Abstract

In the second half of the 20th century, the behaviour of the world's population has undergone drastic changes: it has grown very rapidly, become longer-lived and moved preferably to urban areas. Migration of populations to cities has a special focus because it not only generates socio-economic development, but also causes different economic, social, and environmental issues. How to assess the sustainable development of cities and identify the most representative aspects of sustainability dimensions has become an essential challenge for effective planning and government policies. However, the selection of indicators has scarcely been addressed when Data Envelopment Analysis (DEA) has been used for sustainability assessment. This paper aims to identify the key indicators to determine the eco-efficient performance towards a sustainable development of a set of cities. For this purpose, the joint use of a variable selection method and a DEA model is proposed and applied to a set of 31 Spanish cities. The results obtained identified five social (population percentage at risk of poverty, number of registered gender violence cases, population density), five economic (average household income, average rental price per m2, gross domestic product per capita, public/private vehicle ratio, indebtedness), and three environmental (ozone, NO2, PM10) indicators for the assessment. Moreover, 14 Spanish cities were considered non-eco-efficient, and targets to achieve eco-efficient levels were proposed. Policy makers could focus primarily on the social and economic aspects of non-eco-efficient cities, paying attention, for example, to the risk of poverty of the population, indebtedness of local governments, average rental price and NO2 emissions.