This thesis investigates decentralization on population health through quantitative means in the Philippines during the transition period from 1989-1998. The driving hypothesis is that decentralization provides positive outcomes to health only if certain mechanisms are in place for it to function. The study used a three-pronged approach. At a country-wide level, the fixed effects regression model is used to understand the overall implication of decentralization on health. The study then investigates the regional level using the ordinary least squares model to better understand variations across the country. In both studies, the decentralization indicator used is total LGU expenditures on public goods and the population health indicator is infant death. The study finds that decentralization does not have a statistically significant effect on population health outcomes at a country-wide level. Despite this, four regions’ population health outcomes are found to be positively affected by decentralization. Region VIII, Eastern Visayas, shows the smallest decrease in infant death (p<0.05) followed by Region VI, Western Visayas (p<0.05) and Region II, Cagayan Valley (p<0.01). Finally, Region X, Northern Mindanao, saw the most dramatic decrease in infant death (p<0.01). Mechanisms were then identified that accounted for the positive effect of decentralization on health outcomes. This thesis concludes that mechanisms including champions advocating for health within and outside of the government, citizen inclusion and feedback, strong inter-LGU linkages, and a baseline level of economic well-being is needed for decentralization to be effective in improving health outcomes.