Determinantes de la temporalidad en el mercado laboral ecuatoriano.

Abstract

The differences of diverse activity branches inside the Ecuadorian labor market show the training levels that workers must have to avoid negative effects. Therefore, this study analyzes the role played by personal, labor, and geographical characteristics in the probability of obtaining a temporary employment contract against an undefined one. The analysis is performed by activity branch in order to determine which are the characteristics that would make possible the existence of temporary employment contracts in each of them. In order to achieve this purpose, two stages are developed. First of all, we use a logistic regression model. Then in a second place, we use learning models as a robust technique to check the results. Particularly the technique called support vector machine. In both cases the information used is based on the Survey of Employment, Unemployment and Underemployment, from the National Institute of Statistics and Censuses (INEC). From the above, the mathematical foundations of logistic regression and support vector machine are presented in this paper. It is given particular attention to the most relevant and necessary aspects for the development and implementation of the techniques, it is mainly focused on a binary classification problem. In the context of the logistic regression technique, there are detected differences that exist within the Ecuadorian labor market. Then, the technique of support vector machine are developed in order to contrast the obtained results. In addition, to present the possible advantages and/or disadvantages of using a learning method for individuals classification, it is a widely used procedure currently. Finally, conclusions of this study are highlighted and future research directions are discussed.