The legacy of the reformasi: the role of local government spending on industrial development in a decentralized Indonesia

Decentralization of the government function can be best understood as the devolvement of decision-making from the central government to its lower tier-administration, as described by Litvack and Seddon (1999) in their Briefing Notes for the World Bank. This concept is supposed to bring Pareto growth efficiency according to early decentralization scholars (Tiebout 1956; Oates 1972). These first generation of theorists convinced decision-makers in developed countries and thus propagated a wave of decentralization movement in the 1980s. However, in many developing countries the route to a decentralized governance institution has a quite different context. Large democratic movement in Brazil and the Philippines in the mid-1980s has led to a more decentralized institutions in both countries (Oxhorn et al. 2004). Indonesia followed the same narrative where the centralized-authoritarian rule of then president Suharto met its end in 1998, provided a way for a more democratic bottom-up governance starting in 2001.

Decentralization in Indonesia was part of the nationwide reformasi¹ program. This was done under the background of provincial discontent at around the time (Eaton et al. 2011), specifically in the resource-rich regions. Starting in 2001, provinces and cities/districts are assigned larger fiscal roles that include various governmental affairs excluding executive functions such as foreign and defense affairs. New provinces and cities were also established to meet this demand of a less-centralized state institution. Fiscal decentralization in Indonesia is complimented further with the introduction of direct local elections beginning in 2005. Figure 1 illustrates this change where the share of sub-national against general government spending increased dramatically from 2000 to 2003 during the early phase of decentralization, and then followed by a steady increase until 2017. Consequently, in principle, the sub-national governments are granted larger decision-making abilities.

However, it is important to notice that decentralization transition in Indonesia has not been smooth since the beginning. The transfer of authority from the central to the local government was done in a “big-bang” rush without much preparation and it was directed straight to the lowest level of government, districts and cities, bypassing the province (Nasution 2016). This ill-prepared decision created institutional problems ranging from corruption to the constant lack of capacity building at the local level as pointed out by various studies (Alfada 2019; Henderson and Kuncoro 2004; Kuncoro 2006; Mulyo 2015; Nasution 2016).

Due to the lack of institutional quality at the local level in this first phase, Jakarta slowed down its pace of decentralization with the announcement of Law 32 in 2004. The law introduced the ‘concurrent affairs’ which allowed the active presence of the central government at the sub-national level. Following the law, regulations were set up and various central government’s agencies were re-established in provinces and cities. Some scholars perceived this as an effort to re-centralize political and administrative power (Rahmatunnisa 2015). In between the years, the government has also issued Law 28 on Local Tax that enabled local government to expand their tax base. This law has led to spiking numbers of local regulations in the following years. However, still this does not significantly increase local governments’ capacities especially at the district/city level. Instead, in 2014 Jakarta issued the new decentralization law that (1) maintain the concurrent role of the central government, and (2) provide larger responsibility to the province-level governments at the cost of relatively lower district governments’ role. Despite all the dynamics, in administrative and political sense Indonesia is now a very much decentralized state compared to Suharto’s ‘new order’ era.

It is expected that this dynamic institutional change would bring substantial effect on the economic performance at the local level. Indonesia’s economic progress in the decades prior to this has been hailed as the ‘rising tiger’ due to its unprecedented high growth rate. The progression was largely sustained by the export-driven manufacturing sector, before it was deeply affected by the 1997 Asian crisis. In the subsequent years, the manufacturing share of total GDP had been constantly decreasing, as well as its manufacturing trade specialization (Tijaja and Faisal 2014).

Few studies have explored the link between decentralization on economic development in Indonesia. Setiawan and Aritenang (2019) found a positive impact of decentralization on regional economic growth optimally after 3 years. This confirmed the research of Dartanto and Brodjonegoro (2003), done in the early years of decentralization. They found out that decentralization is positively correlated with local economic growth and the effect is higher in the eastern region. In discussing elite capture, a relevant issue that could affect development under decentralization, Chowdhury and Yamauchi (2010) revealed that village decentralization in Indonesia has not lead to elite capture but rather is representing community’s interests. Contrasting arguments on the impact exists, that highlighted the negative or non-discernible effects of decentralization on the economy (Lewis 2006; Pepinsky and Wihardja 2010; Kuncoro 2006). So far, however, it is still not known whether decentralization can foster industrial development at local level. Industrialization has been regarded as the key towards productive structural transformation in Indonesia and elsewhere (Kniivila 2007; Lin 2010; Otsubo and Otchia 2020). Developing countries employ industrial policy to be able to catch-up with the developed ones. The success of the Newly Industrialized Countries of South Korea, Taiwan, Hong Kong, and Singapore is the famous case of this industrialization strategy by employing export-oriented approach. The same strategy employed by China, through its reform and opening-up (gaige kaifang) program in 1978 that paved the way towards an unprecedented sustained growth for multiple decades.

Considering its importance in sustaining growth and promoting structural change, this paper seeks to fill the gap in explaining the impact of decentralization policy on industrial growth at the sub-national level, using Indonesia as the case study. We understand that theoretically, under neo-classical approach, investment serves as the main channel of industrial growth rather than government spending that, in a decentralized sense, is channelled through the local government. Therefore, we also consider the impact of investment; does higher local government spending expand or diminish investment’s effect on industrial growth? What is the possible underlying cause for the relationship to happen?

The rest of this paper will discuss the literature review on decentralization in Sect. 2, then followed by data description and estimation strategy in Sect. 3. Results and discussion are presented in Sects. 4 and 5, respectively. Section 6 concludes my findings.

Arguments that support decentralization and economic growth have been started at least by Tiebout (1956, 1961) and Oates (1972). Tiebout’s mobile consumer model describes how increasing local budget in public service provisions can benefit individual consumers across-jurisdiction by assuming a free-movement setting. This would incentivize local government as they could enjoy higher tax returns. Oates, meanwhile, argues that fiscal federalism provides Pareto-level efficiency as local government is closer to their constituents thus having a deeper knowledge of local economic needs. These theories are influential and were the driving force behind the decentralization wave that occurred starting in 1980s. Subsequent body of works that support this optimistic view include Musgrave (1969), Weingast (1995), and Qian and Weingast (1997).

In a different manner with the optimistic perspectives above, Bahl and Nath (1986) using a sample of 57 developing countries, found that fiscal decentralization happened along with income growth. Thus, they suggested that for developing countries the policy is more appropriate for the middle and upper-middle income one that are richer. This is, however, not the only prerequisite. Wallich et al. (1995) showed that missing institutional framework has been the source of inefficiency in developing-transitional countries of Eastern Europe. Their study was much related to the concern of the institutionalist approach (Acemoglu and Robinson 2013; Acemoglu et al. 2008; North 1991). The lack of institutional quality in accompanying decentralization programs then became the concern of the later decentralization scholars, (Ahmad et al. 2005; Bird and Smart 2002; Azfar et al. 1999; Litvack et al. 1998). Another institutional work includes also Treisman (1999; 2007a) who revealed that there is no conclusively positive impact of decentralization on economic development for developing countries.

Numerous cross-country empirical evidence analyzing the impact of decentralization programs on growth indicators show negative trends (Colombo and Martinez-Vazquez 2019; Davoodi and Zou 1998; Gemmell et al. 2013; Martinez-Vazquez and McNab 2003; Rodriguez-Pose and Kroijer 2009; Ubago-Martínez et al. 2018; Yushkov 2015). Meanwhile, Martinez-Vazquez and McNab (2006) despite confirming the negative relationship on growth, found that the effect is offset through its positive impact towards macroeconomic stability, measured by consumer price index.

In the case of Indonesia, Kharisma (2013), Kuncoro et al. (2013), and Kis-Katos and Sjahrir (2017) reveal that decentralization has brought positive impact on development. However, for the latter work, in terms of political decentralization they concluded that there are no positive relationships on the public investment in the education, health, and infrastructure sectors. This means that local expenditures in these sectors do not depend on whether the district governments are democratically elected. Another study found out that local government in the post-2001 period was considered to be negatively affecting business, as argued by Kuncoro (2006). He found that firms failed to grow and descents to a process he called ‘informalization’ or downsizing in order to escape from paying local tax and bribes. Elite capture is another concern that often haunting decentralization process. Firdaus (2018) and Lucas (2016) observed this problem at the village level although the findings of Chowdury and Yamauchi (2010) suggest contrary evidence. Meanwhile Alatas et al. (2013) revealed this capture tendency in a targeted welfare program at the district level.

Further studies also showed that Indonesian decentralization does not bring a positive impact on economic performance (Pepinsky and Wihardja 2010). Their estimation method is unique as it involves the creation of “synthetic Indonesia” in which the country was not decentralized. They attributed the problem to local elite capture. On the other hand, Lewis’s finding (2006) pointed out that the negative relationship is also due to the local government’s tax inefficiency. This led to higher dependence on grant transfers from the central government.

Despite some positive findings, many of the studies on Indonesia reflect the cross-country evidence that fail to establish positive relationship. One of the often cited concern is regarding the low capacity of the local government. This situation has been observed by Wallich et al. (1995), Ahmad et al. (2005), as well as Bird and Smart (2002). The problem limits local government in carrying out developmental role. The theoretical evidence, at least in the case of Indonesia, does not support positive moderation effect from local government on investment that translates to industrial growth. Furthermore, this could imply that different agency, i.e., the central government, is responsible in filling the void.

Our analysis uses official data mainly from Statistics Indonesia (Badan Pusat Statistik, BPS). Missing variables due to the unavailability of data are expected when we design this research. The following section discusses data measurement method and descriptive statistics. Then it is followed by elaboration of our estimation strategy.

Our system GMM estimation results are divided into several parts; first we look at the initial impact without covariates before adding in the control variables and the interaction effects between local capital spending data and investments. Following this, we proceed with heterogeneity analysis and robustness check. We bring the findings to formulate our arguments in the discussion part.

Results show that local capital expenditure, by itself, has a positive impact on regions’ industrial growth. Districts and provinces with larger share of capital spending have a higher marginal effect on the sector’s growth. However, simply looking at this relationship without addressing the more important channel of industrialization would result in poor interpretation. Therefore, this research assessed the impact of investments in the sector, and its interaction with local capital spending. The negative albeit insignificant interaction terms across the main specifications suggest that industrial growth could be slower in provinces with high local capital expenditure and high levels of FDI or DDI. These findings are quite puzzling and merit additional discussion.

The results brought up earlier discussions raised by Nasution (2016) that highlighted the weak capacity of the local governments in the case of Indonesia, and the studies of Bird et al. (1995) in the case of democratic-transitional countries. In this case, they have not been able to provide efficient public service to channel private investment for industrial growth. This is an issue that has also been explored previously (see Lewis 2006; Mulyo 2015; Pepinsky and Wihardja 2010).

The weak capacity of the local government has been an inherent problem in Indonesia. There are two sides of this issue. Firstly, decades-long centralized administration under Suharto has impaired much of local institutional capacity. His 1974 decentralization initiative was merely administrative, with political power still largely residing around him. Given this condition, the sudden liberalization of province and district governments following the 1999 Regional Autonomy law has failed to encourage investment growth. Increasing responsibilities that come with fiscal and political decentralization outweigh their capacities and capabilities. Thus, promoting development through structural transformation has not been the primary option of the local leaders. Consequently, their ability to expand their revenue base is limited as productivity was low, which in turn limit their development role especially in the early period of reform (2001–2004).

Under these circumstances, the local government resorted to Jakarta for assistance. Their budgetary structure is composed of large transfer allocated by the Ministry of Finance. At the district level, the share of central government transfer to their own budget was as much as 90 percent in 2001. Meanwhile, their own-source revenue constitutes only 7 percent of their revenue (Nasution 2016).

Secondly, at the same time the central government resolved this capacity gap in the implementation level by introducing the ‘concurrent affairs’ with the 2004 law on Regional Government, and further strengthened in 2014. This paved the way for the reintroduction of central government programs and agencies down to the district level. Currently, most of the strategic infrastructure projects lies within the domain of the central government, or a joint-cooperation between the central and the local government. For the proponent of regional autonomy in Indonesia, this was viewed as a ‘recentralization’ effort (Rahmatunnisa 2015). However, this move has been in line with previous theories addressing the lacking capacity of local government (Ahmad et al. 2005; Bird and Smart 2002). On the one hand, this new institutional arrangement has contributed to the substantial increase, albeit still limited, of sub-national revenue from 11.8 to 19.7 percentage share of GRDP between 2004 and 2019. However, on the other hand, this significant increase in revenue is not reflected on the expenditure side. Local capital spending only increased at merely 1.15 percent from 2.2 in 2004 to 3.35 percentage share of GRDP in 2019.

The dramatic increase in revenue reflects the growth on the productive sector that goes into the provinces’ and districts’ balance through taxes. However, the marginal increase in capital spending serves as a testament on the limited role of local government in promoting industrial development which owes to the two factors described above. Barring other things, capital spending is necessary to facilitate investment through the generation of fix-assets such as roads, bridges, ports, and others in return for larger industrial output.

Another factor worth noting in Jakarta’s decision to make large intervention at the local level is the concern over elite capture. This is quite a common phenomenon in developing countries underwent democratic reform such as in Russia (Blanchard and Schleifer 2001). In Indonesia, the problem has been highlighted by several studies (Firdaus 2018; Lucas 2016; Alatas et al. 2013). This would further prevent capacity building by the district and province governments, thus justified the ‘recentralization’ attempt by the central government.

To sum it up, in a decentralized Indonesia after the ‘reformasi’, the role of the local government in carrying out development project has been limited by both their lacking of institutional capacity as well as the larger presence of the central government. In terms of industrial development, the absent of significant positive interaction terms across all specifications in Table 3 demonstrate this tendency. Although it is not shown in our estimation tables due to the missing variables, we suspect that larger moderation effect on private investment is coming from the central government. Jakarta’s enormous effort on infrastructure development, especially in the last two administrations of Susilo Bambang Yudhoyono and Joko Widodo, contributed to the sustained growth from 2004 to 2019. The positive relationship between local capital expenditure and industrial growth is perceived as a complimentary effect of central government’s program.

Improving the relationship between public and private investment for the benefit of industrial development will remain a major challenge. This burden might be too big of a task for district governments to handle, and less of a challenge for the provinces which has better resources. However, there are several steps that could serve as the alternative to improve local government’s efficiency in this area. First, Jakarta needs to reformulate their ‘concurrent affairs’ strategy. The current strategy of central-local collaboration often does not require substantial role of the local governments, thus has not been able to nurture capacity development effectively, as suggested by our findings. In most cases, local government involvement is limited to the pre-planning and planning stage and much less in the implementation stage. A larger role in the implementation stage would force them to expand their capacity further while at the same time this would reduce the over-dependency issue.

Secondly, from the top-down side Jakarta needs to reduce their direct role at the local level in favor of facilitating and providing assistance to their local counterparts. They should remain responsible for the national strategic development projects and the promotion of inter-provincial linkages, but should not spearhead local development. Facilitation and assistance could be provided through the central government offices located in cities or districts.

Lastly and of no less importance is to prevent elite capture through strengthening institutional/regulatory framework. Capture would discourage capacity development and hindered structural transformation process. This has been one of the concern of Jakarta in the past. Capable local governments would prioritize the development of productive sector.

In order to make this plan work, significant institutional change that alters the “concurrent affairs” concept would need to be carried out. The new collaboration model should empower local governments more. In the long run this would nurture their capacity to better link with industrial investment and thus push for productive structural transformation. This alteration would require political will from all key stakeholders, particularly the central government who must provide all necessary means.

In a decentralized Indonesia, issues regarding local government capacity remained a major institutional challenge. The decision to push for deep decentralization in the early period (2001–2004) exacerbated this situation, lead to Jakarta’s intervention through the second decentralization law that introduced ‘concurrent affairs’. This intervention, in turn, created over-dependency towards the central government, resulting in the negative interaction effect in our finding. To improve this condition, we argue that the central government needs to formulate a new institutional strategy that allows for larger empowerment of its local counterparts. The current model gives too dominant role for the central government in economic affairs, i.e., infrastructure development, and the function is carried out separately between central and local. The collaborative strategy is also meant to improve local governments’ capacity through budget reallocation and Jakarta’s assistance in development projects. Achieving this, however, requires political will especially from the central government who need to incrementally devolve their function to the local government assuming a certain institutional quality has been met. Future research could compliment the finding in this chapter by studying the distributional impact of the central government’s expenditure towards industrial development.