Peter Nijkamp (p), Soushi Suzuki, Karima Kourtit

Over the past decades, the Asia-Pacific Rim has exhibited an unprecedented high degree of economic and geographic dynamics. Clearly, cities in this region display a heterogeneity in terms of economic performance, technological innovativeness, environmental conditions, and cultural recognition and interaction. It is, therefore, interesting to develop an efficiency ranking of the multi-dimensional performance of these large cities so as to identify ‘super cities’. The first aim of this paper is now to undertake a multi-faceted performance ranking of large cities in the Asia-Pacific region by using a DEA (Data Envelopment Analysis). However, there appears to be a wide variety of DEAs in the recent literature. And therefore, a second aim of the present paper is to perform a sensitivity analysis on the type of DEA employed, so as to test the robustness of the base ranking obtained from a standard DEA. A third aim of the paper is related to the question how much the ranking obtained by a DEA is influenced by the internal characteristics of the underlying data system. This leads to a sensitivity analysis of the precision or nature of the data used in the DEA. These three aims of the research will be empirically addressed by using a comprehensive data set on 7 quantitative main indicators regarding economic performance, technological innovativeness, environmental conditions, and cultural recognition and interaction for 13 Asia-Pacific super cities.

Rodger Campos (p), André Chagas

The monocentric city model have received much criticism about how feasible is this pattern with respect to big cities. Urban landscape changing is conditional to economic development movements. Employment decentralization inside city is one of the movements that foster urban landscape modification, influencing wage, housing prices, transport demand and worker commuting. Some theoretical models concerned about multiple centers brought to the debate the urban employment sprawling.
On regarding empirical research, there is no well-established empirical strategy of identification. Initially, the studies use prior region knowledge as identification but results did not fit the data. More objective procedures divide the methodologies as: a) cut-off values: researchers impose some cut-off on the density of jobs and the total number of jobs, b) parametric econometric models and non-parametric models and c) models using spatial statistics.
The cut-off method is very sensitive to the cut-off value and unit of analysis because it uses density of employment. These models receive criticism due to the discretion and sensitivity of the results and the impossibility of generalization to any city.The next two procedures aforementioned are transformations of the pioneer method. The method takes into account the residue of a log density specification of employment, considering the rings that present high employment density and employment level, given their knowledge of the city. The criticism of this approach stems from the estimation of a symmetric density function around the CBD. Shortly, all of these approaches use gross employment density (employment / area) are sensitive to the size of the area used. Thus, larger regions tend to have smaller densities (and vice versa), confounding the identification of subcenters.
To overcome these limitations, a two-stage approach is proposed, incorporating a priori identification of candidates for centrality into the hedonic price model. Thus, in the first stage, it identifies the all candidates for SBD, using an Exploratory Spatial Data Analysis and, in the second stage, we run the Spatial Hedonic Price Model considering as explanatory variables the distances between SBD candidates (identified in the first stage) and housing localization. Thus, in both approaches, the one can test the null hypothesis that each region is or not is an SBD, which is not possible using all other approaches.
As result, we found seven regions that can be considered as SBD. These regions are able to impact housing prices as predicted by policentric theoretical models.