Uma nova forma de medir liquidez: construção e aplicação no mercado brasileiro

• Autor(a) principal: Silveira, Vinicius Girardi da
• Data de Publicação: 2017
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Biblioteca Digital de Teses e Dissertações do UFSM
• Texto Completo: http://repositorio.ufsm.br/handle/1/12331
• topic: Liquidez; Medida de negociabilidade; Análise fatorial de séries temporais; Modelos de precificação com liquidez; Mercado acionário brasileiro; Liquidity; Negotiability measure; Times series factor analysis; Liquidity pricing models; Brazilian stock market; CNPQ::CIENCIAS SOCIAIS APLICADAS::ADMINISTRACAO

This study aimed to construct a liquidity measure using their proxies and assess their applicability in the financial context. To that, this study proposes the creation of a negotiability measure, which is a compendium of negotiability proxies used by the literature. The statistical procedure used to obtain this measure was the time series factor analysis (TSFA), which it is an extension of traditional factor analysis, working with time series instead of cross-section data. The data used for the illustration presented came from the trading of 858 stocks on BM&FBOVESPA from January 2000 to February 2016. As a result, the measure constructed for the market was demonstrated to be consistent with the others and capable, in terms of correlation, of replacing the proxies used in its construction. In addition, it presented intermediate statistics in relation to their peers, which suggests that the measure can show more balanced results. When analyzed the applicability of the measure in liquidity pricing models, was observed that it has an explanatory power similar to the other proxies used. Having as main differential the advantage of reducing the dimensions of liquidity, considering the information contained in all proxies in only one measure. Moreover, the findings suggest no differences between the means of the measures. However, when observed the variance, the negotiability measure showed distinct from the others, presenting intermediate statistics. In this sense, it is possible to conjecture that the negotiability measure tends to present similar results when used in models based on average, as is the case of regressions. On the other hand, it may be more advantageous and accurate in models that consider variance.