Early Impacts of Indonesia’s Investment Reforms A Preliminary Analysis

The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent.


Policy Research Working Paper 10478
The Indonesian government implemented comprehensive investment reforms in 2021 to encourage investment inflows and related positive impacts. Compared to the previous investment regulation in 2016, the new decree removed foreign direct investment restrictions in over 500 business activities. By estimating the difference in difference in difference event study model, this paper empirically assesses the response in (realized and planned) foreign direct investment and realized domestic direct investment to the reforms. The paper also assesses whether there was growth in investments in fully liberalized sectors linked to Sustainable Development Goals as a proxy for the quality of investment. The results suggest that the investment reforms were associated with increases in realized foreign direct investment, realized direct domestic investment, and planned foreign direct investment, especially in fully liberalized sectors, while there was a decline in all three types of investments in the non-liberalized sectors. The results for planned foreign direct investment suggest that the increase in fully liberalized sectors is likely to continue. The results on direct domestic investment suggest a possible crowding-in effect. Among the fully liberalized sectors, the base metal industry was a key driver of growth, and sectors linked to Sustainable Development Goals sectors had mixed results. The findings provide suggestive evidence of the complementary effect of trade reforms, although the analysis does not specify which of the several reforms may have led to the increase. These results are robust when the possible effects of Covid-19 recovery and other macroeconomic factors are controlled for. These results are also robust to alternative event study models. Further analysis would be needed to observe the trajectory in both the quality and quantity of investment going forward, the distributional effects and the needed complementary reforms to ensure sustainable gains beyond the short run.
This paper is a product of the Macroeconomics, Trade and Investment Global Practice and the Finance, Competitiveness and Innovation Global Practice. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://www.worldbank.org/prwp. The authors may be contacted at amontfaucon@worldbank.org.

Introduction
Regulations toward the liberalization of global capital flows have been debated, which reflects the concerns about the loss of national sovereign rights and other potentially negative consequences of free capital movements. Foreign direct investment (FDI), more than other types of capital flows, has historically sparked such concerns because it often involves a controlling stake by large multinational corporations over which domestic authorities, it is feared, have little control. For these reasons, governments have sometimes imposed restrictions on inward FDI. On the other hand, reforms that facilitate foreign direct investment entry and operations appear to have a significant and sizable impact on the level of inward FDI, whereas anti-competitive product market regulations are negatively associated with investment (Mistura and Roulet, 2019).
Before the reforms analyzed in this paper, Indonesia had the third highest restrictions on FDI among the high-and middle-income countries surveyed by the Organisation for Economic Cooperation and Development (OECD) and was the second most restrictive behind all countries in the East Asia region next to the Philippines (World Bank, 2018). Indonesia's negative investment list (Daftar Negatif Investasi -DNI) included widespread foreign equity limits, sectoral reservations for MSMEs, and special licensing regimes. Until early 2021, the DNI applied at least one investment restriction in almost one-third of all economic sectors, and in 20 percent of them it either limited foreign equity participation or prohibited foreign investment altogether. These restrictions significantly reduced both foreign and domestic investments, reduced entry and performance, increased prices in the sectors to which they were applied (World Bank, 2018), and were associated with productivity losses (Genthner and Kis-Katos, 2022).
In 2021, the Government of Indonesia (GoI) implemented its most comprehensive investment reform program in decades. Most direct restrictions on investments were removed through a February 2021 presidential regulation (Perpres) 10/2021, later amended by Perpres 49/2021 on Investment in May 2021 (World Bank, 2022). The new presidential regulation on investment significantly reduced business activities subject to two main restrictions: (i) sectors that are reserved for MSMEs and Cooperatives, or open with requirements to partner/cooperate with MSMEs and cooperatives and (ii) sectors that are open with certain conditions -i.e., max foreign ownership or certain recommendations 1 The reform eliminated foreign equity limits across a wide range of sectors and turned many sectors where FDI was not allowed or restricted by minority shareholding into sectors fully open to FDI. The reforms were intended to boost FDI inflows and attract/encourage the entrance of multinational corporations (MNCs). In this paper, the term "fully liberalized" refers to the removal of all these restrictions in all the business activities or sub-sector within a broader sector, i.e the number of business activities within a sector that has been liberalized (more below).
This study assesses the impact of this reform, specifically making the following contributions: First, we examine the early effects of the regulatory reform on both planned and realized FDI as well as realized direct domestic investments (DDI), by comparing the pre-and post-policy periods and comparing across sectors in response to the most recent version of the regulation (Perpres 49/2021); by analyzing DDI, we test whether foreign firms substitute or crowd-in domestic investments (Barrios et al., 2018). We also assess whether there was growth among fully liberalized sectors in sustainable investment (i.e. investments linked to Sustainable Development Goals (SDGs) as defined by UNCTAD (2022)).
The analysis is done using a difference in difference in difference (DDD) event study model and controlling for the sector and time-fixed effects. As the data is at a more aggregated level than the listed business activities in the reform regulation, the term "fully liberalized" refers to all the sub-sectors or business activities within the broad sector that have had all the restrictions discussed removed (more in the data section). We focus on liberalized non-commodity sectors to highlight investments less likely to have pressure on the natural environment from the expected increase. Therefore, in line with the World Bank's support to the reform. 2 As robustness, we further account for two key events that were taking place around the reform time: the recovery from Covid-19 lockdowns around the globe and related trade reforms, especially on non-tariff measures the government undertook in November 2021 (Montfaucon et al., 2023). The trade policies also serves as a proxy for the institutional landscape which is key to retaining benefits to FDI (Obuobi et al., 2022).
Results suggest that the investment reforms were associated with increases in realized FDI, realized DDI, and planned FDI. Investments increased most in the fully liberalized non-commodity sectors (23 percent for FDI, 25 percent for DDI, and 19 percent for planned FDI relative to the growth of the pre-policy period). On the other hand, there was a decline in all three types of investments in the non-liberalized sectors. Planned FDI results suggest that the increase in fully liberalized sectors is likely to continue and the results on DDI suggest a crowding-in effect. Among the fully liberalized sectors, the base metal industry was a key driver of growth. While there was increased planned FDI in water management, and in some infrastructure-related sectors for realized DDI and FDI, other SDG-related sectors saw mixed results. To account for the institutional context, we assess the effect of significant changes in trade regulations that took place 6 months after the investment reforms. We find that FDI in fully liberalized non-commodity sectors increased following the trade policy changes, but the analysis does not specify which of the changes may have led to this increase. 3 This suggests the complementarity of trade and investment reforms. We find that these results are robust even when possible effects of Covid-19 recovery are controlled for, when we include GDP and nominal exchange rate variables as covariates and the results are robust to alternative event study models.
Early research on FDI traced its origins to international trade theory, which identified the comparative advantage of host countries as the most important determinant of FDI. When foreign companies invest in a country in order to get access to its domestic market and use it as a platform for exporting to other countries, it might have a "home market effect," according to Dunning (1993).
This could lead to increased FDI inflows and improved domestic economic competitiveness. Our main results are in line with this literature. Recently, studies have highlighted the importance of enhancing and sustaining the host-country effect with the government's role, thus shifting the focus to government policies to increase FDI inflows. Aside from the outstanding role of reforms in the host country, institutions in the host country play an important role, as noted by Aitken and Harrison (2017). Countries with strong institutions and effective governance attract more FDI. MNCs prefer countries with the best institutional arrangements, which are characterized by friendly global trade policies (Cantah et al., 2018;Obuobi et al., 2022). Our results on trade policy are in line with this line of literature. Chen (2018) established that investment liberalization can benefit domestic investment when accompanied by complementary policies that promote domestic enterprise growth. Liberalization may harm domestic investment by replacing indigenous enterprises with foreign competitors, hence FDI liberalization measures should be viewed as just one component of a liberalization strategy and must be supported by adjustments to liberalize local markets in order to help maximize the benefits of FDI. Our results suggest that the reforms Indonesia took broadly had a positive impact on domestic investment similar to results by Setiyanto (2023).
Overall, the body of research illustrates the complexities and context-dependence of the effects of investment liberalization on both domestic and international investment flows. More research is needed to fully understand the variables that govern the amount and direction of these impacts, which this study contributes to, with more recent methodology in the difference-in-difference literature and an attempt to closely link reforms to the specific sectors targeted. Therefore, while these results suggest that the reforms were a key role in boosting investments in Indonesia, complementary trade and other reforms would ensure growth is sustained, especially, for example, since restrictions in the services trade sector remain the highest in Indonesia (Lu and Batten, 2023). Further, the broader impacts of the reforms on the economy and various groups are not captured and are beyond the scope of this paper.
The rest of this paper is organized as follows. In section 2 we detail the policy context and data. In section 3 the methodology is discussed and subsequently results are presented in section 4, including from robustness checks. Finally, section 5 concludes. Indonesia removed most investment restrictions through the recent Presidential Regulation on Investment but upheld the protection of select sectors, such as agriculture and transport. The Presidential Regulation eliminated foreign equity limits across a wide range of sectors. Compared to the previous investment regulation in 2016, the new decree removed FDI restrictions in 553 business activities. As a result, the share of sectors that were exposed to foreign equity limit (FEL) fell from 34.1% in 2016 to 7.6% in 2021 (figure 1a). Thus the reform left several sectors fully open to FDI, such as: (i) relatively small activities fully or partially reserved for Micro, Small, and Medium Enterprises (MSMEs) and (ii) sectors considered to be of public interest that require minority or no foreign ownership. With the remaining restrictions, the share of business activities subject to at least one investment restriction decreased from 38.5% to 20.4% of all economic sectors (figure 1b).

Data
Two sets of data are used in this analysis: realized FDI and DDI values, sourced from Indonesia's Ministry of Investment (henceforth BKPM); and planned FDI sourced from fDi markets data which also proxies investor sentiments/expectations. The BKPM data includes all types of FDI while the fDi markets data included in this analysis does not include mergers and acquisitions. BKPM data is at a quarterly level up to September 2022 while fDi market data is at a monthly level up to March

2022.
Klasifikasi Baku Lapangan Usaha Indonesia or KBLI is the official business classification in Indonesia, published by the Indonesia Centre of Statistics with reference to the International Standard Classification of All Economic Activities (ISIC). The regulation was at a highly disaggregated level at business activities of 5-digit KBLI (see list in the Annex Table A1). However, both data are only available at a more aggregated level at the 2-digit KBLI level of aggregation. As such, mapping sectors 1:1 to the list of business activities in the regulation is not possible. Instead, we are able to map the sectors in the regulation to 5-digit KBLI and then calculate the share of 5-digit KBLI that are liberalized within a 2-digit sector.
There are a total of 68 sectors at 2-digit KBLI. Among these, 6 of the sectors belong to the primary sector, 22 to the secondary sectors, and 40 in the tertiary sector, which represents 9 percent, 32 percent and 59 percent respectively (see Table A1

Trends in Realized FDI
Overall, total realized FDI grew by 64.7 percent between the pre-and post-policy periods. Realized FDI for the primary sector, secondary sector, and tertiary sector increased by 73.9 percent, 92.5 percent, and 21.5 percent respectively, for the period after the policy relative to before the policy.
FDI in fully liberalized non-commodity sectors (i.e secondary and tertiary sectors) increased by 40 percent between the pre-and post-policy period compared to just 6 percent in non-fully realized sectors (Figure 2,a). Although realized FDI in the fully liberalized sectors was already higher than in the non-fully liberalized sectors, it increased more steeply and consistently while other sectors remained flat, or growth was more volatile (  Growth was driven mainly by the paper industry, base metals and tobacco processing which had the largest value contribution Figure 3. The largest growth between the pre-and post-policy change period was also evidenced in activities of service providers for buildings and parks; supporting activities for financial services, insurance, and pension funds. Due to the nature of the aggregated data, we are unable to dissect FDI linked to sustainability in Indonesia exhaustively.
However, from the list of 2-digit sectors that are fully liberalized, we see that there was a 190 percent increase in FDI in education in the post-policy relative to pre-policy and a 212 percent increase in water management.

Empirical Methodology
Empirical Methodology To investigate the causal effect of the reform on direct investment, we use a panel data difference in difference in difference (DDD) event study approach. Our research setup is based on the standard canonical difference-in-difference approach, with two time periods (before and after the policy), and two groups: in the first, no sector is treated/liberalized, and in the second, some sectors are treated (liberalized), but some are not (the control group).
Standard DID approaches, however, as proposed by Callaway and Sant'Anna (2019), face backdrops based on potential outcomes and fail to account for treatment effect heterogeneity, which can also vary with observed covariates and time. To address such shortcomings, we use the Sun and Abraham (2021) approach based on a staggered DID setup, which states that once a group is treated/liberalized, it remains treated for the next period, and at the end, some units remain untreated/ not yet treated. In addition to this approach, we implement the DDD framework to account for unobservable group and time-characteristic interactions. We specify the following estimable Equation 1: (1) Where, D k i,t =1{t-G i =k} is defined as an event study dummy variable which takes the value of 1 if the unit i is k periods away from the start of the treatment period t, and zero otherwise.
and D >L i,t are defined comparably, such that, D 0 i,t = 1 if the unit is first treated at time t, and D 1 i,t = 1 if one period has passed since the time of the first treatment (treatment lags) etc. Alternatively, we have D K i,t equal to 1 if the unit i is treated K times from t (treatment leads). Consequently, k=-1 is the reference period, φ s and ω t represents state and time fixed effects and ε st is the error term. In this paper, Equation 1 is estimated using both realized FDI and Domestic Direct Investment (DDI) data, with K and L limited to 4 lags and 5 leads respectively since this is quarterly data; and planned FDI data, with and Ł limited to 13 lags and 10 leads since this is monthly data.
The selection of leads and lags is based on the time period of the data at the time of the analysis. Thus, if we have 13 months of data following the reform for planned investment for instance, then we select the pre-policy period to be an equivalent time period for comparison. The results will first show the average effect, then the effect period by period.
The primary identification issue is using aggregate data to measure the impact of a policy made at a more granular level. As discussed in the data section, while the reform is at the KBLI 5-digit, the unit analysis is at KBLI 2-digit due to data availability. For example, there is no restriction for manufacturing modern ships. However, traditional ship manufacturing is restricted to the domestic market only. Despite the difference in technology and capital intensity, these two manufacturing sectors belong to the same 2-digit KBLI sector. The estimation may pick up various sub-sector shocks (at 5-digit KBLI for example) leading to higher/lower FDI inflow that may not be related to the reform. These potential shocks cannot be dropped using FE estimation used since we are using KBLI 2 digit as the unit of analysis. This is addressed in two ways. First, we present results in FDI regardless of whether the sectors were liberalized or not (but maintaining non-commodities focus). By doing so, the problem oof identification between the reform and the data is avoided entirely. These results will still tell how investors reacted to the reform even if their sectors were not necessarily fully liberalized.
Investors may still decide to increase investment or commence new projects in anticipation that more sectors may be liberalized and will eventually be included. The second way, which allows us to focus on the fully liberalized sectors, is by reporting results on fully liberalized KBLI 2-digit sectors and zero percent liberalized sectors to draw our conclusions (i.e., the two extremes). By doing so and taking the two extremes, there is no variation that exists in the 2-digit sectors that does not exist at the 5-digit sub-sectors. In some instances, results for other levels of liberalization more than zero and below 100 percent may also be discussed for illustration purposes.
For robustness, we include other controls including proxies for the Covid-19 shock and macroeconomic variables to assess if the results remain consistent.

All Sectors
We first present the results of the response of investment to the reform regardless of whether these sectors were liberalized. Table 1 presents the DDD results for realized FDI, without accounting for whether these were liberalized or not. The Table shows the average treatment effect on the treated (ATET) by sector for all sectors. In the period following the policy, realized FDI in the secondary sectors increased by 67.8 percent on average compared to the period before the policy.
On the other hand, FDI in the tertiary sector decreased by 89 percent following the reform.  Note: Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1. The treatment corresponds to 1 while the control group is 0. ATET denotes the weighted average treatment effect on the treated. ATET estimates are adjusted for sector/panel effects and time effects. The sample period is January 2019 to September 2022. The pre-policy period is defined as the period from the first quarter of 2020 to the first quarter of 2021 and the post-policy period is defined as the period from the third quarter of 2021 to the third quarter of 2022. Source: Author estimations Note: Red spike implies the coefficient is significant at 1 percent, 5 percent and/or 10 percent levels of significance. The baseline/reference period is indicated by the solid vertical line in the plot. The sample period covered January 2019 to September 2022, however, the plots of leads and lags covered 2020q2 to 2022q3 for presentation purposes. See Table A2 in annex for the coefficients of leads and lags, which are adjusted for sector FE and time FE. The pre-policy period is defined as the period from the first quarter of 2020 to the first quarter of 2021, and the post-policy period is defined as the period from the third quarter of 2021 to the third quarter of 2022.

Non-Commodity Sectors by Level of Liberalization
The results based on the share of sub-sectors for which the reform opened up to foreign investment show that for the zero percent liberalized sectors (i.e non-liberalized sectors), realized FDI reduced by 85.8 percent for the period after policy as compared to the period before policy (Table 2). For the sectors liberalized fully at 100 percent, realized FDI increased by 22.8 percent for the period after the policy as compared to the period before the policy.
We also show results for the sectors in between, the two extremes, keeping in mind the identification issues previously discussed. We see that sectors for which sub-sectors were liberalized at >=25 percent< 50 percent, realized FDI reduced by 12 percent for the period after policy as compared for the period before policy, while for the sectors liberalized at >=50 percent< 75 percent, realized FDI reduced by 20.7 percent. For the sectors liberalized at >=75 percent< 100 percent, realized FDI increased by 14.2 percent for the period after policy as compared to the period before policy. While these are illustrative, the results suggest that FDi was higher in the sectors for which more sub-sectors were liberalized.
Figure 6 a and b shows the period-by-period coefficients of the response to the policy for the non-liberalized and the fully liberalized sectors. We see that the changes in FDI for the nonliberalized sectors are mostly below zero percent.   Table A3 in the annex for the coefficients of leads and lags, which are adjusted for sector FE and time FE.

Non-Commodity Sectors-Planned FDI by Level of Liberalization
We find that fully liberalized sectors had the largest increase in terms of FDI announcements following the reform. Planned FDI in 100 percent liberalized non-commodity sectors increased by 18.9 percent after the policy compared to the period before the policy (Table 3). On the other hand, non-liberalized sectors had lower planned FDI by 85.6 percent. Once again, we observe negative impact but a move towards more announcements to more liberalized sectors when looking at the sectors between 0 and 100 percent. This observed shift towards liberalized sectors, may suggest that realized FDI may continue to increase in fully liberalized sectors.

Domestic Direct Investment (DDI)
All Sectors  in the period after the policy as compared to the period before the policy. At the same time, we do not find that DDI declined in the secondary sectors where FDI increased following the reform.
Therefore, this may signal that FDI did not crowd out DDI at the broad sector level. Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 Note: The treatment corresponds to 1 while the control group is 0. ATET denotes the weighted average treatment effect on the treated. ATET estimates are adjusted for sector/panel effects and time effects. The sample period is January 2019 to September 2022.

Non-Commodity Sectors by Level of Liberalization
When we analyze the effect based on the levels of liberalization, we find suggestive evidence of crowding in in the fully liberalized sectors. We find that for the zero percent liberalized sectors, realized DDI reduced by 43.7 percent in the period after policy as compared to the period before policy and increased by 25.2 percent in the fully liberalized sectors (  Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 Note: The treatment corresponds to 1 while the control group is 0. ATET denotes the weighted average treatment effect on the treated. ATET estimates are adjusted for sector/panel effects and time effects. The sample period is January 2019 to September 2022.

Drivers of Growth in Fully Liberalized Sectors and Those Linked to SDGs
We then assess the growth in realized FDI, DDI and planned FDI to check the drivers of growth in these sectors and whether FDI can be linked with sustainable development sectors. This will enable us to assess, to the extent possible, the quality in addition to the quantity of FDI. For that, we use the UNCTAD list of "SDG (Sustainable Development Goals) investment" -project finance in infrastructure, food security, water and sanitation, and health (UNCTAD, 2022). Due to the data challenges outlined in the Data Section and the level of aggregation of our data, we are unable to pinpoint exact investments that are not defined at the 2-digit KBLI level. As such, in this section and specifically within the list of fully liberalized sectors, we include Water management, Land transportation, and transportation through pipelines, Warehousing and transportation support activities, Telecommunication, and Education when discussing sustainable investment.
We find that growth in FDI was mainly driven by the paper products industry and base metal industry (Table 6). Related to investments that can be linked to SDGs, only transport and logistics infrastructure in the form of warehousing and transportation support activities saw an increase in the post-policy period. There was either no statistically significant coefficient in all other sectors or a decline, most notable in Water management.
The increase in DDI of the fully liberalized sectors (as reported in autoreftab:DDI l ib)wasmainlydrivenbyBa Finally, the third column in Table 6 reports growth by sector in the fully liberalized sectors for planned FDI. Growth was mainly in Printing and reproduction media for recording media, Base metal industry, Water management, Food and Beverage Supply, Real Estate, Other professional, scientific and technical activities, Office administration activities, and office support activities and other business support activities. In terms of SDG related investments, there was an increase in water management but a decline in infrastructure.
Overall, the sector that saw growth in realized FDI, DDI, and announcements of FDI was the base metal industry. This is no surprise since Indonesia is a major producer of nickel in the global market and has many mines of other minerals, such as copper, bauxite, iron ore, coal, and gold. Therefore, this is an industry that will likely keep growing. Several sectors (about 50 percent) saw growth in the same direction for FDI and DDI, further signaling a possible crowding-in effect.
Finally, while we can expect an increase in FDI in water management going forward due to the increase in planned FDI, and increase in some infrastructure-related sectors, other SDG-related sectors saw mixed results and even a decline (such as education). Further analysis would be needed sector by sector to observe further both the quality and quantity of FDI following the reform.  Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 Note: The treatment corresponds to 1 while the control group is 0. ATET denotes the weighted average treatment effect on the treated. ATET estimates are adjusted for sector/panel effects and time effects. Each industry is estimates separately.

Effect of Trade Policies on FDI
In November 2021, the Indonesian government undertook more regulatory changes by addressing the trade regulatory regime. The three major regulations enacted were Ministry of Trade Regulations 18,19 and 20 of 2021 which removed some non-tariff measures and introduced other new procedures (Montfaucon et al. (2023)). Considering the main policy being analyzed in this paper was 6 months before these trade reforms and that the effect presented in the graphical periodic coefficients shows that response was mostly after this period, we test the effect of these trade policies on FDI. This is also in line with the literature on complementary institutional reforms in addition to FDI reforms to the extent that these regulatory changes improved the trade regime or were perceived to do so.
We use the same methodology, with the reform period adjusted to November 2021 instead of May 2021, and focus on the all the manufacturing sectors, and the fully liberalized non-commodity sectors.
We find that realized FDI in all manufacturing sectors increased by 103.2 percent following these trade reforms (Table 7). In fully liberalized non-commodity sectors, realized FDI also increased by 21.2 percent after the trade reforms compared to the period before the trade reforms. Figure 7 shows the period-by-period coefficients.
While the caveat should be noted that the post-trade reform period is even shorter than that for the investment reform, this may signal the importance of complementary trade reforms and that the shift between the levels of liberalization is still important. Some of the trade reforms were being implemented after they were enacted, which may have had mixed effects. Crucially, this analysis does not distinguish which of these (many) trade regulatory changes may have been the key to this response as that is beyond the scope of this analysis.  (Table 8). Overall, the increase in FDI in the post-policy period holds even when the possible easing of Covid-19 policies is accounted for.
Additionally, we see that planned FDI results remain robust when Covid-19 lockdowns are accounted for with both manufacturing and fully realized sectors still showing an increase relative to pre-policy periods. We also find that these results hold for DDI where there was an increase in DDI for fully liberalized non-commodity sectors following the reform, although there seems to be a decrease in manufacturing DDI when er account for Covid-19, suggesting that DDI was less resilient and may have faced some crowding out in the manufacturing sectors, but not in fully liberalized sectors. The latter is most relevant in regard to the policy analysis thus the liberalization policy may have been a more crucial factor in driving investments.
Following this, it can be concluded that despite the Covid-19 restriction, the effect of the policy reform remained important, and these results are not necessarily being impacted by economic recovery in 2021. Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 Note: The treatment corresponds to 1 while the control group is 0. ATET denotes the weighted average treatment effect on the treated. ATET estimates are adjusted for sector/panel effects and time effects and controlled for Covid-19 global lockdown. The sample period is 2019q1 to 2022q3. Source: Author estimations.

Accounting for Data Seasonality and Endogenous Investment Policy Effect
The FDI investment policy is an endogenous intervention, which means that FDI inflows are likely to be influenced by other industry growth, among other factors. Furthermore, the FDI data exhibit some seasonality. While our model for the main results largely addresses these, we further adjusted for the suspected effect by including GDP and nominal exchange rate variables as covariates in our main estimation. Controlling for GDP and exchange rate, we found that our results remain consistent (Table 8). There is an increase in the fully liberalized sectors in realized FDI, realized DDI, and planned FDI. Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 Note: The treatment corresponds to 1 while the control group is 0. ATET denotes the weighted average treatment effect on the treated. ATET estimates are adjusted for sector/panel effects, time effects, and covariates (Exchange Rate and GDP). The sample period is 2019q1 to 2022q3. Source: Author estimations.

Model Robustness
Finally, we further conduct other robustness to the consistency of the model in subsection A3, where an alternative event study design is used to assess the impact of the reform on FDI growth in the pre-and post-policy periods. The alternative DiD model follows the approach by Clarke and Tapia-Schythe (2021). We denote states as z and time periods as t. Denoting the outcome variable as FDI, the panel event is defined as follows: Where, φ s and ω t represents sector and time fixed effects and ε st is the error term. From equation 1, the lags and leads to the event/policy of interest are defined as follows: The lags and leads are binary variables that indicate the status of the given state based on the number of periods away from the event of interest in the respective time period. Lags J and leads K are included to show the accumulative lags or leads beyond J and K periods.
We find that there was more growth in realized FDI in the more liberalized sectors, in line with our baseline results. This is also the case for planned FDI, where fully liberalized sectors had larger increases (Figure 8). Therefore, we can conclude that the results remain robust and consistent. More detailed results are provided in subsection A3.  Table A8 and Table A9 for the coefficients Source: Author estimations

Conclusion
This study assesses the impact of Indonesia's investment reforms on realized FDI, DDI and planned FDI using a difference in difference in difference (DDD) event study model and controlling for the sector and time-fixed effects. As robustness, we further account for two key events that were taking place around the reform time: the recovery from Covid-19 lockdowns around the globe and related regulatory changes of the trade regime the government undertook in November 2021. These trade policies also account as a proxy for institutional reform which is key to the sustained impact of reforms on FDI.
The results suggest that both realized FDI and DDI increased most in the non-commodity sectors that had a higher share of business activities liberalized by the reform. The increase in DDI in fully liberalized sectors suggests a crowding-in effect. Planned FDI results suggest that the increase in fully liberalized sectors is likely to continue. Overall, among the fully liberalized sectors, the base metal industry was a key driver in growth in realized FDI, DDI, and announcements of FDI, and sustainable investment sectors had mixed results. To account for the institutional context, we account for trade policy changes that took place 6 months after the investment reforms and find that FDI in fully liberalized non-commodity sectors still significantly increased following the trade reforms, although the analysis does not specify which of the several reforms may have led to this increase. We find that these results are robust even when possible effects of Covid-19 recovery as well as other macroeconomic factors are controlled for. The results are also robust to alternative event study models.
The limitation of this paper is it does not account for possible negative externalities of the reforms on various players in the market. However, by focusing on non-commodity sectors, the aim is to capture investments that are not geared to sectors with potentially larger social or environmental impacts.

A3 Event Study Model:Robustness
Below we provide more detailed results from the alternative model.
For the 0% liberalized non-commodity sectors, FDI grew by 93.2 percent before the policy and 126% after the policy. For the non-commodity sectors liberalized >=25% < 50%, FDI growth before the policy was 84.3% and 148% after the policy. For the non-commodity sectors liberalized by >=50 %< 75%, on average, FDI grew by 85.7% before the policy and 161% increase after the policy. Comparing the non-commodity sectors liberalized by >=25 %< 50% and >=50% <75%, there was a 7.2% positive growth rate from an average of 116.2% and 123.4%, respectively.
For the sectors liberalized by >=75% <100%, on average, FDI grew by 39.7% before the policy and 195% after the policy. For the one hundred percent liberalized non-commodity sector, FDI grew by 21.4 % before the policy and an average of 175% after the policy implementation.
Therefore, there was a shift towards more liberalized sectors.