Write a resume about " Exchange Rates and Trade Balances: Effects of Intra-Industry Trade and Vertical Specialization" article. 250 max word count, not formatting requirements.

Abstract: To examine how changes in relative national prices affect trade flows, this study estimates the impact of changes in industry-specific effective real exchange rates on industry-level trade balances, exports, and imports. We analyze the variations in industry-specific intra-industry trade and vertical specialization, which may both govern the long-run relationship between real exchange rates and trade flows. We employ sample information from 13 manufacturing industries across five Asian countries from 2001 to 2015. Rather than country-level aggregate measures, we use disaggregated industry-specific real exchange rates, which provide better measures of relative national prices and may help to uncover different responses that are masked by aggregate data. Fixed-effect estimations reveal that greater price competitiveness, as measured by depreciation in industry-specific effective real exchange rates, increases industry-level trade balances. We find that the elasticity of industry-level trade balances with respect to industry-specific real exchange rates declines as vertical specialization increases. There is also some limited evidence that this elasticity increases as intra-industry trade increases. Taken together, our findings suggest that global supply chains are more important than intra-industry trade in examining the response of trade balances to real exchange rate changes. Importantly, these heterogeneous impacts imply that policies regarding exchange rate management may be of limited potency and will affect different industries in different ways. 1. Introduction

Relative national prices have long been among the key variables in international economics. In particular, economists are interested to know how differences in relative national prices determine the patterns of trade, and how changes in those prices alter trade flows. In this study, we focus on changes in industry specific effective real exchange rates (IERER), which measure relative national prices at the industry level, and their effects on industry trade balances, exports, and imports. We account for the variations in intra-industry trade and vertical specialization in order to explore the heterogeneities that may exist in the long-run relationship between the exchange rate and trade balances. Generally, a depreciation in a country's currency is expected to harm importers as imports become relatively more expensive while benefitting exporters as exports become relatively less expensive (more competitive) on world markets. More nuanced responses are expected across different industries, however. In particular, these responses may dramatically change if there are greater amounts of intra-industry trade or greater vertical specialization. Kharroubi (2011) offers country-level evidence for a sample of 20 OECD countries, highlighting two important channels: 1.) an increase in trade within industries, leading countries to trade substitutable products, increases the sensitivity of the trade balance to real exchange rate movements; 2.) an increase in vertical specialization, leading countries to trade complementary products, reduces the sensitivity of the trade balance to real exchange rate movements. The first channel directly relates to variations in intra-industry trade (IIT). With an increase in IIT in a given industry, a home country increasingly engages in exporting products that are similar to its imports. As a result, import substitutes are more readily available to domestic firms and consumers, who can substitute domestic products for similar imported products when they face an increase in the relative price of imports due to a depreciation in the home currency. To fix intuition, consider two industries in a given country: one with high levels of IIT and one with low levels of IIT. More import substitutes are available in the first industry due to the two-way exchange of similar products (Krugman, Obstfeld, and Melitz, 2012; pp. 164-171). In that industry, consumers are more likely to substitute domestically- 2 produced products for imported goods following a real exchange rate depreciation. In the second industry, however, consumers do not find close substitutes for imported goods, which is why imports in that industry may not fall as much in response to a real exchange rate depreciation. Thus, the trade balance in the industry with high levels of IIT is expected to be relatively more sensitive to changes in the real exchange rate, compared to the industry with low levels of IIT. This implies that in the long-run the elasticity of the trade balance with respect to the exchange rate increases as IIT increases. The second channel relates to variations in vertical specialization (VS). With an increase in VS in a given industry, a home country's exports rely more heavily on imported intermediate goods. As a result, exporters may not benefit from improvements in relative prices following a depreciation in the home currency. Likewise, they may not suffer much from an appreciation of the home currency as their purchasing power of imported content increases, offsetting part of the adverse effects on exports. Again, consider two industries in a given country: one with high levels of VS and one with low levels of VS. In the first industry, exports rely more heavily on imported inputs. In this case, real exchange rate depreciation negatively affects the imports of vertically specialized intermediate goods and, in return, the exports of products that rely on those imported intermediate goods. By contrast, in the second industry, exports do not rely on the imports of intermediate goods. In that industry, exports may benefit from real exchange rate depreciation, as they become relatively more competitively priced on world markets. Thus, the trade balance in industries with high levels of VS is expected to be relatively less sensitive to changes in the real exchange rate, compared to industries with low levels of VS. This implies that in the long-run the elasticity of the trade balance with respect to the exchange rate declines as VS increases. Kharroubi's (2011) findings at the aggregate level corroborate the impacts noted above, showing that in countries with high levels of intra-industry trade, where trade takes place more within industries, the country-level trade balance is more sensitive to changes in the effective real exchange rate. He also finds that in countries where exports rely heavily on imported content, the country-level trade balance is less sensitive to changes in the effective real exchange rate. However, these country-level results may mask important differences across industries. Some industries may rely more heavily on imported inputs, 3 having integrated global supply chains. Such vertical specialization indicates that these industries may respond very differently to changes in the aggregate real exchange rate relative to an industry that has a high degree of intra-industry trade. Thus, rather than relying on country-level variations, we focus on industry level differences. Importantly, we measure the response of industry-level trade balances to industry-specific exchange rate measures, which provide a relevant measure of relative national prices that are measured more precisely than national exchange rates. Because of differences in industry-level relative national prices, there exist significant variations in industry-specific real exchange rate measures in a given country. Also, sectoral heterogeneities may impact the measurement of real exchange rates as different sectors interact with each other in different ways, affecting both the inter-sectoral linkages and the reliance on traded inputs (Patel, Wang, and Wei, 2019). We allow for heterogeneities in the real exchange rate based on industry prices and account for heterogeneities separately in the trade within and between industries by including the IIT and VS measures in our analysis. These heterogeneities are of importance for policy makers as they consider how exchange rate policies may be used to impact global trade imbalances. For example, an industry with high IIT and low VS may respond strongly to real exchange rate movements, thus impacting the trade balance. Conversely, an industry with low IIT and high VS may exhibit very little response to real exchange rate movements. To analyze these heterogeneities, we employ industry-level sample information from a panel of 13 manufacturing industries across five leading Asian exporters (including China, Japan, Korea, Malaysia, and Indonesia) from 2001 to 2015. These countries play an important role in global exporting activities. They exported about $4.15 trillion in 2015, which is equal to 19.52% of world exports and 63.07% of exporting activities in the East Asia and Pacific region in the same year (World Bank, 2019a).1 1 World exports in 2015 were $21.28 trillion. In that year, exports from East Asia and Pacific region were $6.59 trillion, with China's exports at $2.36 trillion, Japan's exports at $0.77 trillion, Korea's exports at $0.63 trillion, Malaysia's exports at $0.21 trillion, and Indonesia's exports at $0.18 trillion. 4 We employ this sample to examine how changes in IERER affect industry-level trade balances, exports, and imports while accounting for the changes in time-varying industry-specific variations in IIT and VS. Our findings suggest that in the long-run an increase in price competitiveness, as measured by a depreciation in IERER, contributes to an increase in industry-level trade balances. We show that this long run relationship is governed by changes in IIT and VS. Without controlling for the variations in IIT and VS, our basic estimation results suggest that the absolute value of the elasticity of trade balances with respect to IERER is equal to 0.83. This implies that a 1% depreciation in IERER is associated with a 0.83% improvement in industry trade balances. The absolute value of the above elasticity declines to 0.79 after we include the controls for variations in IIT and VS. More important, keeping all else at their means, our estimations suggest that the absolute value of the above elasticity increases from 0.79 to 0.95 following a one standard deviation increase in IIT from its mean, and that it declines from 0.79 to 0.38 (statistically, indistinguishable from zero) following a one standard deviation increase in VS from its mean. These point estimates suggest that the absolute value of the elasticity of industry trade balances with respect to changes in the real exchange rate is increasing in IIT and decreasing in VS. The above pattern offers evidence for the presence of heterogenous trade responses to common exchange rate shocks. In contrast to previous work, however, we show that it is the heterogeneities picked up by VS that appear to be more important than those picked up by IIT. As mentioned above, Kharroubi (2011) finds that greater IIT increases the sensitivity of trade balances to aggregate real exchange rate movements. While our coefficients are directionally the same, they are not consistently significant. At best, we conclude that the elasticity of industry trade balances with respect to industry-specific real exchange rate may increase as IIT increases. However, we show more conclusively that the formation of global value chains is quite important to the relationship between industry trade balances and industry real exchange rates. As greater supply chain linkages are made and VS increases, the elasticity declines; this implies that the trade balance becomes less sensitive when exports rely heavily on imported content. The latter effect is particularly important for policy makers. Consider the Electrical Machinery and Apparatus industry as an example. As shown below, it is among the industries with relatively high levels of vertical 5 specialization. It is also among the industries that make large contributions to exporting activities in Asia.2 Real exchange rate swings are expected to have limited effects on Electrical Machinery and Apparatus exports, despite their large share in manufacturing exports, due to the significant reliance of this industry on imported content. This creates a challenge for policy makers who rely on currency intervention for external adjustments as this industry or similar industries with relatively high reliance on imported inputs will face competing price pressures on imported inputs versus exported final goods. The findings of this empirical exercise may in part explain why a common shock to the aggregate nominal exchange rate leads to different reactions in different industries or different reactions at different points in time in a given industry. This makes policy decisions on exchange rate management more challenging as these nuanced responses need to be considered. Taken together with the limited significance of the IIT coefficients, these results point to currency intervention being less potent in addressing global imbalances across countries, particularly for industries in which exports rely heavily on imported inputs. Importantly, the responses depend on the amount of interdependence between industries such that global supply chain linkages help determine how responsive trade balances are to real exchange rate movements. The paper proceeds as follows. Section 2 provides a review of previous empirical findings regarding exchange rate changes and the trade balance at the firm or industry level. Section 3 describes our empirical approach and the data. Section 4 presents the findings of our estimations with conclusions and policy implications in Section 5. 2. Background There have been previous efforts to explore the heterogenous effects of changes in real exchange rates on the trade balance and exporting activities, with an emphasis on differences in productivity, the reliance on 2 To measure the contribution of this industry to manufacturing exports in the countries in our sample, we rely on the median of time-varying share of Electrical Machinery and Apparatus industry's exports in total manufacturing exports. Based on the UN Comtrade data (World Bank, 2017), this share is 37% for China, 13% for Indonesia, 20% for Japan, 28% for Korea, and 47% for Malaysia. This industry is ranked as the highest contributor to exports in China, Korea, and Malaysia and second and third highest in Japan and Indonesia, respectively. 6 imported inputs, and the amount of intra-industry trade. Focusing on the variations in firm-level productivity, Berman, Martin, and Mayer (2012) show that export prices of more productive firms are more sensitive to changes in the real exchange rate, while their export volumes appear to be less sensitive.3 Also, focusing on the reliance on imported inputs, Amiti, Itskhoki, and Konings (2014) show that export prices of large exporters who rely on imported inputs are quite sensitive to movements in the real exchange rate when compared to the prices that are set by small exporters with less reliance on imported inputs. In both of these studies, export prices are measured in domestic currency. Unlike the monotonic relationships reported in the above studies, there are other papers that find a U-shaped relationship between market share and pass-through in export prices; e.g., Feenstra, Gagnon, and Knetter (1996), Yoshida (2013), Auer and Schoenle (2016), and Garetto (2016). The findings by Garetto (2016), for example, suggest that up to a certain threshold of market share or productivity, European automobile exporters are more likely to offset part of the exchange rate movements by adjusting their prices. After that threshold, however, they are more likely to pass on the exchange rate movements to their foreign customers, as they have less incentives to engage in strategic pricing.4 The differences in firm-level responses may also have important industry-level implications. For example, Berman et al. (2012) provide some sector-level evidence, suggesting that the elasticity of export value with respect to changes in the real exchange rate decreases as sector productivity increases. There are also some timeseries studies that document varying industry-level responses to common exchange rate shocks. Bahmani-Oskooee and Hegerty (2010) survey this literature in detail. Much of this literature, however, focuses on the responses to aggregate (country level) exchange rate changes. Oguro (2011) and Demiral (2016) provide support for the importance of IIT variations in governing the relationship between the real exchange rate and the trade balance, without controlling for variations in VS separately. Oguro (2011) suggests that the sensitivity of export quantities to changes in 3 These findings are consistent with theoretical models in which demand elasticity is declining in firms' performance (e.g., Melitz and Ottaviano, 2008). 4 This group of exporters are already among the most productive firms with lowest prices, which may lead them to pass on the exchange rate movements more fully. 7 the exchange rate declines as IIT increases. Exploring a large panel of industries in multiple countries, she finds that the adverse effects of aggregate bilateral real exchange rate appreciation on industry-level exports are less pronounced in industries with high degrees of IIT, compared to those with low degrees of IIT. Also, exploring Turkey's trade balance with a large set of OECD countries, Demiral (2016) shows that incorporating the variations in IIT may weaken the effect of exchange rate depreciation on exports. Focusing on the variations in VS, Ahmed et al. (2015) suggest that global supply chain linkages are responsible for the declining elasticity of manufacturing export volumes to aggregate real exchange rate swings at both the country and industry level. However, Leigh et al. (2017) find no conclusive evidence for the moderating effects of VS variations; even though the share of VS-related trade is gradually increasing, they argue that conventional trade still dominates the relatively small share of trade in intermediate inputs, which could in part justify the obtained evidence in their research. More recently, Bussiere, Gaulier, and Steingress (2020) estimate trade elasticities using product level data from a wide range of countries. They highlight the importance of the import content of exports, suggesting that the magnitude of the estimated elasticities depends on the correlation between exchange rate and unobserved marginal costs which, in return, depend on imports cost. Motivated by these studies, we examine how industry-specific IIT and VS govern the relationship between industry IERER and industry trade balances. The contributions of our paper are threefold. First, compared to Kharroubi (2011), we examine industry-level trade balance variations, exploiting the significant differences in IIT and VS across different industries within a given country. We also employ an industry-specific measure for real exchange rates. We exploit the significant differences that are observed in exchange rates across different industries within a given country. Aggregate variations may mask these differences. Second, compared to Oguro (2011) and Demiral (2016), we control for variations in IIT and VS separately, which enables us to draw clear inference. Also, compared to Leigh et al (2017), we focus on a sample of leading Asian exporters who are heavily engaged in trade of intermediate inputs. In this environment, the control for VS and its interaction with IERER allow us to capture the effects of differences in vertical relationships. Third, we examine the changes in the elasticity of industry trade 8 balances with respect to industry IERER over the entire distribution of IIT and VS. This exercise provides a detailed estimate of the changes in the elasticity based on the heterogeneities across industries. 5. Conclusion In this study, we explore how the variations in intra-industry trade and vertical specialization affect the elasticity of industry-level trade balances with respect to industry-specific measures of the effective real exchange rate. We also explore the effects of the above variations on the elasticity of industry-level export and import intensities. For this purpose, we employ sample information from 13 manufacturing industries across five Asian countries (China, Japan, Korea, Malaysia, and Indonesia) from 2001 to 2015. Our findings are threefold. First, we find that industry-level trade balances become more elastic as intra-industry trade increases. The impact on the trade balances comes predominantly through 28 exports, which are more elastic at higher levels of intra-industry trade, where import substitutes are more readily available. Second, we find that trade balances become less elastic as vertical specialization increases. The channel here is through both exports and imports, which are less elastic at higher levels of vertical specialization, where exports rely more on imported content. Third, we find that the effects from vertical specialization are of greater significance, when compared to intra-industry trade. Growing uncertainties in the global economy may lead policy makers to actively intervene in foreign exchange markets. All else constant, however, the results of this study suggest that such interventions may be of limited potencies when, for instance, exports rely more heavily on imported content. There are, in fact, leading industries in Asia who have relatively large shares in manufacturing exports and at the same time rely heavily on imported inputs. Electrical Machinery and Apparatus industry is an important example, providing the highest share of exports among industries in China while also indicating high VS and low IIT values. Thus, in response to real depreciation in the yuan, there are likely to be limited changes in the trade balance for this industry, perhaps driving limited changes in the aggregate trade balance as well. Overall, our findings imply that real exchange rate swings, altering relative national prices, have limited effects on trade balances for such industries. This presents a challenge to policy makers who, faced with growing uncertainties, may want to rely on currency intervention for external adjustments in these leading industries. Further research that examines the impact of tariffs and other trade barriers on these relationships may be warranted. Incorporating industry-specific tariffs is beyond the scope of this paper but we conjecture that such tariffs are likely to impact not only the industry-specific real exchange rates, but also the global patterns for supply chain linkages. Increases in trade barriers may provide incentives for multinationals to engage in horizontal FDI, lowering IIT. Increases in trade barriers may also provide less incentives for multinationals to engage in vertical FDI, lowering VS.