Classificação de ratings, sustentabilidade e previsão de default uma abordagem utilizando a regressão quantílica
Carregando...
Tipo
Dissertação
Data de publicação
2014-08-29
Periódico
Citações (Scopus)
Autores
Alves Filho, Cy Dy Augusto
Orientador
Perera, Luiz Carlos Jacob
Título da Revista
ISSN da Revista
Título de Volume
Membros da banca
Mendonça Neto, Octavio Ribeiro de
Kimura, Herbert
Kimura, Herbert
Programa
Controladoria Empresarial
Resumo
The literature on analytical methods of accounting and corporate financial analysis models
and cr
edit indicators is
large
, and among the methods of credit risk
classification is the
classification model ratings, through which institutions classifie
s
customers according to their
risk. However, the classical models of modeling credit risk using statisti
cal techniques widely
disseminated, as is the case of simple linear regression, the least squares method, among
others. The quantile regression, evaluated in disseminated by Koenker and Basset (1978) has
it as a
main characteristic
,
analyzing the sample by
the median and allow the analysis of
subpopulations through the quantiles of the sample, which
allows more specific inferences in
according to the needs
of
the study. In recent years the concern with social and environmental
issues have become increasing
present in both the practical means and academia and in
society in general, which brings up the idea of
including the analysis of social indicators in
environmental analysis credit, as already proposed in previous studies. However, the
combined
use
of
ec
onomic, financial, social and environmental indicators, together with
quantile regression, is an innovative proposal, and the subject of this academic study. This
work
is an
exploratory
and
descriptive
study
, objective verify the possible contribution of
t
he
inclusion of
social and
environmental variables, combined with the use of quantile regression
for ratings
classification
and hence prediction of default. To fulfill this goal, we devel
oped a
database on panel, with the
total of 561 observations, consist
ing of data from publicly traded,
its ratings, economic indicators, financial, social and environmental, the years 2007 to 2012
companies. With use of quantile regression was possible to infer that the social environmental
variables are relevant for classi
fication ratings and, consequently, to predict default.
Descrição
Palavras-chave
rating , regressão quantílica , default , sustentabilidade, análise de crédito , indicadores sócio ambientais , indicadores econômico financeiros , rating , quantile regression , default , sustainability , credit analysis , socio environmental indicators , economic indicators financial