During the 1980s Latin American metropolises continued to grow significantly faster than the region's generally already high population growth rates.⁽¹⁾ Historically, the highest levels of urbanization have occurred in more developed countries, but now Latin America is as urbanized as the industrial regions, 70 to 75 percent, while Asia and Africa remain only about 30 percent urbanized (Coraggio 1990, 259). In contrast to developed regions, Latin America has experienced its urban explosion concurrently with economic stagnation. By 2000 the region will have the world's two largest cities, Mexico City and São Paulo, each with over 20 million inhabitants, fully one-third of the world's cities of over ten million in population, and dozens of other cities of more than a million inhabitants. Thus urbanization is of major significance for the welfare of Latin American countries and a huge and multifaceted topic for research.

This paper reviews research by geographers on Latin American cities over the last decade. The scale of analysis is restricted to metropolitan areas and below, leaving interesting trends in Latin American urban systems to others (Portes 1989). The review is structured around geographic research; that is, what geographers working at the intra-urban scale during the 1980s have defined as research problems.⁽²⁾

While the research reviewed in this paper is confined to the themes defined by field research in Latin America by geographers, I have not restricted my review of these subjects solely to what geographers have written. Rather, in the interdisciplinary spirit of Latin American studies, I also consider overlapping work by other social scientists. Arguably, Latin American geography obtains its greatest strengths from interdisciplinary and comparative research, and from operating simultaneously at several geographical scales of analysis (Lawson and Klak 1990).

This paper has three main sections. In the remainder of this introductory section, I comment broadly on the status of Latin American urban research, noting in particular the scarcity of integrated bodies of research. Despite this weakness, the second section is able to identify two major aspects of theoretical progress during the 1980s: 1) research has begun to move beyond thinking about Latin American cities in terms of a series of "dualisms;" and 2) deductive theory has largely been abandoned in favor of a less dogmatic and more interactive approach to theory and empirical investigation. In the third and largest section of the paper, I review Latin American urban research during the 1980s under four broad categories. Research is classified as pertaining to the city under economic crisis, to the spatial organization of the city, to reproduction of the work force, or to formal and informal production. [end p. 283] Taken together, these sections should provide a flavor of recent Latin American urban research and the prospects for the future.

Ideally a review such as this should be organized around a list of the dominant and regionally distinctive intra-urban patterns and characteristics that recent research has uncovered, and the theories developed to explain them. However, at the outset the difficulty in generating this list is notable. Latin American urban research has been less structured around dominant paradigms than its North American counterpart. In the latter, topics such as spatial morphology (Johnston 1982) and production linkages (Scott 1988) define research themes, but Latin American research has been less integrated (Edel 1988).

There are two reasons for this lack of integration, one methodological and the other empirical. First, the lack of integrated research in large part suggests the weakness of Latin American urban analysis as an area of study. We lack systematically comparative research and concept building from both theory and empirical analysis. This is a major shortcoming which we are only beginning to rectify (Klak and Lawson 1990). Another compounding methodological problem is the poor and temporally-inconsistent data for critical variables such as income, unemployment and informal production. Owing to these methodological problems even the most ambitious cross-national studies of urban development are markedly provisional (Portes 1989).

Second, the lack of integrated research reflects the diversity of the subject matter as compared to the pronounced urban morphological and development patterns found in North America, best illustrated by Chicago. Concentric zones and sectors of land use and centrifugal growth have been consistently identified for cities of industrial North America. Even the United States, however, does not follow a single urban development model (Schnore 1965). In Latin America, across dozens of countries with different urban histories and development trajectories, there has not been a dominant and unrelenting force such as industrial capitalism to shape urban development as in the North American Manufacturing Belt (Carrion 1990). To respond to Latin America's urban diversity, research must strive not only to identify generalizable patterns of urban growth and organization but also to investigate the extent of and reasons for urban patterns diverging across cities, countries and regions (Coraggio 1990, 266-287; Unda 1990, 35). Indeed, the singular quest by North Americans for universal models of Latin American urban development is ethnocentric. This essay contributes to a comparative approach to urban research by juxtaposing at several points features of Latin American urban development against those of cities of the North American Manufacturing Belt. In the end, the best research will be analytical sensitive to both the generalizability of findings from a particular case study and the distinctive national or local conditions that those findings reveal (Benton 1986, 47-50; Ward 1989, 58). It is worth noting that there is much room for diverse empirical findings at the same time that the research body is becoming more integrated.

Owing to the lack of integration in Latin American urban research it is possible to identify some important themes but there is a less systematic research tradition and certainly a shortage of comparative studies and concept development. In a following section I identify prominent research areas, but it must be emphasized that many important conceptual interrelationships remain largely unexplored (Lawson and Klak 1990).

RESEARCH PROGRESS DURING THE 1980S: MOVING BEYOND DUALISMS AND DEDUCTION

Despite the methodological and empirical problems discussed here, some research progress has been made over the last decade. Progress is captured by two trends: 1) the move beyond dichotomous thinking about urban activity, especially the formal-informal distinction; and 2) the advance from strict theoretically-deductive interpretations of the city to more empirically-rich and theoretically-contingent analysis. Some illustrations of these two trends follow.

First, as Edel (1988) has noted, Latin American urban research has moved beyond dualistic thinking, such as urban-rural, modern-traditional, formal sector-informal sector, industry-services, state-society, and core-periphery. Research has also moved beyond the continua bridging these poles, which themselves are still one-dimensional characterizations. For example, analytical dichotomies between formal and informal or even between populist and rightist approaches to development policy have been shown to be inadequate for understanding recent debates and struggles in Peru (Bromley 1990). Dichotomous frameworks are being replaced by multi-dimensional ones that are less tidy but more informative. For instance, low income households are being described as "domestic economic archipelagos" (Bromley and Birkbeck 1988, 125). Contrary to the stereotype of the homogeneous "informal sector household," earnings are pooled from a variety of formal and informal sources thereby considerably lessening the risks to the individual members. Overgeneralized dichotomies such as that informal workers live in [end p. 284]/informal housing have also been exposed through detailed surveys in low income households. Surveys have yielded a better understanding of the constraints on and options available to the poor for work and shelter in specific cities (Gilbert and Ward 1985). Conceptualizations about production activities have also moved beyond dichotomies. Subcontracting linkages and the flow of commodities and value between informal work and larger scale, even multinational, enterprise are being explored (see Portes et al. 1989; Lawson 1990).

Second, progress has also been made to the extent that there has been a graduation of theory from reliance on simple explanations such as modernization theory, the culture of poverty, capital logic and state instrumentalism (Benton 1985; Carrion 1990, iv-v). Deductive theories have long histories in Latin American urban studies and, in fact, structural Marxism continues to have a strong influence on urban analysis (Castells 1976; 1978; Landivar 1986), especially in Latin America itself (see Carrion 1985, 1987, 1990; Unda 1990; Coraggio 1990). This is not to reject the contributions to urban analysis by structural Marxism. Arguments such as that of Burgess (1982) sensitized us to the state's class-biases and to the advantages to capitalist development and the status quo of self-help housing. Structural Marxism has provided considerable insight into the overarching power of the state, as representative of capital in setting the parameters of urban development and in negotiations with the poor over land and housing policies.

More recent research, however, has shown important variations in what the poor are able to secure from the state. Researchers have argued that these variations in the power of the poor should not be dismissed through deterministic explanatory frameworks, such as state legitimation-accumulation or state instrumentalism (Gilbert and van der Linden 1987; Benton 1986; Pezzoli 1987; Klak 1990b). As exemplified by Colombian urban research, Marxism has not been abandoned in recent years, but studies have been "anti-dogmatic." Emphasis has been on empirical problems such as urban violence and state-society relations more than the fundamental categories of Marxism (Carrion 1990, 129-130). Research has begun to examine in a more empirically open-ended way the interplay of class forces and the state, the contingencies of political struggles and the interconnections between issues and concepts. An understanding of these issues requires both extensive empirical investigation and iteration with theory. Intertwined with the analytical shift toward empirical details and contingencies has been a greater emphasis on contributing to social change through planning and organizing (Carrion 1990; Pezzoli 1987). These are all ambitious tasks that are largely still before us.

THEMES IN LATIN AMERICAN URBAN GEOGRAPHY

Latin American urban research can be fit into four broad categories: 1) cities under crisis; 2) urban morphology; 3) reproduction; and 4) production. While the categories are rather generic, at least some of the details within each are distinctly Latin American. This owes primarily to the problems facing the region, including a decade-long economic crisis and recession, the draining of resources by foreign debt obligations, overurbanization and a polarized class structure. Of course, similar features are found in poor countries elsewhere, but in Latin America their extent and interaction creates crisis conditions (Portes 1989; Coraggio 1990, 257-315). In the study of Latin American cities, a research goal is to identify and explain the processes both generalizable to the region (and beyond it) and the uniquely local.

It is appropriate that the four categories be given different weighting in this review. The first topic, cities under crisis, is reviewed briefly but is returned to throughout the paper, as research during the 1980s would have found difficulty in avoiding the effects of recession (see Klak 1990a, 582). Because of the overarching effect of recession, different aspects of informal activity are discussed under both "cities under crisis" and "reproduction." The third and fourth categories, reproduction and production are, indeed, important themes in geographical research, but they have been recently surveyed at length elsewhere (Lawson and Klak 1990) and thus are reviewed only briefly here.

Of the four categories, by far the greatest attention is devoted to the second, urban morphology. One reason for this emphasis is that urban spatial structure is a quintessentially geographical area of study. Second, research needs to put greater emphasis on intra-urban geography. I would argue that Latin American urban research by geographers and non-geographers alike would attain substantially greater understanding by a much more systematic treatment of the geographic aspects of the processes under investigation. Instead, much urban research has been placeless. Urban social movements, for example, emerge in selected urban contexts and labor markets from conditions particular to communities confronting the agendas of the state and investors for developing urban land [end p. 285] in distinctive ways. In contrast, little understanding of social movements can be gained from a macro-level analysis (e.g. Walton 1990). Studies often exclude a detailed intra-urban geographical analysis (see Eckstein 1989; Unda 1990, 297-320; Kowarick 1985; Pezzoli 1987). Thus a more thorough review of research during the 1980s with an explicit intra-urban spatial dimension is intended to encourage a greater emphasis on this theme in research during the 1990s.

CITIES UNDER CRISIS

In the 1980s Latin America was devastated by recession at a scale unknown since the 1930s, and considerable research has tracked the urban impacts. Even when "the city under crisis" is not the primary research objective, it is necessarily implied, as virtually nothing has escaped the wrath of economic downturn. Among the recession's principal urban effects were contraction of formal industrial employment, state and IMF-imposed wage restrictions, high inflation and streamlining and de-subsidizing of state provisions. Real wages have plummeted, even for those who have managed to retain formal employment (Kowarick and Campanario 1986, 166; Portes 1989; Becker et al. 1990, esp. p. 32).

In the past, considerable evidence has indicated that informal work has served as a counter-cyclical mechanism, absorbing formal labor made redundant by recession and supplementing falling real formal wages (Conaghan et al. 1990). The economic crisis of the 1980s, however, suggests limits to the informal sector's capacity to absorb and support labor. Labor surplus and redundancy in the informal sector have driven wages down there too, following the pattern produced by layoffs in formal firms. While informal work offers ease of entry, it is typically at a price of inadequate remuneration. As the numbers of informal street vendors swell, work weeks expand, and wages and consumption of basic needs fall (Portes 1989). In Guayaquil, for example, these conditions have pushed a significant portion of working class women to total physical exhaustion (Moser 1989). Even in more industrialized countries such as Brazil and Mexico, at least half of the population is malnourished. In the face of crisis and impoverishment, human survival instincts have moved large numbers of workers to abandon the standard ways of earning a living in the city by turning to subsistence and informal cooperative activity (Portes 1989; Portes et al. 1989).

As Gilbert's (1990) research on Bogotá shows, sustained recessionary conditions also have contributed to the disruption of the role of the state, which in turn has further threatened the availability of items of basic human need. Since 1980, interest charges on foreign loans to state utility companies have drastically increased, while overly ambitious projects and pork barrel allocation of construction contracts have wasted resources and emptied the state coffers. Bogotá residents have been asked to pay an increasing portion of earnings for urban services. Price increases and responses to it, such as increased pirating and refusal to pay, have disrupted urban services and the relationship between the state and urban residents.

In summary, long-term economic downturn has threatened work force reproduction in Latin American cities. Workers have shifted to poorly remunerated informal activities, the state has retrenched and "rationalized" its programs, and the provision of basic needs such as urban services and health care has deteriorated (Klak and Lawson 1990). Threats to the reproduction of the work force also draw into question the legitimacy of the state and the viability of societal reproduction more broadly (Unda 1990, 265-71).

URBAN MORPHOLOGY

Despite the obvious significance of urban morphology to geographers, there has been remarkably little systematic work to identify general spatial features of Latin American cities (Herzog 1990, 71). The literature that does exist, however, is very interesting, in that spatial patterns have been linked to deeper social processes. Urban morphology has thus far been a scattered theme yet is one suggesting fruitful avenues for further research.

Griffin and Ford (1982) proposed a schematic and qualitative urban model of socio-economic and housing sectors and zones, the latter in reverse order of affluence compared to the North American case and originating under Spanish colonialism (Herzog 1990, 72-3). Since the Griffin and Ford model there has been little follow-up work to substantiate, further specify, extend or propose alternatives to their general model of urban spatial organization.

One alternative to Griffin and Ford's (1982) model of urban sectors and zones is the notion of a spatial and class polarization of the city (Portes and Johns 1986; Herzog 1990, 76). In most abstract terms, this conception[end p. 286] implies that the massive economic distance separating affluence from squalor in Latin American cities manifests itself in the location of the residential areas of the rich and poor. Extreme class polarization is manifested in a spatial polarization of urban growth.

In some cities, spatial and class polarization (or at least distancing) is suggested by the fact that although both the rich and poor have migrated outward from the city center in recent decades there has been little mixing of their neighborhoods. Despite simultaneous residential location by the rich and the poor around the fringes of the city the affluent Latin Americans have worked with the state to maintain residential separation from the poor (Ward 1990; Higgins 1990). The Andean capitals of Bogotá and Quito illustrate spatial polarization. In both cases, affluent areas have expanded northward and while poor areas have expanded southward. For Mexican cities Herzog (1990, 86) also reports increased spatial segregation by wealth in recent decades.

Further comparison with the United States urban model may help to interpret Latin American urban patterns.⁽³⁾ Although Latin American cities have grown centrifugally, the municipal balkanization underlying the United States model of suburbanization is much less applicable to Latin America. Although little has been written on the topic, the region generally lacks the municipal balkanization that, in the United States, encouraged capital and the affluent to escape the tax burdens of the central city and create protected and socio-economically homogeneous suburbs (Johnston 1982; Schmidt 1979, 158; Herzog 1990, 76-7). Across Latin America, the great bulk of affluent districts are within the central city. Thus residential growth has been centrifugal in Latin America but not strictly "suburban" as in the United States. Perhaps it is the lack of municipal balkanization, together with the affluent desiring distance from the poor, that has led to the spatial polarization of some Latin American cities.

In those unusual circumstances when Latin American cities undergo balkanization, a predictable result is more class segregation. For example, most urban functions were transferred to local governments in metropolitan Santiago during the 1980s. Rich areas were then cleansed of irregular settlements. Poor areas became more homogeneous as government high-rise housing projects were built in low income peripheral areas (Scarpaci et al. 1988; Portes 1989, 22-23).

Although both rich and poor in Latin American cities have demonstrated a propensity for centrifugal residential growth, the opportunities and decision-making of each group are very different. The affluent of Latin American cities have suburbanized like their United States counterparts, although not with the dynamic and at the scale of United States urban sprawl. Lower commuting mobility and security in Latin American cities help to explain the differences. Latin American cities do not have the "spoke and wheel" networks of highways that are paid for by federal taxes and have made possible urban sprawl by mobile and affluent households in the United States (Unda 1990, 100 cf.). Thus part of the dynamic of massive and unrelenting suburbanization in the United States is not present in Latin America and the trade-off of location for the affluent is not tilted as strongly toward the urban periphery. Another factor dampening the attractiveness of periphery subdivisions surely is the greater security offered by high-rise apartments located in more accessible central areas of Latin American cities. The preoccupation with home security of households of middle income and higher is illustrated by window bars, protective walls topped with jagged glass and guards.

The locational decisions of the Latin American poor are different. John Turner coined two useful terms to capture housing and locational strategies of the urban poor: bridgeheaders and consolidators. These should be thought of as ideal types rather than rigid categories (Conway 1985, 184). Bridgeheaders are newer and poorer urban migrants who take residence near to the urban center, often in crowded and dilapidated tenements for access to the low paying casual work near the bottom of the social hierarchy while avoiding the high transportation cost and time consumption of peripheral residence (Gilbert and Gugler 1982; Benton 1986). The Vila Pinto invasion of poorly drained land one mile from the center of Curitiba, Brazil, illustrates the linkage between residential location and work. Living centrally allows workers to push carts to the city center for recycling, a job made impossible by living on the periphery.

Consolidators were formerly bridgeheaders who have achieved some modest social mobility and have relocated to more spacious individual sites on the periphery when job security and income allows for the commute. That the poor choose to relocate to the periphery should not imply that the geography of land values and residential opportunities is not politicized. The state plays a large role in opening valuable land for investment (Burgess 1982; Benton 1986; Higgins 1990). The geographical pattern of bridgeheaders and consolidators has been borne out in [end p. 287] detailed research on Montego Bay, Jamaica (Eyre 1982) and in other cities. Research documents that the poor living at the urban periphery are not new migrants from the countryside (Ward 1990, 52-55).

Urban renewal programs have a long history in Latin American cities and have served to deprive the poor of access to central jobs, urban services and often better housing than they can find elsewhere (Batley 1983, 104; Gilbert and Gugler 1982; Higgins 1990; Benton 1986). Although many poor eventually choose to relocate to the periphery, this eventuality does not make central housing any less imperative for those surviving as bridgeheaders.

The dynamic of the relation between the state, income groups and the geography of opportunity in the Latin American city contrasts with that of the United States. In the United States the state, through an array of programs and subsidies, facilitates the suburbanization of all but the poor, thereby excluding them from "the American Dream" (Checkoway 1980; Perin 1977). The Latin American state, in contrast, tries to force the poor out of the center and toward the urban edge, directly through housing programs or indirectly for lack of available central land. Yet the Latin American urban poor cling to central residential location despite state efforts at urban renewal, condemnation and demolition and periphery housing schemes (Benton 1986). Further research is needed to determine if and precisely how these inter-American differences are attributable to the geography of urban land values and investment opportunities or to other factors (Ward 1989).

The city of the United States Manufacturing Belt has grown outward and drawn capital away from the central city to the point that the ring of suburbs function almost independently of the central city. The core city is left with a high tax burden, declining population, especially at the high income end, and dwindling tax base (Johnston 1982). The extent of central city exodus varies tremendously across Latin America. Two cities representing extreme cases of urban exodus and the creation of a more North American-like "urban donut" are Managua and Kingston. In both cases, peculiar local events caused vast abandonment of the central city and state housing programs pushed new residential development out to the urban fringes.

Central Managua was leveled by an earthquake in 1972; in total, 80 percent of Managua's housing units were destroyed. The Somoza regime used reconstruction aid to benefit well-connected construction firms and land owners at the urban periphery while it bulldozed efforts by the poor to resettle central Managua informally (Higgins 1990). In Kingston, it was the violence around the 1980 election in Jamaica that left about 900 dead and destroyed hundreds of homes near the city center (Eyre 1984; 1986). The affluent fled to new homes in the hills to the north while state housing programs settled thousands of working and middle class families to the west and northwest periphery. As is typical across the region, however, suburban housing development has not been accompanied by a marked suburbanization of work and basic services for the working and middle classes. Mass transit cannot adequately accommodate the need for daily commuting to the center (Anderson 1988).

In general, Latin American cities demonstrate less clear class-segregated morphologies than United States cities, in which the central city contains most of the working class and the suburbs most of the more affluent. Latin American cities generally lack the incentive of, and power vested in, separate political fiefdoms for the affluent and are without the spoke-and-wheel highway systems of United States cities. For every example of spatial polarization there seems to be just as many striking examples of rich-poor proximity, such as the squatters on the hillsides overhanging the opulence of Ipanema (Scarpaci et al. 1988, 26; Ward 1989, 48). Data assembled by Portes (1989, 22) suggest that Bogotá and Montevideo became more socio-economically mixed and denser during the 1980s. He interprets the mixing and increasing density as adaptations to economic crisis. The middle class seek affordable housing in lower income areas and the poor seek locations accessible to employment. They squeeze into central tenements and new irregular settlements near established elite areas hopeful of work in domestic service to the rich.

The diversity of conditions among Latin American cities is well illustrated by variations in housing tenure and housing geography. About a third of Lima's population lives in invasions, but they are not common in São Paulo (similarly see Gilbert and Ward 1985). Tenure differences among cities cannot be attributed simply to variations in national state policies because some of the greatest variations are found within the same country. In Ecuador, for example, the conditions and geography of popular housing in the two major cities contrast dramatically. In Quito, the poor are less visible because their homes are scattered between the city center, pockets in higher income neighborhoods, the urban periphery and surrounding villages. More than half rent their units, while another quarter live free of housing payments. In Guayaquil, in contrast, two-thirds live in housing without a burden of payment, while less than one-quarter rent. The vast majority of the poor live in Guayaquil's suburbios (JNV et al. 1985, 36; Landivar 1986, 24).

The diversity of trends in growth and segregation suggest the limitations of universal models of Latin American urban spatial structure. There is a need for greater sensitivity to the influence of processes particular to countries, especially national state policies (Gilbert and Ward 1985; Portes 1989; Scarpaci et al. 1988). Local factors, such as working class expectations and demands and barrio organizations, are critical to the pace at which urban renewal can "eradicate" the poor from prime real estate and how much housing assistance, and with how large a subsidy, the poor are able to squeeze from the state (Benton 1985, 10; Banck 1986; Pezolli 1987; Godard 1988; Higgins 1990).

Research on urban morphology has taken as a point of departure that migration, social and spatial mobility and class in Latin American cities diverge from the United States model. However, many themes, such as suburbanization of the affluent, migration of the poor first to central rental housing and then to purchased sites on the periphery, suggest that more systematic study of Latin American urban growth in light of the vast literature on United States spatial patterns and social processes is warranted. More careful comparisons and explanations of similarities and differences between the two urban growth patterns would be fruitful.

REPRODUCTION

Workforce reproduction is primarily the responsibility of the household and the state and can be defined as the social process of regenerating labor power through providing its basic needs, including food, health care, housing, amenities and even cultural environments (Castells 1976; Ward 1990). By this list of components of reproduction it is clear that research on this topic is diverse and interpretations necessarily integrate several dimensions of urban structure. Important research themes include urban land and self-help and other popular housing (Conway 1985; Unda 1990, 207 cf), land invasions in relation to the interests of speculators and government agencies (Mattoso Mattedi 1980-81; Banck 1986), the role of women in self-help housing and community cooperatives (Moser and Peake 1987; Pierro 1989; Moser 1989) and the role of the state in urban reproduction, housing and urban services (Gilbert and Ward 1985; Ward 1990; Klak 1990b).

That research on reproduction is discussed here prior to production is intentional. Referring to research on cities in the United States and Europe, Scott (1988) has insightfully noted a strong bias toward investigating reproduction activities such as residential space and the state's allocation of urban services over production issues such as capital investment and the division of labor. This observation applies equally well to Latin American urban research (Ward 1990). Irregular settlements are arguably the most popular Latin American urban topic (Benton 1986, 33; Unda 1990, 207). The point is not that consumption should be de-emphasized but that more attention needs to be paid to its connections to issues of production, perhaps with an analytical emphasis on how people earn a living rather than how they spend their money.

Another frequently researched topic under reproduction has been termed "clientelism" (Banck 1986). This refers to the dynamics of interaction between the state and squatters and others in substandard housing over land rights, urban renewal and housing assistance (Valladares 1978; Gilbert and Ward 1985; Conway 1985). Much less common are thorough studies of the roles of particular government agencies in urban development, linked to sophisticated state theory (Edel 1988, 173). A model study, however, is Batley's (1983) three-dimensional analysis of Brazil's National Housing Bank, land development in São Paulo and the various housing problems of the working class (similarly, see Benton 1986).

Geographical analysis of reproduction activities has become more sensitive to the connections between urban developments and the international political economy. Forces from outside the region, including global recession, international agencies and aid packages, have had profound and often deleterious effects on both production and reproduction in Latin American cities (see Klak and Lawson 1990).

PRODUCTION

Production encompasses the transformation of inputs into usable goods, both formally and informally. In many countries the urban work force is divided roughly in half between formal and informal laborers. As long as the pitfalls of dualistic thinking are avoided, it is still useful to consider differences between formal and informal work. In particular, the state, through regulations, taxes, credit and infrastructure, interacts differently with formal and informal enterprises. Although the state is often thought to work primarily in the realm of societal reproduction, in fact, far greater state resources are put toward expanding production. Major distinguishing characteristics of [end p. 289] formal and informal activities are the regularity of wages and legality. Formal workers almost all earn the official minimum wage or more, and have better access to state resources (Lawson and Klak 1990). Despite the diversity of wage levels among both formal and informal employees, the latter are on average far worse off. In Bogotá in 1984, for example, more than half of informal workers earned less than the official minimum wage, while only 3 percent of formal workers did (Portes 1989, 25).

Urban informal activity is one of the most researched topics for poor countries. Over 1000 separate studies have been done (Bromley 1990). Research on informal activity typically takes the form of a survey-based study of a particular type of work in a single city (such as Gugler 1988; Portes et al. 1989). What is less understood, however, is the insertion of informal work in the wider economy. For example, as formal workers were laid off during the economic downturn of the 1980s, more work was subcontracted to informal producers (Portes 1989, 26).

In contrast to the abundance of studies of informal production, there has been relatively little urban geographical research on production by formal enterprises. The work that has been done demonstrates the linkages between conditions of production, politics and work force well-being. For example, Storper (1984) has explained how Brazilian regional politics underlie the industrial decline of Rio de Janeiro earlier the century and the simultaneous rise of the São Paulo industrial core. Repressive state policies since the 1964 military coup have kept wage levels in metropolitan São Paulo low enough to discourage decentralization of industry. Similarly, Kowarick and Campanario (1986) have argued that state policies underlying the massive industrial development and urban growth of São Paulo have been at the expense of the working class. There has been a severe deterioration of living standards since 1964 and during the Brazilian "economic miracle".

An important geographical dimension of production is the way in which the world economic order holds a particular role for Latin American cities in the international division of labor. Multinational product assembly has relocated to Latin American urban settings in an effort to offset the high labor reproduction costs and thus declining profits associated with core production sites (Gilbert 1986). Research on informal activity is also beginning to consider the linkages to formal production in the city and beyond (Portes et al. 1989). Thus, in general, geographical work on production in Latin American cities is increasingly emphasizing the inherent geographical and dialectical connectivity between economic activities.

CONCLUSION

This review of Latin American urban geographical research during the 1980s identified a move away from abstract theorizing and toward empirical details. Richly empirical studies pursue one of four themes: recession, morphology, reproduction and production. Whether or not the studies explicitly addressed the economic downturn, virtually all findings have been colored by the impacts of global recession from which Latin America has yet to emerge. Real wages, state urban services and therefore conditions of work force reproduction have deteriorated. Informal work and shelter have expanded but in only partial compensation for losses in formal enterprise and state programs. Social polarization has worsened, but the spatial polarization of residential environments may be decreasing as the impoverished middle class locates with the poor, and as some of the poor invades land near the domestic jobs offered by the affluent.

The four research areas were reviewed in separate sections of this paper because they have been defined and studied as separate topics in the literature. Even the impacts of recession has its own literature. However, there is a need for integration across these themes and case studies of cities. Increased integration would enhance our understanding of each of the four themes of Latin American urban geography. In particular, research during the 1990s should be more sensitive to the intra-urban geography of the topic being investigated, be it subcontracting, social movements, women in the work force or the impacts of recession. Geographers are poised to make important contributions to our understanding of Latin American cities through the use of their skills and historical interest in investigating the spatial details of urban segregation, morphology, production, land development and communities.