The primary objective of this study is to test "the fungibility hypothesis": if categorical grants are executed as mandated by the grantor, the central government. To this end, this study analyzes the data for 15 years (1995-2009) on the grants-in-aid categorized for social welfare programs. Followings are a brief summary of the analysis. Empirical evidence suggests that the hypothesis is generally supported and the extent of fungibility are different from a local government to another. Much of the variances are accounted for by the financial condition of each local government. Financially weak governments are more inclined to divert the grants toward more imminent demands from the points of their own view. On the other hand those metropolitan governments that are in better financial condition have little inducement to reallocate resources against the stipulation attached to the categorical grants. Similar patterns are observed when the effect of population size is controlled by using per capita expenditures instead of the gross expenditures. The fungibility hypothesis was again supported in provinces (Do's). However, Seoul and six other metropolitan governments appear to have significantly different resource allocation patterns: they are not actively pursuing fungibility. Even further, a reverse fungibility pattern is observed: more of their own resources are allocated toward social welfare in their efforts to secure more matching grants. Their stronger financial condition, often represented by high self-sustenancy rates, made this possible. One important implication we can draw from this analysis is that the discrepancy of social welfare services between financially strong governments and weak ones are likely to be expanded in the future. Therefore, a carefully designed reform in categorical grant system is required, especially with regard to differentiated matching rates that reflect the financial conditions of grantees. To sum up, fungibility is a widespread in local resource allocation in response to matching grants-in-aid; and the extent of fungibility is negatively correlated to the financial condition of each local government. "A reverse fungibility" observed among the metropolitan city governments that are in good financial condition.