APPLICATION OF MULTIVARIATE STATISTICAL TECHNIQUES TO FINANCIAL INDICATORS: A CONTRIBUTION DISCLOSURE OF ACCOUNTING METRICS

Abstract

The accounting, the source of economic and financial information of a company, have been major focus of research in recent years. The performance analysis or solvency of companies, from financial indicators, it is a theme present in the daily life of the various economic agents, particularly investors, lenders and financial institutions. However there are questions about reaching their objectives are still enough evident by its users in order to provide them with adequate information to make decisions (LAURENT, 2004, p.10). Therefore, this research attempts, through the financial and economic indicators, develop alternative statistical models in order to help assess the financial performance of manufacturing companies with publicly traded. By means of database Economática® was possible to obtain 24 indicators of the manufacturing sector companies with shares traded on the BOVESPA from 2004-2008. The use of factorial analysis allowed grouping into five factors (indicators) the financial ratios considered important, explaining 76.07% of total data variability. After defined the most important factors, applied statistical techniques of discriminant analysis and logistic regression to identify which model is more effective in the classification of these companies as "efficient" and "inefficient". The results shown in this study, present evidence to the multivariate models can be used as alternative methods to assess the performance of companies with publicly traded.