A major challenge for European and global policymakers in the wake of the financial crisis is re- kindling economic growth. A return to growth will help reduce both government and private debt levels, reduce unemployment, and help alleviate the political and social tensions that have arisen due to stagnant economies or recessions in many nations and problems in the Eurozone. Economic growth, however, is a highly complex and disaggregated phenomenon. Growing GDP at the national level is a function of growth in individual firms and sectors, the creation of new products and services, and actions across complex global supply chains and networks of productive capabilities. Traditionally, economists have analysed growth in an aggregate fashion at the global, national, or industrial sector level. However, recent developments in ICT tools, the availability of “Big Data”, network analysis techniques, and ideas from systems science and complex systems theory enable a much more disaggregated, bottom-up, evolutionary view of economic growth. This disaggregated point of view naturally implies considering also the strong interactions between the various components. This is the challenge of global system science which requires a fundamentally novel attitude in terms of defining the appropriate questions, introducing new methods and algorithms, dealing with immense amount of data and finally turning all these elements into practical recipes which should be relatively simple to understand and implement.
The aim of the proposed project it to make some fundamental progress in this direction by integrating and developing together some very interesting and promising steps which have recently emerged in this context. This is probably the first attempt to develop these ideas coherently for fundamental economics and we expect these ideas to have a large impact in various directions: new questions and challenges economic policy addresses to science and new questions and challenges science faces when investigating economic policy and decision making in a globalized market environment.
The standard approach to analyse systems in terms of their individual parts as independent entities and in a sort of equilibrium framework, has proven to be inappropriate in forecasting systemic crisis or in many of the complex situation corresponding to a strongly interacting globalized market. The example of the recent financial crisis has led to a large interest in the study of interconnected financial systems, with various EU projects. Here we intend to implement this perspective beyond the limits of finance by addressing some basic questions for the very foundations of economic science. Growth and innovation will be considered as an ecosystem with complex interactions. This we can define as economic complexity.
It is doubtful that a single model might address all those issues. Even if one embeds all existing heterodoxies, very few of the policy-relevant questions of the time can be answered. Among the challenges to be faced in order to overcome this limits are the disaggregation and the extension to non-equilibrium phenomena of economic models and the exploitation of the wealth of data at hands thanks to the IT revolution.
Our approach will be to develop some of the most promising research lines in the area of economic complexity and integrate them to form a nucleus for a new direction in economic science. The novelty is in various directions. First the whole approach is data-driven, this requires a careful selection of meaningful data and the introduction of appropriate algorithms to extract the relevant information. This body of experimental information also represents the testing ground for the models. The continuous feedback between data, algorithms and models will provide a firm scientific basis to the whole approach.
 From the point of view of data mining and IT content the situation has peculiar characteristics and consequences. As we are going to see in the various examples described in detail later, the typical situation is to start from some archival data and realize that they can be used to get new information well beyond the motivations of their original collection. This poses new problems and challenges but also leads to great opportunities. Often we will introduce iteration procedures in the spirit of Google Page Rank problem to extract the new information. However, in this way the archival nature of data can be problematic because, even a single error, can propagate and affect many of the results. This requires a new approach to test the validity of the data with respect to this specific question. At the same time the robustness of the methods introduced has to be investigated on the light of this problem. The new results obtained will provide a new perspective on the data and stimulate the collection of new data designed specifically for this new research line.
From the point of view of concepts and methods we will focus directly on the complex structure of the interconnected economic markets. This implies the introduction of appropriate variables which may not even exist for individual elements. These new variables will be the goal of the algorithms and models and finally will be characterized by appropriate new metrics. At the moment this type of questions are analysed mostly in qualitative terms. It is a specific objective of the project to move from qualitative to quantitative results. These are expected to have important implications in the various areas, scientific, technological and social.
In terms of governance and policy making the analysis of a globalized economy has also important implication. The first one is the realization that the regulations and control of an integrated system can only be made at the scale of the entire system. This means for example that national regulations and planning may be ineffective in this new situation. This leads to a limitation of the standard governance but also indicates the instruments that are effective in the new context. Also for companies and technologies we are going to see that their location and interconnections in the global network are fundamental elements for a realistic analysis.