A Benchmark for a Metropolitan Geo-Statistical Framework in Mexico
Total Page:16
File Type:pdf, Size:1020Kb
A Benchmark for a Metropolitan Geo-Statistical Framework in Mexico Roberto Duran-Fernandez Transport Studies Unit Oxford University Working paper N° 1022 July 2007 Transport Studies Unit Oxford University Centre for the Environment http://www.tsu.ox.ac.uk/ A Benchmark for a Metropolitan Geo-Statistical Framework in Mexico (DRAFT: PLEASE DO NOT CITE WITHOUT PERMISSION) Roberto Duran-Fernandez § Oxford University Transport Studies Unit Working Paper No. 1022 July 2007 Abstract This paper presents a benchmark for the definition of standard metropolitan areas in Mexico. It proposes a criterion to identify metropolitan areas based on the analysis of demographic statistics and cartographic data. The paper identifies 69 metropolitan areas that resemble the Standard Metropolitan Areas (SMA) of the US Bureau of Economic Analysis. These metropolitan areas represent 56 per cent of national population and concentrate 77 percent of industrial output. They constitute the basis for a Metropolitan Geo-Statistical Framework in Mexico that enables the analysis of socio-economic at this aggregation level. § Transport Studies Unit, University of Oxford, Dyson Perrins Building, South Parks Road, Oxford, OX1 3QY, England. +44 (0)1865 285192. Contact: sann2513@ herald.ox.ac.uk 1 1. Introduction In this paper we present a benchmark for the definition of standard metropolitan areas in Mexico. We identify 69 metropolitan units that constitute the basis for a Metropolitan Geo-Statistical Framework in this country. The definition of standard metropolitan areas based on commuting patterns has a long tradition in other countries such as the United States (Berry 1977). Unfortunately, the lack of suitable data does not allow us to carry out a similar exercise for Mexico. The National Institute of Statistics, Geography, and Informatics (INEGI) does not publish a comprehensive catalogue of metropolitan areas, and the data that it produces is typically aggregated at state or municipal level. The most notorious exceptions are the labour statistics based on the National Urban Employment Surveys (ENEU), which define 44 integrated urban economic areas. In this paper we present a criterion for the definition of metropolitan units which resembles the criterion used in the US for the definition of standard metropolitan areas by the Bureau of Economic Analysis. The criterion is based on Berry (1977) and relies on the analysis of demographic statistics and cartographic data with geographic information systems. The motivation of this exercise is to propose an analysis framework that reduces and eventually eliminates the aereal problem that an arbitrary geographic aggregation of socio-economic statistics can generate. 2 2. Territorial Political Division According to the constitution of Mexico, the municipality or municipio is the basis for the territorial division in the country. A municipality is governed by an Ayuntamiento an institution elected by popular vote, which constitutes the local government tier in the Mexican federal system. The municipal government has several obligations and responsibilities according to the law. Among them, we can count the provision of water, and sewage systems, the administration of the local cadastre and land policy, the provision of local security and several urban services. This local government is mainly an executive authority that has to follow federal and state legislation. The municipal government does not carry out any judicial function either. These are responsibilities of federal and state judges. The main fiscal resources of the municipal government are federal and state grants and transfers. However, the constitution grants municipalities the right to tax real estate property. Local governments, on the other hand, are not responsible for local schools, a situation that clearly differs from the United States. Municipalities in Mexico are very heterogeneous both in their population and in territorial composition. Their population can range from less than 500 to more than one million. The larger and less dense municipalities are concentrated in the north of the country; on the other hand, in the south of the country we can find small and highly populated municipalities. The territorial division of the country into municipalities is not homogenous, and it was shaped according to which historical and political factors, which not necessarily portrait the spatial distribution of socioeconomic activities (Bassols-Batalla 1993). The arbitrariness of this political division can be illustrated by the following example: Oaxaca State, located in the south, is divided in 570 municipalities, while the northern state of Baja California is divided in only 5 municipalities. Both states have a similar surface (97.5 and 86.7 sq. km. respectively), population is higher in Oaxaca (3.4 versus 2.5 million), but production is considerably higher in Baja California (3.37 and 1.48 percent of national gross domestic product). 3 The territory of a metropolitan zone does not necessarily match the territory of a single municipality. The urban area of some cities may comprise all the territory of a municipality and in some cases, the conurbation extends itself over several municipalities. 3. Statistical Geographic Aggregation Several official statistics produced by INEGI are available at municipal level. This is typically the lowest geographical disagregation level available for any official statistics in Mexico.1 In most of the cases, the municipal aggregation of socio- economic statistics is statistical neutral. However, the aggregation of data in arbitrary geographical units can lead to biased statistics. This problem is especially important when the geographic statistic units exhibit a high degree of heterogeneity. Another problem is the possibility of over (sub) sampling. When a given study area is divided into an arbitrary number of geographical or area units, the statistics employed to describe it may differ according to the size and number of area units into which it has been divided. The nature of this problem is clear in the description given above about the number of municipalities that compose Oaxaca and Baja California states: non- weighted statistics based on municipal observations systematically overestimate the importance of Oaxaca’s observations and underestimate Baja California, generating biased estimators. The essence of the problem is identifying or eliminating differences in parameters, which may be merely attributed to differences in the size of the aereal unit from those differences, owing to truly different relationships (Thomas et. al. 1965). Robinson (1956) proposes as a solution for tackling the aereal aggregation problem the weighting of geographical observations. Ideally, the problem we face should be solved using weights such as, for example, the number of individual agents (i.e. firms, households, or individuals) represented by each geographic unit. 1 An important exception is census population data, which is available at town level. 4 Unfortunately, this possibility is only available for micro datasets. However, other types of datasets can be handled to tackle this problem. The most direct method is the aggregation of data in area units with a high degree of homogeneity. The most natural geographical unit that fulfils this requirement is the city or metropolitan area. Economic geography models suggest that that within the limits of a conurbation, economic agents are more likely to develop strong socio-economic interactions that are reflected in a high degree of homogeneity for certain socio-economic indicators. 4. Definition of the Metropolitan Geo-statistical Framework The arbitrary division of the country in municipalities can be problematic for the definition of metropolitan areas. Fortunately, available statistical and digital cartographic data allows us to identify those municipalities whose territory belong to a particular metropolitan area. This process enables us to build a metropolitan geo- statistical framework, which in principle can serve as the basis for the estimation of geographically neutral statistics and estimators at metropolitan level. Unfortunately, in most cases municipalities contain more than one urban agglomeration. Therefore, a geo-statistical framework based on municipal data can not contain all the urban areas in the country. INEGI defines municipal conurbations as those municipalities where one or more urban agglomerations are unified, such that there is an urban continuity among them. In other words, despite the political or administrative borders, there is no clear separation between streets or built-up areas that determines their limits, The identification and classification of the municipalities into metropolitan areas was made according to the following criterion: First, we identify a subset of metropolitan municipalities. This subset is defined as those municipalities with an urban population of 100,000 or higher. Each of these municipalities constitutes the core of each metropolitan zone. This criterion is uniform for all the municipalities except for Mexico City and Guadalajara, the two largest urban agglomerations. For these conurbations, we define the Federal District 5 and the municipality of Guadalajara as the core for each metropolitan area. If this distinction were not made, we would have two metropolitan areas within the same conurbation. Once the subset of metropolitan municipalities is defined, we follow a contiguity criterion to group municipalities. We identify the conurbation