Back in September
This paper challenges the common international representation of working children as steady job-holders. We take advantage of panel data (the Pesquisa Mensal de Emprego) from Brazil to follow the employment patterns of large numbers of urban children ages 10-16 during 4 months in their lives. Different waves of the panel cover most of the 1980s and 1990s and include thousands of children, making possible generalizations about large-scale patterns. We document a substantial decline in child employment in the 1990s. An analysis of transition rates in and out of employment for young workers shows relatively high volatility in urban employment, with both higher exit rates and lower entry rates responsible for the decline in child employment. Due to this volatility, however, the proportion of urban Brazilian children ever-working in the labor force is substantially higher than the fraction who are workers at any one point in time; we find that movements in and out of employment are "normal" rather than exceptional. An intermittency multiplier summarizes the difference between employment rates in one reference week vs. four reference weeks over a 4-month period. We conclude that intermittent employment is a crucial characteristic of child labor which must be recognized in order to adequately capture levels of child employment and identify child workers.