Abstract: The present study examines the relationship between household poverty and fertility in Coast and Western provinces using both quantitative and qualitative data. Specifically, the study sought to; 1. Estimate levels and trends of fertility in Coast and Western provinces by level of household poverty 2. Test whether the relationship between household poverty status and fertility depends on the context and 3. Test whether the relationship between household poverty and fertility is an artifact of measurement. The first level of analysis used datasets of 1989, 1993, 1998, 2003 and 2008/9 Kenya Demographic and Health Surveys (KDHS), while the second level involved analysis of qualitative data collected in 2011 during the poverty and fertility project and household survey conducted in 2013 in Kwale (Coast province) and Bungoma (Western province).

The relationship between household wealth and fertility was examined over the period of the fertility transition. Clear patterns were discovered in both provinces. In Coast province, fertility among poor household displayed a U-shaped pattern. In Western province, the fertility rate among poor household indicates an inverted J-shaped pattern. Poisson regression models were employed to estimate the effect of household wealth status (poverty) on fertility. Two patterns were observed. The magnitude of the difference in fertility rates between poor and non-poor women declined during the 1989-1998 then widened between 1998-2008/9 period. The increase in the gap between fertility of poor women and non-poor women is attributed to an increase in fertility among poor women. Multivariate regression analyses were performed to estimate the effect of household wealth status on fertility controlling for education and child mortality. Analysis suggests that the effect of poverty on fertility is context specific and depend on the period. In Western Kenya, poverty was found to be significant in the recent periods (2003-2008/9) but not in the earlier period (1989-1998). In contrast, the effect of poverty on fertility in Coast province was significant across most surveys except in 1998. Major differences were also observed in Coast and Western provinces in the effect of poverty on fertility when other factors such as education and child mortality are controlled. The study therefore concludes that the effect of poverty on fertility is context specific and no universal conclusion can be made on the relationship. Despite social and cultural differences, large families are still viewed as important among poor households. Qualitative data sheds some light on possible explanations for the lack of fertility decline observed among women from poor households. In particular, polygyny, old age security, child mortality, lack of spousal communication on reproductive issues and low age at first marriage are important. The relationship between fertility and alternative poverty measure namely; expenditure, wealth index and economic ladder question showed that the estimates of fertility rate did not vary. However, there was a lack of correspondence in the classification of household by different poverty measurement approaches. The study concludes that, in circumstances where income or expenditure data are lacking, asset index provides robust estimates that can be used to make inferences on the relationship between poverty and demographic outcomes.

The findings suggest that efforts by the government to promote family planning services, especially among the poor ought to be accompanied by the assurances to parents that their newly born babies will survive through to old age; transformation of attitudes on larger family size and greater involvement of women in development.