Header image

S17-S2 W on the Focus

Tracks
Special Sessions
Thursday, August 31, 2017
2:00 PM - 3:30 PM
HC 1315.0042

Details

Conveners: Jesus Mur, Ana Angulo / Chair: Eduardo Castro


Speaker

Dr. Elzbieta Antczak
Ph.D. Student
University Of Lodz, Faculty of Economics and Sociology

Building W Matrices Using Selected Geostatistics Tools. Empirical Examination and Application

Author(s) - Presenters are indicated with (p)

Elżbieta Antczak (p)

Discussant for this paper

Eduardo Castro

Abstract

See extended abstract

Extended Abstract PDF

Full Paper - access for all participants

Agenda Item Image
Prof. Davide Fiaschi
Full Professor
Università di Pisa

Mallow Model Averaging for Spatial Matrix Selection

Author(s) - Presenters are indicated with (p)

Davide Fiaschi (p), Lisa Gianmoena, Angela Parenti

Discussant for this paper

Elżbieta Antczak

Abstract

See extended abstract

Extended Abstract PDF

Prof Eduardo Castro
Associate Professor
University Of Aveiro / Govcopp

Insights from Geography, Regional Science and New Economic Geography to define the structure of spatial weighting matrices - W

Author(s) - Presenters are indicated with (p)

Eduardo Castro (p), Arnab Bhattacharjee, João Marques

Discussant for this paper

Davide Fiaschi

Abstract

The analysis of socioeconomic phenomena over geographical space is typical based on the partition of relevant spatial domain into regions defined as a bundle of elementary spatial unit that share some king of similarity or interdependency. Thus, the understanding of structure of a geographical space can be accomplish analyzing: i) spatial patterns, where each region (locality) is defined by a specific set of homogeneous characteristics; and ii) spatial interaction, where the focus is on the interdependencies of each region (location). Here, we are focusing on the spatial interaction, linking the theoretical aspects of spatial analysis and the empirical approach of spatial econometrics. Thus, our main purpose is to present a conceptual framework that can support the way W should is specified.
The theoretical perspectives here considered can be organized according the answers they provide to the following two questions: i) What is a region? (a) Is it a fundamental analytical unit which can be characterized by its specific properties? Or (b) Is it better conceptualized as an aggregation of smaller units which are the basic analytical units? and ii) What drives socio-economic dynamics in a region? (a) Is it institutions which defined the behavior of individual agents? Or (b) Is it the options of autonomous but interrelated agents determined by the optimization of individual decision rules?
These answers defined three main theoretical approaches or disciplines: i) human geography – describes localities as a unique place under a historicist and inductivist perspective. Typically, but not exclusively, such rationale can lead to a general refusal of quantitative measurements and statistical analysis. ii) Regional Economics - mainly concerned with location, follows a methodological approach based on the deduction of patterns. The usual approach to the representation of spatial interactions is to define a spatial weights (interactions) matrix, which typically represents a theoretical and a priori characterization of the nature and strength of spatial interactions, either based on distances or contiguity, or otherwise estimated from the data. iii) New Economic Geography - has established a research agenda directed to the analysis of both locality and location as the outcomes of objective behavioral rules which shape equilibrium landscapes and evolutionary paths. Such paths can, at least theoretically, be deduced and defined as endogenously determined spatial weights (or interaction) matrices.
Taking the above conceptual frameworks it is possible to assume different perspectives regarding the way how spatial interactions and specifically the W matrix is defined.
loading