Segmentación De Municipios Andaluces Según El Coste Efectivo De Los Servicios Públicos: El Caso Del Alumbrado Público

Resumen:La necesidad del uso eficiente de los recursos públicos a nivel municipal exige disponer de información relativa a la forma en que éstos se prestan. Los costes de servicio tanto de municipios como de Diputaciones provinciales vienen condicionados por factores diversos como el planeamiento urbano o la dispersión poblacional; dificultando la comparabilidad. Por ello, en este trabajo se realiza una propuesta para la construcción de índices sintéticos municipales relativos al servicio de alumbrado público en Andalucía a nivel provincial y regional. Posteriormente, se utilizan para modelizar el valor del índice a partir de aquellas variables que tengan mayor capacidad explicativa.Abstract:The construction of municipal synthetic indexes that report on the provision of public services is nowadays a necessity if the aim is to efficiently manage the always scarce budget allocations. This work responds to this demand through a double way, on the one hand, through a methodology for the elaboration of said indicators and, on the other hand, estimating a model that would simulate the value of the indices. For this, and as an example, attention has been focused on the public lighting service in Andalusia and on a double territorial level: provincial and regional, and within the latter, differentiating into two groups according to the population of their municipalities. The construction of the indexes will be based on the information provided by the EIEL corresponding to 2014, which includes both the installed power and the number of light points. To this have been added ratios such as the average power per point of light or per inhabitant, and the points of light per linear meter of track or per inhabitant, thus removing the size effect. All these variables are closely related to each other, given that road length or the number of light points depend on the resident population, which determine the installed power. Since the variables are not independent, that is, they have joint information, by means of multivariate analysis techniques they can be described using a smaller number of dimensions (main components), taking advantage of the interrelationships between the variables for each group of municipalities. From that moment, and through the factorial analysis, a general index can be defined by combining said dimensions, so that only by introducing the initial variables in the expression obtained, the situation of each municipality in terms of public lighting will be quantified. In this way, a classification of the Andalusian municipalities can be made that can explain a certain underlying model in the set of information provided by the individual variables. Within this type of methods, it is worth underline Factorial Analysis, a term used to collect a variety of statistical techniques that allow the evaluation of the underlying interrelationship structure in an observed data matrix. This technique, which is widely used in many disciplines, can be exploratory or confirmatory, depending on the number of factors to be extracted or fixed a priori, respectively. In our case, an exploratory factor analysis will be applied where the Principal Component Analysis will be used. This technique, commonly used by many scientific disciplines, is a simple and quick way to extract relevant information from data sets (in general correlated) in which, apparently, there is no pattern of behavior. Once the factorial extraction has been carried out, those variables that contain more information can be selected, as well as obtaining the factorial scores that will allow the elaboration of a synthetic index that eliminates the subjectivity typical of the classical methods. Specifically, the index is obtained as a weighted average of the scores of the main components for each municipality, where the weights are determined by the square root of the variance of each component, which guarantees that the components with the greatest variance explained have a greater weighting in the index. To the extent that the variables have been typed during the process, the resulting index will take positive or negative values, depending on whether they are above or below the average, therefore, in order to improve the interpretation, the index has been re-scaled to a range 0-100. The first of the detected factors (which we could call "municipal size") has a high correlation with variables such as power, light points, population and length, since the larger the population, the same will be in road length and, therefore, in points of light, which in turn condition the installed power. This relationship is more evident when the municipal population is smaller, since in these cases the lower demand for land (and its greater availability) result in an extensive urban development instead of in height. The second factor (which we could call "density") is at least integrated by the installed power and the points of light per inhabitant. These variables tend to have a negative correlation with the "municipal size", in the sense that the greater the population (and, therefore, its density in the urban nucleus, linked to the urban development in height), the lower the value of these ratios. Finally, the points of light per linear meter of track and the average installed power per point of light have been reduced to a single dimension (which would reflect the concept of "intensity" in terms of public lighting) for the provinces of Córdoba, Huelva and Jaen. Its impact on the public lighting situation is conditioned both by technical requirements (form of arrangement of the luminaires or the type of existing lamps), and by the urban model (width of the roads and dispersion of the buildings). From a regional perspective, it is observed that the municipalities of less than 5,000 and between 5,000 and 20,000 inhabitants follow the general pattern observed above, with the only difference being the greater importance of the power per light point and the smaller of the points of light per inhabitant for the last group. As for the municipalities of between 20,000 and 50,000 inhabitants, almost all the variables related to the public lighting system (in size, density or intensity) are integrated into the first component. After obtaining the corresponding index for each territorial area, it is observed that the highest values at the provincial level (of less than 5,000 inhabitants) are for inland municipalities, in mountain areas and with low population density. This dispersion in the territory means that public administrations are obliged to provide a minimum of lighting service, whose cost (total or in relative terms) is greater than if the population were concentrated in a single urban nucleus. At the regional level and for municipalities of between 5,000 and 20,000 inhabitants, areas with higher index values are appreciated such as the Guadalquivir riverbank in the province of Seville, the metropolitan area of Granada, the south of Cádiz or the northeast of Almeria. Of the rest, the majority corresponds to interior municipalities that constitute first-order regional centers such as Guadix in Granada, Baeza, Bailén and La Carolina in Jaen, or Marchena, Cazalla de la Sierra and Las Cabezas de San Juan in Seville. Finally, for the 51 Andalusian municipalities whose population is between 20,000 and 50,000 inhabitants, the highest rates are found in the Guadalquivir Valley and in coastal areas (especially in Cádiz and Huelva). This greater population is determined to a large extent by the greater extension of said municipalities, as well as by the greater economic activity. However, it would be possible to differentiate them according to presenting a strong specialization in tourist or industrial activities. The indices have been used in the estimation of a model for each territorial area that has made it possible to estimate their value based on those variables that have the greatest explanatory capacity. As a result, it has been observed that in the eastern provinces the main explanatory variable is the number of points of light (either in absolute or relative terms). The smaller size of the municipalities, their greater number and the complicated orography of these provinces, would explain that circumstance, which is consistent with the fact that light points per linear meter of track are also an explanatory variable for Almeria and Jaen, or the length for Granada. On the other hand, for Córdoba and Seville it is the power, and the per capita power for Malaga, which is consistent with its greater specialization in industrial and tourist activities. At the regional level, for the municipalities of between 5,000 and 20,000 inhabitants, the population is the variable with more weight in the index, followed by the power and points of light, both in terms of per capita. On the other hand, for municipalities of between 20,000 and 50,000 inhabitants, the points of light per inhabitant are the only explanatory component of the synthetic index.