WP 42_Eduardo_Zepeda

Insufficient labour income and limited access to employment are critical problems that policy makers must address when designing development strategies in Latin American countries. The persistence of the high incidence of poverty and inequality can be explained largely by the poor performance of labour markets. This Working Paper uses household survey data for Chile, Brazil and Mexico, from the early 1990s to the early 2000s, to examine the link between the growth of labour income, employment creation and the distributional impact of these factors. Through a simple decomposition of the sources of household labour income into earnings per worker and employment per population, the paper evaluates the role of economic, social and demographic factors in contributing to income changes. This decomposition shows that earnings per worker were the single most important determinant of the change in household labour income per capita. The change in earnings had the largest impact on household labour income in five of the eight country periods considered. Changes in the employment to population rate did play a role in determining labour income, but was much less important. Further decompositions show that despite favourable declines in dependency rates, the unfavourable trends of an almost ubiquitous rise of unemployment rates and, at times, the decline of participation rates dampened the contribution of employment to household labour income. The paper also decomposes labour income per capita into 20 equally sized partitions in order to evaluate its distributional pattern. A simple evaluation rule is used to validate whether changes can be considered pro-poor. O f the eight country periods analysed, only three exhibited income changes favouring the poor: Brazil in 1996-2004, Mexico in 1994-1996 and Mexico in 2000-2004. But in two of these, the pro-poor change occurred during economic contractions. In the remaining five country cases, the increase of labour income was associated with a distributional pattern that did not favour the poor. Thus, there was only one period in which labour income not only increased but was also pro-poor. But even in this case, the distribution did not favour the extremely poor. The alternating pattern of change in favour of and against the poor is explained mostly by the change in their earnings. The pattern of change in employment rarely favoured them. But when it did, usually during economic downturns, the rising participation rate of poor workers was the main reason.