ISSN 0024-9521 IJG Vol. 4 , No.1, June 2014 (1 - 11) © 2014 Faculty of Geography UGM and e Indonesian Geographers Association e spatial effect of fiscal decentralisation on regional disparities: the case from Indonesia Adiwan F. Aritenang Received: 05 03 2014 / Accepted: 19 03 2014 / Published online: 30 06 2014 © 2014 Faculty of Geography UGM and e Indonesian Geographers Association Abstract is study investigates the effect of decentralization on regional economic convergence between 1994 and 2004. Using spatial analysis, this research explains the presence of spatial dependence, which it is misleading to ignore it on the common linear regression. e paper argues that there is a strong evidence of spatial autocorrelation among Indonesia districts and regional convergence rate is higher in the decentralization period. Furthermore, the spatial error model suggested that, despite Jakarta’s major economic and political role, a random shock is insignificant if it introduced to the capital city. Abstrak Studi ini mengkaji dampak otonomi daerah terhadap konvergensi ekonomi wilayah di Indonesia antara ta- hun 1994 dan 2004. Menggunakan teknik analisa spasial, penelitian ini menjelaskan adanya keterkaitan spasial. Studi menjelaskan bahwa terdapat keterkaitan spasial antara kota/kabupaten di Indonesia dan terdapat konvergensi wilayah setelah otonomi daerah. Selanjutnya, simulasi spasial menunjukkan bahwa random shock terhadap Jakarta tidak signifi- kan terhadap ekonomi di Indonesia. Keywords: Indonesia; Decentralization; Regional Convergence; Spatial Random Shocks Kata kunci: Indonesia; Otonomi Daerah; Konvergensi Wilayah; Spatial Random Shocks e neoclassical theory lies in the idea of dimin- ishing returns that implies a long run equalization of development. is theory seeks to explain the deter- minants of regional convergence rate with economic growth is regarded as a growth of real terms of per capita income and the variables are used to measure the convergence of the economic growth rate, between ini- tial level of output, y, and its target, y*. Such variables act as additional variables in the conditional conver- gence, while absolute convergence only examines the initial level of output and the level of the target, without the effects of other variables. is theory has inspired numerous studies of regional convergence with several approaches such as cross-section, time series, and panel data analysis. Furthermore, there are also interests in studies the effect of decentralization and trade liberalization on intra-regional scale. e studies are varied in meth- odology and geographical level of analysis with non- spatial analysis such as Rivas [2007] that studies the impact of devolution and Rodríguez-Pose and Bwire [2004] that examine the impact of trade liberalization on Mexico regions. However, following the spatial au- tocorrelation at the regional scale theoretical studies by Anselin [1988, 2007] and Arbia [2006], there is emerg- A. F. Aritenang Agency for the Assessment and Application of Technology email: a.aritenang@gmail.com ing empirical studies and literature on this subject. For example, spatial econometrics analysis with in regional convergence in the US [Rey and Montouri, 1999] and in Turkey [Yildirim et al., 2009], and there are also studies on the spatial effect of trade liberalization such as by Özyurt and Daumal [2013]. While in Indonesia, there are non-spatial studies by Akita and Lukman [1995] and Resosudarmo & Vidyat- tama [2006]. e approach has also been implemented in Indonesia. For instance, a case study of income and consumption in the capital city Jakarta, using the Local Moran analysis shows the patterns in the core city of Jakarta and found that the sub district of Tanah Abang has a negative association, i.e., a low value surrounded by high values [Syabri, 2003]. e paper argues that the sub district is an urban decay as the conversion to busi- ness activity along with other rich sub districts in the core city of Jakarta has not taken place. Another case study of poverty analysis in East Java province by Pus- pitasari, et al [2010] with a LISA cluster map illustrates that severe poverty cluster locations has moved from south to the north and east part of the province be- tween 2003 to 2007. While the study by Karmaji [2010] focuses on the agriculture location. us, there is lack of spatial autocorrelation studies with regard on decen- tralization in Indonesia. is paper aims to examine the presence of spa- tial autocorrelation in regional economic convergence I. Introduction