| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 1.18 MB | Adobe PDF |
Autores
Orientador(es)
Resumo(s)
O propósito desta tese é investigar o fenómeno das falências na indústria do calçado em Portugal, contruindo um modelo capaz de prever este fenómeno com 1 ano de antecedência e analisando que variáveis têm mais influência nesse desfecho. Para tal, uma abordagem baseada em modelos binários Logit, foi usada para construir (e mais tarde validar) 5 modelos baseados numa amostra de 2,073 empresas do sector industrial do calçado em Portugal, entre 2006 e 2018, onde se registou um total de 422 ocorrências que classificamos como falências. Validamos que 2 dos nossos modelos registaram uma capacidade de “discriminação aceitável” ao apresentarem uma AUC entre 0.7 e 0.8, o que também revelou que os rácios ligados à rentabilidade, à alavancagem e à liquidez são os que têm o maior impacto na probabilidade de falência.
The purpose of this thesis is to investigate bankruptcy in the Portuguese shoe manufacturing industry by building a model able to predict it within 1 year, and by analyzing which variables are its main drivers. Therefore, a logistic approach was taken to build (and later validate) 5 models out of a sample of 2,073 Portuguese shoe manufacturing firms, across 2006 until 2018, where there was a total of 422 bankruptcy-like events. We found 2 of our models to have an “acceptable discrimination” ability by presenting an AUC between 0.7 and 0.8 which also revealed that ratios related with profitability, leverage and liquidity are the ones with the most relevant impact in bankruptcy probability.
The purpose of this thesis is to investigate bankruptcy in the Portuguese shoe manufacturing industry by building a model able to predict it within 1 year, and by analyzing which variables are its main drivers. Therefore, a logistic approach was taken to build (and later validate) 5 models out of a sample of 2,073 Portuguese shoe manufacturing firms, across 2006 until 2018, where there was a total of 422 bankruptcy-like events. We found 2 of our models to have an “acceptable discrimination” ability by presenting an AUC between 0.7 and 0.8 which also revealed that ratios related with profitability, leverage and liquidity are the ones with the most relevant impact in bankruptcy probability.
Descrição
Palavras-chave
Falência Indústria do calçado ROC Bankruptcy Shoe manufacturing industry Logit AUC