Submitted by Elba Lopes (elba.lopes@fecap.br) on 2016-01-27T17:03:34Z No. of bitstreams: 2 Renato_Ros.pdf: 443506 bytes, checksum: 7e56f02b1178c2be1cc1a5b8fbe2b165 (MD5) license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) / Made available in DSpace on 2016-01-27T17:03:34Z (GMT). No. of bitstreams: 2 Renato_Ros.pdf: 443506 bytes, checksum: 7e56f02b1178c2be1cc1a5b8fbe2b165 (MD5) license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) Previous issue date: 2011-11-08 / Scientific work using accounting statistical tools with indicative accounting variables were used to study corporate frauds abroad in early nineties. However, studies using with this approach are practically non-existent in Brazil. The purpose of this research was to analyze whether accounting variables used as fraud indicators in foreign studies could be applied to non-financial listed companies in Brazil. For this purpose, based on 220 sentenced lawsuits (Common and Summary Proceedings) containing corporate fraud indications with decisions disclosed by the Brazilian Securities and Exchange Commission (CVM) from 2000 to 2010, an 'analyzed group' was formed with 26 companies with fraud indications available in Econom??tica?? software and a 'control group' with 52 companies without fraud indications. It was extracted 13 fraud-indicating accounting variables most used in foreign researches during the last 10 years. To analyze results, it was applied the descriptive analysis and the logistic regression. From the main results, it can be said that the indicative accounting variables "gross profit/ total asset and "operating gross revenue/ total asset" could be considered relevant fraud predictors in non-financial listed companies in Brazil. This work has the intention to contribute as a helping tool for companies' transparency, ethics and responsibility, as well as to assess capital markets and involved agents, as regulation organs, auditors and corporations as a whole. At the end of this work, it is suggested future scientific works. / Trabalhos cient??ficos com ferramental estat??stico em contabilidade, utilizando vari??veis cont??beis indicativas, deram seus primeiros passos no estudo das fraudes corporativas no exterior no in??cio da d??cada de 90, e com esse enfoque, praticamente inexistem no Brasil. O objetivo desta pesquisa foi analisar se as vari??veis cont??beis utilizadas como indicativas de fraudes em estudos no exterior podem ser aplicadas em empresas n??o financeiras de capital aberto no Brasil. Para isto, baseado nos 220 processos julgados (Ritos Ordin??rios e Sum??rios) com ind??cio de fraude corporativa com julgamentos divulgados pela Comiss??o de Valores Mobili??rios - CVM entre 2000 e 2010, formou-se um 'grupo analisado' com 26 empresas com ind??cios de fraude que continham dados dispon??veis na Econom??tica?? e um 'grupo de controle' com 52 empresas que n??o apontavam ind??cios de fraude. Extraiu-se as 13 vari??veis cont??beis indicativas de fraudes mais utilizadas em pesquisas no exterior dos ??ltimos 10 anos. Para an??lise dos resultados, utilizou-se an??lise descritiva e regress??o log??stica. Dentre os principais resultados obtidos, destaca-se que as vari??veis cont??beis indicativas 'lucro bruto/ativo total' e 'receita operacional bruta/ativo total' podem ser consideradas importantes previsoras de fraudes em empresas n??o financeiras de capital aberto no Brasil. Esta pesquisa visa a contribuir como ferramenta de aux??lio ?? transpar??ncia, ??tica e responsabilidade das empresas, auxiliando mercado de capitais e agentes envolvidos como ??rg??os reguladores, auditores e corpora????es como um todo. Ao t??rmino desta pesquisa, sugere-se novas pesquisas para futuros trabalhos cient??ficos.