Cities are at the forefront of societal change, addressing both to new challenges and new opportunities. Under the pressure of "World city formation" the cities integrate into a "functional city system", and therefore they cannot be fully understand independently of the network in which are embedded. How is this network structured? Which are the main relations that give shape to the dynamic and spatial organization of the network? How does it evolve? The literature about this issue is vast but it only focuses on the node's attributes, without being able to go into functional analyses of the linkages. The research's idea is to analyze the network from a relational point of view, using shock transmissions among cities as a proxy of the edges and to implement traditional macro-econometric tools to measure and map them. Panel Vector Autoregressive model (PVAR) can be applied to study the correlations among cities, and, through "Time Varying" technique, to investigate the dynamic patter of the network. Furthermore "Endogenous Cluster Factor" model will allow finding relevant features shared by the cities, based on the data point of view, as well as the transmission channel of the network.
The new methodology is a novelty in the current research field: a quantity approach directed to estimate the relation among cities has never been done yet. In this sense it can really contribute to a better understanding of the city network.
Once the PVAR model is estimated, according to the function (variable) and importance (coefficient size), the relative position of a city in the system will emerge. For each city it will be possible to have a measure of its capacity to affect the other cities, and, according to the spatial extent of this capacity, it will provide a map their relational geometries (Yeung, 2004).
Additionally the model will provide insight on the structure of the system: its evolution through time (Time Variation); the level of integration, which cities and how much they participate in the shock transmission (Variance Decomposition); which relation are symmetric (sign of the coefficient); the level of reciprocity (different parameter for each relational way); and the resilience of the network and dependency, what happen when a city has a negative shock (Impulse Respond Function).
The research can be use to evaluate current spatial strategies and to help the policy maker in future. In the European context the study can be implemented with the metropolitan areas, whose variables are available at Eurostat database. The research can be used to stress how the European Spatial Development Perspective (ESDP) is reaching its objective of polycentrism. Through this model it's possible to attest the evolution in time and in space of the European urban system: which are the global gateway, the regional leader cities, which cities are more integrated and according to what type of structure. E.g. some cities will more intertwined through the labour market, other through the financial one and other through the services. Moreover it's possible to attest how the MCRs perform in this system, through their links within and outside the region. It will provide information on the length of integration among city-regions and the relative importance of cities in term of intra-regional connectivities, i.e. the degree of polycentrism, and in term of global connectivities, i.e. which cities play the role of gateway for the region.
At global level the analysis will focus on the global cities defined by Sassen and Taylor (1991; 2001). The aim is to define a quantitative measure of capitalism and human flows among these cities. It will be even an occasion to measure the role of cities in the international transmission of economic shock and to measure their contribution to globalization.
In the second phase, the Endogenous Cluster Factor model accepts many possible urban characteristics that may affect the relations and let the data determine which cities share common factors above the other, providing useful clues for the identification of the network transmission channels. Possibly investigating if different structures of the network use different channels, supplying new instruments for policy maker aimed at increase cohesion and integration in the urban network.
The study of transmission channel can be enlarged to the monetary and fiscal policies transmission channel as well. Indeed it is a topic that still needs a strong empirical support, and the disaggregated data at the metropolitan level of analysis may bring new element for understanding the transmission channel and the origin of the heterogeneous effects (Dedola and Lippi, 2001; Neville and Owyang, 2011).
Moreover the high level of flexibility of the model contributes to check which covariates proposed by the theory actually matter: the role of multinational headquarters (Sassen, 1991; Taylor, 2001); the role of proximity such as geographical, institutional, and cultural (Boschma, 2016); the role of morphological and infrastructure endowment and those of performance and social indicators (ESPON projects).