Unit VI Article Review - APA FORMATTING

The Elusive Pursuit of Import Substitution in 21st Century Indonesia*

Dionisius Narjoko Economic Research Institute for ASEAN and East Asia ERIA Annex Office Sentral Senayan II, 6th floor, Jalan Asia Afrika No. 8 Gelora Bung Karno, Senayan, Jakarta Pusat 10270, Indonesia [email protected]

Titik Anas Presisi Indonesia and Padjajaran University Menara BCA, 50th floor Jl. M.H Thamrin No 1, Jakarta 10310, Indonesia [email protected]

Robertus Herdiyanto Presisi Indonesia Menara BCA, 50th floor Jl. M.H Thamrin No 1, Jakarta 10310, Indonesia [email protected]

Abstract This study evaluates the possible impact of an import substitution strategy recently implemented by the Government of Indonesia on the development of the intermediate inputs industry. Based on de- scriptive analysis, the study suggests that the implementation of the strategy was a misguided decision. Experiences from several countries confirm the negative impact of an import substitution strategy. The econometric analysis shows that trade protection impedes firms’ performance. Thus, it would be prefer- able for the government to have an open trade regime and to intensify foreign presence in the industry.

1. Introduction

Indonesia, like many other emerging countries, views the development of a thriving in- termediate inputs industry, including processed raw materials and capital goods, as a cornerstone of its industrialization process. As laid out in the Master Plan 2011–25, the government aims to promote policies that encourage processing of raw materials and semi- processed materials into parts-and-components, to increase domestic value added and to

* This paper draws from the project “Is Import Substitution a Viable Strategy to Develop the In- donesian Intermediate Industry?,” financially supported by the World Bank. The authors thank Miryana Vinka for excellent research assistance. The authors also thank the comments from two discussants of the paper and the conference participants during the Asian Economic Panel Meeting at Keio University, Tokyo, on 22–23 August 2016.

Asian Economic Papers 17:1 © 2018 by the Asian Economic Panel and the Massachusetts Institute of

Technology

doi:10.1162/ASEP_a_00588

The Elusive Pursuit of Import Substitution in 21st Century Indonesia

reduce imports (Rahardja and Varela 2014). This policy objective has not been achieved, as most of the intermediate inputs for domestic production are still imported. Therefore, the government is interested in evaluating the possible impact of an import substitution strategy (ISS) on the development of the intermediate goods industry.

Generally, an ISS aims to replace previously imported commodities with domestic pro- duction and supply.1 Often, the country using an ISS will impose trade barriers to restrict the import of certain products and encourage the domestic production of these products to displace imports. In line with escalating trade barriers, the government commonly of- fers tax or investment incentives to attract foreign investors to inject capital in these sec- tors by establishing plants domestically, either in the form of greenfield investment or joint ventures.

The proponents of ISS admit that the initial costs of domestically produced products that used to be imported will be higher than world prices. Nonetheless, they argue that the strategy will eventually bring down production costs through economies of scale because domestic producers will have the necessary time to learn and improve the production process.

Despite its optimistic objective, ISS raises few concerns as recorded in the literature. First, overly protected industries will make firms operate inefficiently and will prevent national output from reaching its optimum level through misallocation of resources. Further, too much protection can lead to the creation of excessively long-term rents for domestic pro- ducers. Second, as stated, ISS implementation often coincides with tax and investment incentives. Foreign investors will also make use of those incentives, serve the domestic market, and reap the benefits of the high protection wall. Nonetheless, their profits then will be remitted back to their home country. So, the projection that ISS will generate profits and encourage public savings may not be accomplished. Third, ISS implementation may be supported by government subsidies to import capital goods and intermediate products to develop national industrialization. Hence, the large remittance of profits and import of capital goods may lead to balance of payment difficulties.

This paper aims to ascertain the possible costs and benefits of ISS for the development of intermediate input industries in Indonesia. This is necessary considering that the pub- lic costs for implementing the strategy is significantly high but there is no guarantee of a successful result from this strategy. It is worth noting that Indonesia once applied ISS in the 1970s but dismantled it in 1980s because of the lack of public funds to support the

1 See Todaro and Smith (2015) for further explanation about ISS or the other usually adopted indus- trialization strategy, export promotion.

74 Asian Economic Papers

The Elusive Pursuit of Import Substitution in 21st Century Indonesia

strategy.2 The switch toward an export promotion strategy starting in the early 1990s in fact seems to have provided more benefit, as the country manufacturing sector started to grow rapidly only after a number of reforms/deregulations implemented to support the export promotion strategy.3

This paper continues with an explanation of the methodology used by this study in Section 2. Section 3 presents a descriptive analysis, using several statistical measures. Section 4 provides the econometric framework and estimation results. Section 5 concludes the paper.

2. Methodology

2.1 Data and definitions The study utilizes information from the following data sets:

1. The survey of Indonesian medium-sized and large establishments (with at least 20 em- ployees). This survey contains two groups of data sets: a. The “standard” data set is commonly known as Statistik Industri (hereafter SI). Badan

Pusat Statistik (BPS, or Statistics Indonesia) publishes an annual summary report of Indonesian manufacturing industry based on the SI. This study, however, will use the raw data version of the unpublished data sets of the SI. The SI covers a wide range of information including industry codes, a unique plant code, number of em- ployees, value-added, imports, and export values. Industry codes are defined up to the five-digit International Standard Industrial Classification (ISIC) level. The data set covers the years 1990–2013.

b. The raw material or input data set, which provides detailed information on the in- puts used by each plant in producing each product. The types of input are differen- tiated based on locally produced or imported inputs. The information is available in value, quantity, and units. The period covered by this data set is 2000–13.

2. The United Nations Conference on Trade and Development (UNCTAD) Trade Analy- sis Information System (TRAINS) database. This database provides the tariff rates for products at the six-digit level HS code. This study used data for 1990–2013.

3. The wholesale price index. It is utilized to deflate the nominal value-added. The study will use the three-digit level wholesale price index published by BPS.

2 At that time, the profit from oil suddenly disappeared because of the oil crisis in the 1980s. For more on the implementation of ISS in the 1970s, see Aswicahyono (1997) for a discussion of the government support programs in the automotive sector, namely, a local content requirement and a deletion program to substitute foreign-produced parts with locally produced ones. Also, as an impact of various non-tariff measures applied for the ISS, a number of estimates suggested that the effective rate of protection during this period was high (Hill 1996).

3 See Pangestu (1996) for the elaboration of these deregulation packages.

75 Asian Economic Papers

The Elusive Pursuit of Import Substitution in 21st Century Indonesia

The analysis using all these data will utilize the intermediate input definition from the broad economic categories (BEC), which are in Appendix A. Thus, we will make use of the concordance of the HS and ISIC codes to the BEC.4

2.2 Variables The following describes all key variables and the methods to compute them, organized by the issues covered by the study:

• Value-added: Calculated in real terms using the wholesale price index as the price deflator.

• Employment: Calculated as headcount employment. • Labor productivity: Defined as the proportion of real value-added and total labor. • Export and import: To examine export and import trend over time, we use a data set

from World Integrated Trade Solutions focusing on the trade value of Indonesia in five- year intervals from 1990 to 2013, based on six-digit level HS 1988–92.

2.3 Composition of intermediate inputs by source of the inputs (local or imported, across plants in manufacturing industries and over the time)

The composition utilizes the information from the detailed raw material and input used by plants in SI data, and therefore it is only available for 1996 and 2001–10.

2.4 Ownership pattern (i.e., private local, state-owned, or foreign) of plants producing intermediate inputs

This study examines the pattern of ownership across intermediate inputs during 1990–2013 over five-year intervals. The descriptive analysis draws from ownership share information available in the SI data set. This study classifies the ownership into three categories (pri- vate local, state-owned, and foreign) and, to precisely define the controlling role in joint ventures, it defines three joint venture groups (foreign-government, foreign-domestic, and foreign-government-domestic). All joint ventures are considered foreign plants in this study.5

4 Instead of using the widely known concordance from the United Nations Statistics Division, the concordance utilized in this study is taken from the Ministry of Trade of Indonesia that concords the six-digit HS 2007 to the five-digit Indonesia-specific product code (or known as the Klasifikasi Baku Lapangan Usaha Indonesia, KBLI). KBLI is the more disaggregated version of ISIC rev. 4 and the BEC code.

5 Previous studies suggest that the share of foreign ownership does not necessarily reflect the extent of control. Aswicahyono and Hill (1995), for example, showed that local partners often play a relatively minor role even when they hold most of the equity; this is particularly true for matters related to finance and technology.

76 Asian Economic Papers

The Elusive Pursuit of Import Substitution in 21st Century Indonesia

2.5 Difference in quality between locally produced and imported intermediate inputs This study compares the information from SI data of the locally produced and imported input used by plants, and therefore it is available only for 1996 and 2001–10. Conceptually, price is assumed to reflect the quality of the product, and the average price approach is adopted to control for annual price fluctuation. To maintain coherence with regard to the input unit, we use only those data with unit of kilogram for the reason that it is the most used unit in the data set. Thus, the price of locally produced intermediate input i for every year t (Pl,i,t) is defined as the following, using information of plant p in industry i in SI data:

Pl,i,t = ∑n

p−1 (output)l,p,i ∑n

p−1 (quantity)l,p,i . (1)

Meanwhile, price of imported intermediate input (Pm,i,t) is defined similarly, with m denoting import:

Pm,i,t = ∑n

p−1 (output)m,p,i ∑n

p−1 (quantity)m,p,i . (2)

2.6 Trade protection over time and across industries in manufacturing The patterns cover both final goods and intermediate inputs. This study uses nominal rate of protection (NRP) for intermediate input industries (NRPi) or for final goods indus- tries (NRPj), with i and j denoting intermediate input and final goods sector, respectively. The NRP measures the degree of protection (ideally, from tariffs and non-tariff measures [NTMs]) in a certain product or sector without considering the impact of protections in other sectors that support its production process. In calculating the NRP, import tariff rates of products from the UNCTAD TRAINS database were averaged per sector of the input– output table. NTMs are too complex to be numerically calculated into the NRP; thus, only tariff rates were accounted for in this procedure.

3. Descriptive analysis

The following subsections present the results and their analysis based on the methodology defined in Section 2.6

As stated earlier, the scope of the study covers trends over time related to real value-added, employment, productivity, allocative efficiency, and exports and imports. In addition, we highlight the trends in terms of value, growth, and share. Figure 1 presents the trend in value-added and employment. Looking at the pattern across the group of intermediate

6 Unless otherwise stated, the source of the tables and figures in this section are based on the au- thors’ computation from all data used for this study.

77 Asian Economic Papers

The Elusive Pursuit of Import Substitution in 21st Century Indonesia

Figure 1. Trend in intermediate inputs Indonesian manufacturing, 1990–2013

Source: Authors’ computation.

Note: See those of Table 2.

inputs (Figure 1a), output is dominated by BEC 220 (industrial supplies) consistently over time from the 1990s to the 2000s.

There is no clear pattern over time for BEC 121 (food and beverages, mainly for industry), BEC 210 (industrial supplies, primary), and BEC 420 (parts and accessories of capital goods [except transport equipment]), and their local production is similar. There is, however, an increasing pattern for BEC 220 and BEC 530 (parts and accessories of transport equipment),

78 Asian Economic Papers

The Elusive Pursuit of Import Substitution in 21st Century Indonesia

Table 1. Real value-added for broad economic category 220 (thousands of IDR), 1990–2013

ISIC Rev. 2 1990–94 1995–99 2000–04 2005–09 2010–13

321 37,298 29,036 16,616 17,550 16,530 331 31,058 18,955 12,906 6,336 4,664 341 10,767 8,088 9,532 9,150 10,220

Source: Authors’ computation.

Note: IDR = Indonesian Rupiah; ISIC = International Standard Industrial Classifica- tion; ISIC 321 = manufacture of textiles; ISIC 331 = manufacture of wood and wood products; ISIC 341 = manufacture of paper and paper products.

especially over the last 10 years or so. The escalating pattern may be driven by higher pro- duction by a few industries under BEC 220. Examining the data in more detail in Table 1, we find the top three outputs under BEC 220 are in ISIC 321 (manufacture of textiles), ISIC 331 (manufacture of wood and wood products), and ISIC 341 (manufacture of paper and paper products).

In terms of employment, cross-section pattern is the same as that of output (Figure 1b). Most employment is dominated by industries in BEC 220. The pattern over time for each group reveals two interesting observations. First, there is an increasing pattern for BEC 420 and BEC 530, suggesting expansion of these sectors. Second, in contrast, there is a declining pattern for BEC 220, suggesting improvement in labor productivity, given the increasing output pattern observed in Figure 1a.

Table 2 presents the annual growth rate of output, employment, and productivity of all intermediate input groups from the 1990s to the 2000s. Several observations are worth not- ing. First, output growth has been steady at about 8 percent per year since the mid 1990s for all intermediate inputs (Table 2a). This is much lower (i.e., halved) than the growth in the early 1990s, most likely due to adjustments to the 1997–98 economic crisis.

Second, employment growth fell significantly, from 7 percent per annum in the first half of the 1990s to just below 1 percent per annum in 2000s (Table 2b). This is a mas- sive contraction, again, most likely the result of the more rigid labor regulation after the 1997–98 crisis.

Third, the sharp employment contraction seems to have improved industry performance, albeit unfavorably from a social point of view. Indeed, productivity has increased substan- tially since the crisis, even reaching about 16 percent per annum during 2010–13 (Table 2c).

Fourth, the figures indicate an unexceptional recent performance of BEC 420, and espe- cially for BEC 530. Output and productivity growth for these two sectors were about 20 percent and 30 percent per annum during 2010–13, respectively.

79 Asian Economic Papers

The Elusive Pursuit of Import Substitution in 21st Century Indonesia

Table 2. Growth of intermediate inputs Indonesian manufacturing, 1990–2013 (%)

a) Real value-added

BEC 1990–94 1995–99 2000–04 2005–09 2010–13

121 9.3 0.2 10.7 21.1 0.7 210 29.4 42.6 −11.1 30.6 9.5 220 19.7 9.1 6.2 7.2 3.4 420 26.1 21.7 10.7 6.2 21.6 530 19.1 4.2 34.2 3.0 33.8 All intermediate inputs 18.5 7.9 8.1 7.1 8.4

b) Employment

BEC 1990–94 1995–99 2000–04 2005–09 2010–13

121 3.4 0.3 −1.8 2.4 −3.6 210 1.0 7.6 −3.5 17.2 −0.3 220 6.9 1.0 0.6 −1.4 0.9 420 18.0 2.6 0.1 8.6 1.1 530 8.3 −3.3 6.9 2.8 4.1 All intermediate inputs 7.0 0.8 0.7 −0.2 0.8 c) Labor productivity

BEC 1990–94 1995–99 2000–04 2005–09 2010–13

121 6.0 2.8 13.4 18.8 6.6 210 20.0 29.3 −8.2 13.7 10.5 220 11.9 7.9 5.6 9.4 2.8 420 8.2 15.1 10.9 −0.7 20.3 530 12.1 6.4 25.1 0.6 28.1 All intermediate inputs 8.3 8.9 10.3 4.8 15.9

Source: Authors’ computation.

Note: BEC = Broad Economic Category; 121 = Food and beverages, processed, mainly for industry; 210 = Industrial supplies, nec (not elsewhere classified), primary; 220 = Industrial supplies, nec, processed; 420 = Parts and accessories of capital goods (except transport equipment); 530 = Parts and accessories of transport equipment.

Table 3 illustrates the relative importance over the time and across intermediate input groups. As suggested, BEC 220 dominates national production, but the production of BEC 530 is quite important and has become relatively more important since the mid 1990s (Table 3a). The output share of BEC 530 in total national production increased from 6.8 per- cent in 1995–99 to 22.7 percent in 2010–13.

It is worth noting other important observations concerning BEC 530. First, the figures suggest that there is a sector under BEC 530 that is quite capital-intensive. This is in- ferred from the share of the sector’s employment that has been consistently about 10 percent of the total employment for all intermediate input sectors (Table 3b). Second, in terms of productivity, BEC 530 contributes almost half of the total productivity of all intermediate input sectors (Table 3c). In addition, BEC 420 is the second-largest con- tributor to overall productivity of all intermediate input sectors. Its importance has declined over time, however, due possibly to a rapid increase in the importance of BEC 530.

80 Asian Economic Papers

The Elusive Pursuit of Import Substitution in 21st Century Indonesia

Table 3. Share of intermediate inputs Indonesian manufacturing, 1990–2013 (%)

a) Real value-added

BEC 1990–94 1995–99 2000–04 2005–09 2010–13

121 8.1 5.1 4.3 4.5 4.4 210 0.6 1.1 0.7 0.9 1.3 220 77.2 79.6 76.9 75.8 64.3 420 5.7 7.4 6.8 5.7 7.4 530 8.4 6.8 11.3 13.1 22.7 All intermediate inputs 100 100 100 100 100

b) Employment

BEC 1990–94 1995–99 2000–04 2005–09 2010–13

121 7.9 7.3 6.7 6.8 6.3 210 1.1 1.3 1.1 1.6 2.0 220 82.0 81.6 81.9 79.0 77.3 420 4.0 5.0 5.2 6.4 7.0 530 5.0 4.8 5.2 6.2 7.4 All intermediate inputs 100 100 100 100 100

c) Labor productivity

BEC 1990–94 1995–99 2000–04 2005–09 2010–13

121 18.1 13.2 11.2 12.8 11.0 210 10.0 16.1 11.1 11.1 10.6 220 16.7 18.5 16.5 18.4 13.2 420 25.1 26.5 22.7 17.1 16.7 530 30.1 25.7 38.4 40.5 48.5 All intermediate inputs 100 100 100 100 100

Source: Authors’ computation.

Note: See those of Table 2.

Figure 2 shows the contribution of locally produced versus imported inputs. The contribu- tion from imports tends to be persistent over time at about 25–30 percent, albeit there is a declining trend over the past decade.

Figure 3 illustrates the composition of intermediate input sectors. There are some variations, but the basic trend holds—that is, a declining importance and share of imported inputs. The declining pattern cannot tell us precisely whether firms oper- ating in Indonesia rely on locally produced inputs, however. It may have been dis- torted by, for example, origin of the imports, or other factors that constrain the flow of imported inputs.

A more formal analysis to confirm the story will be done later using econometric method- ology. At this point, however, one may speculate on two potential explanations. First, the declining share may be the result of the increase in the tariff rate of intermediate in- puts. Second, local firms may not be interested in sourcing from imports because they are not aiming to produce intermediate inputs of the highest quality. Theory suggests that high-quality inputs are necessary for export, and many Indonesian manufacturers are not exporters.

81 Asian Economic Papers

The Elusive Pursuit of Import Substitution in 21st Century Indonesia

Figure 2. Composition of intermediate inputs by source, Indonesian manufacturing, 1996–2010 (%)

Source: Authors’ computation.

Table 4 presents the ownership share of the intermediate sectors. The population of inter- mediate input producers in Indonesian manufacturing is dominated by private local (as defined by percent ownership share) companies (Table 4a). This pattern persists from 1990 to 2013. The private local share declined significantly during 2000–04 but soon returned to the higher levels seen in the previous period.

The decline matches the increase in the share of government ownership in the same period, suggesting some acquisition by the government. Evidently, this was temporary, as the share of government ownership subsequently returned to the previous period’s level. The temporary reallocation between private and government ownership may have reflected a bailout policy by the government, which benefitted some local private companies that were severely affected by the 1997–98 crisis.

The foreign ownership shares, although concentrated in a few BECs, are small. Foreign ownership is considered important for industries in BEC 420 and 530, particularly in the post-crisis period, and this was characterized by above-average ownership shares. This indication is in line with the fact that production networks are growing along with the pro- liferation of trade agreements regionally since the early 2000s.

Table 4b displays the ownership share in terms of real value-added. State-owned enter- prises are important for BEC 121, although the trend has been declining in the past five years or so. Lastly, foreign ownership is essential in BEC 220, 420, and 530, starting in the mid-1990s.

82 Asian Economic Papers

The Elusive Pursuit of Import Substitution in 21st Century Indonesia

Figure 3. Composition of intermediate inputs, Indonesian manufacturing by source and broad economic category, 1996–2010 (%)

Source: Authors’ computation.

Note: See those of Table 2.

This section attempts to contrast the quality of locally produced and imported intermedi- ate inputs. Figures 4 and 5 display the patterns for all intermediate inputs and by BEC. In general, imported inputs are of much higher quality and this is consistent over 1996–2010. The figures for each BEC are in line with that of the general pattern, except for BEC 420, where local inputs are of better quality than imported ones. An interesting observation in

83 Asian Economic Papers

The Elusive Pursuit of Import Substitution in 21st Century Indonesia

Table 4. Ownership share by broad economic category in Indonesian manufacturing, 1990–2013 (%)

a) In terms of plant population

Private local BEC 1990–94 1995–99 2000–04 2005–09 2010–13

121 81.7 84.6 68.5 88.6 88.3 210 84.8 88.8 74.2 87.4 87.0 220 89.2 89.3 70.3 85.9 84.8 420 79.2 79.6 65.8 80.2 79.6 530 89.0 83.5 62.5 75.0 74.0 Total 87.9 88.1 69.6 85.3 84.3

State-owned enterprise BEC 1990–94 1995–99 2000–04 2005–09 2010–13

121 17.3 14.3 30.1 9.5 9.6 210 14.3 9.1 23.1 8.2 8.6 220 5.5 3.6 22.0 4.2 3.3 420 6.1 2.1 17.7 2.6 1.6 530 2.9 2.6 20.3 2.2 1.5 Total 6.7 4.5 22.3 4.6 3.8

Foreign BEC 1990–94 1995–99 2000–04 2005–09 2010–13

121 1.0 1.1 1.4 1.9 2.1 210 0.9 2.1 2.8 4.4 4.4 220 5.3 7.0 7.7 9.9 11.9 420 14.7 18.2 16.6 17.2 18.8 530 8.1 14.0 17.2 22.8 24.6 Total 5.4 7.4 8.1 10.1 11.9

b) In terms of real value-added

Private local BEC 1990–94 1995–99 2000–04 2005–09 2010–13

121 43.1 63.5 50.1 51.9 50.2 210 64.5 49.7 60.0 76.1 83.7 220 72.7 72.8 60.0 59.4 62.6 420 69.9 63.4 53.7 52.1 38.5 530 65.2 58.4 29.8 44.3 39.2 Total 72.2 72.6 59.7 59.2 61.7

State-owned enterprise BEC 1990–94 1995–99 2000–04 2005–09 2010–13

121 55.3 34.6 46.4 41.1 27.7 210 35.1 43.7 27.5 7.4 8.8 220 9.7 6.4 19.9 7.7 7.6 420 5.7 1.2 7.2 1.4 2.8 530 14.2 10.3 17.1 1.9 1.2 Total 10.3 6.6 20.0 7.9 7.6

Foreign BEC 1990–94 1995–99 2000–04 2005–09 2010–13

121 1.6 1.9 3.4 7.0 22.0 210 0.5 6.6 12.5 16.4 7.5 220 17.6 20.8 20.0 32.8 29.8 420 24.4 35.4 39.1 46.5 58.7 530 20.5 31.3 53.1 53.8 59.6 Total 17.5 20.8 20.3 32.9 30.6

Source: Authors’ computation.

Note: See those of Table 2.

BEC 121, 210, and 530 is that the quality of imported inputs is significantly superior to that of local inputs, particularly in the case of BEC 530.

84 Asian Economic Papers

The Elusive Pursuit of Import Substitution in 21st Century Indonesia

Figure 4. Quality of intermediate inputs, Indonesian manufacturing by source, 1996–2010 (thousands of IDR)

Source: Authors’ computation.

Note: IDR = Indonesian Rupiah.

Figure 5. Quality of intermediate inputs, Indonesian manufacturing by source and broad economic category, average 1996–2010 (thousands of IDR)

Source: Authors’ computation.

Note: See those of Table 2. IDR = Indonesian Rupiah.

The most pragmatic approach to evaluate trade protection policy is to look at the nominal tariff rates, depicted in Figure 6 for Indonesia during the period 1996–2014. At the onset of the 1997–98 crisis, nominal tariff rates were high, with an average of 17 percent in 1996.

85 Asian Economic Papers

The Elusive Pursuit of Import Substitution in 21st Century Indonesia

Figure 6. Effectively applied tariff (simple and weighted average) (%)

Source: UNCOMTRADE Database, retrieved from WITS (World Integrated Trade Solutions).

Note: See those of Table 2.

Crisis-support assistance from the IMF forced the government to undertake unilateral lib- eralization across the board. Tariff rates then dropped to only 6 percent and continued to decline until 2008–09, when they increased slowly but steadily up to 2014. This pattern is also observed at a more disaggregated level. Among the five identified BECs, BEC 220 and 121 have the highest applied tariff rates, and BEC 210 and 530 had the lowest rates. It is encouraging to see that, in general, the tariff rate is relatively low, which means the inter- mediate inputs sectors are relatively open and supportive to production networks.

86 Asian Economic Papers

The Elusive Pursuit of Import Substitution in 21st Century Indonesia

4. Econometric analysis

An econometric approach is adopted to generate new evidence on the impact of ISS mea- sures (i.e., tariffs and NTMs) on the performance of industries, focusing on intermediate input industries based on the objectives of the paper. Because the impact will eventually affect final goods, however, the study will also assess the impact on final goods industries. The study estimates the following equation for the impact of tariffs (nominal tariff, NRP) on the performance of an intermediate industry,

(Performance)i,t = αi + β1NRPi,t + Zi,t + εi,t , (4)

where is the performance variable considered by this study. The subscript i denotes the in- termediate input industry and subscript t denotes the year. Industry i is defined at the five- digit ISIC level; this is because many of the performance variables are computed using SI data (ISIC-based) whereas NRP is drawn from the UNCTAD TRAINS database (HS-code based). αi refers to intermediate-input and time-invariant effect to capture the unobserved characteristics of the industry.

Zi,t is the set of control variables at the industry level that are introduced to control for: (i) the degree of product market competition, or (ii) other industry characteristics that change over time. Proxies considered for product market competition are:

• Concentration ratio (CR4), or • Measures of firm dynamics (e.g., entry and exit rate).

Meanwhile, the time-variant control variables included in the model are:

• Foreign ownership (measured by the share of foreign ownership in terms of value-added),

• Skill intensity (measured by the ratio of non-production to production workers).

Equation (5) is estimated using fixed-effect panel data model for three periods: 1991–2013 (excluding 1997–2000 [the period of deep crisis]), 1991–96 (pre-1997–98 crisis), and 2001–13 (post-1997–98 crisis). An important justification for the pre- and post-crisis period estima- tion is the finding from a few studies (e.g., Aswicahyono, Hill, and Narjoko 2011; Narjoko 2014) that suggests a change in the structure of Indonesian manufacturing after the crisis (i.e., since the early 2000s).

Performance variables considered by this study are the following:

• Output, calculated by aggregating real value-added across plants in industry i. The wholesale price index is used as the price deflator.

87 Asian Economic Papers

The Elusive Pursuit of Import Substitution in 21st Century Indonesia

• Employment, measured as headcount employment. • Productivity, proxied by labor productivity (LP). LP is defined as real value-added

per labor. • Product quality, proxied by price of a product defined at the six-digit HS code level. This

uses the statistics produced earlier in the section on descriptive analysis. The estima- tion could be done separately for the group of locally produced inputs (using as Pl,i,t the dependent variable) or imported input (using Pm,i,t as the dependent variable).

• Trade variables: export or import, measured as the value of the export or import. • Industrial agglomeration: entry and exit rates of firms. Entry rate is defined as the num-

ber of new plants entering an industry in period t as a proportion to total firms in period t − 1. Exit rate is defined as the number of plants that exit an industry in period t as a proportion to total firms in period t − 1.

The analytics of the ISS suggest that protectionism as the policy tool to support the strategy has a negative impact on the performance of the industry (the failure of the infant indus- try approach, vested interests, etc.). Therefore, a negative relationship between NRP and each of the performance variables, or β1 < 0 , is expected. Nevertheless, a positive rela- tionship, or β1 > 0 , may occur if intermediate input industries exist and the final goods (downstream) industries buy the output from industries. Thus, the impact of protection is ambiguous at the end.

Table 5 presents the econometric exercise for equation (5). Because the major interest of this exercise is the impact of trade protection on the intermediate inputs industry, we only shows the estimated coefficients of NRP against the performance indicators of industry (i.e., output, employment, and productivity) that serve as the dependent variable.

In general, NRP has a significant negative relationship with the three performance indi- cators. This confirms that heightened trade protection measures, in this case import tariff rates, is associated with lower output, employment, and productivity levels. Neverthe- less, there is a clear difference in the magnitudes of the coefficients between the pre- and post-crisis periods. The relationship is weaker in the pre-crisis period. This may confirm the different nature of Indonesian industrialization policy in the two periods. In the pre- crisis period, exchange rates were fixed by the government. Thus, if there were import tariff rate hikes, their impacts were more predictable than in the post-crisis periods when the exchange rates were floated.

The general policy direction in the pre-crisis period was export-orientation and it is widely known that export products require better-quality imported intermediate inputs. The sup- porting policies during this period are characterized by investment-promotion policies and NTM deregulation. Market competition, either externally from countries in the region or internally within the domestic market, was lower than in the post-crisis period. Hence,

88 Asian Economic Papers

The Elusive Pursuit of Import Substitution in 21st Century Indonesia

T ab

le 5.

Im p

ac t

o f

n o

m in

al ra

te o

f p

ro te

ct io

n o

n in

te rm

ed ia

te s

in d

u st

ry p

er fo

rm an

ce s

(l o

g )

O u

tp u

t i ,t

(l o

g )

E m

p lo

y m

en t i

,t (l

o g

) R

ea l

P ro

d u

ct iv

it y

i, t

W h

o le

W h

o le

W h

o le

B E

C p

er io

d P

re -c

ri si

s P

o st

-c ri

si s

p er

io d

P re

-c ri

si s

P o

st -c

ri si

s p

er io

d P

re -c

ri si

s P

o st

-c ri

si s

A ll

in te

rm ed

ia te

in d

u st

ry −0

.0 94

2* *

−0 .0

14 0*

* −0

.1 74

** −0

.0 09

25 **

−0 .0

03 99

+ −0

.0 13

4+ −0

.0 20

8* *

−0 .0

07 21

** −0

.0 31

7* *

B E

C 12

1 −0

.0 18

7* *

0. 00

10 6

−0 .1

35 **

−0 .0

01 39

0. 00

23 2

0. 00

04 4

−0 .0

00 1

−0 .0

01 47

−0 .0

84 6*

*

B E

C 21

0 −0

.1 35

** −0

.0 68

4* −0

.2 27

** −0

.0 03

02 −0

.0 27

8* −0

.0 44

8 −0

.0 20

7+ −0

.0 23

6 −0

.0 29

7 B

E C

22 0

−0 .1

60 **

−0 .0

69 1*

* −0

.2 06

** −0

.0 16

1* *

−0 .0

25 4*

* −0

.0 21

0* *

−0 .0

42 5*

* −0

.0 25

5* *

−0 .0

26 7*

B E

C 42

0 −0

.2 01

** −0

.1 08

** 0.

15 7*

−0 .0

24 3+

−0 .0

26 8

0. 08

94 *

−0 .0

50 8*

* −0

.0 68

2* *

0. 14

3* *

B E

C 53

0 −0

.1 52

** n

.a .

−0 .0

52 5

−0 .0

25 5

n .a

. −0

.0 43

4 −0

.0 36

2+ n

.a .

0. 03

88

S o

u rc

e: A

u th

or s’

co m

pu ta

ti on

.

N o

te :

S ee

th os

e of

T ab

le 2.

N R

P =

N om

in al

R at

e of

P ro

te ct

io n

. **

S ta

ti st

ic al

ly si

gn ifi

ca n

t at

th e

1 pe

rc en

t le

ve l;

* s ta

ti st

ic al

ly si

gn ifi

ca n

t at

th e

5 pe

rc en

t le

ve l;

+ st

at is

ti ca

ll y

si gn

ifi ca

n t

at th

e 10

pe rc

en t

le ve

l.

89 Asian Economic Papers

The Elusive Pursuit of Import Substitution in 21st Century Indonesia

Table 6. Impact of nominal rate of protection on intermediates industry trade indicators

(log) Exporti,t (log) Importi,t BEC Whole period Pre-crisis Post-crisis Whole period Pre-crisis Post-crisis

All intermediate industry −0.0319** 0.000798 −0.100** −0.0459** −0.0166** −0.250** BEC 121 0.00799 0.00981+ −0.0744* −0.0186** −0.00128 −0.103** BEC 210 −0.0880** −0.0439+ −0.173** −0.0841** −0.162** −0.312** BEC 220 −0.0574** −0.0491** −0.112** −0.0667** −0.0354** −0.299** BEC 420 −0.160** −0.0978** 0.0243 −0.0496** −0.0555** −0.120* BEC 530 −0.0953** n.a. −0.0353 −0.0448 n.a. −0.13 Source: Authors’ computation.

Note: See those of Table 2. NRP = Nominal Rate of Protection. ** Statistically significant at the 1 percent level; * statistically significant at the 5 percent level; + statistically significant at the 10 percent level.

despite higher import tariffs than in the post-crisis period, industry performance might not have been hampered, because they presumably were able to maintain their competitive- ness to serve export markets.

Comparing the impact in terms of the coefficient’s magnitude between the three indicators, the relationship with output is the strongest and the impact on employment is the weakest. This shows the important role of imported intermediate inputs on the industries’ domestic production. In contrast, many factors contribute to employment policy at the firm level, and trade protection measures seem to have only a relatively smaller weight in job creation or termination policies.

Moving on to the results across industries, most industries show comparable results be- cause NRP has a negative relationship with all the three indicators. On the other hand, BEC 420 shows rather anomalous results, particularly in the post-crisis years. NRP has a posi- tive relationship with all three performance indicators. This may show the benefits of im- port substitution measures for that particular industry. Domestic producers in the industry can maximize the higher protection from imported products to boost domestic production. This higher output production is presumably achieved by better use of capital because the magnitude of the coefficient of productivity is higher than that of employment.

The impact of NRP on trade indicators presented in Table 6, particularly exports and im- ports, clearly shows a negative relationship between NRP and those indicators in general. The relationship seems to be stronger for imports than exports, however, confirming that tariff rates have a direct impact on imports. Moreover, imports of BEC 210 and 220 suf- fer the most, suggesting a high level of price elasticity of demand of imported products in those industries.

It is also important to examine the impact of NRP on product quality. Table 7 displays the regression results on imported product prices and locally produced prices. The impact on imported products are more varied across industries, although none is statistically

90 Asian Economic Papers

The Elusive Pursuit of Import Substitution in 21st Century Indonesia

Table 7. Impact of NRP on product quality

(log) Price importi,t (log) Price locali,t BEC post-crisis post-crisis

All intermediate industry 0.00927 −0.124* 121 −0.0146 −0.031 210 −0.234 −0.23 220 0.0567 −0.147+ 420 0.011 −0.338+ 530 −0.203 −0.206 Source: Authors’ computation.

Note: See those of Table 2. NRP = Nominal Rate of Protection. * Statistically significant at the 5 percent level; + statistically significant at the 10 percent level.

Table 8. Impact of NRP on firms’ entry

[(Entry rates)i,t ] BEC Whole period Pre-crisis Post-crisis

All intermediate industry −0.00504 −0.00094+ −0.0211 BEC 121 −0.00182 −0.00129* 0.0216 BEC 210 0.00122 −0.00002 0.0163 BEC 220 −0.0122* −0.0017 −0.0380+ BEC 420 0.00847* 0.0101+ 0.00716 BEC 530 0.00199 0.0139 0.0189

Source: Authors’ computation.

Note: See those of Table 2. NRP = Nominal Rate of Protection. * Statistically significant at the 5 percent level; + statistically significant at the 10 percent level.

significant. On the other hand, an increase in NRP is associated with lower prices of lo- cally produced inputs. This may suggest that local products are unable to substitute for imported products that are becoming more expensive because of higher import duties. It may also show the opposite, however—namely, that domestic producers utilize imported inputs in their production as well. Domestic producers have adjusted their product specifi- cations due to more expensive inputs, which may reduce the quality of their products.

The NRP, as its name suggests, offers protections to firms in the domestic market from for- eign producers. Head and Ries (1999) state that lower import duty rates encourage more competition in domestic markets because foreign firms have more access to domestic mar- kets. More intense market competition through lower tariff rates, or vice versa, influences the extent to which a firm can generate profits. As high-cost firms would generate lower profits, they would lose from market competition and, in the worst-case scenario, would have to exit the market.

Table 8 displays regression results of NRP on the entry of plants in intermediate industries across the periods. The results are inconclusive. Most estimates are statistically insignifi- cant. For those that are statistically significant, a negative relationship shows that higher NRPs (or greater protection) do not necessarily provide an incentive for new firms to es- tablish operations. This suggests that establishing intermediate industries requires high (or

91 Asian Economic Papers

The Elusive Pursuit of Import Substitution in 21st Century Indonesia

substantial) initial capital, which most likely comes from the cost of acquiring the technol- ogy embodied in machinery. The incentive from greater protection is only evident in BEC 420 (parts and components except those of transport equipment).

5. Conclusions

This study assesses the impact of recent ISS policies on the development of intermediate industries in Indonesia. The rationale for such a policy was to reduce an excessive reliance on the imports of intermediate inputs and further assumes that there has been little devel- opment or growth of local intermediate industries prior to the implementation of the recent ISS in the country.

The analysis presented in the paper, especially the descriptive analysis, suggests that the decision to apply import substitution on intermediate input industries was misguided. Evidence shows that, contrary to the underlying assumption behind this policy, there has been robust development of intermediate input sectors in the past few decades even with- out government intervention protecting these local industries. As the descriptive anal- ysis shows, national parts and components industries (especially those defined by BEC 530) grew rapidly during this time and started to gain importance in the domestic struc- ture of locally produced intermediate inputs. It is suggested that the robust performance is the result of foreign presence in the industries. Hence, it may be inferred that the ma- jor factor contributing to the past performance is a liberal direct investment regime, as Indonesia’s FDI regime has been quite open since the 1990s and became even more so in the mid-2000s.

The study confirms the theory and experience of other countries on the negative impact of import substitution policies. The econometric analysis shows that trade protection in- hibits the performance of the intermediate input industries. Thus, having a closed trade regime will retard the growth and development of the industries instead of promoting them, and is therefore not recommended. Moreover, firms that import their inputs would become worse off and may exit the market. This is another reason why the government’s decision is misguided. The descriptive analysis shows that having an open trade regime is useful to allow local final-good production access to some specific inputs that are not yet produced locally.

Overall, the decision to apply import substitution to local intermediate input industries is unlikely to be optimal. It may not achieve the objective of having more supply or more- developed intermediate industries. Instead of protecting these industries, a better ap- proach may be to maintain or intensify the foreign presence in the industries. It is clear from the analysis that intermediate input sectors with a growing presence of foreign invest- ment grew rapidly since the 1990s.

92 Asian Economic Papers

The Elusive Pursuit of Import Substitution in 21st Century Indonesia

References

Aswicahyono, Haryo. 1997. Transformation and Structural Change in Indonesia’s Manufacturing Sector. In Waves of Change in Indonesia’s Manufacturing Industry No. 42, edited by Mari Elka Pangestu and Yuri Sato, pp. 1–25. Chiba, Japan: Institute of Developing Economies.

Aswicahyono, Haryo, and Hal Hill. 1995. Determinants of Foreign Ownership in LDC Manufactur- ing: An Indonesian Case Study. Journal of International Business Studies 26(1):139–58.

Aswicahyono, Haryo, Hal Hill, and Dionisius Narjoko. 2011. Industrialisation After a Deep Economic Crisis: Indonesia. Journal of Development Studies 46(6):1084–1108.

Head, Keith, and John Ries. 1999. Rationalization Effects of Tariff Reductions. Journal of International Economics 47:295–320.

Hill, Hal. 1996. The Indonesian Economy since 1966: Southeast Asia’s Emerging Giants. Hong Kong: Cam- bridge University Press.

Narjoko, Dionisius. 2014. Did Better Firms Enter the Indonesian Manufacturing Sector after the 1997– 98 Crisis? Bulletin of Indonesian Economic Studies 50(3):371–391.

Pangestu, Mari. 1996. Economic Reform, Deregulation, and Privatization: The Indonesian Experience. Jakarta: CSIS.

Rahardja, Sjamsu, and Gonzalo Varela. 2014. Nothing to Fear but Fear Itself: Evidence on Imported Intermediates in Indonesia. Economic Premise Note Series No. 138 (March). Poverty Reduction and Economic Management (PREM) Network of the World Bank.

Todaro, Michael Paul, and Stephen Charles Smith. 2015. Economic Development. Boston, MA: Pearson.

Appendix A. Classification of broad economic categories

1. Capital goods 41* Capital goods (except transport equipment) 521* Transport equipment, industrial 2. Intermediate goods 111* Food and beverages, primary, mainly for industry 121* Food and beverages, processed, mainly for industry 21* Industrial supplies not elsewhere specified, primary 22* Industrial supplies not elsewhere specified, processed 31* Fuels and lubricants, primary 322* Fuels and lubricants, processed (other than motor spirit) 42* Parts and accessories of capital goods (except transport equipment) 53* Parts and accessories of transport equipment 3. Consumption goods 112* Food and beverages, primary, mainly for household consumption 122* Food and beverages, processed, mainly for household consumption 522* Transport equipment, non-industrial 61* Consumer goods not elsewhere specified, durable 62* Consumer goods not elsewhere specified, semi-durable 63* Consumer goods not elsewhere specified, non-durable

Source: United Nations Statistics Division.

93 Asian Economic Papers

Copyright of Asian Economic Papers is the property of MIT Press and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.