Luiz Henrique Paloschi Tomé

The objective of this research was to analyze the multidimensional rural development and the determinants of the price of agricultural land in Brazil from 2006 to 2017, in addition to the relationship between rural development and land prices. In the first essay, the objective was to measure and analyze rural development in Brazilian micro-regions in 2006 and 2017 from a multidimensional definition of development. In the second essay, the objective was to analyze the existence or not of a convergence process of rural development in Brazilian micro-regions between 2006 and 2017, evaluating the spatial dependence in this process. In the third essay, the objective was to analyze the behavior and determinants of agricultural land prices in Brazil and their spatial dependence by measuring a panel data model that incorporated the multidimensional rural development index (IDRM). According to the results, the average IDRM for 2006 was 0.2208 and for 2017 it was 0.2644, showing an increase in Brazilian rural development in the period. The Midwest region was the one with the highest IDRM in both 2006 and 2017, followed by the South, Southeast, North and Northeast regions, with the same composition in the ranking in both years. The North region had the highest percentage growth in the index, followed by the South region. In both 2006 and 2017, the micro-regions with medium or higher IDRM are concentrated mainly in the Southeast, South and Midwest, while those with development below the medium level have a strong presence in the North and Northeast. The global Moran’s I calculated on the 2006 and 2017 IDRM data showed the positive spatial relationship for rural development. There was also the existence of an extensive high-high cluster in the states of the Southeast, South and Midwest and a low-low cluster predominantly located throughout the interior of the states of the Northeast. In the models of absolute and conditional β-convergence, the coefficient associated with the initial level of IDRM was negative and significant, proving the existence of a process of spatial convergence of rural development in Brazilian micro-regions from 2006 to 2017 and the tendency to reduce inequalities, with a half-life of 12.73 years in the absolute β-convergence model and 10.32 years in the conditional β-convergence model. There was positive spatial autocorrelation of the non- modeled effects and the overflow of shocks from neighbors on the local IDRM. It was also observed that micro-regions that had other micro-regions with high levels of IDRM in their surroundings in 2006 showed, on average, higher rates of growth of IDRM. In addition, the presence of family farming in neighboring regions has been shown to positively influence rural development. As for the price of land, it was possible to observe higher prices in the states of the South and Southeast and lower prices in those of the Midwest, Northeast and North. The global Moran’s I calculated on the land price showed the positive spatial relationship for the variable. There was a high-high cluster in the southern states and São Paulo and a low-low cluster in Roraima, in addition to low-low clusters in some states in the Northeast. As for the econometric model, Moran's I of the residues and the AIC and BIC information criteria, indicated the spatial autoregressive combined model (SAC) as the most appropriate. The variables asset valuation, demographic density, Rural Territorial Property Tax (ITR), rural credit and IDRM, have shown to positively influence the price of agricultural land in Brazil. It was observed that increases in the prices of neighboring regions positively influence the price of local land, although with little effect, in addition to the existence of negative spatial autocorrelation of the non-modeled effects, which highlights the overflow of shocks from neighboring regions on the price of local agricultural land.