The Economic Organization of Bolivia: An Analysis of Commodity Flows

THE ECONOMIC ORGANIZATION OF BOLIVIA: AN ANALYSIS OF COMMODITY FLOWS Robert B. South* Among the characteristics often attributed to nations in the third world is a lack of economic and social integration in space. (2) Underdeveloped countries are viewed as composed of regional islands and a purported goal is the creation of a "national economic space articulated by an interdependent system of metropolitan places, areal functional specialization, and national markets." (2) Implementation of regional development programs has been particularly noteworthy among South American nations where settlement patterns have traditionally been oriented to peripheral areas. Brazil, for example, has attempted to shift development priorities from coastal to inland areas as exemplified by the movement of the nation's capital with a concomitant thrust of road building in western states. Andean nations have promoted the marcha al Oriente, in part to help relieve population pressures in highland areas, but also to occupy and develop areas outside the effective national territory. (3) Regional development problems have been a major concern for the Bolivian government. Questions of spatial economic organization have particular relevance to a nation which has lost more territory to neighboring countries through conquest than remains within the national borders. (4) Considerable internal turmoil has had regional undertones and the Bolivian government is concerned about the extent of national integration. The spatial component of development manifested in regional development programs has become increasingly important for national growth objectives. Development does not affect all parts of a nation equally and a basic prerequisite for the formulation of regional development programs is a better understanding of the organization of the nation's space economy. (5) The objective of this research is to analyze transaction data in order to provide some insight into the * Dr. South is assistant professor of geography at the University of Cincinnati . This paper was accepted for publication in February, 1976. 10Southeastern Geographer spatial structure of the Bolivian economy. Specifically, it is a regionalization study using commodity flow data to clarify some of the complexities of economic space. (6) Delimiting a nation's spatial structure within a third world context has implications of practical significance and theoretical importance. Areas which share overlapping regions can be potential locations for a number of government supported facilities as exemplified by large scale milling operations. Regional boundaries can highlight different ecological zones and point to possible locations for government planned bulking facilities or transshipment centers. Within a theoretical context , a regionalization of transactions over time can provide a comparative measure of changes in regional structure resulting from subnational planning objectives. Transactions which initiate economic integration may be identified, thereby facilitating the process of spatial transformations from regional isolation to national economic integration. In summary, this paper provides a comparative basis for other studies conducted in other settings, and helps to fill a void in what is known about Bolivia and the spatial organization of an underdeveloped economy. (7) This paper departs from most commodity flow studies in that a variety of commodity types are utilized in an analysis of economic organization, and the research is conducted in Latin America, a cultural setting which has received little attention in the literature concerned with transactions and spatial structure. (8) THE SETTING. Based upon relief, Bolivia can be divided into two distinct regions: a highland area to the west consisting of a bifurcated Andean Cordillera and the Altiplano and an extensive lowland area to the east, the Oriente (Figure 1). In addition, a third region is often recognized which can best be described as transitional. This region of intermountain basins and valleys (The Yungas) is situated between the western highland and eastern lowland. Highland Bolivia comprises 30 percent of the nation's territory but contains 80 percent of the population. With one exception Bolivia's most populous cities are located in the southwestern one-third of the nation. This population pattern is a result of several factors, but primarily it is a historical consequence . Spaniards settled particular areas of the Altiplano, or high- Vol. XVI, No. 1 11 \PUERTO VASOUEZ BOLIVIA LEGEND ROAD RAILROAD PUEHTO '^ RURRFNARJIi)IIF ( CARANAVI'V^ DESAGUADEP uaqu_7t.viacha TUNAR ???????_?< C SANTA CRUZ ROBORE PUEBTO POTOS ERTC ause». MONTES LLAZO* L^ Figure 1. Bolivia: Surface characteristics. lands, in their search for minerals. Other areas were occupied because they were on favorable transportation routes or because they were potentially productive agricultural areas which could supply the mining towns. All major cities in Bolivia are colonial cities, and the population to a large extent, has remained concentrated in these areas. Concerned with underpopulation in eastern Bolivia, the government has sponsored a number of colonization programs in the Yungas and Oriente in order to shift population from highland areas 12Southeastern Geographer and to promote regional economic growth. Some of these agricultural resettlement programs have had considerable regional impact as exemplified by the approximately 12,000 families near Villa Tunari who are currently engaged in cash crop production. (9) Agriculture is expanding beyond the traditional producing areas of the Cochabamba valley and the Altiplano, and some of the newer areas are noted for large scale specialized production. For example, most of Bolivia's sugar cane is produced near Montero and the shipment of sugar cane to mills near Santa Cruz constitutes the most important commodity movement within the nation. Southeastern Bolivia is emerging as an important producer of other national resources . Both forestry and petroleum production are largely concentrated in the Oriente. Mineral production, however, continues to be concentrated along the Cordillera Oriental-Central. Minerals constitute nearly 90 percent of Bolivia's foreign exchange earnings and are exported via the railnet to Chilean and Peruvian ports. (10) Connecting resources with population and markets is the problem currently confronting the Bolivian transportation system. Historically the railroad tied the mines with export points and now is facing decreasing traffic with antiquated equipment. The trucking industry is attempting to connect the old occupied areas with the new but operates over an archaic roadnet. Summarizing the Bolivian landscape provides a sense of perspective to the study but a number of events, namely revolution (1952), the resulting agrarian reform ( 1953), and colonization programs have had spatial impacts upon the traditional regionalization of the nation. TRANSACTION DATA. Commodity flow data were obtained from a transportation survey conducted in Bolivia during 1967-1968. (11) These data consist of the origin and destination of commodities, in hundredweight per week, for railroads and truck carriers. These two modes carried approximately 106,800 tons of cargo per week over the transportation network representing over 94 percent of the mineral and general cargo traffic in Bolivia. Trucks handle about 63 percent of the traffic, and railroads about 31 percent (Table 1). The highway and railroad commodity survey yielded over 274,000 pieces of information encompassing a great variety of commodities and a complex array of origin-destination patterns. To reduce the Vol. XVI, No. 1 13 TABLE 1 CARGO MOVEMENT IN BOLIVIA, 1967-1968 Class and Mode Annual Ton—Km (000) Percent of Subtotal Excluding Pipelines Percent of Total Minerals by highway46,500 Other cargo by highway549,700 Minerals by rail84,900 Other cargo by rail208,700 River transport, all cargo11,400 Lake Titicaca, all cargo31,400 Air transport, all cargo13,400 Subtotal945,900 Pipeline, crude petroleum1,923,000 Pipeline, petroleum products176,000 Total3,044,900 4.9% 58.1 9.0 22.1 1.2 3.3 1.4 100.0% 1.5% 18.1 2.8 6.9 0.4 1.0 0.4 31.1% 63.1 5.8 100.0% Source: Daniel, op. cit., footnote 11, Vol. 3, p. 21. complexity of this information the data were aggregated in the field by employing a two-fold classification procedure; the first was geographic and involved designating flows by centroids, and the second was a commodity classification which aggregated similar commodities. Geographic classification involved the subdivision of Bolivia into zones determined by economic-demographic criteria which separated major urbanized areas. The most populous urban place within each zone was then designated as a centroid, and all traffic to and from a zone was considered to originate and terminate with the centroids. This is not an unreasonable assumption considering the concentration of economic activities in the largest Bolivian cities and that major bulking operations generally occur in these urban areas. (12) The centroid classification identified 14 import-export points and 31 internal urbanized areas for a total of 45 centroids (Figure 2) . A commodity classification was designed to group commodities which were similar in "the nature of material from which they originate and by the amount of processing they have received." (13) For example, an agricultural commodity such as wheat would be classified as origin-agriculture; unprocessed, wheat; semiprocessed, flour; 14 Southeastern Geographer Riberolto BOLIVIA Commodity Group I Rurrenobaque .|V_- Io ???\ ^^_ I»uaqui ¦ f^% £ I ^^^ I VJ f /jnqui-m Villa Tunori ^^1^^ """""IrVi1IChO/?\___- — f ~-.______ ^^«^^ I #\Choror__ / ' \ \Coehobon_JO BelloA/i-ta ~-___.SonJo! ^ \ \ __,'' «\V^ / · , ^^—— R, \ I #Ur,cìi 1 Rio Muloto ' ' I I Potoi! / / V^uRi 200300400 Functional Regions Nodal Regions — — km Figure 2. A functional regionalization. Agricultural and forestry products. processed, bread. This system yielded a 16 commodity classification system (Table 2) . Three commodity types were not analyzed in this study because of limited tonnage. These were live plants, crude oil, and miscellaneous products. Commodity flow data for all modes were then aggregated and organized into 16 matrices corresponding to the commodity types. Each matrix consisted of 45 rows and columns, and individual cells Vol. XVI, No. 1 15 within each matrix contained the flows from origins to destinations (centroids), for each commodity type in hundredweight per week. ANALYTICAL STRATEGY. The organization of the nation's space economy is identified in a five step analytical approach. First, flow data are organized by commodity type in a dyadic matrix in which flows from one place to another are treated as observations. Thus the original 16 data matrices are reduced to one dyadic flow matrix which consists of ? rows representing dyads in the study area (l-n,2-n—45-n), and m columns (1-16) representing the different commodity types. Second, similarities in flow patterns by commodity type are isolated by a components analysis of the dyadic matrix. (14) Commodity flows are then reorganized in origin-destination matrices where each matrix contains similar flow patterns as revealed by the components analysis. The origin-destination matrices within each group are converted to transition probability matrices and summed. (15) The identification of functional regions by Markov chain analysis constitutes the next step in the data analysis. Markov chain analysis abstracts a matrix of mean first passage times (MFPT) from the summed transition probability matrices. The Fn element of the MFPT represents a measure of functional distance reflecting the propensity for any two nodes to interact. (16) Similarities in functional distance values in the rows and columns of the MFPT matrix are identified to initiate the regionalization procedure. (17) The results of the Markov chain analysis are then subjected to a hierarchical grouping algorithm constituting step four in the data analysis. Functional regions are delimited on the basis of minimizing within group variance for MFPT's. Regions are then mapped in the final step concluding the analytical strategy. EMPIRICAL FINDINGS: COMMODITY GROUPS. Four factors, explaining approximately 65 percent of the variance in the data, were extracted in a components analysis of the dyadic matrix. The first factor was associated with agricultural products, including unprocessed commodities, semiprocessed, and processed agricultural goods. Semiprocessed and processed forestry products also had high loadings on Factor I (Table 3). Dominant flow patterns were associated with major agricultural producing regions and with zones noted for the 16 Southeastern Geographer TABLE 2 COMMODITY CLASSIFICATION SYSTEM Raw and Unprocessed Semiprocessed Processed and manufactured 3 tic Codel Wheat and other grains Raw and baled fibers Fresh fruits Legumes and vegetables Seeds, nuts Code 2 Flour Other milled grains Sugar, semi-refined and refined Dried fruit Coffee, cocoa, tea Code 3 Food: canned fruits, vegetables, juices, etc. Tobacco Beverages Vegetable oils Confections Code 4 1—t I Live animals¦g Wool and other fibers¦*