An investigation of dynamic poverty in turkey with random effects multinomial logit model

It is important to examine the dynamic structure of poverty, because of the fact that, poverty can vary according to differences both in the welfare level and situation at a point in time. The examination of the dynamic structure of poverty by panel data allows both the learning of the chronic and transitory dimension of poverty and the examination of the profiles of chronic and transient poor families. It is also possible to investigate the factors influencing both of people who entries poverty at a point in time and people exist from poverty utilizing panel data. Therefore, this thesis aims to provide the profile of poverty examining the determinants of poverty taking into account the dynamics of poverty; poverty outflow, transitory poverty and chronic poverty in Turkey using Turkish Statistical Institute, Income and Living Conditions Survey (SILC) panel micro data. While the determinants of poverty in Turkey were estimated with random effects logit model, the determinants of the dynamics of poverty (chronic poverty, transitory poverty) were estimated employing the multinomial logit random effects model. Multinomial logit random effects model was estimated separately by classical approach and bayesian approach.