Csis.or.id

ERIA Research Project Report 2008, No. 1 DEEPENING
EAST ASIAN
ECONOMIC INTEGRATION
Edited by
JENNY CORBETT
SO UMEZAKI
TABLE OF CONTENTS
Deepening Economic Integration in East Asia: Services Liberalization toward the ASEAN Economic Community Investment Climate Study on ASEAN Member Countries Background Data for Constructing Quantitative Measures Firm-level Analysis of Globalization: A Survey Kazunobu Hayakawa, Fukunari Kimura, and Tomohiro Machikita Gains from Fragmentation at the Firm Level: Evidence from Japanese Multinationals in East Asia Fukunari Kimura, Kazunobu Hayakawa, and Toshiyuki Matsuura Learning-by-exporting in Korean Manufacturing: Chin Hee Hahn and Chang-Gyun Park Vertical and Horizontal FDI Technology Spillovers: CHAPTER 10. Plant Entry in a More Liberalised Industrialisation Process: An Experience of Indonesian Manufacturing during the 1990s CHAPTER 11. Maximizing Benefits from FTAs in ASEAN Daisuke Hiratsuka, Kazunobu Hayakawa, Kohei Shiino, and Seiya Sukegawa CHAPTER 12. A Survey of Micro-data Analyses in Indonesia CHAPTER 13. A Survey of Micro-data Analyses in Vietnam: CHAPTER 14. A Survey of Micro-data Analyses in Australia CHAPTER 12
A Survey of Micro-data Analyses in Indonesia
HARYO ASWICAHYONO
Centre for Strategic and International Studies Globalizing corporate activities, deepening economic integration and its impact on the performance of local firms has increasingly become a subject of extensive discussion. As a result, a growing body of literature on this subject has developed rapidly. This paper provides a survey and evaluation of this literature. It seeks to find in the literature answers to important questions such as: why some firms export abroad and others do not, why some firms fail to survive under intense pressure from globalization, whilst others do and why some choose to invest abroad rather than export. Since MNE is becoming more important, it is also necessary to survey the impacts of the presence of MNE and exporting activities on domestic firms. However, analysis on the subject for developing countries is sparse. This paper, therefore, pays particular attention to the empirical micro analysis of Indonesian firm performance Introduction
Interest in the impact of globalizing corporate activities and deepening economic integration on the performance of local firms has developed over the last decade. The interest has led to a new and rapidly expanding body of literature on the subject. As a result, the literature has generated new insights on why some firms export abroad and others do not, why some firms fail to survive under intense pressure from globalization, whilst others do and why some choose to invest abroad rather than export. Another strand of literature seeks to answer the question whether the presence of MNE and exporting activities has a positive impact on domestic firms. In short, the new literature sheds light on the key drivers of globalization and the impact of the phenomenon on local firms’ It goes without saying that the impact of globalization and economic integration differs between countries, depending on the stages of economic development, industrial structures and policy environment within which firms operate. The extent to which important questions posed above can be answered also depends on the nature and the quality of the available dataset. It is therefore important to review the literature in the context of the specific conditions surrounding a country. This paper reviews the rapidly growing literature of micro-data analyses on the themes in Indonesia. Export Decision
Recent empirical literature on international trade has put heterogeneity as its core aspect. One important aspect of this heterogeneity is firms’ participation in international trade: some firms export abroad and others do not. It raises an obvious question: what are the determinants of firms participation in exports. For example, using a sample of 650 Columbian firms throughout the 1980s, Roberts and Tybout (1997) found a significant impact of sunk costs on the decision to export. In their paper, they detected the presence of sunk costs by testing if the previous export activity of the firm could be used to explain its current status, and found that previous participation in exporting increases the probability of current export activity by up to sixty percent. Another important hypothesis is whether sunk costs can be affected by spillovers from other firms. Aitken, Hanson and Harrison (1997), for example, suggest that a firm in an export intensive sector may find its cost of entering the foreign market reduced by the export activity of other firms. Moreover, Aitken et al. hypothesise that such spillovers would be even larger from multinational companies. The reason they suggested is that the presence of MNC might operate as a “natural conduit for information about foreign markets, foreign consumers, and foreign technology” to domestic firms. Aitken et al tests this hypothesis empirically on a sample of Mexican firms from 1986-1990, and found that multinational firms do have a positive spillover effect on the probability of domestic firms Sjoholm and Takii (2003) focus their analysis on the role of foreign network on export participation of Indonesian firms. In particular, they hypothesized that foreign contracts may increase the likelihood of exports. The model they use is profit maximizing firms decision to export (or not to export) under the existence of sunk entry costs. They use firm level Indonesian manufacturing panel data between 1990 and 2000. The dataset contains 197,195 observations for 26,987 plants during 1990-2000. From the model, they derive time- specific dummy variables and time-variant plant-specific variables, while the amount of foreign contacts was captured by a dummy variable on foreign ownership and a dummy variable for imports of intermediate goods. In addition, they include several control variables, such as public ownership, labor productivity, capital intensity, scale economies, and 3- digit ISIC level industry dummy variables. The paper employed OLS, fixed effect, and GMM approach. Even though, the authors prefer GMM methods due to its superiority in terms of unbiasedness and efficiency, they include OLS and fixed effect model in order to compare the result with previous studies. The study confirms the relatively high export orientation and flexibility of foreign- owned plants, in the sense that foreign-owned plants that began their operation in Indonesia by producing only for the domestic market are more likely than domestically-owned plants to start exporting. Inclusion of plant specific variables such as size, capital intensity, and labor productivity do not change the result of the study. A similar study regarding the role of foreign ownership in exporting can be found in Ramstetter and Takii (2005). Instead of using industry dummy variables, they ran regression to 13 Indonesian manufacturing industries from 1990 through 2000 separately. The foreign influence is captured by dummy variables for minority, majority and wholly- foreign plants, while the control variables are capital intensity, production-worker intensity, size and vintage. There is not much discussion on the econometric technique that they use except that they utilized Tobit estimator. The results confirm the previous study. They found that differences between MNCs and local plants remained positive and statistically significant even after the influences of factor intensities, plant size, and plant vintage were accounted for. Second, the results indicated that the size of these differences was usually reduced by accounting for the influences of factor intensities, plant-size, and plant vintage. Third, heavily-foreign MNCs tended to have the highest export propensities, but differences among foreign ownership groups were statistically insignificant in about half of the industry-period combinations examined. Statistically significant differences among foreign plants were concentrated in the mid- to late-1990s and in five industries, textiles, plastics, basic metals, metal products, Narjoko and Atje (2007) is another literature on the decision to export. The paper tries to explain the sluggish export performance during and after the crisis. It uses the same firm level panel data from Industrial statistics for the period of 1997-2004. They estimate to set of equations. They use the Probit model to estimate the first equation that explains a plant’s probability to export. The second equation explains the growth of a plant’s export propensity. To avoid selection bias for the second equation, they employ Heckman’s two- steps estimation procedure, within which, the computed inverse Mills ratio in the first equation is added in the explanatory variables of the second equation. The explanatory variables for the two equations are more or less similar to the previous study: previous year export, capital intensity, skill intensity, labor productivity, foreign ownership, size, age, and imported –input dependence. However, since the primary concern is to analyze the impact of the crisis on a firm’s export performance, they add financial leverage as an explanatory The major findings from the econometric analysis can be summarized as follows. First, being an exporter in the previous year significantly increases the probability to remain as an exporter in the current year. However, the result suggests the impact of exporting history was higher in the recovery rather than in the crisis and early recovery period (1999-2000). Second, the results strongly support the self- selection hypothesis, where firms need to be efficient to compete in highly competitive export markets. Third, in general, firms with some foreign ownership are suggested to have a higher chance to participate in exports compared to their domestic counterparts. However, the importance of this effect is shown to have been much weaker during the period 2002-04. Finally, the paper shows that the extent of financial constraint does not matter in determining the export supply response of Firm’s Survival
The second important question is what determines survival of firms under globalization. There are three channels through which globalization may influence firms’ survival: reduction in trade cost, competition from importing products, and foreign capital Studies that look into the impact of foreign capital shares on firms’ survival are rare, even for developed economies, and most of the studies show inconclusive results. For example, Bernard and Jensen (2002) found that U.S. multinationals are substantially less likely to close than other U.S. plants over five year intervals. However, after controlling for establishment characteristics, they found the opposite result. Gorg and Strobl (2003) found that Irish plants with majority foreign ownership were less likely to survive. Ozler and Taymaz (2004) found similar results to that of Bernard and Jensen (2002) for developing countries. Their analysis of foreign and domestic establishments in the Turkish manufacturing industry for the period 1983-96 indicates that foreign establishments have a higher survival probability. However, when the establishment characteristics are controlled for, domestic establishments have the same survival probability. Bernard and Sjoholm (2003) pose an intriguing concern that reliance on foreign nationals may be a risky development strategy as foreign firms are likely to be less rooted in the local economy and may be quicker to close down production. Using data from 1975- 19891 and employing semi parametric estimation of the hazard function, they found that plants with some foreign ownership are far less likely to close down compared to completely domestic- owned plants. However, the difference in firms’survival rate is not the result of plants nationality of ownership, but is caused by plant characteristics. Controlling for size and productivity, they reveal that foreign ownership is associated with increased probability of closure. Using information on changes of ownership, they further test the result and found that foreign ownership, rather than unobserved plant characteristic, is associated with the lower survival rate. Narjoko and Hill (2007) analyze firms’ survival within an economic crisis situation. They focus their attention on export and ownership variables by regressing the percentage change in RVA on export and foreign ownership over the period 1998-2000. To test whether the level of foreign ownership matters, they include foreign ownership dummy and interaction term between the dummy and foreign ownership level. Various control variables such as size, age, financial leverage, import dependence, industry concentration, import penetration, trade protection, and dummy variables representing region and industry are also included in their equation. Finally, to avoid censoring bias, they employed Heckman’s two-step estimation technique. The result is consistent with studies of other crisis episodes, foreign ownership and prior export orientation are found to be highly significant determinants of survival and 1 The reason for not using the more recent data is that beginning in 1990, plants were excluded from the sample if their size fell below 20 employees. Thus it is not possible to know if a plant exits in the 1990s because of closure or because of decrease in size. recovery. The effects of firm size are ambiguous, a result which tends to refute the popular notion that smaller firms are more adaptable in times of crisis. The industry in which firms are located, in particular its factor proportions, is also found to be significant. FDI Spillovers
Multinationals are distinguishable from local firms because the proprietary technology that MNCs owned allows them to compete successfully with local firms. Moreover, the entry of foreign firms disturbs market equilibrium and forces local firms to take action to protect their market share and profits. Those two factors may create various externalities that benefit local firms. The channel through which these externalities spillover to domestic firms are: increased competition, labor turnover, or through demonstration. Spillover effect has been tested by a large number of papers and summarized by Gorg and Greenway (2004). They found that robust empirical support for positive spillovers is, at best, mixed. There are two explanations for the mixed results, first the positive competition effect from the presence of MNC may be outweighed by the negative impacts of the decrease in production per firm. Second, the heterogeneity of MNC and domestic firms makes it difficult to pin point the impact of MNC since not all types of MNC provide spillover benefits, and not all types of domestic firms have the capacity to obtain spillover The first important question that needs to be answered is whether foreign firms have a higher level of productivity compared to domestic firms. Applying simple regression techniques to 1991 manufacturing establishment data and after controlling for skill level, capacity utilization, scale and industry dummy, Blomstrom and Sjoholm (1998) found that foreign establishments have comparable high levels of labor productivity. By adding minority and majority owned foreign affiliates dummy, they also show that both minority and majority- owned foreign affiliates are more productive than domestic firms. Moreover, the coefficients for those two dummy variables are similar in size and a chi-square test can not reject the hypothesis of equal coefficients. A more detailed and longer time period (1975-2000) analysis by Takii and Ramstetter (2004) confirm the early findings. They found that MNCs generally had much higher average labor productivity than local plants and these differentials persisted after accounting for electricity consumption per worker, size and vintage. They also found that there was also a large variation in productivity differentials across industries and time, with statistically significant differentials most common in chemicals, metal products, and transportation machinery. However, unlike the previous findings, they found significant differences between minority, majority, and heavily-foreign MNCs, with lower labor productivity for minority and heavily-foreign MNCs compared to the majority-foreign The second question relevant to the themes is whether the presence of foreign firm affiliates creates positive externalities that spillover to domestic firms. Blomstorm and Sjoholm (1998) examined the spillover hypothesis by testing whether labor productivity in local firms varies with the degree of foreign production in an industry. They found a positive and statistically significant coefficient of FDI, which suggests that domestic establishment benefit from the presence of foreign establishments in the same 5-digit industry. Given the previous finding of no labor productivity differences between minority and majority- owned foreign affiliates, they expect a larger spillover from minority- owned foreign affiliates. But, regression results contradicted their expectation; the degree of foreign ownership of an establishment did not seem to affect the amount of intra-industry Tomohara and Takii (2005) provide an analysis of another channel through which positive externalities from foreign establishment’s spillover to domestic establishments. Using the generalized method of moments they estimated the dynamic model using panel data to examine whether foreign direct investment benefit workers employed by domestic companies in a host developing country. They found that the MNCs had positive externalities on the wage level of domestic companies. They also found that employees in domestic companies enjoyed increased wages through two spillover channels, those resulting from increased productivity and those resulting from equity concerns. Export Spillovers
In the theoretical IO and trade literature, it is often argued that participation in export activities may be beneficial to a country because of increased productivity through reallocation of resources from inefficient firms to more efficient firms (self-selection hypothesis), and through learning by exporting. Moreover, the benefit may be magnified if we can find the existence of productivity spillovers from exporting. Sethupathy (2007) argued that there are three channels for export spillover. First, under economies of agglomeration, the knowledge accumulated from learning by exporting could easily spillover to other firms in the same industry (horizontal spillovers). Second, a highly competitive international market forced the exporting firms to use higher quality inputs. This could result in exporting firms sharing knowledge and technology with their upstream partners in order to improve the inputs that they receive (upstream spillovers). Finally, the exporting firm’s improved productivity could lead to higher quality input for its downstream partners, which in turn could have a positive effect on downstream The export spillover hypothesis has been tested through various studies. Alfarez and Lopez (2006) summarized the studies and stated that: “these studies either do not find evidence that export activity increases the probability of exporting (e.g. Clerides et al., 1998; Barrios et al., 2003; Bernard and Jensen, 2004) or find that only multinational exporters generate spillovers (e.g. Aitken et al., 1997; Greenaway, et al., 2004; Ruane and Sutherland, 2004). The effect of exporting activity on export intensity of exporters is also not clear. While some find a positive effect of exporting activity by multinationals on export intensity (e.g. Greenaway, et al., 2004) others find a negative effect (e.g. Ruane and Using plant-level data from Chile, Alfarez and Lopez (2006), in general, found upstream export spillover. However, distinguishing between foreign and domestic firms, they found heterogeneity in the channel of spillovers. Exporting by foreign-owned plants generates positive spillovers in all directions, while domestic exporters increase productivity of their suppliers and, to a lesser extent, that of plants in the same sector. Sethupathy (2007) used two-step methodology to estimate the effect of exposure to exporting on productivity. First, he estimated plan level productivity. Second, using GMM he regressed the TFP result with lag TFP, dummy for whether the plant became an exporter in the previous period, and three proxies to measure the extent of exposure to horizontal, upstream, and downstream relationship. He found productivity gains to downstream firms of approximately 2.5-5.0% during the period 1990-1996. However, he did not find the presence of spillovers upstream or horizontally. Industrial Demographics and Productivity Growth
Mainstream economic literature usually uses representative plant approach in the estimation of TFP growth. However, if all plants were identical, the only source of TFP growth would be productivity improvements occurring through simultaneous productivity improvement within plants. However, this simplification is in stark contrast to the fact that plants differ in various characteristics such as, size, age, factor proportions, technology, as well as productivity levels and growth rates. Moreover, representative plant approach masks the microeconomic dynamic due to changing plant demography. Using yearly Industrial data 1975-1995 at the plant level Vial (2008) decompose TFP growth into intra-plant TFP growth, market share reallocation among incumbents and plant turnover effect. The paper used three decomposition methods: (i) TFP aggregated with market shares, (ii) TFP aggregated with market shares, TFP relative to the average, (iii) TFP aggregated with market shares and time- average market shares, TFP relative to the

Source: http://www.csis.or.id/sites/default/files/atch/2009-03%20Aswicahyono,%20H.%20''A%20Survey%20of%20Micro-data%20Analyses%20in%20Indonesia''%20ERIA%20Research%20Project%20Report%202008.pdf