Three Essays on Economics in Big Data Scenarios
Big Data, ESG, Exchange Rate Predictability, Consumption Inequality, Economic Complexity.
This work is composed of three studies on economics in big data contexts. The first analyzes the impact of ESG (Environmental, Social, and Governance) news on the stock returns of the leading Brazilian companies, using an unprecedented ESG Terms Dictionary specifically developed for this study to select and classify news according to the standards of the Sustainability Accounting Standards Board (SASB). The research indicates that only news with financially material content influences stock returns. The second study explores the highfrequency predictability of the Brazilian exchange rate (at intervals of 1, 5, and 15 minutes), employing both machine learning techniques and traditional linear regression for forecasting. Two types of exercises are conducted: one with contemporary predictors and another in "real" out-of-sample testing. Although the out-of-sample forecasts show some efficacy when the rate of B3's foreign exchange futures contracts is considered as a predictor, the real out-of-sample predictability proves to be limited, highlighting almost instantaneous adjustments in exchange rates. The third study, currently underway, aims to measure consumption inequality at the municipal level using data from the payment methods of the Brazilian Payment System (SPB) of the Central Bank. Furthermore, as an application, we study the relationship between inequality and economic complexity, this being the first assessment of this kind for Brazilian municipalities.