Urbanization in the Caribbean Basin: Social Change during the Years of the Crisis Alejandro Portes Jos Itzigsohn Carlos Dore-Cabral The Johns Hopkins University June 1993 To the memory of Derek Gordon, University of the West Indies. Abstract This paper presents findings of a comparative study of recent Caribbean and Central American urbanization focused on three subjects: 1) the evolution of urban systems and urban primacy; 2) class polarization of space in the larger cities; 3) the informal sector and its role vis--vis the rest of the urban economy. Earlier theories bearing on these issues and more recent counterarguments are grounded on empirical material from the larger, continental countries of Latin America. The study seeks to examine the applicability of these notions to the smaller nation-states and cities of the Caribbean Basin. In particular, the analysis highlights similarities and differences in the evolution of these three subjects during the economic downtown of the late nineteen seventies and eighties. Findings from the five countries selected for study indicate both general agreement with recent trends observed elsewhere in Latin America and significant differences across countries. Implications of these complex set of results for theories of Latin American urbanization and national development are discussed. Urbanization in the Caribbean Basin The purpose of this article is to examine empirically several recent hypotheses on the evolution of Latin American urbanization on the basis of data from five countries of the Caribbean Basin. The hypotheses were advanced by the senior author in an earlier article in this journal (Portes 1989). They concern three major aspects discussed in the literature on Latin American cities: changes in urban primacy, spatial polarization within the largest cities, and the urban informal economy as a countercyclical mechanism. The hypotheses contradict much of the scholarly consensus on the character of Latin American urbanization, summarized in a number of earlier publications (Beyer 1967; Hardoy 1975; Portes and Walton 1976; Roberts 1978; Portes and Johns 1989). The empirical research on which our earlier article was based was conducted primarily in three South American countries--Chile, Colombia, and Uruguay--and in their respective capitals. Although secondary data from the rest of Latin America was also used, the main thrust of the findings came from detailed studies of these three countries. They yielded several surprising departures from past scholarly expectations, including our own. As a whole, they did not indicate that past descriptions of Latin American urbanization were wrong but that something had changed in the process during the last two decades as a consequence of the profound transformations induced by the economic crisis. Yet the theoretical propositions suggested by these novel results could be challenged as inapplicable in other settings. They rested on weak empirical grounds, based on a limited number of field studies and on incomplete secondary data series. The earlier article cautioned against these limitations which rendered its conclusions tentative. For this reason, we decided to undertake a replication in a different regional setting, namely the smaller countries of Central America and the Caribbean. In contrast with its South American counterparts, these smaller countries possess distinct characteristics that may render generalizations based on the earlier materials invalid. This new study of Caribbean Basin urbanization recommended itself for two reasons: First, because it furnished an opportunity to test our hypotheses in a setting quite different from that which provided their original empirical basis. Second, because it afforded a chance to examine urban processes in a sub-region where they have been less studied in the past, in comparison with the larger Latin countries. The following sections summarize the theoretical conclusions of our earlier research and the particular characteristics of Caribbean Basin countries as an introduction for presenting new findings on urban trends in the area. This empirical material will be used to seek answers to two broad questions: 1) the extent to which past generalizations about Latin American urbanization apply to countries of the Caribbean Basin; 2) the extent to which significant variations exist in how the process has taken place within this region itself and, if so, their main determinants. 1. Theoretical Review The abundant literature on Latin American urbanization until the late nineteen seventies or so painted a fairly coherent picture of its evolution. This literature, which set the stage for our earlier research, emphasized the uniformity that the process had acquired throughout the continent. First, the population of Latin America was becoming rapidly urbanized but the process was distorted by the region's underdevelopment. The migration of the rural population toward the cities did not occur in a gradual manner, but as an explosive influx toward a few receiving centers. In most countries, a single city played simultaneously the role of political capital, main site for industry and commerce, and place of residence of the dominant classes. The condition of urban primacy--gigantic heads of dwarfish national bodies--was not new to Latin America, but the rural migrant flows of the mid-twentieth century accelerated it and suggested an inexorable increase in the disparity between the largest metropolitan agglomerations and the rest of the urban system (Breese 1966; Beyer 1967; Hardoy 1975; Portes and Johns 1989). Second, within the large cities, rapid demographic growth combined with highly unequal income distribution to produce other distortions. The advent of the automobile allowed the wealthy to escape the peasant crowds by moving to remote suburban locations. The political power of urban elites then compelled city governments to extend services to these areas. At the opposite extreme, increasing rents and housing scarcity drove the poor to create their own shelter solutions in irregular settlements. These were also built in the urban periphery, but in directions generally opposite to the exodus of the well-to-do from the central city. The outcome of these centrifugal forces was growing spatial polarization of the classes: rich and poor lived in increasingly different worlds even if they formally shared the same city (Amato 1969; Hardoy, Basald#a, and Moreno 1968; Portes and Walton 1976). Third, the disintegration of traditional agriculture in Latin American rural areas took place without creating sufficient capacity to absorb labor either in the new modernized farms or in urban industry. The first type of scarcity caused rural outmigration, while the second led to the growth of a vast "marginal mass" in the cities that survived by inventing employment in the fringes of the urban economy (Nun 1969; Garc!a 1982). Yet unemployment remained low because the urban poor could not afford not to work in the absence of welfare protection. Instead, the typical profile of Latin American cities featured low rates of unemployment combined with high rates of casual or informal employment that often absorbed half or more of the urban labor force (Tokman 1982). Regardless of the label applied to it, irregular work was perceived by many analysts as a countercyclical mechanism. This view was frequently discussed by economists associated with the International Labor Office (ILO) and its Latin American offshoot, the Regional Employment Programme for Latin America (PREALC) (Bairoch 1973; Lagos and Tokman 1983; PREALC 1982; Marshall 1987). Together, accelerating urban primacy, spatial polarization of the classes, and high informal employment constituted central features of Latin American urbanization prior to the 1980s. The research literature describing these features also provided a fairly coherent explanation of their causes on the basis of a common condition of external dependency. Industrialization as it took place in the region was highly centripetal in its consequences since the largest industries, many subsidiaries of transnational corporations, concentrated in the main urban centers. Together with the consistent decline in traditional agriculture, this concentration naturally gave rise to rapid rural migration toward the few places where industrial employment could be found. But industrialization under foreign control created a mismatch between the resource endowments of these countries, abundant in labor and short in capital, and the labor saving character of imported technologies (Eckstein 1977; Tokman 1982). The inability of urban industry to absorb the mass of rural migrants gave rise in turn to a growing segmentation between a sector of "modern," protected, and relatively well-paid employees and a vast informal economy where most migrants survived on the basis of invented jobs of minimal productivity (PREALC 1981; Marshall 1987; Portes and Johns 1989). The poverty of most rural migrants due to their lack of suitable employment barred them from access to market- provided housing and created the conditions for the emergence of vast shantytowns in the periphery of most large cities. Their sheer number in turn led elite sectors to escape the city toward ever more remote suburban enclaves. These twin processes accelerated the spatial polarization observed, with almost monotonous regularity, in most large cities of the region (Leeds 1969; Goldrich 1970; Cornelius 1975; Eckstein 1977). Our 1989 study did find plenty of evidence of urban primacy, spatial polarization, and a large urban informal economy, but along with them, there were notable departures from past trends. First, the seemingly inexorable increase in urban primacy had decelerated and even reversed itself during the last decade in a number of countries; second the great physical distances separating rich, middle-class, and poor in most metropolitan areas appeared to have diminished significantly in several cities as a result of novel rearrangements of the urban population; third, a number of urban labor markets registered vast increases in open unemployment questioning the assumed "countercyclical" role of the informal sector during economic downturns. Open unemployment rather than informal employment emerged as the key adjustment mechanism at the height of the crisis, contradicting the earlier assumption that in the absence of welfare coverage, the poor must find some form of employment. These trends represent not only empirical departures from past theories, but contain important lessons for their revision. Each trend appears to reflect, in its particular way, the rapid adjustment of Latin American countries to the debt-induced economic crises of the mid- 1970s and early 1980s and their changing insertion in the global economy. We summarize next the specific ways in which this societal adjustment affected each aspect of urban development and formalize the three alternative theoretical propositions that they suggest. Beginning with the region-wide economic downturn prompted by the increase in petrol prices in 1973, Latin American countries turned increasingly toward export promotion as a means to alleviate balance-of-payments deficits and service a growing foreign debt. The process accelerated during the early 1980s when a second major increase in oil prices was accompanied by the resistance of international banks to cover the deficit with fresh loans. After the Mexican debt moratorium of 1982, country after country embarked in a painful process of economic adjustment under close monitoring by international financial organizations. The details are well-known and have been examined at length in the specialized literature (Massad 1986; ECLAC 1988; Inter-American Development Bank 1990). Less well noticed have been the effects that the rapid shift from the previous import- substitution model of development to the new export- oriented model have had on civil society. An unanticipated consequence of this shift was the re-channelling of domestic migration flows toward the new growth areas created by export agriculture, export fisheries, and export platform industries. Together with the decline in employment opportunities in the old import substitution industries, concentrated in the large cities, the new migration patterns led to the rapid growth of many secondary cities and to the slowing down of growth in several metropolitan areas. Hence, export- oriented development (EOD) may reduce or even arrest urban primacy to the extent that the new export industries are located away from the major cities and hence induce a centrifugal pattern of domestic migration. The argument can be formalized in the following proposition: I. The greater the shift from import- substitution toward an export oriented model of development, the greater the probability of secondary city growth and a decline in urban primacy. The economic adjustment programs inspired by international finance organizations to deal with the debt crisis led to the exacerbation of the already marked income disparities in most Latin American countries (Iglesias 1985; PREALC 1987). In the large cities, this trend suggested the acceleration of the pattern of spatial polarization present even before the onset of the crisis (Portes and Johns 1989). The contrary results, reported in our earlier article, emerged as the outcome of two previously unnoticed processes. First, middle-class urban groups, hard-pressed by the economic situation breached the geographic divide separating them from poor settlements in search of affordable housing. In Bogot this phenomenon took the form of a massive displacement of middle-income groups toward the south of the city, formerly the preserve of the urban working-class and the marginalized groups (Cartier 1988). Second, there was a simultaneous growth of irregular or squatter settlements near high-income residential areas. This displacement was prompted by the search by the poor for some sort of employment--usually informal vending and oddjobbing--whose markets are among higher-income groups. The result of these twin processes was the partial rearrangement of the urban spatial order leading to greater intermingling of the classes. Kowarick, Gambier Campos, and De Mello (1990), who observed the same phenomenon in S~o Paulo, labelled it "perverse integration" because the spatial convergence between rich, middle-class, and poor was not prompted by a better distribution of income but rather by the generalized impoverishment of the urban population. Similar events reported in cities as dissimilar as Montevideo, Lima, and Rio de Janeiro (Portes 1989; Kowarick, Gambier Campos, and De Mello 1990) suggested a regional trend summarized in the following proposition: II. Increases in poverty and income disparities produced by economic adjustment programs led to a reduction of spatial polarization in Latin American cities as an outcome of the survival strategies of both middle-class and low-income sectors. The crisis of the 1980s brought a contraction of formal employment and a significant reduction of urban wages which, according to earlier explanations of the role of the informal economy such as those advanced by ILO and PREALC analysts, should have produced large increases in irregular employment. Our 1989 study did find that the urban informal sector expanded in most countries but that, relative to the magnitude of the economic downturn, the growth was modest. According to estimates by PREALC, it averaged about 20 percent for Latin America during the early 1980s. Underemployment, another indicator of informality, did not increase significantly in most countries with information available (Portes 1989: 24-27). By contrast, open urban unemployment increased rapidly, reaching record levels in Colombia, Peru, Honduras, Chile, and Venezuela. For Latin America as a whole, open unemployment increased from an unweighted average of about 6 percent of the urban economically active population in 1974 to some 14 percent in 1984 (ECLAC 1986: 23). The significant growth of open unemployment at the height of the crisis provided evidence against dualistic theories of Latin American labor markets which assumed that the informal sector would function as a compensating mechanism to absorb surplus labor. Its failure to do so supported instead the argument advanced originally by Roberts (1976, 1978) and expanded in Portes and Walton (1981) concerning the profound articulation of formal and informal activities as part of unified urban economies. Informal producers and vendors--defined as those who operate outside the pale of state regulation--do not live in a world apart from the firms which are so regulated. Instead the two types of enterprises are closely intertwined through a variety of arrangements, providing goods and markets for each other (Bener!a 1989; Fortuna and Prates 1989). For informal entrepreneurs, in particular, the larger firms are a key source of demand both directly, through subcontracting agreements, and indirectly, through the buying power of formal workers' wages. When formal firms cease to exist as during the economic crises of the 1980s, demand for informal goods and services also drops. Although more people are available to engage in informal activities, their remunerations rapidly approach zero as labor supply vastly outstrips demand. The outcome is the rise in open unemployment observed in a number of countries during these years. The argument can be formalized in the following third proposition: III. The formal and informal sectors are integral parts of the same urban economies. Hence informal employment functions only imperfectly as a countercyclical mechanism. During severe recessions, open unemployment will rise reflecting the limited absorptive capacity of irregular activities. 2. Urbanization in the Caribbean Basin The purpose of this study is to examine the validity of these hypotheses in a sub-regional context different from that which gave rise to them. For this purpose, we selected the countries of the Caribbean Basin--defined as the nations of the Central American isthmus and the larger island-nations of the Caribbean Sea. In the past, theories of Latin American urbanization based on the experience of the larger countries have been applied by extension to smaller ones, such as those of the Caribbean Basin. Our study focuses explicitly on urban trends in these small countries and thus excludes the larger coastal nations of Colombia, Mexico, and Venezuela. As we saw already, Colombia was one of the countries included in our original research. We also excluded from consideration the English, French, and Dutch-speaking islands of the minor Antilles. In these very small nation-states where the urban population is concentrated in a single city, it hardly makes sense of speaking of urban primacy or spatial polarization. From the remaining universe, Cuba and Puerto Rico were also excluded because of their special political circumstances. The countries finally selected for study are Costa Rica, the Dominican Republic, Guatemala, Haiti, and Jamaica. Although not encompassing the entire sub- regional universe, these countries carry great weight in its total population and are illustrative of the great diversity of historical experiences found in the area. The comparative logic underlying the selection of these countries follows Przeworski and Teune's (1970) method of maximum differences according to which similar processes observed in dissimilar settings are indicative of a more general trend than those registered in similar situations. For example, if the same trend in urban primacy is observed in geographically and culturally diverse countries, this carries greater weight for the empirical evaluation of Hypothesis I than similar results observed only in culturally close neighboring nations. In addition to regional coverage and significance, the selected countries represent opposite extremes of the region's historical colonization patterns, ethnic composition, economic development, and political stability. It is beyond the scope of our article to narrate their very diverse histories, but a word may be said about their present differences. These are summarized in Table 1. The five countries include the richest and most politically stable nation among those eligible (Costa Rica) and the poorest and most unstable (Haiti). Politically, there are two stable democracies, although with different political regimes (Costa Rica and Jamaica). The other countries range from a strong presidentialist incipient democracy (Dominican Republic) to fragile political systems with repeated military interventions (Guatemala and Haiti). In terms of size, they include the largest Central American country (Guatemala) and, with the exception of Cuba, the largest island country (Dominican Republic). Capital cities include the largest urban agglomeration in the Caribbean Basin, again with the exception of Havana. ------------------------------------------------------ Table 1 about here ------------------------------------------------------ The urbanized population ranges from less than a third of the total in Haiti to over half in Costa Rica and Jamaica. Given other economic and political differences, it is possible to expect that urbanization patterns would also differ across the five national settings. What these differences are and how they bear on the above theoretical propositions is the purpose of our study. Results presented in the following sections are based on detailed reports prepared by research teams in each country. The collaborative character of the study called for the same three broad areas to be covered by each report so that systematic comparisons could be conducted and inferences drawn. These sources are cited extensively in the next sections; they are complemented by additional secondary data to provide the most up-to- date portraits of urban systems in the region. 3. Urban Primacy The evidence from the five countries on the first hypothesis is mixed. There has been a deceleration of primacy in some nations, but not in others. The observed intra-regional differences tend to agree, however, with the logic of the hypothesis. The shift toward EOD is clear and is exemplified by the rapid growth of export production zones (EPZs) throughout the Caribbean Basin. Table 2 presents illustrative data for four of the five countries studied. The consistent growth of EPZs and export assembly plants documented by these figures is a direct consequence of the search for new sources of foreign exchange, coupled with the favorable tariff regime created by the new Caribbean Basin initiative. ------------------------------------------------------ Table 2 about here ------------------------------------------------------ Without doubt, the principal stimulus for foreign industries to transfer operations to the Caribbean has been low labor costs. In 1988, the hourly manufacturing wage was estimated at U.S.$ 1.00 in Costa Rica, $0.61-0.88 in Guatemala, $0.44-0.88 in the Dominican Republic, and $0.36 in Jamaica. Costs of labor and overhead associated with the assembly of women's garments were estimated at $4.75 per unit in the United States, $2.20 in Hong Kong, and $1.66 in the Caribbean (Schoepfle and Perez-Lopez 1989: 135-136). Guatemala, the single country excluded from the table because of the absence of EPZs has recently established a new large zone in Puerto Barrios and has begun fostering the growth of export garment assemblies on the basis of extremely low labor costs (Prez-S inz 1992). Yet the effects of export platform industrialization on Caribbean primate cities are not even because they have depended on three additional factors: 1) the physical location of the export zones; 2) their relative viability; 3) the growth of other foreign-oriented sectors, especially export agriculture and tourism. These factors determine variations of urban development around a common pattern dominated by the tendency to re-channel internal migration toward new areas of tourism and export manufacturing. Of the five countries under study, Jamaica is the one that has experienced most clearly a reduction in primacy. As shown in Table 3, the primacy index declined from 7.2 in 1960 to only 2.2 in 1990. This result was associated with the expansion of the tourist industry in the northern coast of the island, the revival of bauxite production in the interior, and the growth of satellite cities in the vicinity of the Kingston metropolitan area (KMA). The latter process is exemplified by the rapid population increase of Spanish Town to the northwest of the capital. Under the effect of expanded tourism, cities like Montego Bay and Ocho Rios also grew at a very fast clip during the last decade. With the arrival of Edward Seaga's Jamaica Labor Party to power in 1980, the island became transformed into an export-oriented open economy, a process completed under Seaga's successor Michael Manley. This led to growing capital investments both in tourism and export platform industrialization (Gordon and Dixon 1991). The decentralizing potential of EPZ industries is partially neutralized because the largest export production zone is located in Kingston itself. However, a second EPZ is located in Montego Bay and, along with the phenomenal growth of tourism in this and other northern cities, has led to significant reductions in the former hegemony of Kingston. ------------------------------------------------------ Table 3 about here ------------------------------------------------------ The Dominican Republic has experienced a similar rapid expansion of both tourism and export industries. By 1985, receipts from tourism had surpassed the sum of the three main traditional agricultural exports, that is sugar, coffee, and tobacco. About the same time, receipts from the export zones also started to climb rapidly. In the Dominican case, the centrifugal potential of export oriented industrialization was not partially neutralized as in Jamaica because most of the EPZs are located outside the capital city of Santo Domingo. Unfortunately, the last available census figures are for 1981, that is before the massive growth of export manufacturing took place and, hence, can not be used to assess the demographic effect of the process. Nonetheless, it should be noted that the fastest urban growth during the last intercensal period (1970-81) took place in La Romana, precisely the city where the first EPZ was established and which was also the recipient of significant tourism investment. The rate of growth of Santo Domingo declined somewhat during the same period falling behind that of La Romana's. By 1981, however, the capital's primacy remained unchanged, absorbing 23 percent of the total population and 45 percent of the urban population (Lozano and Duarte 1991). Although no census data are available at this time, there are clear indications that the growth rate of several Dominican secondary cities during the 1980s has been explosive, driven by tourism and the continuing expansion of export manufacturing. Tourist development has centered in the north coast city of Puerto Plata, in a pattern very similar to Jamaica's Montego Bay, and in all-inclusive resorts in La Romana. As seen in Table 2, export-oriented industries expanded five-fold between 1973 and 1990 and their labor force grew by 1500 percent. This rapid expansion has continued unabated during the early nineties. By 1992, for example, employment in export manufacturing plants was estimated at 134,100, a 21 percent increase in relation to just two years earlier. With the exception of the San Cristobal export zone, located near the western fringe of the capital, most export-manufacturing and ancillary economic activities are located in secondary cities such as Santiago, La Romana, and San Pedro de Macor!s (Guarnizo 1992: Ch. 2). These cities have experienced a rapid expansion of their urbanized perimeters as well as notable densification. A recent study of Santiago, the country's second largest city concludes: The city has grown in a chaotic way to the east in massive settlements in the direction of Puerto Plata; to the south, advancing rapidly toward rural zones, and to the southeast toward the mountains (Santana 1992:44). This and similar reports indicate that, despite the dearth of official data, urbanization patterns in the Dominican Republic are likely to follow those in Jamaica, with an upsurge of smaller urban centers and a relative deceleration of the capital's primacy. Costa Rica is another country that has been making significant investments in tourism and export platform industrialization. Unlike the Dominican Republic, however, neither sector has yet surpassed the dominance of traditional export agriculture. In addition, export assembly remains a small percentage of total Costa Rican manufacturing, still dominated by import-substitution industries. As in other countries, these industries cluster in the largest city, in this case the metropolitan area of San Jos (AMSJ). Despite these trends, urban primacy which had been increasing steadily up to 1980 declined in the following decade. As shown in Table 4, this decline was accompanied by a halving of the rate of growth of San Jos along with very rapid increases in the population of smaller cities. Among the latter are the two port cities of Punta Arenas and Limon suggesting that new export-oriented investments in these areas are beginning to have significant demographic consequences (Lungo, Perez, and Piedra 1991). In the Costa Rican case, however, there is a powerful countertrend to a more balanced urban system because the largest sites of export assembly manufacturing are located in the country's central valley, in close proximity to the AMSJ. Despite government efforts to locate the EPZs in coastal cities, most export industry converged in the central valley which also concentrates a large portion of the tourist infrastructure. As seen in Table 3, Cartago, a central valley city, grew very rapidly during the last decade and so did Alajuela and Heredia, all urban areas in close vicinity to San Jos. Their relative locations are portrayed in Figure 1. Combined with the continuous outward expansion of the capital, the growth of these satellite cities threatens to arrest the reversal of primacy, recreating it on a larger scale. The contours of a new "mega-city" comprising the 31 central valley cantones or municipalities and concentrating the majority of the national population have begun to emerge (Lungo, Perez, and Piedra 1991). Hence, despite the recent weakening of San Jos's primacy and the rapid growth of Costa Rican coastal towns, there is a clear danger that the decentralizing potential of the new export-oriented industries may be lost as they reinforce rather than weaken the expansionary tendency of the capital. ------------------------------------------------------ Figure 1 about here ------------------------------------------------------ Guatemala has lagged behind most of its neighbors both in the establishment of export assembly industries and in the development of tourist infrastructure. Prolonged political instability and generalized violence have conspired against successful investments in either sector. In this context, the country's entry into the new export-oriented model of development has depended so far on the expansion of its agricultural exports. As noted by Amaro (1990: 13-29), coffee exports grew by almost 800 percent between 1950 and 1981 and cotton also experienced a boom beginning in the mid-century. Guatemalan urbanization patterns have reflected, with some lag, these tendencies in export agriculture. The urban system is one of the most primate in Latin America due to the weakness of secondary cities. In 1980, the population of Guatemala City was more than 7 times the combined total of the next three cities. Nevertheless, figures in Table 4 show that the second city, Quezaltenango, grew at twice the rate of the capital during the last intercensal period, corresponding to its role as the principal coffee center. The two "cotton" cities, Escuintla and Retalhuleu also grew at a fast clip during the period of this crop's expansion and leveled afterwards. The banana center, Puerto Barrios, stagnated during the last two decades following the collapse of this sector. Congruent with the absence of any significant economic innovation, at least until the mid-eighties, Guatemala's urban system has not experienced any significant transformation. Primacy remained unchanged and the overall rate of urbanization declined during the last intercensal period (Table 4). The only signs of dynamism were the relative fast growth of Quezaltenango and the rapid increase of the population in municipalities adjacent to the capital city. The suburban towns of Mixco and Villa Nueva grew from a combined total of 15,000 in 1964 to 186,000 in 1981. These municipalities by themselves are larger today than the sum of the next three cities outside of the Guatemala Metropolitan Area (AMG). Without them, the AMG's primacy would have declined from an index of 7.6 to 6.1 (Prez-S inz 1991: 23). These results are congruent with the suburbanization of the metropolitan population and growth of satellite towns observed in other countries. The question for the future is whether the establishment of new EPZs and investments in tourism in the smaller cities would introduce a second dynamic, so far absent in Guatemalan urban development. ------------------------------------------------------ Table 4 about here ------------------------------------------------------ The dynamic role of export-oriented manufacturing is much more visible in Haiti which was one of the first countries to take advantage of the Caribbean Basin Initiative (Schoepfle and Perez-Lopez 1989). However, the effects of the new industries have been the opposite of those anticipated by the hypothesis of decelerating primacy. Political instability, lack of a suitable infrastructure, and fear of AIDS have all but dismantled Haiti's tourist sector. Goods assembly, the country's main export earner, is concentrated in the late 1980s in a single EPZ located right by Port-au-Prince's airport. This location accelerated rural migration to the capital, already stimulated by land scarcity and soil erosion (Manigat 1991; Miller 1984). The result is that Haiti continued to experience sustained increases in primacy along the pattern typical of earlier Latin American urban development. Port-au-Prince's annual rate of growth during the last two decades, 8.8 percent (Table 3), is the highest of all the cities studied. The evidence in this section indicates that urban primacy is not declining everywhere, but that the underlying forces identified as responsible for its decline in the larger Latin American countries also operate in those of the Caribbean Basin. Effects of the new export-oriented model of development on the urban system are not uniform because they depend on the location of the new industries and their capacity for employment creation. When sizable tourist and export manufacturing projects are located away from the primate city, the urban system responds along the lines predicted by the hypothesis (Jamaica and the Dominican Republic); when these same sectors are located in or near the capital city, primacy is exacerbated (Haiti); in situations where export-oriented development is in its early stages, the urban system remains unaltered (Guatemala). The analysis also identified a second dynamic concerning the rapid growth of satellite towns and suburbs, a tendency which runs contrary to the decentralizing potential of export growth and which may negate its effect by giving rise to future larger cities. Costa Rica is the principal case in point, as the rapid expansion of the capital is linking fast with satellite towns where export assembly plants have clustered. The two forces affecting the evolution of the urban system reinforce each other in this instance, leading to the possible emergence of a new megalopolis in the country's central region. 4. Spatial Polarization Our analysis of patterns of spatial distribution in the capital cities of the five countries studies does not support the hypothesis of a uniform reversal of class polarization. However, the changes observed accord with the underlying rationale of the hypothesis. These cities were generally less polarized than their larger South American counterparts at the onset of the economic crisis, in part because local elites were not sufficiently large to occupy vast expanses of territory. Instead, they created protected enclaves in an urban landscape dominated by low-income neighborhoods and squatter settlements. This portrayal varied, of course, with the level of economic development, the topographical characteristics of each city, and, above all, the policies of the national government. The most polarized of our five cities is Kingston whose social configuration resembles an inverted ice- cream cone, with shantytown and working-class neighborhoods at the base and elite settlements in the upper reaches of the Liguanea plain where the city is located. This pattern was already observable in Colin Clarke's (1975) study of the city in the 1960s and remained essentially unchanged during the following two decades. Debates among Jamaican urban specialists during the seventies centered on the evolution of Kingston's "transition zone" of middle housing between the low lying shantytowns and the elite high grounds (Norton 1978; Knight and Davies 1978). A second debate was on the extent to which an "inner ring" of popular settlements close to the wealthy foothills subverted Clarke's portrayal of spatial polarization. As shown in Figure 2, the presence of these shantytown areas in the mid-seventies did not really alter the overall tendency toward spatial segregation. The latter was strongly correlated with different population densities. Norton (1978: 100) noted that 41 percent of Kingston's residential area in the early seventies was occupied by 6 percent of the population living at densities of 0.1 persons per room while 75 percent of the population occupied only 33 percent of the residential area at average densities of 2 persons per room. Class differences were also superimposed on ethnic characteristics. As Gordon and Dixon (1991) observed, the white/mulatto/black composition of Kingston's population overlapped significantly with the elite/middle-class/poor residential locations in urban space. ------------------------------------------------------ Figure 2 about here ------------------------------------------------------ Since the economic crisis of the mid-seventies, there has been two novel tendencies which correspond fairly well to those observed in the larger South American cities. First, the consolidation of new elite enclaves around the shopping and businesses districts of New Kingston and Constant Spring Road to the north has been partially countermanded by the rapid growth of "inner-ring" irregular settlements around the same areas. A Statistical Institute of Jamaica report summarizes the resulting configuration as: ... an erratic arrangement of residences; close juxtaposition of residences containing opposite socioeconomic groups and the dispersed clusters of low-income residences throughout the higher income areas; an increased tendency towards peripheral location of lower income residences as they rival higher income groups in search of accommodation (cited in Gordon and Dixon, 1991: 33). The creation of these "suburbs of the poor" (Norton 1978) has been followed by the still more notable displacement of middle-income and impoverished groups out of Kingston proper and into the adjacent St. Catherine Plain. The movement was accommodated in planned housing estates such as Portmore, Enson City, and others in the road to Spanish Town as well as in new large shantytowns. Low-lying Portmore, across the Bay from Kingston grew from a community of 5,000 in 1970 to 73,400 in 1982, equivalent to an annual growth rate of 25 percent. As shown in Table 5, this new frontier of suburban projects and squatter settlements led to the absolute decline of the population of Kingston proper, already observable in 1982 along with the growth of St. Catherine. The spread of population across Kingston Bay was the outcome of two sets of forces: First, efforts by the poor to escape the growing violence in central Kingston, caught in the throes of factional warfare among rival political parties. Second, the closure of opportunities for affordable middle-income housing in the established northern suburbs. While the first phenomenon is unique to Jamaica, the second is quite similar to that observed in Bogot , S~o Paulo, and other South American cities. In both instances, the impact of the economic crisis forced large sectors of the urban middle class to seek housing solutions in areas regarded previously as physically or socially unacceptable. The difference is that while in South American capitals the physical displacement of middle-income groups took place in the direction of established working-class areas, in Kingston both the middle classes and the poor moved together toward previously unsettled land. Along with the expansion of the "suburbs of the poor" toward the north, this movement led to greater intermingling in urban space and hence, to a partial reversal of the pattern of class polarization. ------------------------------------------------------ Table 5 about here ------------------------------------------------------ A similar trend is observable in Santo Domingo, although with other variants. During the period of import substitution industrialization (ISI), the Dominican capital grew rapidly giving rise to a new industrial zone and working-class settlements toward the north and east of the Ozama River and the displacement of elite sectors to the west. Sustained economic growth under ISI during the 1960s and early 1970s generated new wealthy groups and an urban middle class capable of fueling demand for luxury housing (Guarnizo 1992: Ch. 2). This demand led to a rapid inflation of land prices in the northwest quadrant of the city and the emergence of exclusive residential neighborhoods like Naco, Los Jardines, and Arroyo Hondo. South of them, toward the Caribbean Sea, grew more affordable middle-class housing developments such as Miramar and Mirador Norte (Lozano and Duarte 1991). The same period witnessed the growth of a vast conglomeration of slum dwellings and squatter settlements north of downtown and east of the Ozama River around the area of Los Mina. The apparent polarization of urban space into a western frontier of elite and middle income developments and a northeast zone of marginalized population was interrupted by the fateful decision of the national government to establish a new industrial zone in the area of Herrera, in the western fringe of the city. This industrial zone and the working class neighborhoods that grew around it effectively "bracketed" the upper income residential developments, converting them into an elite enclave in the midst of a poor city and limiting their possibilities of expansion. This situation was going to have important consequences for urban development in the years that followed. As in Kingston, the economic crisis in Santo Domingo saw the displacement of poor population toward high income areas in search of employment and better housing. The movement took the form of increasing occupation of publicly-owned land in the interstices of middle- and upper-income developments as well as the rapid growth of the western fringe around the Herrera industrial zone (Lozano and Duarte 1991). The process reduced spatial polarization and, more significantly, further encapsulated upper-income neighborhoods between two vast rings of impoverished population. The wealthier groups responded by seeking even more remote and exclusive locations in the northwest area, in development like Arroyo Manzano and Altos de Arroyo Hondo. The emergence of these expensive suburban divisions represents a clear attempt to maintain social distance from the poor, but even here physical space has been increasingly contested by squatter settlements moving north from the settled working-class areas. These trends are illustrated in Figure 3 that portrays the relative location of socio-economic strata in Santo Domingo according to quality of dwellings and relative densities in 1990. The map indicates two noteworthy features: First, the spread of irregular squatter settlements throughout the entire urban area. Although concentrated north and east of downtown, squatter settlements are also found close to the rich suburban developments in the city's northwest. Second, the existence of a sizable middle-income area east of the Ozama River and close to the largest poor settlements. The rapid growth of this area in neighborhoods such as Los Trinitarios and Villa Faro is a recent phenomenon that corresponds well to the trend already observed in other Latin American cities. In the Dominican case, this trend was prompted by the closure of the western fringe to middle-class expansion and by the occupation of the adjacent northwest by the wealthy. Predictably, urban densities and rents increased in the old established middle-class areas. In response, some middle-class groups breached the symbolic divide separating them from working-class areas in search of affordable housing. In Santo Domingo, this decision became known as "crossing the bridge" (across the Ozama) toward the new residential developments in the east. This process is essentially the same as that observed in Bogot and, other South American cities. Along with the displacement of the urban poor toward the west, it has produced a visible reintegration of the city, partially reversing its past class polarization. ------------------------------------------------------ Figure 3 about here ------------------------------------------------------ Port-au-Prince offers the most dramatic example of transformation of the urban spatial order. In the Haitian capital, reversal of class polarization has been more marked than in Kingston or Santo Domingo and it has been entirely due to the migration of the rural poor. There is no observable movement of middle-class sectors in search of new housing, in part because the urban middle class is so small. Instead, there has been a thorough invasion of urban space by impoverished masses migrating from the countryside (Manigat 1991). The counterpart of Port-au-Prince's sharp increase in primacy, described above, is the emergence of vast "bidonvilles" such as Cite Soleil and the densification of formerly working-class areas to unbelievable levels (Duquella 1989). The invasion of the city by the rural poor has not been limited to the creation of irregular settlements but extends to the occupation of free space in residential districts. No upper-class area, no matter how remote or exclusive, has escaped this invasion. The occupation of urban space by the migrant poor is not limited to vacant land, but extends to the streets, several of which have been turned into informal markets blocking vehicular traffic. The fragile infrastructure of public services has been swamped by this rapidly growing population, making access to water, drainage, and electricity a privilege accessible only to a few. By 1988, 72 percent of the city's population lacked running water and 92 percent had only access to basic latrines (Manigat 1991). In this context, the gradual displacement of rural migrants toward formerly elite areas, such as Petionville, Laboule, and Tomassin has been prompted not only by the need for some form of employment, but also by the search for basic services. Deprived of legal access to these services, the poor have simply appropriated them through widespread "pirating" of water, electricity and, in some areas, even the TV cable. In 1988, 80 percent of dwellings in Port-au-Prince had access to electricity, but the majority acquired it through clandestine hookups. What has emerged then is a silent struggle between the classes for open space, basic services, and even the streets. Armed guards are commonly posted in front of the residences of the well-to-do seeking to prevent further encroachments by the poor. Port-au-Prince today is best described as a few islands of wealth surrounded by a sea of poverty. It represents an extreme case of "perverse integration" (to use Kowarick's term) prompted by the exacerbation of the underlying social processes. In Haiti the economic crisis of the late seventies and eighties took place in a context of declining productivity and employment in agriculture and where no other industry emerged to take its place. Neither the new EPZ close to the capital, nor the city itself could cope with the material needs of the vast new wave of rural migrants. As in other cities affected by the crisis, the poor gravitated toward those areas where some wealth existed but, in this case, such areas are much smaller and the needs and numbers of those seeking refuge in them much greater. The two Central American capitals represent exceptions to the patterns outlined above. Neither city has experienced significant reversals of spatial polarization, in part because neither was highly polarized to begin with. Reasons accounting for this lesser segregation are quite different, however. The Costa Rican capital is the most socially integrated of the five cities studied because higher levels of economic development were combined with state policies that sought to reduce economic and spatial inequalities. State intervention in San Jos's development led to a relatively homogeneous urban space, lacking either distinct elite enclaves or a shantytown belt. This outcome was the result of three types of state policies: a) subsidized provision of housing for lower-middle and low-income sectors; b) rapid response to land invasions through the relocation of invaders in state housing projects; c) dispersion of these projects throughout the metropolitan area. As Lungo, Perez, and Piedra (1991: 117) note: The state has been the central agent in the development of metropolitan San Jos. Its actions promoted the extension of the urban infrastructure of roads, water, and electricity and the construction of good quality housing for the lower middle-class ... These actions generated a more homogenous urban space where state projects were interspersed with those of the private sector preventing the radical spatial segregation found in other Latin American cities. The economic downturn of the early eighties, coincided with the administration of Rodrigo Carazo (1978-82) and led to a contraction of employment, wages, and the ability of the state to intervene in the housing market. Urban shantytowns and land invasions proliferated, concentrating in specific areas and threatening to initiate a process of spatial polarization. Both the economy and the state investment capacity recuperated, however, by the mid-eighties. During the administrations of Luis A. Monge (1892-86) and Oscar Arias (1986-90), efforts were made to reestablish spatial equilibrium. The government gave priority to the provision of housing solutions through new programs of subsidized credit for the middle-class, resettlement of land invaders in new projects, and prevention of further land invasions through an innovative sites-and-services program. By 1990, ... the tendency for a marked segregation of urban space that emerged in the early eighties disappeared because the revamped housing market offered solutions which, although small, were accessible to an impoverished middle class. (Lungo, Perez, and Piedra 1991: 126) While San Jos is primarily a "middle-class" city, Guatemala City is characterized by the opposite, namely the feebleness of its middle-income sectors. Guatemalan society is sharply divided between a wealthy elite and a mass of impoverished population, mostly Indians. For the country as a whole, 1.25 million families or 83.4 percent of the total were classified as living in poverty in 1986-87; of these, 949,000 or 64.5 percent lived in conditions of extreme poverty. In the capital city, the situation improves but still 56.3 percent are poor or very poor (Prez-S inz 1991: 53; Ruiz 1990). This class structure has produced an urban landscape dominated by poor working-class settlements and shantytowns, with a well-defined elite enclave in its midst. Guatemala City is administratively divided into some twenty-five "zones" that facilitate a more refined analysis of its spatial structure. In 1985, only 5 of the 25 zones that comprise the metropolitan area could be classified as middle-income or higher. The others had average family monthly incomes of 800 quetzales (US$ 120 in 1985) or less. One of the middle-class residential zones (#14) combined a population with average family incomes exceeding US$ 300 with a very poor 36 percent of families whose incomes did not reach US$ 45. More remarkable still, the middle- and upper-income enclaves were not in the suburbs, but in relative close proximity to the central city. Most of these areas are located south of downtown in low-density residential developments such as Santa Clara and Tivoli, established during the late 1940s (Prez-S inz 1991: 31). Figure 4 portrays the spatial distribution of socioeconomic strata in the Guatemala Metropolitan Area. Two facts are immediately apparent: First, shantytowns are frequently located in central areas, in close proximity to elite residential neighborhoods. Zones 10 and 14, for example, are actually shared by low-density residential housing and high-density working-class and squatter settlements. Second, this spatial mix has not prompted upper-income groups to flee the city; there is no "frontier" of suburban elite developments anywhere in the urban periphery. ------------------------------------------------------ Figure 4 about here ------------------------------------------------------ The generalized and increasing poverty in Guatemala City during the last decade and the close and visible presence of well-to-do neighborhoods should have prompted the most deprived sectors to invade elite areas in search of better employment opportunities. This phenomenon, observable in Kingston, Santo Domingo, Port-au-Prince, and larger South American cities did not materialize in Guatemala City as a direct consequence of the level of state-sponsored violence. An entrenched elite willing to employ any means in defense of its privileges presents a major challenge to any popular attempt at revindication. While working-class occupations of land near high-income neighborhoods is an everyday affair in other cities, it would be quite unthinkable in the elite urban zones of Guatemala in the face of violent opposition (Jonas 1991). This political situation helps explain as well the peculiar absence of residential segregation apparent in Figure 4. There is no spatial segregation of the classes in part because their social segregation is so vast as to render the proximity of the poor more a convenience than a threat to privileged groups. With the state means of violence at their disposal and a population intimidated by years of repression, there is little incentive for the well-to-do to escape the city. Instead, poor settlements function mostly as a convenient source of domestic labor and personal services of various sorts for elite areas (Roberts 1978). In synthesis, the five capital cities provide evidence in favor of the hypothesis of reversal of spatial polarization but also indicate significant variations and exceptions to the process. The comparison highlights the significance of the state in countermanding the tendency of urban elites to create exclusive residential areas and of impoverished groups to try to settle near them in search of better material opportunities. In the two Central American capitals selected for study, these tendencies are absent due in large part to state action: In one case, state policy attempted to meet those basic needs that lead the urban poor to invade elite suburbs; in the other, state repression foreclosed the possibility of such popular actions and, hence, removed one of the major incentives for elite flight from the city. The second trend leading to the reduction of spatial polarization, that is the movement of middle-class groups into formerly low-income areas, is observable in two of our five cities (Kingston and Santo Domingo). In the other three, either the middle class is too small (Port-au-Prince and Guatemala City) or it comprises the majority of the urban population and has been the recipient of state programs to meet its housing needs (San Jos). 5. The Informal Economy and Unemployment Our analysis of urban labor markets and of the role of the informal sector seeks to understand the extent to which irregular forms of employment played a countercyclical role at the height of the economic crisis. This analysis, guided by hypothesis III above, is limited to four countries excluding Haiti. This last country is exceptional in that formal sector employment barely exists. Furthermore, the dearth of reliable statistics makes it impossible to detect short-term changes in a labor market that is overwhelmingly informal. According to the International Labor Office, only 7.7 percent of the Haitian labor force in 1987 held jobs that could be considered in any sense regulated or formal, the bulk of the figure being formed by 32,000 state employees (Manigat 1991). Among the remaining countries, the best available information comes from Costa Rica. Time series data on the evolution of the labor market in metropolitan San Jos are presented in Table 6. During the crisis years of 1981-84, open unemployment rose sharply while informal employment expanded only moderately. According to the figures, unemployment in the metropolitan area of San Jos (AMSJ) rose 66 percent in 1980-81 and another 36 percent in 1981-82 coinciding with an aggregate decline of real GNP of 9.6 percent in this two-year period. Informal employment expanded more slowly, by 8.2 percent between 1980-81 and 2 percent in 1981-82. Unemployment remained high during the next three years and then declined rapidly with economic reactivation in the late eighties. By 1989, unemployment was approximately half of what it had been at the beginning of the decade. By contrast, informal employment remained steady at about a quarter of the labor force. These contrasting trends indicate that adjustment to the economic crisis of the early eighties occurred primarily through open unemployment rather than a massive expansion of the informal sector. ------------------------------------------------------ Table 6 about here ------------------------------------------------------ This is not the whole story, however, because additional evidence shows a parallel rise in unstable and lower-paid jobs in both sectors of the San Jos economy. Female labor market participation grew during the decade in connection with the proliferation of export assembly plants in the AMSJ. Maquilas as these plants are known locally, more than tripled their employment between 1984 and 1990, reaching 40,000 in the latter year. Most of these jobs are unstable and low-paid, employing mainly young women. Part-time employment also grew, to a high of 37.5 percent of the economically active population by the mid-eighties. While the minimum wage recuperated in 1989 to its full value at the beginning of the decade, earnings of skilled and white-collar workers remained below the 1980 figure (Lungo, Perez, and Piedra 1991). The informal sector registered a parallel evolution in its internal structure. The better remunerated and more stable category of informal entrepreneurs declined by 45 percent between 1980 and 1986, recuperating only during the last years of the decade. Less stable and lower paid jobs such as odd-jobbing and other forms of informal self-employment grew, however, throughout these years. As seen in Table 7, they rose 43 percent between 1980 and 1989 representing two-thirds of all urban informal employment in the latter year. These figures suggest a complex adjustment process of the Costa Rican urban labor market. Governmental intervention was instrumental in checking the rise of open urban unemployment during the 1980s and preventing the deterioration of the minimum wage. Simultaneously, however, there was a rise in lower paid, more unstable employment in both the formal and informal sectors. Part-time work and low paid "maquila" employment became more common among formal workers, while self-employment (much of it in marginal odd-jobbing) became predominant in the informal sector. This development is consonant with our pro-cyclical hypothesis since it shows that instead of counteracting trends in the formal economy, the informal sector followed its evolution. In contrast to San Jos, the labor market in Guatemala City has always been highly informal. Among the five cities studied, only Port-au-Prince has a larger proportion of the economically active population (EAP) in the informal sector. During the 1980s, however, figures for informal employment in Guatemala City barely budged. Estimates based on the measurement criteria employed by the ILO indicate that the informal sector comprised 30 percent of the urban EAP in 1980 and 33 percent in 1989. Alternative figures, based on the proportion of the EAP working without legally mandated coverage yield a much higher estimate for 1980, 54.3 percent, but coincide in showing little change during the decade. As in San Jos, adjustment to the crisis took place mainly through a significant rise in open unemployment. Two sets of estimates are presented in Table 7. They coincide in showing that urban unemployment more than quadrupled between 1980 and 1983 and then remained at a high level during the entire decade. ------------------------------------------------------ Table 7 about here ------------------------------------------------------ Unlike the Costa Rican case, no significant government intervention took place to bring down record unemployment, a pattern of official policy congruent with that observed for low-income housing in the preceding section. Hence, the presence of a large urban informal sector where average wages are lower than among protected workers did not successfully cushion the effects of the economic downturn. The quadrupling of open unemployment in Guatemala City during the 1980s is also congruent with our hypothesis concerning the limits of the informal sector as a countercyclical mechanism. As Table 7 shows, the internal structure of the informal sector in Guatemala City appears to have evolved in a direction contrary to that of San Jos. Comparing estimates available for 1980 and 1989, the relative representation of informal entrepreneurs seems to have increased while that of the self-employed declined. As in San Jos, entrepreneurs earned incomes significantly above the official minimum and their presence led to an improvement of average earnings in the informal sector. Yet, despite its apparent decline, self-employment still represented the majority of the Guatemalan informal labor force in 1989. The Caribbean island nations of Jamaica and the Dominican Republic experienced a similar evolution of their urban labor markets in the 1980s. Estimates for both countries indicate an expansion of informal employment, especially in street vending and odd-jobbing, along with a contraction of both formal employment and earnings. Despite the scarcity of data, figures for the Dominican Republic are consistent in showing a significant rise in the jobless population, a sharp decline in formal sector wages, and an increase in self-employment. As shown in Table 8, the proportion of the population entering the labor force rose as real industrial and public sector wages declined steeply. Most of the increase in participation concentrated in the cities, as indicated by the rising representation of the urban labor force in the national EAP. Self-employment increased during the last half of the 1980s, but the proliferation of these informal activities did not prevent a sharp rise in open unemployment which doubled between 1977 and 1991. It appears that efforts of Dominican households to compensate for declining wages with an increased labor offer met with a shortage of remunerated employment in both the formal and informal sectors. While street vending and similar activities rose, they could not absorb a rapidly increasing labor supply leading to record levels of open unemployment. ------------------------------------------------------ Table 8 about here ------------------------------------------------------ In parallel fashion, informal employment and unemployment increased simultaneously in the Kingston Metropolitan Area during the economic crisis. The unemployment rate reached record levels between 1983 and 1985 and declined afterwards. This apparent improvement may mask, however, a rise in the number of discouraged workers. As shown in Table 9, labor force participation actually dropped, suggesting a high quit rate among prospective workers. As in San Jos, the internal structure of the informal sector evolved in the direction of an apparent decline of small entrepreneurs and a rise in informal self-employment. Informal street vendors increased their participation in the Kingston labor force among both males and females, while entrepreneurial activities in small-scale services and agriculture declined. The overall trend was toward a simultaneous deterioration of employment opportunities and earnings in both segments of the Kingston labor market. ------------------------------------------------------ Table 9 about here ------------------------------------------------------ Despite the scarcity and imperfections of official data, the figures indicate that urban labor markets in the smaller countries of the Caribbean Basin adjusted to the economic crisis in a manner similar to their larger South American counterparts. In both instances, adjustment involved a combination of declining wages, declining formal employment, growing informality, and record levels of unemployment. The growth of unemployment in all countries, clearly points to the limitations of the informal sector as a labor absorptive mechanism. Contrary to the dualistic conceptualization of Latin American labor markets proposed by analysts of ILO and its Latin American affiliate PREALC, there is no hydraulic relationship between formal and informal sectors in which excess supply in one is automatically absorbed by the other. The overall pattern of results seems more congruent with the alternative view of unified urban economies in which both types of activities coexist. In accordance with hypothesis III, open unemployment rises as a logical consequence of economic contraction which reduces employment opportunities and earnings in both formal and informal enterprise. 6. Summary and Conclusions Overall, these results support the conclusion of our earlier study that something significant has changed in Latin American urbanization during the last two decades. Though the three dimensions of urbanization examined do not exhaust by any means all that there is to the topic, they capture important aspects highlighted by a voluminous research literature. The new Caribbean Basin findings presented in this article do not support such simplified conclusions as "primacy is declining everywhere" or "class polarization is everywhere diminishing." Indeed, the evidence from one or more countries runs contrary to such assertions. Instead, what the pattern of results show is support for the theoretical logic beneath each hypothesis: Primacy decelerates or not depending on whether the decentralizing potential of the new outward-oriented model of development is actualized; spatial polarization declines or not depending on whether middle-class and poor sectors are able to implement novel strategies to cope with economic emergencies; the informal sector can absorb more or less labor depending on the state of the economy and the success of governmental efforts to re- activate it. The contrasting empirical trends observed in the five countries, added to those included in our earlier South American research point toward a second important conclusion. This is the deeply flawed earlier portrayals of the urban "explosion" in Latin America and the uniform characteristics of peripheral urbanization. Such generalizations may have played a useful role in an earlier stage of conceptual development, but are out of step with present findings. Contemporary Latin American urbanization cannot be understood on the basis of blanket notions about peripheral societies, whether they stem from orthodox neo-classical or unorthodox world-system and dependency theories. Instead, the weight of the evidence points to an alternative theoretical approach which combines global trends with specific national realities. There are a number of similarities among Latin American countries which preclude the opposite extreme of asserting that only idiosyncratic national traits count. A more useful level of analysis focuses on the changing insertion of these countries in the global economy along with characteristics specific to each, in particular their prior level of development and the character of their state systems. In an attempt to reassert the primacy of the state in the analysis of national development, Evans (1989) has proposed a typology of state systems, situated in a continuum from "predatory" to "developmental." Much of what he has to say about the quality and effectiveness of state apparatuses bears directly on these results. All of the countries studied confronted a similarly adverse external environment during the early 1980s and all moved in the direction of promoting exports in order to reactivate their economies. However, in some instances, record levels of unemployment lasted throughout the decade (Guatemala), while in others they were promptly brought down by effective state intervention (Costa Rica). Similarly, the decentralizing potential of the new export industries was fully actualized in some cases (Dominican Republic), while in others it either did not exist (Guatemala) or it aggravated the preexisting urban concentration (Haiti). The effects of the crisis were also felt differently within each capital city. In some instances, the absence or ineffectiveness of state programs led to the exacerbation of the "perverse integration" observed in larger South American cities (Kingston, Port-au-Prince). In others, spatial polarization was kept at bay by strong state intervention. However, the character of the latter ranged all the way from effective middle-class and low- income housing programs that reduced the impact of economic inequality (San Jos) to vigorous state repression that prevented popular sectors from implementing their own unorthodox solutions (Guatemala City). These conclusions converge and reinforce similar trends in the analysis of other aspects of national development. They support an emerging consensus that the proper analytic focus is not at the level of sweeping theories of global trends and cycles nor at that of unique national or local traits, but in the interaction between them. The empirical results in these pages illustrate the advisability of this middle-range perspective. Endnotes References Amaro, Nelson. 1990. Descentralizaci"n y Participaci"n en Guatemala. Ciudad de Guatemala: ICEP. Amato, Peter. 1969. An Analysis of the Changing Patterns of Elite Residential Areas in Bogot , Colombia. Latin American dissertation series. Ithaca, N.Y.: Cornell University. Anderson, Patricia. 1987. Informal Sector or Secondary Labour Market? Towards a Synthesis. Social and Economic Studies 36:3. Bairoch, Paul. 1973. Urban Unemployment in Developing Countries: The Nature of the Problem and Proposals for its Solution. Geneva: International Labor Office. Bener!a, Lourdes. 1989. Subcontracting and Employment Dynamics in Mexico City. In The Informal Economy: Studies in Advanced and Less Developed Countries, ed. A. Portes, M. Castells and L.A. Benton, 173- 188. Baltimore, Md.: The Johns Hopkins University Press. Bener!a, Lourdes, and Marta I. Roldan. 1987. The Crossroads of Class and Gender: Homework, Subcontracting, and Household Dynamics in Mexico City. Chicago: University of Chicago Press. Beyer, Glenn H. 1967. The Urban Explosion in Latin America. Ithaca, N.Y.: Cornell University. Breese, Gerald. 1966. Urbanization in Newly Developing Countries. New York: Prentice-Hall. Cartier, William. 1988. Urban Processes and Economic Recession: Bogot in the 1980s. Paper presented at the seminar on Latin American Urbanization During the Crisis. Center for Latin American and Caribbean Studies, Florida International University. Clarke, Colin. 1975. Urban Development and Social Change 1692-1961. Berkeley: University of California Press. Consejo Nacional de Zonas Francas de Exportaci"n. 1992. Data gathered and compiled by Jos Itzigsohn. Santo Domingo, R.D. Cornelius, Wayne. 1975. Politics and the Migrant Poor in Mexico City. Stanford: Stanford University Press. Corporaci"n de la Zona Franca de Exportaci"n. 1991. Data gathered and compiled by Jos Itzigsohn. San Jos, Costa Rica. Direcci"n General de Estadistica y Censo de Costa Rica. 1991. Encuesta nacional de hogares. Data compiled by Jos Itzigsohn. San Jos, Costa Rica. Dominican Documentation Center. 1991. Series estad!sticas sobre empleo. Compiled by Edwin Croes. Duquella, A. 1989. La population et les besions de logement au Haiti. Mimeograph. Port-au-Prince. Eckstein, Susan. 1977. The Poverty of Revolution, the State and the Urban Poor in Mexico. Princeton: Princeton University Press. Economic Commission for Latin America and the Caribbean (ECLAC). 1986. 1986 Statistical Yearbook of Latin America and the Caribbean. New York: United Nations. ------. 1988. Preliminary Overview of the Latin American Economy, 1987. Notes on Economy and Development 455-456(December). ------. 1992. 1992 Statistical Yearbook of Latin America and the Caribbean. New York: United Nations. Encyclopaedia Britannica. 1991. Book of the Year. Britannica World Data. Nations of the World. Evans, Peter B. 1989. Predatory, Developmental, and Other Apparatuses: A Comparative Political Economy Perspective on the Third World State. Sociological Forum 4(December):561-587. Fortuna, Juan Carlos, and Suzanna Prates. 1989. Informal Sector versus Informalized Labor Relations in Uruguay. In The Informal Economy: Studies in Advanced and Less Developed Countries, ed. A. Portes, M. Castells and L.A. Benton, 78-94. Baltimore, Md.: The Johns Hopkins University Press. Garc!a, Norberto E. 1982. Growing Labor Absorption with Persistent Unemployment. CEPAL Review 18:45-64. Goldrich, Daniel. 1970. Political Organization and the Politicization of the Poblador. Comparative Political Studies 3(July):176-202. Gordon, Derek, and Cheryl Dixon. 1991. Urbanization in Kingston, Jamaica; Years of Growth and Years of Crisis. Paper presented at the seminar on Urbanization in the Caribbean in the Years of the Crisis. Florida International University, Miami, May 29-June 1. Guarnizo, Luis E. 1992. One Country in Two: Dominican- Owned Firms in New York and the Dominican Republic. Ph.D. dissertation. Department of Sociology, The Johns Hopkins University. Hardoy, Jorge E. 1975. Urbanization in Latin America. Garden City: Anchor Books. Hardoy, Jorge, Raul Basald#a, and Oscar Moreno. 1968. Pol!tica de la tierra urbana y mecanismos para su regulaci"n en America Latina. Buenos Aires: Editorial del Instituto. Iglesias, Enrique. 1985. The Latin American Economy during 1984: A Preliminary Overview. CEPAL Review 35(April):7-44. Instituto de Estudios Dominicanos. 1992. Data gathered and compiled by Carlos Dore-Cabral. Santo Domingo, R.D. Inter-American Development Bank. 1990. Economic and Social Progress in Latin America, 1990 Report. Baltimore: The Johns Hopkins University Press. Jamaica Statistical Institute. 1973. Jamaica Population Census 1970. Kingston. ------. 1982. Jamaica Population Census 1982. Kingston. Jonas, Susanne. 1991. The Battle for Guatemala: Rebels, Death Squads, and U.S. Power. Boulder: Westview Press. Knight, Pauline, and Omar Davies. 1978. An Analysis of Residential Location Patterns in the Kingston Metropolitan Area. Social and Economic Studies 27:4. Kowarick, Lucio, Ana Mar!a Gambier Campos, and Maria Cecilia de Mello. 1990. Los Percursos da Desigualdade. In S~o Paulo, crise e mudanca, ed. R. Rolnick, L. Kowarick and N. Somekn. S~o Paulo: Brasiliense. Lagos, Ricardo, and Victor Tokman. 1983. Monetarismo global, empleo y estratificaci"n social. El Trimestre Economico 50(July-Sept.):1437-1474. Leeds, Anthony. 1969. The Significant Variables Determining the Character of Squatter Settlements. America Latina 12(July-Sept.):44-86. Lozano, Wilfredo, and Isis Duarte. 1991. Proceso de urbanizaci"n, modelos de desarrollo y clases sociales en Rep#blica Dominicana. Paper presented at the seminar on Urbanization in the Caribbean in the Years of the Crisis. Florida International University, Miami, May 29-June 1. Lungo Ucls, Mario, Marian Perez, and Nancy Piedra. 1991. La urbanizaci"n en Costa Rica en los '80: el caso del area metropolitana de San Jos. Paper presented at the seminar on Urbanization in the Caribbean in the Years of the Crisis. Florida International University, Miami, May 29-June 1. Manigat, Sabine. 1991. L'urbanisation de Port-au-Prince durant les annees de crise. Paper presented at the seminar on Urbanization in the Caribbean in the Years of the Crisis. Florida International University, Miami, May 29-June 1. Marshall, Adriana. 1987. Non-Standard Employment Practices in Latin America. Discussion Paper #6, Labour Market Programme. Geneva: International Labour Office. Massad, Carlos. 1986. Alleviation of the Debt Burden: Historical Experience and Present Need. CEPAL Review 30:17-34. Mesa-Lago, Carmelo. 1991. Social Security and Prospects for Equality in Latin America. Discussion Papers no. 140. Washington, D.C.: World Bank. Miller, Jake C. 1984. The Plight of Haitian Refugees. New York: Praeger. National Statistical Office. 1987. Rep#blica Dominicana en Cifras. Santo Domingo: ONE. ------. 1990. Rep#blica Dominicana en Cifras. Santo Domingo: ONE. Norton, Ann. 1978. Shanties and Skyscrapers: Growth and Structure of Modern Kingston. Working Paper #37. Kingston: Institute of Social and Economic Research. Nun, Jos. 1969. Superpoblaci"n relativa, ejercito industrial de reserva y masa marginal. Revista Latinoamericana de Sociolog!a 5(July):178-235. Prez-S inz, Juan Pablo. 1991. Ciudad de Guatemala en la decada de los ochenta: crisis y urbanizaci"n. Paper presented at the seminar on Urbanization in the Caribbean in the Years of the Crisis. Florida International University, Miami, May 29-June 1. ------. 1992. Informalidad urbana en America Latina: enfoques, problem ticas e interrogantes. Caracas: Editorial Nueva Sociedad. Portes, Alejandro. 1989. Latin American Urbanization during the Years of the Crisis. Latin American Research Review 23:7-44. Portes, Alejandro, and Michael Johns. 1989. Class Structure and Spatial Polarization: An Assessment of Recent Urban Trends in Latin America. In Lost Promises: Debt, Austerity, and Development in Latin America, ed. W.L. Canak, 111-137. Boulder, Colo.: Westview Press. Portes, Alejandro, and Richard Schauffler. 1993. The Informal Economy in Latin America: Definition, Measurement and Policies. Population and Development Review 19 #1 (March). Portes, Alejandro, and John Walton. 1976. Urban Latin America, the Political Condition from Above and Below. Austin: University of Texas Press. ------. 1981. Labor, Class, and the International System. New York: Academic Press. PREALC. 1981. Din mica del subempleo en Amrica Latina. Santiago de Chile: International Labor Office. ------. 1982. Mercado de Trabajo en Cifras: 1950-1980. Santiago de Chile: International Labour Office. ------. 1986. Cambio y polarizaci"n occupacional en Centroamerica. Costa Rica: Editorial Universitaria Centroamericana. ------. 1987. Ajusta y deuda social: un enfoque estructural. Santiago de Chile: International Labour Office. Przeworski, Adam, and Henry Teune. 1970. The Logic of Comparative Social Inquiry. New York: Wiley. Roberts, Bryan R. 1976. The Provincial Urban System and the Process of Dependency. In Current Perspectives in Latin American Urban Research, ed. Alejandro Portes and Harley L. Browning, 133-150. Austin: Institute of Latin American Studies Publications Series, University of Texas. ------. 1978. Cities of Peasants: The Political Economy of Urbanization in the Third World. London: Edward Arnold. Ruiz, C. H. 1990. Situaci"n de la pobreza en Guatemala en la decada de los ochenta. Documento. Ciudad Guatemala: IDESAC. Santana, Julio. 1992. Restructuraci"n neoliberal, zonas francas y proceso de urbanizaci"n: el caso de Santiago, Rep#blica Dominicana. Working paper. Santo Domingo: Latin American School of Social Sciences (FLACSO). Schoepfle, Gregory K., and Jorge F. Perez-Lopez. 1989. Export Assembly Operations in Mexico and the Caribbean. Journal of Inter-American Studies and World Affairs 31(Winter):131-161. Tokman, Victor E. 1982. Unequal Development and the Absorption of Labour: Latin America 1950-1980. CEPAL Review 17:121-133. Trejos, Juan Diego. 1991. Informalidad y acumulaci"n en el area metropolitana de San Jos, Costa Rica. In Informalidad Urbana en Centroamerica: entre la acumulaci"n y la subsistencia, ed. J.P. Prez-S inz and R. Menjivar Lar!n. Caracas: Editorial Nueva Sociedad. Unikel, Luis. 1972. La dinamica del crecimiento de la Ciudad de Mxico. Mxico, D.F.: Fundaci"n para Estudios de Poblacion. United Nations. 1988. Prospects of World Urbanization. New York: United Nations. Valdez, Cristobal. 1988. Modelo de desarrollo urbano y organizaci"n interna del espacio en Santo Domingo, D.N. Santo Domingo: IDDI-Fundaci"n Friedrich Ebert. Yunn, Rafael Emilio. 1985. La isla como es. Santiago: Universidad Cat"lica Madre y Maestra.