ABSTRACT
This study measures dependency in northeastern Sonora, Mexico, as the country underwent change through three different development models. It is a test of the relationship between dependency theory and development at the local scale. The goal is to assess how well people benefitted from development policies of import substitution compared to those of the previous era of liberalism and the current era of neoliberalism. The copper mining region in northeastern Sonora is an ideal case study because changes in capital ownership at the mines are directly related to national development policies. Indicators reveal positive benefits during the era of import substitution, when resources were shared more equitably. Initial measures for the current free-trade era give contradictory indications as to the local benefits or drawbacks.

INTRODUCTION
Copper has been mined and smelted in northeastern Sonora, Mexico, for over a century. Copper mining ushered in the industrial era in this region and fundamentally changed social relations. U.S. capitalists purchased claims at Nacozari in 1895 and Cananea in 1896 (Figure 1), began large-scale production, and initiated the creation of an industrial proletariat (Heyman 1991). Changes in relations of production in the copper industry in northeastern Sonora are directly related to changes in national economic strategy. This study uses measures of dependency as indicators of development under three different development models (liberalism, import-substitution, and neoliberalism) to test which model best benefits people in the local region.

DEPENDENCY AND DEVELOPMENT
Dependency theory was developed by Latin American scholars in the 1960s as they studied the perpetual state of underdevelopment throughout most of Latin America. The theory was popularized in the English-speaking world by André Gunder Frank (1969), who described a metropole-satellite structure in which the satellite (periphery) remains in a perpetual state of underdevelopment because of its links to the capitalist core. Central to Frank's conception of dependency were global connections, which Wallerrstein (1974) clarified by tracing the development of the 'World System' from the 16th century. This structural view of dependency was explicitly geoographical, with regions or countries dependent on the geographical core, and it emphasized distribution nettworks as the focal point of dependency relationships. External forces were used to explain underdeveloppment in the periphery, while internal class relations, local players and culture, and the role of the state were de-emphasized.

Proponents of an alternative view of dependency claimed peripheral dependency occurred through the capitalist mode of production, not through distribution (Laclau 1971). This view emphasized the social relations of production and class structure internal to the peripheral region (Kay 1989). The 'mode of production' school believed dependency occurs when international capital interacts with and modifies capital accumulation in peripheral regions. Dependency was seen as " ... a series of relationships among classes of unequal power" (Johnson 1983: 232). Analyses of underdevelopment, while [end p. 17] acknowledging the role of international capital, stressed local class, cultural, and political conditions, rather than emphasizing only global structural forces (Stem 1988; Smith 1984; Cardoso and Faletto 1979).

Policies of import substitution, in which the state played an active role in economic and social relations and established high tariffs to foster national industries, were advocated by most dependency theorists-the' structural' school because they sought to break circulation ties with the core, and the 'mode of production' school because they saw the state as an agent to mediate relations of production.

Dependency theory has largely fallen from grace in part because as a theory it had little explanatory power. Dependency has been called a fact which needs explaining rather than a theory which explains (Halperin-Donghi 1982). The problem is the link between dependency and development, which is not at all clear. Even early dependency theorists rejected the claim that dependency created underdevelopment; rather, some development did occur, but it was highly uneven and unequal (Cardoso 1973). Furthermore, if dependence occurs with increased relations between the core and the periphery, development should occur when autonomy from the outside is achieved, which is clearly not true (Bernstein 1979).

The neoliberal development model adopted in Latin America today advocates free-trade policies that run exactly counter to those advocated by dependency theorists. Typical programs include liberalization of trade and industrial policy to emphasize exports, privatization of state-owned enterprises, and deregulation of foreign investment (Ros 1994). Mexico adopted neoliberal programs in 1982 because it faced recession, hyperinflation, and finanncial chaos. State subsidies for investment, production and consumption, an overvalued currency through much of the 1970s and early 1980s, and increased interest rates caused excessive indebtedness, which forced radical macroeconomic adjustment (Loaeza 1994; Urquidi 1994).

In spite of the problems Mexico and other developing countries faced in the 1970s and 1980s, few studies have measured the benefits of import substitution programs at the regional level to assess their effects on the level of development and to test dependency theory (Johnson 1983). The local results of national development policy directly shape the [end p. 18] construction of place. At the regional level, characcteristics which define dependency, such as a lack of diversity of economic production, concentration of wealth, and simplified social relations, can be used as indicators of development. The relationship between dependency and development is more clear at the local or micro level than on a global or national scale, where unequal development and contradictions muddy the picture.

Characteristics of Regional Dependency
Although Cardoso and Faletto (1979) stress the uniqueness of each dependent situation, several characteristics of regional dependency exist. Four measures are used to analyze the state of Sonora and the northeastern copper region (Table 1). These measures are all interrelated, and they link the structural view of dependency based on geographic regions with the alternative view that class structure and social relations of production determine dependency. The first two measures relate to the urban system in the region and are adapted from central place theory. A dependent region has an external focus and lacks a complex arrangement of urban centers. A simplified commercial arrangement inhibits a rank-size distribution of cities serving ever increasing functions. The resulting urban system in the periphery is therefore characterized by a near uniform size and function pattern with direct external control.

An external focus usually develops a dendritic arrangement of connections which funnel out exports (often one staple product). Debates about dependency theory have concentrated on this dendritic arrangement, which also existed in the pre-capitalist system of mercantilism; a system in which the 'mode of production' school claims dependency did not exist (Laclau 1971). Smith (1978) has refined the argument concerning marketing systems, dependency, and modes of production. She argues that in a situation of dependency, commerce must be based on capitalist competition. Administered trade does not permit capitalism to flourish; however, a dendritic system can be representative of peripheral dependent areas as long as the core is based on capitalist compeetition. A commercial marketing system must exist at the higher levels of the hierarchy.

The final two measures of dependency concern class relations. Dependency also is characterized by a marked concentration of wealth, which tends to become more concentrated over time. Local elites benefit from their role as middlemen to foreign capitalists and their links to exterior control. Wealth becomes politically controlled, with the local elites and foreign capitalists compounding their assets and forcing small local producers out of the market. The majority of workers and residents do not prosper. Often consumption is increased, usually of goods produced in the core region, which further transfers wealth to the core.

The final characteristic of dependent regions, largely a result of the previous characteristic, is a simplified social arrangement. Without the vibrant internal economy, and with reliance on one export staple, jobs and social status are very much homogenized, with the exception of local elites who are links to external capital. An external focus generates a formal region, characterized by homogeneity of one or several variables. In this case, a relatively homogeneous product is produced and a simplification of internal economic and social relationships are generated. Van Young (1994) refers to this arrangement as the funnel model, as opposed to a central internal focus, which he calls the pressure-cooker model. This view links social relations to regional characteristics of dependency. The funnel model creates a formal region, with homogeneity of jobs and city functions, while the pressure-cooker model creates a functional region with complex social relationships, a solar arrangement of connections, and a complex hierarchy of urban functions.

CLASS AND COPPER IN SONORA

Changes in the ownership of capital in the Mexican copper industry can be used to define three distinct periods of industrial development: (1) The American Capitalist Era (1895-1961), (2) The Mexican Era (1961-1982), and (3) The Free-Trade Era (1982-present). American investment began during the Porfiriato (1876-1911), when liberalism was the dominant economic model. The American era was characterized first by neo-colonialism and enclave economies, with a slow transition towards decreasing foreign control and the increasing role of the state (Sariego 1988). This era started with the classic company town development and company control over all aspects of life. All goods and products went through Arizona, with a virtual independence from the Mexican nation (Ruíz 1988). Favoritism granted to foreign capitalists by local elites created a clear situation of dependency in Sonora during the initial copper era. A simplified class structure of foreign management and largely Mexican laborers existed. [end p. 19]

Intermediaries in a contradictory position between capital and the proletariat--crew chiefs, technicians, lesser administrators--often become the 'front line' in struggles between classes (Johnson 1983). Because there is no objective classification which defines the class of middle managers, technicians, small employers, and semi-autonomous wage earners, their positions get translated into class allegiances as the social relations of production work themselves out in each situation (Wright 1980). During the American era, American middle management had strong allegiiance towards capital through ethnic bonds. Another intermediary position was the contract labor boss, a person who organized work crews to meet company quotas. While almost always Mexican, these people were in a position to exploit labor, pocket bonus payments received when quotas were exceeded, and make job insecurity high. They, too, alienated themmselves from wage-laborers and allied with capital. Eliminating these positions became a central theme in subsequent labor disputes.

Slowly the Mexican government broke foreign control by restricting foreign ownership and allying with labor. The ruling Partido Revolucionario Institucional (PRI) was based in part on the organized inclusion of industrial labor (Dresser 1994; Peña and Trinidad Chávez 1985). For the next two decades, a complex mixture of state control, private capital, the labor union, and foreign capital managed copper production (Contreras Montellano 1986). This was the era of import substitution, when the state mediated relations between capital and labor in a corporatist system (Garza Toledo 1994). Labor gained unprecedented power in Cananea (Nacozari was closed from 1949 to 1979), and many intermediaries allied with labor. As management became weaker and more ineffectual during the labor era and government guaranteed production levels, a capital class focused solely on profit-making dwindled. Corruption increased as decisions became politically controlled rather than guided by market forces. Supply contracts often included kickbacks to those in charge (Epler 1989).

With the change to the neoliberal model, union strength at Cananea was a key holdout of the old corrporatist model, and the government saw the union's collective contract as an impediment to change. As a result, federal troops occupied Cananea in 1989, closed the mine, and declared it bankrupt and up for sale. A telling indication of a labor-management alliance was that, upon armed takeover of the facilities, all employers, including management, were terminated and thirty-two company employees were arrested. With the subsequent reopening, management positions were cut in nearly equal proportions to laborers (Epler 1989).

Since 1988, the federal government has completely privatized the copper industry, now controlled by one Mexican-owned corporation, Mexicana de Cobre, which is controlled in turn by the billionaire Jorge Larrea. Organized labor has lost influence with the federal government, while at the same time the state has withdrawn from many social support programs. New strategies of flexible production, team [end p. 20] participation, and total quality management can be seen as an attempt to win allegiance of the intermediaries to the side of capital, and the labor union at Cananea has resisted these new policies. Because Nacozari was closed for 30 years during the height of labor power, Mexicana de Cobre has stepped into a labor vacuum and established unilateral control over relations of production, as well as social relations in the town.

MEASURING DEPENDENCY

1. Urban Hierarchy
A series of tests were performed to measure the level of dependency in the copper mining region of Sonora. I used a census-defined region (INEGI 1991b), which encompasses both mining centers, to measure dependency (Figure 2). Because dependency exists between a 'core' and a 'periphery' several tests compared the copper mining region with the state of Sonora to evaluate the level of dependency of the copper mining region in its larger context.

The urban hierarchy is first plotted for the top fifteen cities in the state to see how closely it conforms to a rank-size distribution. Deviations from a slope of negative one indicate a flat urban hierarchy or a top-heavy system. The urban system for the state of Sonora indicates the structural arrangement in which the copper region exists, and serves to measure regional complexity.

The state of Sonora had a flat urban hierarchy in 1949, the late American era (regression slope = -.82) (Secretaría de la Economía Nacional 1947). Little internal complexity existed at the time, as the state was emerging from a long period of American domination (Tinker Salas 1992). The state passed through a rank-size distribution during the Mexican era, and has tended toward a top-heavy system ever since (1970 slope = -1.15, 1990 slope = -1.23) (INEGI 1984, 1991a; Secretar 161a de Economía 1953; Secretaría de Industria y Comercio 1971).

The optimal rank-size urban distribution during the Mexican era suggests a clear hierarchy of urban functions. Today's tendency towards a top-heavy system shows the increase of Hermosillo as the key focus for the Sonoran economy, both for export [end p. 21] agriculture from the coast and for its increasing role as an external focus for foreign mining interests. Since 1990, over 100 U.S. mining firms have opened offices in Mexico, nearly all in Hermosillo (Peterson 1992).

State planners determine which cities lack critical functions based on isochrones, or time to the nearest urban center for various central goods. In general, all of La Sierra (the more traditional and underdeveloped mountainous eastern Sonora) is today lacking intermediate city functions and is dependent on external provisions. The Programa estatal desarrollo urbano, 1992-1997, reveals an imbalance between Hermosillo and its peripheral regions (Secretaría de Infraestructura Urbana y Ecología 1991). Sonora again lacks the complex arrangement of urban centers which are characteristic of an internally vibrant, mature economy. Hermosillo is rapidly increasing functional control in the state and becoming the focus for renewed external ties.

2. Flow of Goods and Single-Product Reliance
The pattern of the flow of goods is difficult to measure and can influenced by factors other than a condition of dependency--in this case, the local topography of northeastern Sonora. The region consists of a series of north-south trending valleys and mountain ranges, with an outlet at the north into Arizona and funneled to the south into Hermosillo. Even in a mature regional economic system, commerce would be channeled through key centers and interregional trade inhibited. But trade was more than inhibited in the boom years of the American controlled era, it was monopolized. All copper was shipped to the United States for smelting, the distriibution of goods was initially administered through company stores, and a reliance on U.S. consumer products developed early. Because all copper was exported and most consumer goods were imported, the mining boom through to the end of the 1920s brought little benefit to the region outside of the actual mining nodes (Ramírez and León 1985). By 1910, Sonora had the second largest amount of foreign capital invested in any state in Mexico, yet lagged far behind the more developed states of the Republic in industry (Ruiz 1988). Mining centers in early twentieth-century Mexico typically did not draw their hinterland into regional economic development (Sariego et al. 1988). A dendritic relationship existed to and from Arizona and the mining industry dominated the Sonoran economy. [end p. 22]

Regional development requires less dependence on a single primary product. The importance of other economic sectors can be shown with location quotients, a ratio of ratios measuring regional employment in a certain sector to a base measure (such as the country) of employment in that sector. Values greater than 1.0 indicate that a sector produces more than its regional consumption and therefore is interpreted as producing for export. Values less than 1.0 indicate a region imports goods or services in that sector. Location quotients between the copper mining region and the state and between the state and the country were calculated to measure dependency at two scales. At the state level, very high location quotients during the American era for mining show the continued importance of this sector for the Sonoran economy (Table 2). The only sector considerably below the national average is manuufacturing, indicating a lack of regional development in the sector considered central to local development (Sheahan 1991).

In contrast to the American era, during the late Mexican era the mining sector drops significantly below the national average, and an increased imporrtance is seen in government and commerce. This is indicative of more local control at the state level and a more vibrant commercial activity, rather than administered trade networks. The Free Trade era reveals a renewed emphasis on mining and a rise in finance. Both of these likely are related to increased foreign investment and development associated with the NAFTA. Although mining is again central to the economy, this in itself is not indicative of dependency and underdevelopment. Many of the other sectors are close in employment proportions to the nation overall, indicating a more diverse economy than seen earlier in the century.

Location quotients for the six municipios that comprise the copper region in northeastern Sonora show that mining by far has dominated the economy. The low point in mining dominance was reached during the Mexican era in 1980, the same year in which manufacturing was at its highest. A lesser reliance on mining and the greater importance of manufacturing suggest a more robust local economy. However, the majority of manufacturing employment occurs at border maquiladoras in Agua Prieta and Naco. Border maquiladoras have been shown to promote few backward linkages which spawn regional development (Shaiken 1990), and in fact can be considered another form of dependency on foreign capiital with few local benefits. Nevertheless, they do provide employment options and create wage competition to the benefit of labor. For this very reason, Mexicana de Cobre has successfully opposed assembly plants locating in the Nacozari area and has maintained a monopoly on wages (Montano 1995).

Finally, during the Free Trade era we see the copper region return to a strong emphasis on mining, with manufacturing above average but all other sectors below state averages. Commerce has progressively declined in importance at the regional level across the time periods analyzed as the copper region loses out to trade at the state level. Government has declined as both state and federal organizaations reduce their size, and finance has increased as the region is increasingly drawn into the global economy. In short, in both the Mexican and Free Trade eras, the copper region has been externally dependent in all economic sectors except mining and manufacturing, and these two are also externally controlled and serve a global market. By this measure, the region has not yet developed into a mature autonomous economic unit.

3. Distribution of Wealth
The distribution of wealth is indicated by three measures, the first simply being the concentration of capital ownership over the three periods. Initially, Phelps-Dodge and the Cananea Consolidated Copper Company were sole owners of the two big production centers. During the Mexican era, ownership at Cananea was shared among four groups. Mexicana de Cobre owned the complex at Nacozari after it reeopened in 1979, but the federal government also initially had much investment. Most recently ownership has concentrated to one principal source, Mexicana de Cobre (Figure 3).

The second measure of wealth distribution conncerns the equity of federal spending, measured during the Mexican era. If the distribution of population is a base measure, spending should be distributed in a similar manner to be equitable. Since a calculation of equity varies with geographic scale, the distribution of federal funds is shown by municipio and by the twelve economic regions identified in the Censos Económicos for Sonora (INEGI 1991b). Lorenz curves show that the spending is more equitable at the regional level than for the individual municipios, where disbursement is more concentrated (Figure 4). Areas between the 45 degree line and the plotted lines are proportional to concentration in funds allocation.

A Gini coefficient reveals the degree of innequality of federal spending and can be interpreted as the percentage of funds which need to be [end p. 23]

A final measure of wealth distribution is calculated by the ratio of total wages paid in mining sectors at the state level and at Cananea to value added, a measure of profits earned, across the three time periods. By far the peak value in terms of percentages of wealth shared with the employees occurred at the height of the Mexican era from 1975 to 1985 (Table 3). In 1988, when the government took control of the facility and initiated privatization, the percentage again dropped. The mining industry in Cananea and non-ferrous mining in the state are reported for comparison (Nacozari was closed during much of the time when data are available).

4. Social stratification
Two final measures are used to test the degree of social stratification. The first compares job categories across the three time periods to check for occupational diversity. For all economic sectors in the copper mining region, professionals and technicians have steadily increased in percentage (Table 4). Office and administrative workers have also had a general increase in percentage of workers, and laborers (artesanos y obreros) have gone from a low of 31.4 percent in 1960 to 39.7 percent in 1990. Much of this latter class is associated with maquiladora employment in the border towns of Naco and Agua Prieta.

In the Cananea mining industry, the low point of total employed in 1980 apparently contradicts the prevalent story that employment was bloated during the Mexican era (Table 5). Employment did increase throughout the 1980s--there were already 2,500 workers by 1982, and 3,700 in 1989 at the time the government seized the plant and declared it bankrupt. The late 1980s were years with a bloated workforce, when productivity was said to be one-fifth that of the Phelps Dodge operations at Morenci, Arizona (Epler 1989). Major layoffs took place after the privatization in 1989, yet by 1,990 more people worked at Cananea than recorded in any of the previous decennial censuses. In general, the percentage of upper-level jobs (professionals, technicians, directors) increased throughout the three eras, while the percentage of common laborers in the mines and smelters decreased. Operations today are highly mechanized and computer controlled. Further disaggregation of job description in 1980 and 1990 show the labor force increases are associated with increasing transportation needs as open pit operations expand, superrvisory positions, and a flexible workforce under the category of helpers (ayudantes y similares). Security personnel, clerks, and technicians also registered sizeable gains.

The division of labor shows increases in jobs associated with managing the expanding business, but relatively small gains in the actual workers doing the mining. Automation and technological increases have largely accounted for the change in labor emphases. Microeconomic adjustment strategies in the United States have resulted in more flexible and temporary workers who do not receive benefits or job security. The increase in ayudantes y similares in Mexican mining may represent the same process. Contreras Montellano (1986) asserts that modernization has created more need for skilled labor, but the increased division of labor has resulted in the total exclusion of workers from decisions in the production process.

A second measure of social stratification uses Mexican census calculations on the degree of economic diversity and indicators of social well-being in Sonora in 1980 during the Mexican era. The index of social well-being, calculated using principal component analysis on 15 variables on income, education, health, housing, and employment, scores Cananea first in the state, Nacozari fourth, Naco fifth, and Agua Prieta eighth out of 70 municipios (INEGI 1990). Similarly, using an index of economic diversity based on the concentration of nine economic sectors in each municipio, Nacozari ranked first, Cananea second, and Naco third. These calculations certainly indicate that the copper region of northheastern Sonora was among the most well-off in the state, and they suggest that indeed mining centers may serve as growth poles as neoliberal investment strategies predict. Industry advocates claim the increased surge of American investment in Mexican mining has contributed to the revitalization of the country's economy (Mining Engineering 1993). The 1994 economic crisis in Mexico led many to question that assertion. [end p. 25]

[end p. 26]

The generalized results from the above analyses paint a fairly clear picture, but with some doubt for the current era (Table 6). Evidence supports a situation of dependency during the American era, when wealth was concentrated in external hands and a simplified social arrangement existed in the copper mining region. These indicators also suggest that during the Mexican era the federal government seriously attempted to promote regional development and a more equitable share of the wealth. At the local scale, import substitution policies served the region well. Because the state allied with labor and funded many social programs, the majority in the towns benefitted.