Abstract
This paper focuses on one of the most determinant processes in business failure assessment: Variable selection. After a preselection of variables based on previous empirical literature, we perform a statistical variable selection on a sample of small firms using both mean and median differences. As the resulting variables differ in each test, we have performed a varied group of business failure assessment methods (linear discriminant analysis, quadratic discriminant analysis, logistic discriminant analysis, k-th nearest-neighbor discriminant analysis, logit, and probit) to identify the implications of using one test or the other. Our results show that the nature of the sample determines not only the statistical variable selection test, but the most appropriate methods to assess business failure, which constitutes our main contribution. Additionally, we contribute new evidence on the addition of qualitative information (payment incidents), with previous evidence for SMEs being scarc ...
This paper focuses on one of the most determinant processes in business failure assessment: Variable selection. After a preselection of variables based on previous empirical literature, we perform a statistical variable selection on a sample of small firms using both mean and median differences. As the resulting variables differ in each test, we have performed a varied group of business failure assessment methods (linear discriminant analysis, quadratic discriminant analysis, logistic discriminant analysis, k-th nearest-neighbor discriminant analysis, logit, and probit) to identify the implications of using one test or the other. Our results show that the nature of the sample determines not only the statistical variable selection test, but the most appropriate methods to assess business failure, which constitutes our main contribution. Additionally, we contribute new evidence on the addition of qualitative information (payment incidents), with previous evidence for SMEs being scarce.------------------------------------Este trabajo se ocupa de uno de los procesos más determinantes en la evaluación del fracaso empresarial: la selección de variables. Tras una preselección de variables basada en los resultados empíricos de la literatura previa, llevamos a cabo una selección estadística de variables sobre una muestra de empresas pequeñas, utilizando tanto diferencias en medias como diferencias en medianas. Como las variables resultantes difieren con el test, hemos utilizado un variado grupo de métodos de evaluación de fracaso empresarial (LDA, QDA, LogDA, KNNDA, logit y probit) con el fin de identificar las implicaciones de usar uno u otro test. Nuestros resultados muestran que la naturaleza de la muestra determina no solo el test de selección estadística de variables, sino también los métodos más apropiados para evaluar el fracaso empresarial, lo que constituye nuestra principal contribución. Además, el trabajo proporciona nueva evidencia sobre la adición de información cualitativa (incidencias de pago), siendo escasa la evidencia previa para pymes.
