Spatial issues in economics and econometrics

Abstract: This thesis consists of four self-contained papers, divided into two parts where the first part considers issues in economics and the second discusses spatial issues in econometrics. In paper [1] the different waste disposal options, with emphasis on recycling and incineration, is examined and evaluated, using earlier research. This paper concludes that the literature reveals a range of results regarding the economics of waste paper recycling and incineration, respectively, and that the conceptual and empirical basis on which to determine efficient waste paper policy is still seriously incomplete. The effort in paper [2] is to identify and analyze determinants of inter-country differences in recovery and utilization rates, respectively. The paper concludes that the degree to which policy can affect these rates are limited since relative waste paper recovery and use are largely market-determined, and consequently depend on long-standing factors such as population density and competitiveness in the world market for paper and board products. Paper [3] explores the differences in inferences that one would draw from different econometric models in a spatial econometric setting. The study notes that ordinary least squares to a very large extent produce biased estimates due to spatial correlation in the data set. Hence, one would draw very different inferences from ordinary least squares and general spatial model estimates. Finally, paper [4] builds on paper [3] but explores some Bayesian estimation methods, i.e. heteroscedastic models, which take into account non-constant variance or spatial outliers. The data set used in these studies were limited because of censoring. The objective of this paper was to obtain Bayesian estimates that account for outliers and sample censoring. We found that ignoring the spatial autoregressive nature of the data, outliers and sample censoring would produce different inferences than the Bayesian models.