• Entrar
    Ver item 
    •   Página inicial
    • Teses & Dissertações
    • Teses & Dissertações
    • Ver item
    •   Página inicial
    • Teses & Dissertações
    • Teses & Dissertações
    • Ver item
    JavaScript is disabled for your browser. Some features of this site may not work without it.

    Previsão da receita corrente liquida dos entes federados : análise da acurácia do modelo governo vs modelo holt-winter

    Thumbnail
    Visualizar/Abrir
    R - D - FRANCISCO XAVIER SOARES FILHO.pdf (1.923Mb)
    Data
    2014
    Autor
    Soares Filho, Francisco Xavier
    Metadata
    Mostrar registro completo
    Resumo
    Resumo: Apresenta a perspectiva historia da gestão publica brasileira, a questão da transparência das informações fiscais do Estado e do sistema orçamentário brasileiro. Identifica o Sistema de Coletas de Dados Contábeis dos Entes da Federação (SISTN) como fonte de informações e geração de conhecimento para a tomada de decisões, especialmente no tocante aos avanços da gestão fiscal a luz da Lei de Responsabilidade Fiscal (LRF). Explora as informações contábeis consolidadas no Anexo III de Relatório Resumido de Execução Orçamentaria, com foco na variável Receita Corrente Liquida (RCL), para obtenção de previsões e analise da acurácia. A pesquisa objetiva fornecer previsões da RCL mais confiáveis aos 399 municípios paranaenses, com a utilização do método estocástico de previsão Holt-Winters Aditivo e Multiplicativo. Para tanto, utiliza-se exclusivamente softwares livres, tais como o site Wessa.net, R-Project, Sistema Operacional Ubuntu e planilha eletrônica Libre Office Calc, para tabular, calcular e analisar as previsões das series temporais pelo modelo de previsao Holt-Winters. Analisa a serie histórica da Receita Corrente Liquida (RCL) dos municípios do Paraná, no período de jan-2006 ate dez-2012, e mensura a precisão dos modelos de previsão dos Governos Municipais vs modelo Holt-Winters Aditivo e Multiplicativo. Compara os resultadas das previsões, e o poder preditivo de cada modelo, através do calculo do Erro Médio Percentual Absoluto entre os valores previstos e observados e também através do teste estatístico U de Theil. Na analise dos resultados, o MAPE médio das previsões dos governos municipais paranaenses para a RCL do período em analise foi de 12,05%, bem acima dos 4,25% e 4,33% respectivamente dos modelos Holt-Winters Aditivo e Multiplicativo. Apenas 8,06% das previsões apresentaram valores menores do MAPE para as previsões dos Governos Municipais do Paraná em comparação ao modelo Holt-Winters. Resultado semelhante também foi identificado com o teste estatistico U de Theil, com a media de 1,2605 para os Governos Municipais e a media de 0,4794 para o melhor modelo Holt-Winters. Cerca de 3,6 bilhões de reais num período de 6 anos deixaram de compor os orçamentos municipais nos municípios paranaenses analisados com previsão nas respectivas leis orçamentarias em virtude da falta de precisão do modelo de previsão da RCL dos Governos Municipais. Por fim, com base nos dados da pesquisa, foi possível constatar que a utilização do modelo de previsão Holt-Winters Aditivo e/ou Multiplicativo apresentaram melhores resultados de poder de predição para o comportamento da RCL nos municípios do Paraná, inclusive com desempenho superior aos modelos adotados pelos Governos Municipais do Paraná no período de jan-2006 ate dez-2012.
     
    Abstract: Presents historical perspective of the Brazilian public management, the issue of transparency of tax information from the State and the Brazilian budgetary system. Identifies the Accounting System Collects Data from loved Federation as a source of information and knowledge generation for decision making, especially regarding progress in fiscal management in light of the Fiscal Responsibility Law. Explores consolidated in Annex III Summary Report on Budget Execution, with focus on the variable Current Net Revenue to obtain accurate forecasts and analysis of accounting information. The research aims to provide more reliable predictions of the Current Net Revenue 399 municipalities, with the use of stochastic forecasting method Holt-Winters Additive and Multiplicative. For this purpose, it is used exclusively free software, such as Wessa.net site, R- Project, Ubuntu Operating System and Libre Office Calc spreadsheet to tabulate, calculate and analyze the predictions of time series forecasting model by Holt-Winters. Analyzes historical series of net current revenue of municipalities of Parana, in the period Jan-2006 to Dec-2012, and measures the accuracy of the predictive models of Municipal Governments vs Model Holt-Winters Additive and Multiplicative. Compare the resultadas forecasts and predictive power of each model by calculating the Middle Absolute Percentage Error (MAPE) between predicted and observed values and also by statistical test Theil U 2. In analyzing the results, the average MAPE of forecasts of Parana municipal governments for the RCL of the period was 12.05 %, well above the 4.25% and 4.33 % respectively of the models Holt-Winters Additive and Multiplicative. Only 8.06 % of the forecast showed lower values of MAPE for the forecasts of the municipal government of Parana against the Holt-Winters model. A similar result was also identified with the Theil U 2 statistical test, with an average of 1.2605 for Municipal Governments and the average of 0.4794 for the best Holt-Winters model. About 3.6 billion dollars over a period of six years no longer included in municipal budgets with estimates in their budget legislation because of the lack of precision of the prediction of RCL Municipal Governments of Parana model. Finally, based on survey data, it was found that the use of the forecasting Holt-Winters additive and/or multiplicative model presented a better predictive power for the behavior of the Current Net Revenue in the municipalities of Parana, including superior performance to models adopted by the municipal government of Parana in the period Jan-2006 to Dec-2012 .
     
    URI
    https://hdl.handle.net/1884/35411
    Collections
    • Teses & Dissertações [10894]

    DSpace software copyright © 2002-2022  LYRASIS
    Entre em contato | Deixe sua opinião
    Theme by 
    Atmire NV
     

     

    Navegar

    Todo o repositórioComunidades e ColeçõesPor data do documentoAutoresTítulosAssuntosTipoEsta coleçãoPor data do documentoAutoresTítulosAssuntosTipo

    Minha conta

    EntrarCadastro

    Estatística

    Ver as estatísticas de uso

    DSpace software copyright © 2002-2022  LYRASIS
    Entre em contato | Deixe sua opinião
    Theme by 
    Atmire NV