Abstract
This paper employs a panel of 23 local governments in Taiwan over 1998–2010 to re-estimate the redistribution effects of intergovernmental fiscal transfers by considering a self-financing resources of local government as the transition variable in panel smooth transition regression models. Empirical results show that the income (or tax revenues) redistribution effects of fiscal transfer policies are nonlinear and vary with time and across local governments. The grants from central government can improve income and tax revenues distribution of local governments; however, the centrally allotted tax revenues have inverse effects and the total fiscal transfers have ambiguous effects. The total fiscal transfer is a proper policy instrument for improving income redistribution, and the grants for improving tax revenues redistribution. However, high self-financing resources ratios are harmful for these redistribution effects.
Notes
1. For more detailed discussions on intergovernmental fiscal transfers, see Boadway and Shah (Citation2007).
2. Another possible channel is the balance between the central government’s capacity and its incentive to support the local government. If the tax revenue is very low, the central government’s disposable revenue is very limited, which means the central government lacks the capacity to support the local government, leading to a low fiscal transfer. If the tax revenue is very high, the central government lacks the incentive to support the local government, also leading to a low fiscal transfer, thus forming an inverted U-shape. We deeply appreciate one of the reviewers for providing this useful explanation.
3. For more details on the utility of PSTR models, see Fouquau, Hurlin, and Rabaud (Citation2008).
5. If the amount of centrally allotted tax revenues of the region is relatively large, the central government may reduce the grants to the region, especially when the central government has a large domination on transferring grants. Thus, it is probable that there exists a substitution between these two components. To investigate a more complete picture of fiscal transfers, this paper performs the redistribution effects of total fiscal transfers. We appreciate one anonymous reviewer providing this suggestion.
6. The case m = 1 corresponds to a logistic PSTR model, and m = 2 refers to a logistic quadratic PSTR specification. In addition, it is easy to extend the PSTR model to more than two regimes. For the introduction of PSTR models, see González, Teräsvirta, and van Dijk (Citation2005).
7. For more details, see Fouquau, Hurlin, and Rabaud (Citation2008).
8. There are 25 local governments in Taiwan, but we abandon the Kinmen County and Lienchiang County due to their incomplete data.
9. Again, we appreciate the anonymous reviewer providing this suggestion.
10. In fact, we also conduct a second test of non-remaining linearity with direct effects in which the transition variable is used as an explanatory variable, and the result supports the fact that the lagged transition variable has no direct effect on the dependent variable.