Zurich—Switzerland (June 11-13, 2014)
Spatial econometrics is a well-consolidated body of methodologies for the analysis of externalities, spillover and interactions with applications in so many diverse scientific fields such as regional economics, transportation, criminology, public finance, industrial organization, political sciences, psychology, agricultural economics, health economics, demography, epidemiology, managerial economics, urban planning, education, land use, social sciences, economic development, innovation diffusion, environmental studies, history, labor, resources and energy economics, food security, real estate, marketing, and many others. The VIII World Conference of the Spatial Econometrics Association is an annual conference of the association to promote the development of theoretical tools and sound applications of the discipline. The conference offered a forum for discussing methodological advances and empirical results in all applied fields and encouraged such knowledge and good practice in academic and research institutions and in the society at large.
The focus of my research is road transportation. Road Traffic Crashes (RTC) are a global scourge characteristic of our technological era, whose list of victims insidiously grows longer day by day. The objective was to propose models that account for spatial effects to explain the dynamics of RTC. Thus, spatial autoregressive (SAR) model, SAR model with SAR disturbances (SARAR), and SARAR model with additional endogenous variable (SARARIV) were calibrated. To allow for comparability, the Traditional Classical Regression Model (TCLRM) that do not account for spatial effects was calibrated. The parameter estimates for the exogenous variables, that is, population; travel density; land area were positive, while, that of major road length was negative. The estimated was 0.37; 1.37; 1.20 with p values equal 0.06; 0.00; 0.00 for the SAR, SARAR and SARARIV models respectively. This indicates RTC were clustered around LGAs rather than the expected random distribution. The estimated was -1.43; -1.18 with p values equal 0.07; 0.19 for the SARAR and SARARIV models respectively. This suggests, an exogenous shock to one LGA will cause moderate changes in the neighbourhood. The highest concentrations and hotspots were found to be in Egbeda, Oluyole and Akinyele LGAs. The log – likelihood, Akaike Information Criterion and Schwarz Criterion indicated the proposed SAR models where better fits in comparison to the TCLRM. Accounting for spatial effects can positively impact model fitting and provide a means for linking RTC with neighbourhood characteristics. Thus, the proposed SAR models are more informative, efficient and valid. The framework should enable the orientation of safety and injury prevention policies.