**4. Conclusions**

Analysis of the energy market is complicated. It involves the relationship between forecasting models and uncertainty, distinctly regarding the stochastic behavior of variables. The present paper is aimed at policymakers, offering a forecasting tool that deals with grouped time series. It also proposes a new forecasting approach, based on hierarchical modeling of the energy generation in Brazil.

The present paper introduces the use of trace minimization procedures (MinT) to aggregate and disaggregate forecasts based on the ARIMA and ETS models. MinT models performed better than the classic linear approaches, such as OLS and WLS. The MinT models also have high reliability for short predictive horizons. It is noteworthy that both hierarchical procedures and forecasting methods influence the predictive values of power generation in Brazil. Despite its advantages, the optimal reconciliation approach also has some limitations. This method could be unduly influenced by the sample period, and thus its ranking might change for other periods.

Therefore, the use of other predictive models, such as those based on analogs, machine learning, and other hybrid techniques, for example, is recommended. For future research, fine-tuning forecasts of the "south" electrical subsystem, as well as testing the accuracy of the hierarchal methods by using new forecasting approaches, is also recommended.

Finally, the present study contributes to the energy planning processes of different agents, given that understanding energy generation patterns is singularly important for minimizing risks and supporting reliable production planning. Good forecasts for future energy generation can support operational arrangements since energy supply and demand impact spot market sales prices.

**Author Contributions:** Both authors made substantial contributions to the analysis presented in the paper. T.S.G. took lead responsibility for proposing the methodology and for drafting the manuscript and M.A.C. for revising it critically. M.A.C. supervised the project. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by [National Council for Scientific and Technological Development—CNPq] gran<sup>t</sup> number [141740/2019-1].

**Acknowledgments:** The authors would like to thank the National Council for Scientific and Technological Development (CNPq) and Companhia Energética Integrada (CEI) for supporting this research.

**Conflicts of Interest:** The authors declare that there is no conflict of interest.
