Análisis de series de tiempo y regresión logit del desempleo juvenil en el Ecuador, periodo 2013 – 2018

Abstract

The objective of this paper was to analyze youth unemployment in Ecuador, using time series and logit regression techniques, for the period 2013-2018. The necessary information was taken from the National Institute of Statistics and Census (INEC), collected through the National Survey of Employment and Unemployment (ENEMDU). The time series technique allowed for the analysis of information from Ecuador's 24 provinces concerning time, focusing on identifying the magnitude of youth unemployment in the country and the changes that it presents over time. Besides, the Box-Jenkins methodology was applied, in which it was detected that the time series were stationary, identifying SARIMA models and the partial fulfillment of theoretical assumptions, the best-estimated model was the national youth unemployment rate SARIMA (1,0,0)(1,0,1)4 with an AIC= 70.67. The logit model of 2013 was selected as the best probabilistic model, which shows that the variables selected for the study were important. It also indicates that the determinants of youth unemployment among Ecuadorian youth are related to the sociodemographic variables of educational level and age that have the greatest weight in the model. It was jointly determined that if an Ecuadorian youth meets the characteristics of being a woman, being 26 years old, not being the head of household, being without a partner, attending classes, having a higher academic level, the income is less than the Unified Basic Salary of $357.50, do not specify the economic sector and labor occupation group and do not have social security, it will have a 94.7% probability of being unemployed. In general, the evolution and future changes of the youth unemployment rate studied should be discussed with government agencies to make decisions on contingency measures to avoid an increase in the rate.