Spatial Econometrics for Cross-Sectional Data
Theory, Algorithms, and Programming in R
Preface
These lecture notes have been prepared to provide an intermediate and advanced treatment of spatial econometrics for cross-sectional data. Their purpose is to offer readers a solid theoretical foundation while also providing the practical tools required to understand, implement, and apply the main estimators used in this field. I hope to expand their scope in the future by incorporating additional topics, including spatial panel-data models and nonlinear spatial models.
The philosophy of these notes largely reflects the way in which I learned econometrics during my own training. I believe that understanding theoretical results is essential, but I also believe that an important part of the learning process consists of understanding how estimators are constructed and programmed. Implementing an estimation procedure from its basic components makes it easier to see the logic connecting the assumptions, the statistical results, and the estimation algorithms. In other words, it helps open the “black box” that statistical software functions often represent.
This approach also seeks to develop programming skills that allow readers to adapt existing methods, identify potential errors, and work with models for which specialized functions or software packages are not yet available. For this reason, whenever possible, each theoretical chapter is followed by an applied chapter devoted to programming the estimators under study. The objective is not to replace specialized software, but rather to understand the methods well enough to use such software in an informed and critical manner.
The preparation of these notes began in 2016, when I taught Spatial Econometrics in the graduate programs of the Department of Economics at Universidad Católica del Norte. After a break of several years, I resumed the project in 2022 and have since continued to revise, expand, and reorganize its contents.
These notes are primarily intended for master’s and doctoral students who have previously completed courses in econometrics and statistics and who possess some basic knowledge of R programming. Nevertheless, they may also be useful to researchers seeking an introduction to spatial econometrics or wishing to deepen their understanding of the relationship between its theoretical foundations and computational implementation.