Rural roads, education, and agriculture: A micro-econometric evaluation study using Trinidad and Tobago data
This dissertation evaluates the effects of infrastructure investments on agriculture and education in the rural communities of a small Caribbean economy. Specifically, it provides an empirical test of the causal impact of an access-road rehabilitation program on school attendance, attainment, agricultural labor supply, input choice, crop harvest, yield, and commercialization, and it investigates the program's targeting effectiveness. Data are drawn from population censuses taken in 1990, before the program started, and 2000 and from an agricultural census of 2004. Appearing shortly after the program's end, the latter two datasets provide ex-post measurements that permit assessment of program impact. The study illustrates that evaluations of infrastructure investments are possible even where program-specific primary data are lacking, and that commonly available non-program-linked data (such as censuses), while not ideal, may provide the raw information required. Identification relies on propensity score matching (PSM) techniques to create a set of analytical samples that comprise statistically comparable groups of program and non-program communities. Results indicate that the program was effective in extending benefits to poor agricultural communities at the national level but that sub-national targeting was less successful. Road rehabilitation also fed increases in school attendance, of up to 16 percent and 18 percent for children in early childhood and advanced secondary levels, and promoted gains of up to half a year in attainment. Agricultural operations in the domestic food-crop sector were also affected, with a reallocation of labor supply away from fulltime farming, consolidation of ownership among larger operators, and greater use of newer technologies, particularly in the production of certain cash crops. Robustness checks, including use of alternative welfare measures to assess targeting, and varying sample composition and PSM weighting schemes to measure impacts, left the overall conclusions unaltered. These findings represent new empirical evidence on the impact of infrastructural improvements in the Caribbean. They reveal that social and productive sector constraints can be ameliorated with targeted programs, although benefits may accrue unevenly. And they suggest a role for complementary public investments to buttress potential gains to all communities, and to increase the progressivity of distribution of these gains amongst weaker ones.