Labour Market Dynamics in SA: Evidence from Kwa-Zulu Natal Province

Six years into South Africa's fledgling democracy one is prompted to ask: what has been achieved, if anything? In this paper, I will attempt to provide some answers, if tentative, concerning economic mobility as it pertains to labour markets in KwaZulu-Natal using the KwaZulu Income Dynamics Study (KIDS) data. To do so I adopt both univariate and multivariate techniques. Univariate estimates of earnings mobility are presented under Markovian assumptions, first in the form of transition matrices and then in the form of first order autoregressive models. To shed light on other correlates of mobility, I then turn to an analysis of transitions between labour market states using a multinomial logit framework. Key findings are: (i) females experience a 61% increase in their transition probability from self-employment to being "out of the labour force"; (ii) race appears to be insignificant in predicting transitions out of unemployment and into employment, but is significant in predicting jobloss: in short, it would appear that being an African person has become unimportant for getting a new job, but is still important for losing one; (iii) belonging to a revolving credit association increases one's chances of finding a new job by 45%; (iv) years of experience (proxied for by age in years) is significant for finding a new job if you were unemployed in 1993, but for new entrants into the labour force, education replaces experience; and (v) an additional year of education is important for remaining employed and increases one's chances of moving up the earnings distribution by 5%.