Processos estocásticos em finanças: uma abordagem para investimentos em ações de empresas do setor elétrico listadas na B3

Abstract

The analysis of stochastic trends in asset pricing, especially in the electricity sector, is essential to reduce uncertainties, ensure robustness in models and adapt them to real behavior. This involves the use of time series and stochastic volatility models, in addition to the analysis of stationarity and points outside the curve, known as "jumps". This work seeks to verify modeling assumptions, including stationarity, and improve pricing considering stochastic characteristics and jumps in the electricity sector. In this context, the central problem was to explore how the stochastic characteristics of the sector's asset prices can improve pricing, including jumps. The general objective is to investigate these characteristics to use in modeling and analyzing price dynamics, considering the historical series and accounting statements of specific shares at B3. The research, of an empirical-analytical nature, initially collected data from 64 companies in the electricity sector, composing a final sample of 34 organizations, after analysis of data presence. The results indicate that a stochastic approach to pricing contributes to reducing uncertainty and adapting models. Most stocks in the electricity sector show stochastic behavior, mitigated by the first difference. However, some stocks remain stochastic even after differentiation. Stocks with stochastic behavior reached jumps above and below the established limits, without a single trend in relation to the regulation of the electricity sector. In short, the modeling considering stochastic processes and jumps was partially guaranteed, being improved with adjustments to the Poisson mean reversion models, MRM, resulting in statistically significant values for the coefficient of determination R² and improving the predictive capacity.