We study on how a larger local government budget on infrastructure does not reflect into good quality of road in forty-one district/city across Indonesia given the fact of low infrastructure quality and low government spending on infrastructure. This study excels its preceded studies done by Tanzi and Davoodi (1997) at country level. The methodology used is a combination of quantitative and qualitative approach since our main research query is to seek facts on why a larger government spending is not followed by better quality of infrastructure. Our empirical strategy proved that corruption has aff ected government spending directly reflecting cost of corruption has been included as infrastructure project cost. We argue that corruption does aff ect quality of infrastructure negatively through government spending. Given the fact of Indonesian decentralization system enabling local election at district/city level it gives more room for decision-making process that has never been conducted before under centralized system. Corruption tends to occur in situation where the monitoring effort of central bureaucrat does exist while participation of local civil society is low. Although a district has already implemented e-procurement system intended to reduce opportunity of corruption incidence, it still exists through hidden-agreement between bureaucrat and such contractor who agree upon the project announcement process. They agree to set bid-price at lower level so that the contractor can win the project. This unfair process is suspected to happen since the budgetary process taking place at legislative where political bargaining is massive.