This paper aims to study the effect of recomposed institution quality to extreme income inequality. Findings reveal aggregated institutional quality of World Governance Indicators (WGI) have anomalies, distorted by its individual components' incongruent relationships with income inequality. The study covers period from 2010 to 2017 and applies quantile regression method due to rejection of normality of residuals and present of data clustering. Total of 43 countries are selected based on availability of data. WGIs do not always have negative relationship with income inequality. The recomposed WGI-plus and WGI-minus are all significant at correct sign, except insignificant for one case. These findings contribute six implications. Firstly, the WGI has subconsciously set democracy and free market as "good quality" institution, yet findings of positive relationship reveal this is not completely true. Secondly, the positive findings in control of corruption signal possible serious structural flaws regarding policies, perception, and its conceptualization. Thirdly, middle-income countries have relatively more anomalies. Fourthly, relatively more insignificant results of certain WGI components in middle-income countries cast doubt on their system of separation of power, prompting critical review of political will and governance effectiveness towards inclusiveness. Fifth, the significant results of the recomposed WGI enhance call for not aggregating all components of institution quality in future research and policy making decision. Sixth, the classic school that propagated free market is not effective to reduce inequality. Keynesian economies, especially targeted fiscal expenditure helps in middle-income but not highincome counties.