain means maintaining the system of local taxationin place, rejecting the transfer, and keeping a system of overlapping taxationbetween levels of government, even if that system is inefficient from anoverall economic and revenue-maximization point of view. In the case of aunitary imposition, the rejection strategy means resisting a national army,leading perhaps to civil war

n the extreme case, the establishmentof a hegemonic party as in Mexico or dominant political machines such asthe Peronist Party in the Argentine provinces offers long political careerswhile eliminating electoral threats

Such a bargain, I argue, can be based on two types of sidepayments: financial transfers from the federal to the state governmentsand assured political careers to local politicians. Financial transfers andassured careers are only viable if the national government can crediblycommit to providing them, an issue that is explored in the next section.Whereas financial transfers can be understood in a relatively straightfor-ward manner as side payments, careers as concessions require some furtherexplanation

Of course, the problem is that too much centralization can also beinefficient in terms of foregoing a close match between government policiesand citizens’ preferences. State governments are reluctant to accept limitson their fiscal authority, and federal regimes vary widely in the degree offiscal centralization they exhibit

This body of work suggeststhat the success of federal systems depends on being able to strike a balancebetween states being strong enough to protect their rights and having afederal government that is strong enough to provide goods and services tothe constituent jurisdictions

In Brazil, stategovernments instead kept fiscal authority, so the system remained highlydecentralized. A credible threat by the most powerful states to challengethe federal government always remained in place. Not even the militaryrulers successfully centralized the system to the extent that they envisaged.Instead, in order to stay in office, they had to construct a ruling coalitionthrough a crafty combination of respecting the fiscal authority and polit-ical autonomy of the powerful states while constructing a redistributiverevenue-sharing transfer system

In Mexico, a hegemonic party provided an institutional solution to thedilemma of commitment encountered by regional politicians in the pro-cess of state-building. This institutional solution tempered the centrifugalforces unleashed by the Mexican Revolution and set the stage for a periodof moderate growth with political stability. The hegemonic party struc-tured a highly disciplined system of progressive ambition. Local politi-cians were willing to empower the federal government to pursue a cen-tralized national “developmentalist” strategy. 8 The PRI system protectedlocal politicians from electoral challenges and ensured them attractive polit-ical careers. Local politicians surrendered their fiscal authority becausethey were protected from competition in local electoral and economicmarkets

At the beginning of the 20th century, state (provincial) and local gov-ernments around the world maintained a large degree of fiscal authority.The authority to levy excise, sales (turnover), income, or inheritance taxesdid not belong exclusively to national governments. But over the course ofthe century, most national governments centralized and obtained exclusiveauthority over these taxes while local and state governments instead gottransfers from the central level.

When fiscal bargains are struck between local and national politiciansfor the creation of revenue-sharing systems, the locus of authority overtaxation is shifted away from state governments

In all cases, political representation conditioned the constructionof regional coalitions that determined the specific way in which fiscal bar-gains unfolded.

The book argues that fiscal centralization occurs when national politi-cians use the power of the central government to protect regional politiciansfrom challengers and electoral threats in exchange for financial resources. Inturn, regional politicians are willing to forgo fiscal authority.

The basic premise of the theoretical framework ofthis book is that fiscal centralization can be best studied as the consequenceof a bargain struck between self-interested regional and national politicians.

is book explores the politics offiscal authority, focusing on the centralization of taxation in Latin Americaduring the 20th century. The first half of the book explores this issue ingreat detail for the case of Mexico. The political (and fiscal) fragmentationassociated with civil war at the beginning of the century was eventuallytransformed into the highly centralized regime we associate with Mexicotoday

Finally, the state may fail altogether in its attempt at penetration. Disen-gagement or lack of engagement of the state in the local arena will result in littletransformative effect on the society — and limited effects of the society on thestate. Failures to engage in arena struggles in even the most remote parts of thecountry can affect the state in the capital city by denying state components thereresources and support from the larger society

hird is existing social forces’ incorporation of the state. In this type, thepresence of the state’s components spurs adaptation by dominating social forces,but does not produce radical changes in the pattern of domination.

Second is state incorporation of existing social forces. In this type, the state’sinjection of new social organization, resources, symbols, and force into an arenaenables it to appropriate existing social forces and symbols in order to establisha new pattern of domination.

These struggles and accommodations in the junctures between components ofthe state and other social forces have produced a range of outcomes. We cancapture these in four ideal types of results. First is total transformation. Here,the state’s penetration leads to the destruction, co-optation, or subjugation oflocal social forces and to the state’s domination.

whether particular social forces can create an integrated domination.That is, can they prevail within given arenas to produce resources and support —a material base and a normative framework — that can be used to dominatelocally and then be carried into other domains to create society-wide domina-tion? Or do the struggles in the arenas result in a pattern of dispersed dominationby limiting the creation of authoritative, legitimate forces that can dominatebroadly across society

Frances Hagopian, forexample, demonstrates how in Brazil, the authoritarian military regime found ithad to reinstate accommodations with local traditional, oligarchic elites after ithad instituted a political system of domination that it believed had rid Brazilianpolitics of these very forces. “The military was no more successful at cleansingthe political system of patronage politics than it was at purging the state of thetraditional political elite.” The old patrons’ ability to manipulate resources inorder to achieve domination in local arenas forced the state’s leaders to seek animplicit coalition with them.

Even in as established a democracy as India, as Kohliargues, integrating frameworks of authority are difficult to find today. In fact,the opportunities provided by democracy for mobilization have opened the wayfor new groups, especially the lower and lower-middle strata, to expand theirparticipation in politics substantially.

To conclude this section, there has been an unfortunate tendency in socialscience to treat the state as an organic, undifferentiated actor.

rprising that states seldom generate a single, homogeneous response to anissue or problem, or even necessarily a varied but coordinated set of responses.The different constellations of forces each part of the state faces mean thatvarious units have diverse histories of their own, leading to differing degrees ofesprit de corps, purposefulness, and insularity. Political outcomes — the formu-lation and implementation of the state’s policies — reflect the aggregation of aseries of different actions based on the particular calculus of pressures experi-enced by parts of the state at each level.

he commanding heights. At the pinnacle of the state is the top executiveleadership.

The agency's central offices

he dispersed field offices. A notch higher are the regional and localbodies that rework and organize state policies and directives for local consump-tion, or even formulate and implement wholly local policies.

The trenches. Here stand the officials who must execute state directivesdirectly in the face of possibly strong societal resistance. They are the taxcollectors, police officers, teachers, foot soldiers, and other bureaucrats with themandate to apply state rules and regulations directly.

An anthropology of the state, then, allows us to dismantle it analytically andto discern the distinct structural environments of its different components andthe interaction among them. One possible way to disaggregate the state is tobreak it down into four levels, which differ markedly in the kinds of pressuresthey face from other state components and from nonstate actors. From bottomto top, they are

“My principal interest,” writes Mann, “lies in those centralized institu-tions generally called ‘states,’ and in the powers of the personnel who staffthem, at the higher levels generally termed the ‘state elite.’ ”?But states surely consist of far more than this. As in ideology and policyformulation, policy implementation also reflects the state’s engagement withother social forces.

State leaders attempt to create an aura of invincibility about the state.

n short, throughout the territory they claim to govern, most political leadershave maintained that the state should have primacy. In some instances, that hasmeant privileging powerful social groups with which state leaders are allied aswell as the organizations those groups dominate, such as markets or churches.But commonly the quest is for the state to exercise control directly — to im-pose its own systems of meaning and boundaries for acceptable behavior cen-trally on its subjects, in everything from sexual unions to labor-management re-lations.

n thespecific instance of Shanghai, Elizabeth Perry notes how the Chinese Commu-nist Party (and, later, state) changed in character as the result of incorporatingthe most skilled components of that city’s working class. Similarly, as ResgatKasaba demonstrates, the nineteenth-century Ottoman state’s engagement withprincipally non-Muslim merchants fundamentally changed its goals and charac-ter, drawing it into many new roles and procedures.Problems with existing conceptions of the state go beyond lack of interest inthe changing foundation upon which state goals are built; problems exist evenon the issue of capabilities, which is the heart and soul of such definitions.There is a troubling tendency of authors to take too seriously actual states’abilities to make binding their decisions for people. This penchant to exaggeratecapabilities has stemmed from states’ near ubiquity in the struggles and accom-modations occurring in arenas of domination and opposition, as well as fromthe presumptions of state officials themselves

In brief, my aim is todraw the attention of those concerned with state—society issues to the appro-priate focuses of analysis.

Often, state- or society-driven initiatives have been provoked by the funda-mental changes associated with the great transformation — the growth of cities,the increased use of fossil fuels and other technological innovations, the declineof agriculture in terms of total domestic production, and so on. These changeshave swept beyond Europe to every nook and cranny of the globe. Capitalismand the model of the strong European state have sent reverberations throughevery continent, precipitating massive dislocation and mixtures of appropriationof new ideas and methods, reactions against them, and their adaptations tolocal circumstances.

The point is that toglean the patterns of domination, one must focus on the cumulation of strugglesand accommodations in society’s multiple arenas.Such a focus is possible only by first conceptually breaking down states andsocieties and the junctures between them. In some cases, the numerous strugglesmay move a society toward integrated domination, in which the state as a whole(or possibly even other social forces) establishes broad power and in which itacts in a coherent fashion. In other instances, the conflicts and complicities inthe multiple arenas may lead to dispersed domination, in which neither the state(nor any other social force) manages to achieve countrywide domination and inwhich parts of the state may be pulled in very different directions.

Following the IndustrialRevolution, Karl Marx, Max Weber, Emile Durkheim, and many others devotedthemselves to what Karl Polanyi would later call the great transformation.' LikeHobbes, they too focused on the state, now in its relationship to the momentoussocial and economic changes overtaking European societies.

States and other social forces may be mutually empowering. Finally, weurge scholars to eschew a state-versus-society perspective that rests on a viewof power as a zero-sum conflict between the state and society.

Social forces, like states, are contingent on specific empirical conditions. Ifit is important to resituate states in their social setting, we also need to considerthe adequacy of the categories in which we are accustomed to conceptualizepolitically salient social structures and social actions. This is our third claim.We adopt the view that the political behavior and the power capacities of socialgroups are contingent, at least in part; in other words, the political action andinfluence of a social group are not wholly predictable from the relative positionof that group within the social structure.

States must be disaggregated. A second, related claim follows from thefirst. If states have to be viewed in their social contexts, it is important to studynot only the peak organizations of states and key social groups, often located atthe center of the polity in the capital city, but also state-society interactions atthe periphery.

We suggest instead that a state’srelative effectiveness is a function of the varied forms in which state—societyrelations are interwoven.

They also have blurred the development of specific explanations; for example, astate’s autonomy from society has come to be mistakenly viewed by some as asource of that state’s effectiveness.

States are parts of societies. States may help mold, but they are also continu-ally molded by, the societies within which they are embedded. Once the state’simportance has been emphasized, therefore, the intellectual attention immedi-ately shifts to issues of why states do what they do, under what circumstancesstates are effective, and why states differ in their respective roles and effective-ness.

he present volume shares these “first assumptions.” The more recent state-oriented literature, however, which constitutes a subset of the larger body ofWeberian scholarship of the 1970s and 1980s, has pushed some statist claimstoo far.

develop a state-in-society approach,

he present volume continues the dialogue withthese intellectual traditions by offering a state-in-society perspective. Proceedingboth at a general and at country-specific levels, the contributors to this collectionhope to persuade others to move in several related but new theoretical directions:to go beyond “bringing the state back in” by resituating the study of states intheir social setting and thus adopting a more balanced state-in-society perspec-tive; to disaggregate states as objects of study, both as an end in itself and as ameans toward a better understanding of states and political change; to rethinkthe categories used to conceptualize the evolving and fluid nature of socialforces in developing countries; and to be continually sensitive to the mutuallytransforming quality of state—society relations

I have proposed a political and sociological theory of state building thatemphasizes the absence of labor-repressive agriculture and a collective, but exclusionary,elite political hegemony. Both conditions are necessary if state building is to proceed.Without the former, agrarian elites, typically the dominant political force at the initia-tion of state building, cannot tolerate the centralization of coercive authority. Withoutthe latter, elites would lack an incentive to cooperate across whatever other issues maydivide them to strengthen the state.

State expansion was thus not a response to an anarchic environment, contra belli-cist expectations. Indeed, given the technological constraints on its neighbors in themid- to late nineteenth century, Chile was, as Blakemore has called it, effectively an“island nation.” 100 Hemmed in by the Andes to the east, the massive and scarcelypopulated Atacama Desert to the north, and the Pacific Ocean to the west, none ofChile’s neighbors had the logistical capacity to launch a serious attack on Chileaninterests.

The critical point is not that the oligarchy spoke with a single voice in the Portalianera (it did not) but rather that it created a system in which its internal conflicts couldbe managed in a way that did not lead to the destruction of the administrative structureof the state.

Indeed, its diversity was also its strengthfrom the perspective of state building. There were no marked social or economic bar-riers to entry into agriculture for newly wealthy elites operating in other sectors of theeconomy. Instead, agrarian property ownership quickly became the gateway to entryinto upper-class social circles. 88 It also paved the way for participation in nationalparliamentary politics, for it brought with it clientelistic influence over the votes ofhacienda peasants in a system of typically quite restricted suffrage—and by the latenineteenth century, electoral registries also came under the control of the wealthiestlocal taxpayers, quite often a local landlord.

that these were voluntary tenancies—peasants could legallyand practically leave at will. And indeed, even the traditional hacienda system in Chilealways had a substantial migratory and seasonal component (afuerinos) that providedsupplemental labor during harvest periods.

ndeed, the Chilean state was not even born as a territori-ally unified entity; for fifty years after independence, its southern provinces were cutoff from the core of the state by an unconquered and well-defended Indian nation thatspanned the breadth of the country.The factor that opened the door to political development—though itself notsufficient—was the absence of a labor-repressive agrarian social structure.

Even as of the 1950s, the creation of a basic meritocratic civilservice was not yet a reality; the possibility was only just entering serious politicaldebate. Administratively, the central government lacked technical competence, paid itsemployees little, did not coordinate among its agencies, and was governed by exces-sive and contradictory rules.7

Unlike Chile’s Central Valley, however, Spain’s conquest of Peru did not eradicatethe indigenous population or its independent communities, though impoverish them itdid. And it was the survival of these communities, down to the present day, that madethe ongoing violent conflicts over the control over land both possible and inevitable.

The war laid bare the fundamental social problem rooted in the countrysidethat made Peruvian political modernization impossible: the continued reliance on anenormous, and rebellious, semiservile labor force—the indigenous population of theSierra and indentured Chinese laborers on the coastal plantations. Indeed, class andracial politics very quickly overwhelmed “national” loyalty as the Chilean invasionarrived. That is, elites were generally far more concerned that peasants remain undercontrol than they were with contributing to the national defense against the Chileaninvaders.

Guano wealth thus underwrote the expansion of the Peruvian state, whosebudget ballooned eightfold in the span of twenty-five years. According to Peloso,nearly half of the guano revenues went into the expansion the military and civilianpersonnel of the central state; much of the rest was effectively transferred to domesticand foreign bondholders. 48 Resource wealth thus contributed to ending internal con-flict, but was in large measure dissipated as rents rather than invested in institutionaldevelopment. That said, it still marked an administrative improvement relative to theera of greater natural resource scarcity

Postindependence Peru thus remained deeply fragmented by race and class, dis-tinctions that were at the very heart of the system of taxation and social organization.But even the omnipresent threat of indigenous insurrection was insufficient to promotecooperation among members of the Creole elite, which was also profoundly dividedalong regional lines.

Peru’s colonial history suggests that it should have had a well-developed administra-tive and coercive infrastructure (for its time). 39 The Viceroyalty of Peru was a locus ofSpanish control in lower South America, and its coercive infrastructure had to copewith the possibility of insurrection among the largest indigenous population south ofthe Mayan regions of Central America.

Directly measuring state administrative capacity is a difficult task; it is muchmore so than, for example, measuring levels of economic development. In this article,I take as an indicator of state capacity and the penetration of public power objectivemeasures of the difficult task that virtually all states must pursue: taxation for the pro-vision of basic public goods. While states have discretion over a wide range of policychoices, all require tax resources.

The approach to hypothesis testing here is comparative historical in method, utilizinga cross-time, most-similar-systems design. The justifications for such an approach areboth theoretical and practical. In the first case, arguments about the relationships ofwar, wealth, and politics to state building are cast here (and by most other scholars) inpath-dependent terms. T

Thus, even a better measure of the size of the contemporary resourcesector is not enough to permit a valid test of the natural resources–political underde-velopment linkage. It becomes clear that any such design must encompass a longhistorical sweep, and indications of resource intensity of economic life must be con-sidered at the start rather than at the culmination of the process of political developmentand state building

Scholars of state formation—by emphasizing the historical development of actually existing states—have falleninto the trap of selection bias by virtue of the fact that the states for which war sparkedcollapse (or that likely responded to external threats with what Hui calls “self-weakeningexpedients”) were often dismembered and thus selected out of the contemporarysample.25 While for those that survived, war appears to have been an important motiva-tor for institutional improvement, to correctly assess its causal weight, one must focuson the universe of states present at the initiation, not dénouement, of the state-buildingprocess

An unexplored implication of this article is that there may be an elective affinitybetween authoritarianism or oligarchic dominance and the selection onto an initialstate-building path. It would not, however, preclude subsequent democratization

Equally important arethe linkages between the state executive and the upper classes as a whole. The criticalfactor here is that elites beyond the governing faction must have achieved meaningfulpolitical incorporation.

specially the more repressive “feudal” variety—as inimi-cal to democratization, for it, like military service, would threaten their fundamentalinterests.

The emphasis here is spe-cifically on the mechanism by which labor is recruited and employed in production—isit an inherently coercive process or can workers leave the farms free of any practicalor de jure encumbrances?

Hearkening back to Moore, I contend that it is the absence of labor-repressiverelations of production that is critical to the long-run development of not regimes buteffective governmental institutions.

What we must focus on, then, are the conditions that shape thedevelopmental trajectories that states face. The critical junctures that set the stage forpolitical development or underdevelopment seem to occur early in the process of eco-nomic modernization. In such periods, agriculture is typically the dominant economicsector, and the social relations that govern its operation are thus critical to the possibil-ity of effective state building. It is my contention that where a local elite organizes alabor-repressive agrarian economy, effective political development, even in the face ofwar or wealth, is unlikely. 16

But at its core, it turns on a simple and compelling idea, sum-marized by Ross: in such contexts, politicians seize control over the ability to distributethe natural resource windfall and then “divert state assets into patronage, corruption,and pork barrel funding . . . [and] once they hold the ability to reshape resource institu-tions to their advantage, they may use the opportunity to create additional, allocablerents to meet their patronage and corruption needs.” 14 The implication is that rents,once available, help consolidate political control, which then makes possible the fur-ther corruption of the state bureaucracy in a downward spiral of malgovernance, slowgrowth, cronyism, patronage politics, and institutional decay. The critical implicationof this dynamic—though not always explicitly stated—is that where rents are notavailable, politicians must use other means to maintain support, including moderniza-tion of the state and economy or the provision of public goods

Tillyhas among the best statements of this perspective, arguing that “preparation for war . . .involves rulers ineluctably in extraction. War builds up an infrastructure of taxation,supply, and administration that itself requires maintenance and often grows faster thanthe armies and navies that it serves.”7 What most bellicist accounts are missing, how-ever, is a sufficiently clear statement of the causal mechanism that complements theiressentially functionalist theoretical structure. 8 Why do elites necessarily respond toexternal threat with institution building? Or for those who do not assume effectiveresponses, what explains the path taken? 9

state building requires rev-enues that are even more-than-usually difficult to obtain from such elites, given servileeconomies’ notoriously low productivity levels. Third, the sorts of large- and long-runinvestments (in, for example, infrastructure, education, and military professionaliza-tion) that central governments must make as part of any state-building effort requirethat leaders and the elites they represent have relatively lengthy time horizons

But where labor-repressive agrarian systems predominate, local control over coercive resources is thesine qua non of elite survival. Were central governments (and thus potentially urbanactors) to gain monopoly control of the means of violence, effective continued enforce-ment of rural political control could not be guaranteed, and the economic and socialviability of agrarian estates would be threatened.

If, for example, states initially providesome public goods reasonably efficiently, this can encourage societal (and elite) expec-tations of public efficacy, can induce coordination around public provision as a solutionto collective problems, and may create subsequent pressure to provide public goods ofdifferent types.

Theformer condition makes possible the centralization of authority and revenue extractionessential if state building is to proceed, while the latter renders it politically viable byensuring that fundamental elite interests are not harmed in the process and that thecollective benefits (and short-term rents) that development entails are well distributedamong the upper classes.

the absence of labor-repressive agriculture and exclusion-ary but collective elite political dominance