Measuring Quality of Life in Spanish Municipalities

Measuring quality of life in municipalities entails two empirical challenges. First, collecting a set of relevant indicators that can be compared across the municipalities in the sample. Secondly, using an appropriate aggregating tool in order to construct a synthetic index. This paper measures quality of life for the largest 237 Spanish municipalities using Value Efficiency Analysis (VEA) to derive comparative scores by combining the information contained in 19 partial indicators. VEA is a refinement of DEA (Data Envelopment Analysis) that imposes some consistency in the weights of the indicators used to construct the aggregate index. The indicators cover aspects related to consumption, social services, housing, transport, environment, labour market, health, culture and leisure, education and security. The results show that the Northern and Central regions in Spain attain the highest levels of quality of life, while the Southern regions report low living conditions. Education is the variable that requires the largest improvement in low performing municipalities, followed by health and culture facilities, pollution and crime. Population density, growth and ageing seem to relate positively to quality of life. Key Words: Quality of lifewelfaremunicipalitiesDEAVEA Acknowledgements We gratefully acknowledge financial support for this research by the Ministerio de Educación y Ciencia (Plan Nacional de I+D+I: SEJ2007-67001/ECON). A preliminary version of this paper was published as a working paper N417 in the working papers series of Fundación de las Caias de Ahorro (FUNCAS). Notes 1 Using the lowest quantity of an input, for instance. This problem is also present and intensified in variants of DEA such as FDH. 2 Also, different studies deal with different territorial levels of analysis (nations, counties, regions). 3 Some authors also point to factor analysis as a valid aggregating methodology (Somarriba & Pena, Citation2009). 4 The DEA approach tries to reduce inputs to the minimum possible because they imply a cost in production. It also tries to increase outputs to the maximum because they have a positive value for the productive firm. In our setting, city drawbacks imply a cost of living in the municipality and should be reduced to a minimum, while advantages imply a benefit for citizens and should be increased to the frontier maximum. Thus, the parallelism is clear and the applicability of DEA to our research setting is granted. Throughout the paper we will refer indistinctly to inputs–drawbacks and outputs–advantages. 5 We describe the dual DEA programs instead of the more usual primal specifications because we will use the weights of inputs and outputs in these dual programs to perform the VEA analysis. Anyway, the primal specification would, of course, reach exactly the same solutions and provide the same performance indicators. 6 We used the software LINGO to solve the DEA and VEA programs of this research. While many packages are pre-programmed to solve DEA, we are not aware of any that can solve VEA. However, any mathematical programming software can be used to solve (4). 7 In the computation of this index, INE uses class marks that go from 0 (unemployed) to 3 (entrepreneur). 8 To compute this index, La Caixa takes into account the population, number of phones, automobiles, trucks and vans, banking offices and retail activities. In order to make this index comparable across municipalities we divided it by the population and multiplied by 25 000. 9 To make the numbers comparable we divided the total number of facilities by the population and multiplied by 25 000. 10 For the computation of the index, INE uses class marks that go from 1 (illiterate) to 10 (PhD). 11 This index, elaborated by INE, ranges from 0 to 100 and takes into account factors of the buildings, such as the age of construction, tumbledown status, hygienic conditions, running water, accessibility, heating, etc. 12 The raw data distinguish between these two destinations or the variable is the arithmetic average of both. We must also indicate that INE does not compute an index associated with these variables. Instead the report includes the percentage of people on seven intervals that go from 'less than 10 min' to 'more than 90 min'. We took mark classes in the mean of the intervals (90 for the last interval) and weighted each class mark by the percentage of the population within the interval. The weighted sum can be interpreted as the average time employed to get to school or work and is the variable used in this paper. 13 In the other dimensions, it is around the mean, although far from the best performers. 14 Boadilla del Monte is a municipality in Madrid that excels in many dimensions (education, socio-economic conditions, housing, pollution). In contrast, its citizens must incur costly hours driving to schools or work and the level of facilities (health, cultural, etc.) is relatively low. 15 Security was the main variable to account for by citizens, with an average weight of 18 per cent, then the labour market (15 per cent), housing (13 per cent) and health services (12 per cent). 16 Therefore, it is the only one that can be used as MPS. Barcelona, Madrid and Valencia could not be considered as the MPS because the VEA program would not have a feasible solution because the cities are not on the DEA frontier. 17 Other good candidates to be the MPS were Vitoria, Getxo and San Sebastian. However, we were not able to find the independent support of other studies, as with Pamplona. We repeated the VEA analysis with these municipalities as MPS and found no important differences. 18 In the DEA program, Boadilla del Monte assigned a weight 0 to communications and time to work or school. Although it still is a good place to live, it is no longer a referent (frontier) under the VEA formulation. 19 We believe that using the VEA frontier for this purpose would be misleading since some reference points are, in reality, unattainable and are just taken to approximate the comparative value of the variables of a municipality. 20 Note that the DEA value reported before is equiproportional and, therefore, is limited by the variable that admits the lower improvement. Table 5 adds to these equiproportional improvements the slacks that the linear program finds in all the quality of life dimensions. 21 Caja España also provides on its webpage a municipal database, but most of the information is taken from the INE statistics. 22 Only one municipality with a population over 25 000 was excluded because data on journey times and university studies were not reported in the INE database. This municipality is La Vall d'Uixo (Castellón).