Modelling The Number of Unemployment in East Java: Negative Binomial Regression Approach

Unemployment is one of the benchmarks for the success of development in a country and affects sustainable economic growth in an area, including in East Java. The government has made lots of effort to overcome high unemployment, such as holding job fairs every month. However, in East Java, the unemployment rate in 2019 still exceeds the ideal unemployment rate, which is around 2-3 percent. Besides, there is no significant change in the unemployment rate in the last three years during 2017-2019. Therefore, this study aims to model the number of unemployment in East Java by using Negative Binomial regression. In other words, this study investigates how certain factors affect the number of unemployment in East Java. The Negative Binomial regression model is employed in this study as an alternative from the Poisson regression model because the number of unemployment is a count data and, in many cases, is overdispersion. That is, the comparison between the expected value is not the same as the variance. This research will contribute to the East Java Provincial Government or related labor agencies to overcome high unemployment. The finding shows that factors such as regional minimum wage and the number of enterprises significantly affect East Java's unemployment in 2019. Besides, the Negative Binomial regression model with only significant explanatory variables is the best model for modeling the number of unemployment with the lowest AIC value.