Infrastructure Upgrades and Rural–Urban Connectivity: Distance Disparities in a Tri-National Frontier in the Amazon

Abstract

New infrastructure reorganizes spaces by modifying connectivity among locations. We report findings pertaining to rural–urban connectivity in the southwestern Amazon, a tri-national frontier with varying levels of infrastructure due to recent highway paving. We compare distances to nearest markets and distances to primary markets reported by community leaders. The findings show many differences, so we model distance disparities featuring variables derived from location theory. The models show that communities without road paving are more often connected to more distant markets. During the process of infrastructure upgrades, regions experience adjustments in connectivity among locations that deviate from equilibrium expectations of location theory.

La nueva infraestructura reorganiza los espacios modificando la conectividad entre las localizaciones. En este escrito reportamos hallazgos relacionados con la conectividad rural–urbana en el Amazonas sudoccidental, que es una frontera tri-nacional con variados niveles de infraestructura debido a la reciente pavimentación de las carreteras. Comparamos las distancias a los mercados más cercanos y distancias a los mercados primarios según informes de líderes comunitarios. Como los hallazgos muestran muchas diferencias, modelamos las disparidades de distancias a la luz de variables derivadas de la teoría locacional. Los modelos muestran que las comunidades sin pavimentación de sus vías a menudo están conectadas con mercados situados a mayor distancia. Durante el proceso de actualización de la infraestructura, las regiones experimentan ajustes en la conectividad entre localizaciones que se desvían de las expectativas de equilibrio asumidas por la teoría locacional.

Key Words:

• Amazon
• connectivity
• infrastructure
• road
• rural

关键词:

• 亚马逊
• 连通性
• 基础设施
• 道路
• 农村

Palabras clave:

• Amazonas
• conectividad
• infraestructura
• carretera
• rural

Notes

1. We use informants as the unit of analysis rather than communities. This is because some informants gave differing markets in their responses. We weighted our informants based on the size of their communities and the number of informants per community. The analysis therefore reflects communities in terms of their size, without losing information from different community leaders. One might question the use of data from informants to draw conclusions about communities. We suggest that community leaders serve as informants insofar as such leaders represent community members and adjudicate relationships involving community members, including relationships with buyers. We recognize that there is some theoretical debate about relations among individuals and aggregate social entities (cf. Giddens 1984), including over the question of whether it is only individuals who exercise agency by taking decisions or whether social aggregates such as communities can also exercise agency (Hindess 1986; Long 2001).

2. To obtain urban populations for the market towns, we consulted demographic censuses for Bolivia, Brazil, and Peru. Both Brazil and Peru conducted population counts in 2007, and we employ municipal and district urban populations for that year (IBGE 2008; INEI 2008). The most recent Bolivian census was in 2001. We therefore used official projections (INE 2008) to obtain an estimated 2007 urban population for Cobija, the only officially designated urban area. Although Bolivia does not recognize other market centers as officially urban, census data do list communities as populations attached to central places. We therefore used those 2001 population figures, projected to 2007, to obtain population estimates for other central places aside from Cobija. In Acre, Brazil, the towns of Brasiléia and Epitaciolândia are literally side by side, so we combined their populations to constitute a single market area. Cobija is also across a river from Brasiléia and Epitaciolândia, but we consider Cobija separately. Bolivian and Brazilian informants differentiated Cobija from Brasiléia and Epitaciolândia.

3. One could argue that this finding is an artifact of our not having combined Brasiléia and Epitaciolândia with Cobija. We therefore reran the analysis using the combined populations of Brasiléia, Epitaciolândia, and Cobija. The results are very similar to those reported, and the interaction terms have similar coefficients and levels of significance.