Income, Inequality and Poverty Convergence at the Local Government Level in Decentralized Indonesia.

• Author: Aji, Priasto

1. Introduction

   Over the last three decades, more than seventy-five countries have undertaken decentralization reforms to devolve or delegate revenue sources and important service delivery responsibilities to lower levels of government (Muriu 2013). The transfer of rights and responsibilities to lower levels of government not only includes administrative or fiscal tasks, but can also be political--given that decentralization reforms are often accompanied by the establishment of local-level elections (Ahmad and Mansoor 2002). Despite its widespread adoption, the impact of decentralization on regional inequality is still a subject of debate. Numerous studies have investigated this issue and there is no consensus.

   On one hand, some researchers have argued that decentralization can increase regional inequality because it weakens the capability of the central government to lessen disparities by allocating additional financial transfers to lagging areas (Prud'homme 1995; Peterson 1995; Sacchi and Salotti 2014). Furthermore, subnational governments may not have the fiscal and bureaucratic capacity to effectively attract investment or catalyse growth (Sacchi and Salotti 2014). Prud'homme (1995) argues that decentralization can increase disparities as the poor in rich regions over time will most likely become better off while the poor in poor regions remain poor or even become worse off. He also argues that decentralization has negative implications for inequality, because the poor will tend to move out of areas with lower benefits, while the rich will move to areas with lower taxes. Studies by Morelli and Seaman (2007) and Neyapti (2006) confirm that decentralization has led to an increase in regional inequality.

   On the other hand, there are also arguments for the opposite causal effect. Decentralization measures are often thought to increase the efficiency of resource allocation and to allow for the utility-maximizing behaviour of policy makers in local governments (Tselios et al. 2012). This is because subnational governments are likely to have an information advantage over central governments when it comes to responding to the needs and preferences of locals (Ezcurra and Pascual 2008; Oates 1972; Tiebout 1956; Tselios et al. 2012). A study by Tselios et al. (2012) finds that decentralization is related to lower levels of regional inequality.

   Another study by Siburian (2020) concludes that fiscal decentralization--namely, the devolution of fiscal authority from national to local governments--reduced income inequality across provinces during the 2001-14 period. According to the study, there are three possible reasons for this. First, decentralization granted local governments with autonomy to design development programmes that match the unique characteristics of a given municipality or regency and distribute more resources within it in a more targeted, balanced fashion. Second, decentralization enables local governments to provide public services that match the needs of their citizens. Third, decentralization motivates local politicians to effectively allocate local public goods and services because they are keen on being re-elected by better serving the voters.

   Suharyo (2003) contends that Indonesia's simultaneous implementation of democratic reforms, along with sweeping decentralization measures, was a fundamental transformation of the country's institutional arrangements that influences not only relations between levels of government, but also the way all levels of government interact with communities and other development stakeholders. Suryahadi and Izzati (2018) find that, since the 2001 reforms, Indonesia has achieved significant progress in improving people's welfare, as seen from the increase in per capita income and a reduction in the poverty rate. However, this has been accompanied by an increase in inequality. Taken together, these mean that the improvement in welfare has been faster for those in the higher income population compared to those in lower income groups. This is different from developments during the New Order era, where the increase in per capita income and the reduction in poverty rate were not accompanied by an increase in inequality.

   This increase in vertical inequality raises a question of whether it has been coupled with a divergence in economic development outcomes across provinces and local governments in Indonesia or whether they still converge to the same level of development in the long run. Using provincial and municipality/regency level data from Java and Sumatra from 1999 to 2008, Vidyattama (2013) finds insignificant convergence between 1999 and 2008. However, when the analysis is conducted separately for Java and Sumatra, he finds significant GDP (Gross Domestic Product) per capita convergence in Sumatra and insignificant divergence in Java.

   The issue of convergence, i.e., whether lower income entities will eventually catch up with higher income entities, is a long-standing debate in development. Examples include convergence in incomes between rich and poor parts of a country; in plant and firm size in industries; in economic activities across different regions (states, provinces, district or cities) within the same country; in asset returns and inflation rates across countries in a common trade area; in political attitudes across different groups; in wages across industries, professions and geographical regions (Quah 1996). Although some economic theories predict convergence in income and other development outcomes, the empirical evidence has been subject to debate (Barro and Sala-I-Martin 1992).

   Based on the traditional neoclassical growth model, originally set out by Solow (1956) and Swan (1956), the growth rate of income distribution generally tends to equalize across economies. This means that there is not only convergence in per capita income but also convergence in income inequality (Tselios 2009). By extension, this also means that there is convergence in poverty rates. The process of convergence occurs when poorer regions experience higher growth rates in per capita income than richer ones and either high-inequality regions experience falling levels of inequality or a more rapid fall in inequality in high-inequality regions compared to low-inequality ones.

   This paper addresses this question by examining whether there has been a convergence or divergence in per capita income, inequality and poverty across local governments--namely municipalities and regencies--over the last two decades in Indonesia following the 2001 decentralization reforms. Furthermore, Indonesia provides an interesting case to examine the issue of convergence, in particular regarding inequality, as the country has experienced a significant increase in vertical inequality. Theoretically, convergence in levels of inequality across local government areas during a period of increasing inequality is still possible if the rate of increase in inequality in local governments with high levels of inequality is lower than among local governments with low levels of inequality.

   Using a method of convergence analysis that takes into account spatial correlations across neighbouring local government areas, this study uses spatial econometric models with eighteen years of local government-level panel data covering the 2002-19 period. The findings indicate that there has been a conditional convergence of per capita income, inequality and poverty rates across local governments in Indonesia. This means that the levels of per capita income, inequality and poverty across local governments will converge, but the gaps will not be completely eliminated due to differences in local governments' endowment factors, such as natural resources, human resources, physical capital, financial capital, social capital and institutions. Findings from heterogeneity analysis indicate that these conditional convergence trends occur both across municipalities (kota) and regencies (kabupaten) as well as across local governments in Java and across those outside Java.

   The structure of this paper is as follows. The second section discusses economic growth, inequality and poverty trends in Indonesia over the last two decades. The third section discusses the theoretical and empirical framework pertaining to convergence. The fourth section explains the method and data used in this study. The fifth section sets out and analyses descriptive statistics on income per capita, inequality and poverty at the local government level. The sixth section presents and discusses the empirical results, including the heterogeneity analysis. Finally, the seventh section provides the conclusion and policy implications of this study.

2. Growth, Inequality and Poverty Trends in Indonesia

   During the three decades from 1970 to the late 1990s, Indonesia had a high, albeit relatively volatile, rate of economic growth, averaging around 6 to 7 per cent annually. Inequality, however, was remarkably stable at a Gini Ratio of around 0.32-0.33 during this period, except for a brief increase in 1978. Hence, up until the late 1990s, inequality did not rise despite high economic growth. Boediono (1997) argues that this was due to the implementation of the New Order government's development strategy of Three Pillars of Development (Trilogi Pembangunan), which consists of stability, growth and equality.

   However, this changed after the Asian Financial Crisis and the ensuing economic reforms. The crippling financial crisis, which hit Indonesia in 1997, developed into an economic crisis and soon was followed by a political crisis in the following year. In 1998, the Indonesian economy contracted by more than 13 per cent. This culminated in the toppling of the authoritarian New Order government, which had been in power for more than three decades.

   Following this, Indonesia enacted important democratic reforms, holding free elections regularly to elect its national and local...