Aplicação do "CASH FLOW AT RISK" e de cenários de stress no gerenciamento dos riscos corporativos do setor de distribuição de energia elétrica

• Autor(a) principal: Flavia Vital Januzzi
• Data de Publicação: 2010
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Repositório Institucional da UFMG
• Texto Completo: http://hdl.handle.net/1843/BUBD-9A5HDY
• topic: Testes de Stress; Modelagem VAR; Modelagem ARIMA; Simulação; de Monte Carlo; Backtesting; Fluxo de caixa em risco; Administração de risco; Fluxo de caixa; Energia eletrica Distribuição; Administração

Due to the relevance of the cash flow at risk (CF@R) to manage corporate risks, the present study proposes a comparison of two estimation approaches: the first one is based on Autoregressive Integrated Moving Average Models (ARIMA), and the second one appliesVector Autoregressive Models (VAR) with exogenous variables, for the following brazilian energy companies: Companhia Energética de Minas Gerais (CEMIG), Centrais Elétricas de Santa Catarina (CELESC), Companhia Energética do Ceará (COELCE), CompanhiaParanaense de Energia (COPEL) and AES Eletropaulo. In order to analyze the performance, eight observations were excluded (from the third quarter of 2007 to the second quarter of 2009), to calculate the root mean squared error, remaining then 38 observations to estimate theproposed models. Considering the difficulties with the survey and measurement of the Brazilian distribution energy sector risk factors, the results indicate that the second approach has presented a better predictive performance, in the analyzed series. Nevertheless, both models encountered difficulties to capture situations regarding the readjustment of negative annual rates, high increases on the energy purchase rates, increases in the installed capacity, and laborcontingencies. By backtesting the CF@R estimates, 1.000 scenarios were simulated for the risk factors, generating the same number of estimations for the operational cash flow for each quarter. Considering 5% significance level and the interval of eight quarters analyzed, the mean CF@R overestimate the risk for COELCE and underestimate the cash flow risk for CELESC. For the remaining companies the Cf@R measures were consistent, considering a 5% significance level. Beside these analysis, stressed scenarios were simulated, considering theextreme values obtained from the distribution of the risk factors. In some situations, the stressed test estimated a decrease of 62,7% when compared with values in normal scenarios. Those analyses are useful to measure the risk associated to low financial liquity situations, thatpledges the enterprises capacity to manage possible loans and to create or to maintain investments projects.