Newton Fund: Rio-UK+BigData - Modeling and Improving Rio de Janeiro State Economic Policies using Big Data Analytics Tools

Rio de Janeiro State Treasury Office have a high scalable (large amount of taxpayers) and dimensional (several descriptions of each taxpayers, such as its economic activity, how many employees, etc.) database. Extract knowledge from this dataset to provide support to the decision maker is a hard task, mainly when classical database and data mining techniques are employed. This difficulty can turn fiscal policies, which the main idea was to bring development for some economic sector, harmful to others. For example: evaluate how tax revenues coming from Goods and Services Conveyance Tax (ICMS) is affected by including food and beverages in Basic Food Basket; how is the impact of growing tax rates for Oil & Gas industry in inflation and unemployment rate, and so on.
Through Big Data Analytics tools, i.e., an adaptation and development of new database and data mining methods for large scale problems, we can tackle this enormous challenge. In this project our main objective is to provide Big Data Analytics solutions to support decision making for Rio de Janeiro State Treasury Office. As specific aims, we will:

* Propose a model to simulate fiscal policies made by Rio de Janeiro State Treasury Office. This model can generate several scenarios to evaluate the effects of a fiscal policy (e.g. increase of food and beverages taxes) in terms of other economic variables: total tax revenue, inflation rate, unemployment, etc., before its deployment in the state. For example, given an increase of 3% in ICMS Foods and Beverages products, will this increase state tax revenues? How much this will be? What is the expected impact of this policy in inflation rate or economic activity? The answers of these and other questions can help design fiscal policies less risky (i.e., harmful for other economic sectors) and more effective, such as preventing taxpayers default, developing economic sectors, and so on.

* Establish an optimization methodology to settle taxes rates at optimal levels in Rio de Janeiro State. The optimization method must consider multiple objectives (maximize tax revenues, develop more some economic sectors, etc.), in which some are intrinsically contradictory (inflation rate control and tax revenue growth). Also, this methodology must be adequate for constrained and non-convex problems, and where partial information is available, i.e., situations under uncertainty. Thus, this methodology will use information from Fiscal Policies Simulator (addressed earlier) and from other sources, in order to build the objective function and meet the policy maker demands and constraints.

An example of a possible application is to measure the recent settlement of Nissan Motor Company at Resende city (a city of Rio de Janeiro State) and its implications in terms of tax revenues (direct and indirect). Another possibility is to evaluate the possible reduction of royalties paid by Petrobras Oil & Gas Company, etc. With these information at hand, the Rio de Janeiro State Treasury Office can improve its fiscal policies, perform better economic stimulus and elaborate hedge mechanisms against systemic risks or economic decline (e.g., inducting investments in economic activities negatively correlated to the more essentials actually, etc.).
Both methodologies, if correctly employed, can generate savings (more effective management of public resources, improve policy makers productivity, etc.), increasing taxes revenues (through setting a suboptimal level for taxes), a better evaluation of tributary benefits from society and a healthier business environment. Also, these benefits can improve economic planning in the State, where the gain in resources can be used to develop public Education, Health, Safety, Science and Technology, and so on.

This partnership between UCL and Pontifical Catholic University of Rio de Janeiro (PUC-Rio -- Brazil) will enable:

* Support and training of PUC-Rio research group with knowledge from Big Data Analytics and Financial Computing developed by UCL research group.

* Applying UCL knowledge and tools in a real Big Data problem: modeling economic policies in Rio de Janeiro State.

* Education and training of human resources. Also, this project can be an initial step for others research and development projects forthcoming.
From RCUK perspective, this project will enable to evaluate its investment in research and facilities carried out in UCL Big Data Analytics and Financial Computing group. Finally, from FAPERJ perspective as a government funding agency, all technology transferred from an international partner, can demonstrate to society that research resources are being applied in favor to generate growth and social-economic development for Rio de Janeiro State.