This preview shows page 1. Sign up to view the full content.
Unformatted text preview: ed States Strong negative correlation Joulfaian and Marlow (1990) United States Strong negative correlation Grossman and West (1994) Canada Strong negative correlation Ehdaie (1994) Cross-country comparison Strong negative correlation 154 Commission on Fiscal Imbalance Box -2: Empirical Studies on the Impact of Fiscal Decentralization on Economic Growth
Little research has been done on the impact of fiscal decentralization on economic growth. Until recently the debate
over the merits of fiscal decentralization had been on theoretical grounds of efficiency gains and the empirical studies
that have analyzed the impact of fiscal decentralization on economic growth have only appeared recently. Interestingly,
these studies generally find that fiscal decentralization is associated with slower economic growth.
Studies on The Impact of Fiscal Decentralization on Economic Growth
Study Unit of Analysis Findings Davoodi and Zou (1998) Cross-country comparison Significant negative relationship Xie, Zou and Davoodi (1999) United States Significant negative relationship Zhang and Zou (1998) China Significant negative relationship Serious methodological issues confront efforts such empirical studies. First, there is no consensus about specification
of an empirical model for growth studies. The literature on economic growth suggests that growth is a complex
phenomenon with multi-dimensions (Levine and Renelt, 1992). Growth studies are usually criticized on the grounds of
a possible model misspecification (Martinez-Vazquez and McNab, 1997). Second, the fiscal decentralization variable
used in these studies does not represent the multidimensionality of the issue. Without controlling for subnational
governments’ autonomy over expenditure and revenue decisions and whether subnational officials are democratically
elected, the expenditure share of subnational governments as a fiscal decentralization variable means very little in
representing the level of decentralization. Third, regression coefficients may very well be the product of spurious
correlation. In regression analysis models, the cause-and-effect relationship runs directly from explanatory variables to
the dependent variable. However, if both dependent and independent variables are determined simultaneously, the
distinction between dependent and explanatory variables becomes dubious. If there is no unidirectional cause-andeffect relationship, the dependent variable is determined by explanatory variables, and some of the explanatory
variables are, in turn, determined by output. Given that there has been extensive research on the role of economic on
growth on fiscal decentralization (Oates, 1985; Pommerehne, 1977; Kee, 1977; Bahl and Nath, 1986) and very little
research on the causation line from fiscal decentralization to economic growth, it is highly suspected that the
regression coefficients reported on these studies is a mere reflection of spurious correlation. Therefore, the theoretical
underpinnings of the relationship between fiscal decentralization and economic growth still...
View Full Document