Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
WORKING PAPER NO: 13/32 The Role of Extensive Margin in Exports of Turkey: A Comparative Analysis August 2013 Altan ALDAN Olcay Yücel ÇULHA © Central Bank of the Republic of Turkey 2013 Address: Central Bank of the Republic of Turkey Head Office Research and Monetary Policy Department İstiklal Caddesi No: 10 Ulus, 06100 Ankara, Turkey Phone: +90 312 507 54 02 Facsimile: +90 312 507 57 33 The views expressed in this working paper are those of the author(s) and do not necessarily represent the official views of the Central Bank of the Republic of Turkey. The Working Paper Series are externally refereed. The refereeing process is managed by the Research and Monetary Policy Department. THE ROLE OF THE EXTENSIVE MARGIN IN EXPORTS OF TURKEY: A COMPARATIVE ANALYSIS1 Altan ALDAN2 Olcay Yücel ÇULHA3 Abstract Turkey successfully increased its share in the world exports in the last decades. We examine the role of extensive margin, in other words, new export products and destinations, on Turkish export performance between 1993 and 2011, in comparison with some other countries. Our results suggest that, Turkey was quite successful in extending its export products and markets compared to other developing countries. The success of Turkey in extensive margin mostly comes from entering new markets. Nevertheless, the share of Turkey’s export basket in world’s exports is still comparatively low as of 2011. Turkey still has important opportunities to increase her exports via extensive margin. Keywords: Turkey, exports, extensive margin, comparative studies of countries. JEL classification: F10, F14, F19, O57. 1 The authors would like to thank an anonymous referee for helpful comments, and to Utku Özmen, S. Tolga Tiryaki, Çağlar Yüncüler for their generous help in technical issues. And a special thanks to Murat Üngör for contributing with valuable discussions and suggestions. The views expressed in this paper belong to the authors only and do not represent those of the Central Bank of the Republic of Turkey or the World Bank 2 ALDAN: World Bank Turkey Country Office, Ankara, Uğur Mumcu Caddesi 88, 06700 Ankara, Turkey ▪ [email protected]▪ 3 ÇULHA: Central Bank of the Republic of Turkey, Research and Monetary Policy Department, İstiklal Caddesi 10, Ulus, 06100 Ankara, Turkey ▪ [email protected]▪ 1 1. INTRODUCTION In recent decades the international flow of trade has grown to unprecedented levels. Since 1950, world trade has increased over 20-fold (by value), far exceeding the growth rate for population or GDP. For example, since 2005, world merchandise trade has grown by 3.7% annually, while GDP has risen by 2.3%.4 Parallel to the rise in exports, there is a growing recognition of importance of entry to new export markets, such as new destination countries and products. The world economy has witnessed the expansion of trade, especially in the number of exchanged varieties, the so-called extensive margins. For example, Hamano (2012) shows that from 1980 to 2000, the average number of items exported into and imported from the U.S. increased by 38.1% and 13.4%, respectively.5 One particular aspect of the rise in international trade is that the economic emergence of developing countries, such as China and India, has dramatically increased global trade flows overall; and exports are perceived to be crucial in having a sustainable and high growth in most of the developing countries. Consequently, the volume of world trade and share of developing countries in the world exports increased dramatically in the last decades. In this regard, Turkey was one of the most successful countries in export growth, increasing its share of 0.43% in world exports in 1993 to 0.76% in 2011.6 In this paper, we analyze the importance of extensive margin, in other words, new export products and destinations, on Turkish export performance between 1993 and 2011, in comparison with some other countries. The experience of Turkey in the last decades is interesting for studying extensive margin. First, it joined customs union with the European Union at the end of 1995, which led the country free entry to one of the biggest markets in the world. Second, Turkey witnessed two major financial crises in 1994 and 2001, which caused export orientation in order to compensate the decline in the domestic demand.7 Furthermore, Turkey implemented a successful structural transformation program after the crisis of 2001 in 4 http://www.wto.org/english/res_e/statis_e/its2012_e/its2012_e.pdf Hamano (2012) studies the average growth of 4-digit Standard International Trade Classification (SITC) of exported and imported goods in U.S. bilateral trade with 15 OECD countries. 6 Turkey, with per capita GDP (at current US$) of $10,524 in 2011, is an upper middle income country with a population around 75 million and a GDP around US$0.8 trillion, making it the 18 th largest economy in the world as of 2011 (The World Bank, World Development Indicators Database, online access). 7 Turkey’s total exports reached $135 billion in 2011, the peak level since the foundation of the modern Republic. “Turkish Exports Strategy for 2023” was initiated by the Ministry of Economy and Turkish Exporters Assembly in 2009. The main purpose of this strategy is to reach 500 billion dollars of exports volume in 2023, the centenary anniversary of the Turkish Republic, with an average of 12% increase in exports annually. Becoming one of the world’s 10 largest economies in 2023 and taking 1.5% share from the world’s trade are also being targeted. 5 2 order to boost productivity and competitiveness.8 We find that Turkey was quite successful in extending its export products and markets compared to other developing countries. To put simply, export growth may happen due to the two margins: (i) extensive and (ii) intensive margins. In general, extensive margin refers to export growth due to new firms entering into the export market, new goods exported or new countries destined as export markets or a combination of these. On the other hand, export growth can come from increasing exports of existing firms, goods and markets, which is referred as the intensive margin. In recent years, there is a considerable literature on measuring the contributions of these margins and relating them to some economic variables. Different trade theories have similar predictions that as countries become bigger and richer they export more. However, there is not a consensus on the sources of higher exports.9 For example, the Armington model assumes that all countries produce and export a single variety (Armington, 1969). Hence, exports can grow only by exporting more of the single export good, i.e. the intensive margin. On the other hand, the Krugman model assumes endogenous number of export varieties which is proportional to per capita GDP of the countries (Krugman, 1981). In addition, all countries export the same quantity per variety. As a result, all export growth comes from the extensive margin. The Melitz model with heterogeneous firms and fixed cost of exporting finds that only productive firms will export (Melitz, 2003). As firms become more productive, more firms will enter the export market. Hence, Melitz model has a room for extensive margin in export growth. Building on Melitz’s (2003) model with heterogeneous firms, Helpman et al. (2008) and Chaney (2008), among others, developed trade models that explicitly consider the decision to export and therefore explicitly model the extensive margin of trade.10 Sources of export growth might also be interesting for policymakers as governments in developing countries usually concern about the vulnerability arising from export concentration (Cadot et al., 2012). Concentration of exports in a small group of products might increase volatility in terms of trade, which indeed may cause volatility in income (Jansen, 2004). Hence, policymakers might prefer export growth coming from the extensive 8 The Turkish economy grew at an average annual rate of 6.9% between 2002 and 2007. There were contractions in 2008 and 2009 due to the global crisis. After contracting by 4.8% in 2009, the Turkish economy rebounded quite rapidly and recorded a real growth of 9.2% in 2010 and 8.8% in 2011. 9 See Hummels and Klenow (2005) and the references therein for a comprehensive discussion on the sources of export growth in different trade models. 10 In a recent study, Santos Silva et al. (2013) focus on estimating the extensive margin of trade; and study the estimation of models for the number of sectors exporting from country j to country i. 3 margin in order to avoid possible risks on growth path from export prices or change in the composition of world import demand. Given the importance of the issue, there is an expanding literature, which focuses on the correlation of extensive and intensive margins with economic variables. Felbermayr and Kohler (2006) find that GATT- or WTO-membership increases world trade via extensive margin. Helpman et al. (2008) propose a methodology to decompose the effects of trade barriers on extensive and intensive margin. Markusen (2013) extends the gains from trade literature by including the gains via extensive margin explicitly. Cadot et al. (2011) decompose Theil index of export concentration into extensive and intensive margins and analyze the relationship between export concentration and per capita income. They find that export concentration increases with per capita income up to a threshold where specialization starts to increase. Cadot et al. (2011) also argue that it is the extensive margin that causes inverse U-shaped distribution of Theil index with respect to per capita income. The analysis of extensive and intensive margins has been conducted in alternative dimensions in different studies. For example, Felbermayr and Kohler (2006) and Helpman et al. (2008) define extensive margin on country basis. On the other hand, Hummels and Klenow (2005) and Kehoe and Ruhl (2013) define the extensive margin on product basis. Hence, exports classified as extensive margin in one approach can be considered to be intensive margin in another other one. Evenett and Venables (2002) and Besedes and Prusa (2011) use the broadest definition of extensive margin by using product-country export lines as unit of analysis.11 In this definition, exports of traditional products to new markets or new products to traditional export markets are considered to be extensive margin. There are several ways to measure extensive and intensive margins. One method is directly decomposing export growth due to existing, new and disappearing goods, where the contribution of existing goods are defined as intensive margin and the contribution of others are defined as extensive margin. Using this method, Amiti and Freund (2010) decompose Chinese export growth and Berthelon (2011) decompose Chilean export growth into intensive and extensive margins. Brenton and Newfarmer (2007) use a similar methodology for a list of developing countries but they extend the analysis to product-country space. Studies based on 11 See, also, Hillberry and McDaniel (2002), Eaton et al. (2004), Dennis and Shepherd (2007), Berthou and Fontagné (2008), Helpman et al. (2008), Hillberry and Hummels (2008), and the references therein for alternative definitions of the extensive margin based on different levels of aggregations. 4 firm level data also use similar decomposition, such as Eaton et al. (2007) for Colombian firms. Another methodology to measure extensive and intensive margins is based on the literature on the variety of goods in trade, starting with Feenstra (1994). Feenstra and Kee (2004) define a country’s export variety to US, as the share of total US imports that are exported by the country. Hummels and Klenow (2005) adjust Feenstra and Kee’s export variety definition in order to get extensive and intensive margin definitions. They define extensive margin as the ratio of total worldwide exports of a country’s export basket to the total worldwide exports and intensive margin as the share of a country’s exports to the worldwide exports in the country’s export basket. Using this definition, they analyze a cross section of countries in 1995 and conclude that differences in exports between larger and smaller economies mainly come from the extensive margin. In this framework, export growth comes from the extensive margin if the share of country’s basket of export goods in world’s exports is increasing. Kehoe and Ruhl (2013) criticize the definition of new good as the goods that were not exported at all in the beginning of the analysis period. Instead, they argue that goods that were exported with very small amounts should not be considered as export goods. Hence, they introduce the evolution of the exports of initially least traded goods as an indicator of extensive margin. They argue that such an indicator might capture the effect of structural changes or trade agreements on the evolution of the extensive margin. They analyze several countries and argue that extensive margin is the leading factor in export growth of developing countries while there is no such observation for developed countries. Furthermore, they find that structural reforms and trade agreements have significant effects on the extensive margins whereas business cycles do not have such an effect. Besedes and Prusa (2011) criticize the decomposition methodologies that use a static framework and compare two points in time. Static approaches, as in Amiti and Freund or in Kehoe and Ruhl, define the goods that are exported at the end of the sample period but not exported at the beginning, as new goods. This way of defining ignores the dynamics between these two points. Indeed, some of the new goods would be considered as traditional export goods or some traditional export goods would be considered as new goods if sample period changes slightly. They propose an alternative way of decomposing export growth which takes the survival rate of export relationship into account. 5 We, basically, follow the method suggested by Hummels and Klenow (2005). In doing so, we extend the method in several ways. First, we adapt the definition of non-traded goods following the critics of Kehoe and Ruhl (2013). In addition, we present the evolution of the share of initially least-traded goods as in Kehoe and Ruhl (2013). Second, following the critics of Besedes and Prusa (2011) and in order to avoid dependence on choice of initial year, we calculate extensive margins in an annual basis instead of just comparing start and end years. Finally, we extend the Hummels-Klenow and the Kehoe-Ruhl methodologies to product-country space. Our results suggest that, the success of Turkey in terms of extensive margin mostly comes from entering into new markets instead of exporting new products. In fact, Turkey’s extensive margin in terms of products does not change much during 1998-2011 and Turkey is still behind all countries considered. On the other hand, extensive margin is constantly increasing throughout the period and Turkey seems to surpass Mexico and catch up with the Czech Republic. Nevertheless, there is still a large room to increase exports through extensive margin, both in terms of products and geography. This paper is a contribution to the recently growing literature on extensive margin in Turkish exports. However, still very few studies are available. For example, Üngör (2011) follows Kehoe and Ruhl (2013) and analyzes the evolution of the share of the least traded goods in Turkish exports. Aldan and Çulha (2012) apply the same methodology and extend it to product-country space and for Turkish exports to EU and MENA regions, as well as to the world. The paper proceeds as follows. In the next section, we define the concepts and measures of extensive and intensive margin within the Hummels-Klenow and the Kehoe-Ruhl frameworks. Section 3 introduces data with descriptive statistics. Section 4 presents the results of the analysis. Section 5 concludes. 2. MEASURES OF EXTENSIVE AND INTENSIVE MARGINS In the Hummels-Klenow framework, the share of a country’s exports in world’s total export is the product of extensive and intensive margins. The extensive margin is the share of country’s basket of export goods in world’s export basket, whereas the intensive margin is the share of country’s exports in world’s exports in country’s basket of export goods. Formally, extensive (EM) and intensive (IM) margin of a country can be formulated as; 6 ∑ ∑ ∑ ∑ , (1) , (2) where the value of country c’s and world’s exports are denoted by is the set of goods exported by country c while the set exported in the world. Therefore, is a subset of and , respectively. contains all goods that are . Equations (1) and (2) can be used to compare countries’ extensive and intensive margins for a given year. On the other hand, it is possible to decompose growth of exports of a country between years t and t+n into contributions of extensive and intensive margins (Kehoe and Ruhl, 2013) as; (3) , where is the growth rate of exports, is the growth rate of extensive margin and is the growth rate of intensive margin. The measures of extensive and intensive margins in Equations (1) and (2) and their contributions in export growth in Equation (3) might be sensitive to several definitions. First, one should make a choice on the set of Klenow (2005), , the export basket of country c. In Hummels and is the set of goods that has a positive export record. However, they admit that a country exporting tiny amounts of many goods might misleadingly have a high extensive margin. Evenett and Venables (2002) classify a good with export value below 50,000 dollars as non-traded. Kehoe and Ruhl (2013) argue that such fixed thresholds might be misleading due to different sizes of economies. They suggest excluding the least traded goods from . They sort goods according to export values in the beginning and form 10 sets of goods which account for 10 percent of total exports. The value of the export of the good with highest exports in the first quintile (i.e., in the least exported goods) gives the threshold value in choosing the set . The advantage of this method is that it allows threshold values vary across countries. In order to check robustness of their results, they also sort goods based on share in world’s exports. That is, they sort goods according to the ratio of , where is defined as 7 . (4) Kehoe and Ruhl (2013) introduce the evolution of the share of initially least traded goods (goods not included in in the beginning) in exports as an indicator of the extensive margin. If the composition of exports change over time in such a way that the share of least traded goods (with initial share of 10 percent) increases considerably, it should be considered as a sign of importance of the extensive margin. On the other hand, if the share of least traded goods is still around 10 percent at the end of the analysis period, it is a sign of no contribution of extensive margin. The second issue to be dealt with is missing information in using only the beginning and the end of the analysis period. Decomposition of export share growth in Equation (3) is based on static comparison of extensive and intensive margins at time t and t+n. Suppose that a good started to be exported in t+1 and is exported every year between t+1 and t+n. Then, in the framework of Equation (3), the contribution of this new export item will be counted as extensive margin whereas in a dynamic framework it should be counted as intensive margin after t+1. Therefore, we analyze year-on-year transitions instead of just comparing years t and t+n and obtain a time series of extensive margin. This approach enables us to have threshold values for least traded goods varying within time as well as across countries. Furthermore, such a time series indicator might be informative on the response of extensive margin to structural reforms, global conditions or trade agreements comparable to the least exported goods indicator of Kehoe and Ruhl (2013). A third problem may arise due to the changing composition of world trade within time. Extensive margin of a country might change due to two reasons in Equation (3); either the country increases its product space of exports or the products that the country initially exports increase their share in world exports. In other words, extensive margin of a country might seem increasing even with no change in its basket of export goods because of increase in world demand to the products in the basket. Feenstra and Kee (2007) propose averaging world exports over time in Equation (1) in order to filter out effects of world demand composition and obtain time invariant comparison set of products. Finally, the measures of extensive and intensive margin are heavily dependent on the unit of analysis. In the analysis based on products as unit of measurement, only new export products are taken into account as the extensive margin. In this narrow definition, the effect of 8 exporting a traditional product to a new market is considered as intensive margin. On the other hand, if the analysis is extended to product-country space as in Besedes and Prusa (2011), such an entry will be counted as extensive margin. We conduct our analysis both on product and product-country spaces. This approach enables us to compare the sources of extensive margin, i.e., whether it comes from new products or new markets for traditional products. 3. DATA AND DESCRIPTIVE STATISTICS We use annual export data, covering the period of 1993-2011. We take 1993 as the beginning year, since the formation of new countries after the collapse of Communist block has ended in this year. We use data from the United Nation’s COMTRADE database at fourdigit level of the Standard International Trade Classification (SITC Revision3). In fact, Hummels and Klenow (2005) suggest using as much disaggregated data as possible, since low level of aggregation may cause some newly traded commodities to be ignored. We have detected some inconsistencies between the data obtained from the COMTRADE database and the Turkish Statistical Institute, at 5-digit level.12 In addition, the summation of 5-digit data does not add up to the total exports for other countries as well.13 As a result, for the sake of accuracy, we have chosen four-digit level for our quantitative analysis. Given the fact that Turkey aims export growth through manufacturing, as most of the developing countries with no abundance of natural resources, we use data on manufacturing exports; which is measured as the SITC codes 5 to 8 with chemicals (5), manufactured materials (6), machinery and transport equipment (7), miscellaneous manufactured articles (8). We compare the results for Turkey with those for several countries. First, we compare the results for Turkey with the U.S. (the largest economy in the world) and China (the biggest exporter in the world). The U.S. is the largest economy in the world with around US$15 trillion as of 2011; and China is the second largest economy in the world with around US$7.3 trillion as of 2011.14 Moreover, the U.S. remains the world’s biggest trader in merchandise, with imports and exports totaling US$ 3,746 billion in 2011; and China is the world’s biggest 12 The 5-digit COMTRADE data cover only 67.2% of TURKSTAT 5-digit data on average, during 1993-2011. The coverage ratio also varies between years ranging from around 75% in 1994 to less than 63% in 2004. 13 The 5-digit COMTRADE data cover only 67.3% of total exports in COMTRADE for Turkey on average, similar to the comparison of TURKSTAT and COMTRADE data. The corresponding average coverage ratios, of 5-digit COMTRADE data, for other countries are as follows: 75.9% for China, 71.3% for the Czech Republic, 75.1% for India, 74.7% for Korea, 61% for Mexico, and 72.2% for the U.S. Detailed information is available upon request. 14 The World Bank, World Development Indicators Database (online access). 9 exporter in merchandise trade, with exports around US$ 1,898 billion in 2011.15 In addition, we add countries to the comparison group that have manufacturing growth higher than the world average. We also check GDP and export figures in order to exclude very small countries not comparable with Turkey. As a result, we end up with the U.S., China, India, Czech Republic, Korea and Mexico as comparison countries. Table 1 presents information for GDP, per capita GDP and manufacturing exports of Turkey to the world in comparison with some countries. Except for the U.S., export growth in manufacturing in all countries between 1993 and 2011 was higher than the world average, most significantly China, Czech Republic and India. Except for Mexico, all countries performed better than the world average in terms of per capita GDP growth. Czech Republic has the smallest economy of around 30% of that of Turkey. We still include Czech Republic since we would like to compare Turkey with transition countries in Eastern Europe; and Czech Republic is the largest Eastern Europe country with available data. Table1: Selected Indicators GDP (billion $, current prices) 1993 United States 2011 6,582.9 14,991.3 Annual Growth (Avg. %) 4.7 GDP per capita ($, in 2000 prices) 1993 2011 28,834.6 37,691.0 Annual Growth (Avg. %) Manufacturing exports (billion dollars) 1993 1.5 356.2 2011 Annual Growth (Avg. %) 966.5 5.7 Turkey 180.4 775.0 8.4 3,731.7 5,741.4 2.4 11.0 104.2 13.3 China 440.5 7,318.5 16.9 536.4 2,639.5 9.3 73.8 1,771.9 19.3 39.3 217.0 10.0 4,754.4 7,926.9 2.9 11.6 142.7 14.9 India 284.2 1,848.0 11.0 332.6 837.7 5.3 16.3 187.8 14.5 Korea, Rep. 362.1 1,116.2 6.5 8,247.9 16,684.2 4.0 38.7 247.1 10.9 Mexico 403.2 1,153.3 6.0 5,107.4 6,288.3 1.2 76.3 473.3 10.7 24,999.8 69,981.9 5.9 4,645.1 6,103.0 1.5 Czech Republic World 2,668.4 11,510.9 8.5 Source: Worldbank, World Trade Organization. Table 2 displays the number of export products and product-country export lines in 1993 and 2011. All countries export most of the whole 717 products in SITC 4 digit classification both in 1993 and 2011. Turkey has the lowest number of products in both years but the difference with other countries declined considerably between 1993 and 2011. In the case of product-country pairs, the differences between countries are more significant. In 1993, Turkey was exporting in 14,161 export lines (ranked 6th among 7 countries), quite small compared to the U.S. with 61,507 lines. In 2011, all countries increased the number of export lines and Turkey is one of the most successful countries after China and India. In sum, the 15 The World Trade Organization, International Trade Statistics 2012. 10 number of products and product-country pairs exported provides evidence for Turkey’s relative success in the extensive margin of export growth. Table 2: Number of Products and Product-Country Pairs Exported Products Product-Country Pairs 1993 2011 1993 2011 China 715 707 35,998 92,304 Korea 707 708 27,416 46,472 Turkey 673 702 14,161 52,631 India 697 707 23,820 62,596 Czech Rep. 702 706 18,158 40,017 USA 714 708 61,507 77,241 Mexico 709 705 11,578 25,842 Source: UN COMTRADE, Authors' Calculations. 4. RESULTS We start with the methodology developed by Kehoe and Ruhl (2013), in which evolution of the share of least exported products in exports is used as an indicator of the extensive margin. The upper panel of Figure 1 shows the share of the least exported products that had 10 percent share in 1993. Products are sorted according to their 1993-1995 average export values in order to minimize the dependence of sorting to base year choice following Kehoe and Ruhl (2013). Figure 1(a) is based on the baseline sorting of products. We sort products according to export share of the products in Turkey’s total exports. Figure 1(b) presents the results of the alternative sorting based on export share of products in world’s exports, as in Equation (4). The share of the least exported goods in Turkey increased to 25% according to baseline sorting. This increase is in line with the developing economies, China, India and Czech Republic. There is not much increase in the share of the least traded goods for the U.S. and Mexico. It is not surprising not to have a significant rise in the share of the least traded goods in a developed country like the U.S., where structural transformation in exports is not expected according to Kehoe and Ruhl (2013). On the other hand, it is noteworthy that Mexico did not follow other developing countries in extending the share of initially least traded goods. When we sort according to the share in world trade in Figure 1(b), Turkey seems to be the most successful country with China in transforming its exports in favor of new goods. Unlike baseline sorting, Korea seems to perform better than the U.S. and Mexico. 11 Figure 1: Share of Exports of Least Traded Products and Product-Country Pairs Share of Total Exports (%) 1.a: Product line, baseline sorting Turkey China Korea India USA Mexico Czech Rep. 35 30 25 20 15 10 5 2011 2009 2007 2005 2003 2001 1999 1997 1995 1993 0 1.b: Product line, alternative sorting Turkey China Korea India USA Mexico Czech Rep. 40 30 20 10 2009 2011 2009 2011 2009 2011 2007 2005 2003 2001 1999 1997 1995 0 1993 Share of Total Exports (%) 50 1.c Product-country line, baseline sorting Share in Total Exports (%) 60 50 Turkey China Korea 40 India USA Mexico Czech Rep. 30 20 10 2007 2005 2003 2001 1999 1997 1995 1993 0 1.d Product-country line, alternative sorting Turkey China Korea India USA Mexico Czech Rep. 50 40 30 20 10 2007 2005 2003 2001 1999 1997 1995 0 1993 Share in Total Exports (%) 60 12 When we extend the analysis to the product-country space, share of the least traded product-country export lines in total exports rose to 55% in 2011 in Turkey, surpassing all other countries presented in lower panel of Figure 1. India, Czech Republic and China also have dramatic increases in the export share of initially least exported goods. Again, the U.S. and Mexico have the lowest changes in least traded export lines. Unlike the analysis in product line, Korea seems to increase the share of its initially least traded export lines; a sign of success in entering into new markets. The analysis of the share of the least traded goods/export lines points out that extensive margin is quite important in export growth for Turkey, similar to China, India and Czech Republic. Turkey is more successful in exporting traditional export goods to new destinations compared to other countries. Furthermore, there seems to be a decline in the growth rate of the least traded goods/export lines for Turkey in recent years. However, this is not peculiar to Turkey; similar patterns can be observed in China, India and Czech Republic. We now turn to the extensive margin analysis based on Hummels and Klenow (2005). Again, we first start with the analysis based on products. In Figure 2, we provide the evolution of extensive margin, as defined in Equation (1). The series reported in Figure 2(a) are the results of a direct replication of Hummels and Klenow (2005). In Figure 2(b), we adjust the extensive margin series by using average export values of goods across years, as suggested by Feenstra and Kee (2007), in order to account for any change in composition of world exports. The results show that all countries export almost all of the goods exported in the world. Indeed, if we account for world export composition, Czech Republic seems to be the exporting all product varieties traded in the world. This is not surprising since we include all goods with positive export values as exported goods. Hence, following the critique of Kehoe and Ruhl (2013), we calculate the extensive margin of goods with export values bigger than threshold values. 13 Figure 2: Extensive Margin in Products, Threshold=0 2.a: Original Hummels-Klenow 1,00 0,99 0,98 0,97 2009 2007 2005 2003 2001 1999 1997 1995 0,95 1993 0,96 2011 Turkey China Korea India USA Mexico Czech Rep. 2.b: Hummels-Klenow adjusted by Feenstra and Kee 1,00 0,99 0,98 0,96 2007 2005 2003 2001 1999 1997 1995 1993 0,95 2011 0,97 2009 Turkey China Korea India USA Mexico Czech Rep. In Figure 3, we replicate Figure 2 with the adjustment suggested by Kehoe and Ruhl (2013). We sort the products according to their average export values between 1993 and 1995 and define the threshold as the level of export value of the first good that would not be included in the least traded goods. A Comparison of Figures 2 and 3 reveals that a caution is needed in interpreting results of extensive margin analysis. In 1993, Turkey was exporting 96% percent of all goods exported worldwide. However, most of these exports were composed of goods that account for only 34% percent of the worldwide exports. Figure 3 depicts that Turkey had the lowest extensive margin in 1993, followed by India and China. As of 2011, it seems that Turkey converged to other countries and caught up with Mexico and Korea. China seems to be the most successful country in terms of increasing the extensive margin in this period. 14 Figure 3: Extensive Margin in Products, Constant Threshold (10 percent) 3.a: Original Hummels-Klenow 1,0 0,9 0,8 0,7 0,5 0,4 2009 2007 2005 2003 2001 1999 1997 1995 1993 0,3 2011 Turkey China Korea India USA Mexico Czech Rep. 0,6 3.b: Hummels-Klenow Adjusted by Feenstra and Kee 1,0 0,9 0,8 0,7 0,4 2007 2005 2003 2001 1999 1997 1995 1993 0,3 2011 0,5 2009 Turkey China Korea India USA Mexico Czech Rep. 0,6 The methodology suggested by Kehoe and Ruhl (2013) enables to have thresholds specific to countries, which takes sizes of countries into account. On the other hand, the thresholds in this method are constant over time which might induce problems when analyzing countries with strong export growth. As we see in Table 1, total exports of all countries increased very fast throughout the period. As an extreme example, Chinese exports in 2011 are 23 times higher than the exports in 1993. Hence, in order to have time varying threshold levels, we sort products every year according to export values and exclude the least exported goods which constitute 10 percent of exports every year. We present the results in Figure 4. 15 Figure 4: Extensive Margin in Products, Time-varying Threshold (10 percent) 4.a: Original Hummels-Klenow 0,8 0,7 0,6 0,5 Turkey India Czech Rep. 0,4 China USA Korea Mexico 2011 2009 2007 2005 2003 2001 1999 1997 1995 1993 0,3 4.b: Hummels-Klenow adjusted by Feenstra and Kee 0,8 0,7 0,6 0,5 2011 2009 Korea Mexico 2007 2005 China USA 2003 2001 1997 1995 1993 0,3 1999 Turkey India Czech Rep. 0,4 Figures (2)-(4) show that the analysis of extensive margin is heavily dependent on threshold selection. In Figure 2, all countries export almost all varieties traded in the world. In Figure 3, all countries except for the U.S. increase their extensive margin fast. However, Figure 4 depicts that this is the result of threshold level losing importance as countries become bigger exporters. For example, in 2011, China seems to be exporting almost of all the traded goods in the world, whereas 90% of Chinese exports are composed of goods that constitute 75% of world exports. Using the threshold based on 1993-1995 average, China seems to be the exporter of 97% of goods that are traded in the world. Indeed, if we extend the time period and assume Chinese exports will continue growing, we could have China exporting all products at the end year as the threshold value would be totally unbinding. Figure 4 suggests that Turkey’s extensive margin increased very rapidly between 1993 and 1998. The Customs Union Agreement between Turkey and the European Union, which came into effect on 31 December 1995, might have some influence in this rise, especially in the considerable rise in 1996. On the other hand, the role of extensive margin on Turkish 16 export growth seems limited between 1998 and 2011. Interestingly, there seems to be a decline in 2002 and a stop in the rise of the extensive margin in 1995, the years after major economic crises. There seems to be a gradual increase after 2002, which, however is not enough to catch up with other countries. India, which had similar extensive margin in 1993 increased its extensive margin fast and caught up with Korea, Mexico and the Czech Republic. The increase of extensive margin of Chinese exports was rather modest up to 2005 when there is a jump and considerable rise afterwards. The jump in 2005 might be due to the removal of quotas to Chinese textiles and wearing apparels products as a result of China’s entry into World Trade Organization. Finally, we see no significant change of the extensive margins of the exports of US, Korea, Mexico and the Czech Republic. The analysis based on products suggests that Turkey’s export growth was not mainly due to the extensive margin of products. In addition, Turkey lags other developing countries in terms of product varieties exported, which indeed suggests that there is a room for Turkey to increase its exports via new products. We now extend the analysis to the product-country space. Figure 5 displays the results without any threshold level. Figure 5: Extensive Margin in Product-Country Export Lines, Threshold=0 5.a: Original Hummels-Klenow 0,9 0,8 0,7 0,4 2007 2005 2003 2001 1999 1997 1995 1993 0,3 2011 0,5 2009 Turkey China Korea India USA Mexico Czech Rep. 0,6 5.b: Hummels-Klenow adjusted by Feenstra and Kee 0,9 0,8 0,7 0,4 2007 2005 2003 2001 1999 1997 1995 1993 0,3 2011 0,5 2009 Turkey China Korea India USA Mexico Czech Rep. 0,6 17 In 1993, Turkey was exporting in only 45% of active export lines. In other words, Turkish exporters were unable to export even with tiny amounts to more than half of the active export markets. China, U.S and Korea were the most successful countries in exporting to many product-country export lines. In 2011, Turkey seems to converge to other countries even surpassing Mexico. We continue with the analysis based on threshold adjustment by Kehoe and Ruhl (2013) in Figure 6. Figure 6: Extensive Margin in Product-Country Export Lines, Constant Threshold (10 percent) 6.a: Original Hummels-Klenow 0,9 Turkey Korea USA Czech Rep. 0,8 0,7 0,6 China India Mexico 0,5 0,4 0,3 0,2 0,1 2011 2009 2007 2005 2003 2001 1999 1997 1995 1993 0,0 6.b: Hummels Klenow adjusted by Feenstra and Kee 0,9 Turkey Korea USA Czech Rep. 0,8 0,7 0,6 China India Mexico 0,5 0,4 0,3 0,2 0,1 2011 2009 2007 2005 2003 2001 1999 1997 1995 1993 0,0 Turkey was mostly exporting to only around 10 percent of worlds’ product-country export lines in 1993. Only Mexico in the comparison group had such a low extensive margin. Turkey’s extensive margin increased parallel to the Czech Republic, lower compared to China 18 and India. There are slight increases in the extensive margins of US, Korea and Mexico. Note that, Mexico seems to be negatively affected by the world export demand composition after 2000 when we compare Figures 6(a) and 6(b). Finally, Figure 7 presents the results of the analysis based on threshold levels obtained by excluding the least exported product-country export lines for each year. Figure 7: Extensive Margin in Product-Country Export Lines, Time-varying Threshold 7.a: Original Hummels-Klenow 0,6 0,5 0,4 0,3 0,2 2011 2009 2007 2005 Korea Mexico 2003 2001 China USA 1999 1995 1993 0,0 1997 Turkey India Czech Rep. 0,1 7.b: Hummels-Klenow adjusted by Feenstra and Kee 0,6 0,5 0,4 0,3 0,2 2011 2009 2007 Korea Mexico 2005 2003 2001 China USA 1999 1995 1993 0,0 1997 Turkey India Czech Rep. 0,1 Similar to the results presented in Figure 6, Turkey increased its extensive margin steadily. This time, the overall rate of increase is equal to that of India and only China seems to perform better than Turkey in increasing the extensive margin. However, when we exclude the rise in 2005, increase in the extensive margin in Chinese exports is almost parallel to that of Turkish exports. It should be noted that, there is no rise in the extensive margin of India after 2003. Czech Republic increased its extensive margin continuously but slowly compared 19 to Turkey when we account for changes in the world demand composition and hence Turkey seems to have caught up with the Czech Republic. Finally, there seems to be no increase in the extensive margins of US, Korea and Mexico throughout the period. The analysis of extensive margin based on Hummels and Klenow (2005) suggest that Turkey was quite successful in increasing extensive margin between 1993 and 2011. As for the products, extensive margin of Turkey increased fast between 1993 and 1998. After 1998, there seems to be no considerable increase in the extensive margin. On the other hand, Turkey was more successful in export growth through extensive margin when we consider productcountry space, comparable to successful East Asian countries like China and India. Nevertheless, it can be concluded that Turkey has room to increase its exports via extensive margin both in product and product-country space when we consider the level of the extensive margin in 2011. 5. CONCLUSION Turkey appears to be one of the most successful countries in terms of export growth in the world during 1993-2011. Export diversification is perceived to be crucial in maintaining export growth and reducing external demand risks in many developing countries and Turkey is no exception. In this paper, we have analyzed one aspect of export diversification, namely extensive margin. We have applied two recently developed measures in the literature by Hummels and Klenow (2005) and Kehoe and Ruhl (2013) to Turkey with some adjustments and extensions. In addition, we have compared Turkey with similar countries whose export performance is better than the world average. We have applied several adjustments to assess the robustness of the results. Our results suggest that caution is needed in interpreting the findings. First, comparing only the beginning and the end years of a sample ignores the dynamics throughout the period and might be misleading. Second, the results heavily depend on the definition of traded goods; i.e. the choice of threshold levels for export values in order to decide whether a good is traded or not. Finally, changes in the composition of world demand might distort results for some countries. Turkey seems quite successful in increasing the extensive margin between 1993 and 2011. Turkey’s performance is comparable to outperforming countries like China and India and it is better than the Czech Republic with respect to some measures. On the other hand, there is not a significant rise in the extensive margins of developed countries like US and Korea. 20 Surprisingly, extensive margin of Mexico, with similar initial values to Turkey, does not seem to be changing considerably. We argue that the topic of this paper provides a productive research area for future studies. For example, one can study the relation between policies and extensive margin in Turkey. More specifically, the factors behind stagnation of extensive margin in products can be studied and policy recommendations can be suggested. In addition, the factors behind the success in entering into new markets can be studied. Finally, further research may focus on survival in export markets following Besedes and Prusa (2011) which will complement the research on extensive and intensive margins. REFERENCES Aldan A. and Çulha O. (2012), “The Role of Extensive Margin in Export Growth: Turkish Case,” Central Bank of Turkey, Research Notes in Economics 2012-36 (in Turkish) http://www.tcmb.gov.tr/research/ekonominotlari/2012/eng/EN1236eng.pdf Amiti M. and Freund C. (2010), “The Anatomy of China’s Export Growth,” NBER Chapter in “China’s Growing Role in World Trade”, 35-56, National Bureau of Economic Research. http://www.nber.org/chapters/c10451.pdf Armington, P. S. (1969), “A Theory of Demand for Products Distinguished by Place of Production,” IMF Staff Papers 16(1), 159-178. Berthelon M. (2011), “Chilean Export Performance: The Role of Intensive and Extensive Margins,” Central Bank of Chile Working Papers No 615. Berthou, A. and Fontagné, L. (2008), “The Euro Effects on the Firm and Product-Level Trade Margins: Evidence from France,” CEPII Working Paper No. 2008-21. Besedes T. and Prusa T .J. (2011), “The Role of Extensive and Intensive Margins and Export Growth,” Journal of Development Economics, 96(2), 371-379. Brenton P. and Newfarmer R. (2007), “Watching More Than Dıscovery Channel: Export Cycles and Diversification in Development,” The World Bank Policy Research Working Paper No 4302. Cadot O., Carrere C. and Strauss-Kahn V. (2011), “Export Diversification: What’s Behind the Hump?” The Review of Economics and Statistics, 93(2), 590-605. Cadot O., Carrere C. and Strauss-Kahn V. (2012), “Trade Diversification, Income and Growth: What Do We Know?” Journal of Economic Surveys, http://onlinelibrary.wiley.com/doi/10.1111/j.1467-6419.2011.00719.x/pdf Chaney, T. (2008), “Distorted Gravity: The Intensive and Extensive Margins of International Trade,” American Economic Review, 98(4), 1707-1721. 21 Dennis, A. and Shepherd, B. (2007), “Trade Costs, Barriers to Entry, and Export Diversification in Developing Countries,” The World Bandk Policy Research Working Paper No. 4368, Washington, D.C. Eaton J., Kortum S. and Kramarz F. (2004), “Dissecting Trade: Firms, Industries, and Export Destinations,” American Economic Review, 94, 150-154. Eaton J., Eslava M., Kugler M. and Tybout J. (2007), “Export Dynamics in Colombia: FirmLevel Evidence,” NBER Working Paper No 13531. Evenett S. J. and Venables A. J. (2002), “Export Growth by Developing Countries: Market Entry and Biletarel Trade Flows,” Paper Presented in CEPR European Research Workshop in International Trade, Munich. Feenstra R. (1994), “New Product Varieties and the Measurement of International Prices,” American Economic Review, 84(1), 157-177. Feenstra R. and Kee H .L. (2004), “On the Measurement of Product Variety in Trade,” American Economic Review, 94(2), 145-149. Feenstra R. and Kee H.L. (2007), “Trade Liberalisation and Export Variety: A Comparison of Mexico and China,” The World Economy, 30(1), 5-21. Felbermayr G. J. and Kohler W. (2006), “Exploring the Intensive and the Extensive Margin of World Trade,” Review of World Economics, 142(4), 643-674. Hamano M. (2012), “The Harrod-Balassa-Samuelson Effect and Endogenous Extensive Margins,” Journal of the Japanese and International Economies, http://dx.doi.org/10.1016/j.jjie.2012.06.004 Helpman E., Melitz M. and Rubinstein Y. (2008), “Estimating Trade Flows: Trading Partners and Trading Volumes”, Quarterly Journal of Economics, 123(2), 441-487. Hillberry, R. and McDaniel, C. (2002), “A Decomposition of North American Trade Growth since NAFTA,” Working Papers 15866, United States International Trade Commission, Office of Economics. Hillberry, R. and Hummels, D. (2008), “Trade Responses to Geographic Frictions: A Decomposition Using Micro-Data,” European Economic Review, 52, 527-550. Hummels, D. and Klenow, P.J. (2005), “The Variety and Quality of a Nation’s Exports,” American Economic Review 95(3), 704-723. Jansen, M. (2004), “Income Volatility in Small and Developing Economies: Export Concentration Matters,” WTO Discussion Paper No 3. Kehoe, T. J. and Ruhl K.J. (2013), “How Important is the New Goods Margin in International Trade?” Journal of Political Economy, 121(2), 358-392. Krugman, P. (1981), “Intra Industry Specialization and Gains from Trade,” Journal of Political Economy, 89(5), 959-973. 22 Markusen, J. R. (2013), “Expansion of Trade at the External Margin: A General Gains from Trade Result and Illustrative Examples,” Journal of International Economics, 89, 262-270. Melitz, M. (2003), “The Impact of Trade on Intra-Industry Reallocations and Aggregate Industry Productivity,” Econometrica, 71(6), 1695-1725. Santos Silva J. M. C., Tenreyro S. and Wei K. (2013), “Estimating the Extensive Margin of Trade”, http://personal.lse.ac.uk/tenreyro/EM_2j.pdf The World Bank. (2013), “World Development Indicators Database.” http://databank.worldbank.org/ddp/home.do The World Trade Organization. (2012), “International Trade Statistics 2012.” http://www.wto.org/english/res_e/statis_e/its2012_e/its2012_e.pdf Üngör, M. (2011), “Dynamic Changes in Comparative Advantage of Turkey: Some Observations,” Unpublished Manuscript, available at www.muratungor.com/research 23 Central Bank of the Republic of Turkey Recent Working Papers The complete list of Working Paper series can be found at Bank’s website (http://www.tcmb.gov.tr). Alternative Tools to Manage Capital Flow Volatility (Koray Alper, Hakan Kara, Mehmet Yörükoğlu Working Paper No. 13/31, July 2013)) How do Banks' Stock Returns Respond to Monetary Policy Committee Announcements in Turkey? Evidence from Traditional versus New Monetary Policy Episodes (Güray Küçükkocaoğlu, Deren Ünalmış, İbrahim Ünalmış Working Paper No. 13/30, July 2013) Some Observations on the Convergence Experience of Turkey (Murat Üngör Working Paper No. 13/29, July 2013) Reserve Option Mechanism as a Stabilizing Policy Tool: Evidence from Exchange Rate Expectations (Ahmet Değerli, Salih Fendoğlu Working Paper No. 13/28, July 2013) GDP Growth and Credit Data (Ergun Ermişoğlu, Yasin Akçelik, Arif Oduncu Working Paper No. 13/27, July 2013) Yield Curve Estimation for Corporate Bonds in Turkey (Burak Kanlı, Doruk Küçüksaraç, Özgür Özel Working Paper No. 13/26, July 2013) Inattentive Consumers and Exchange Rate Volatility (Mehmet Fatih Ekinci Working Paper No. 13/25, July 2013) Non-core Liabilities and Credit Growth (Zübeyir Kılınç, Hatice Gökçe Karasoy, Eray Yücel Working Paper No. 13/24, May 2013) Does Unemployment Insurance Crowd Out Home Production? (Bülent Güler,Temel Taşkın Working Paper No. 13/23, May 2013) Home Ownership and Job Satisfaction (Semih Tümen, Tuğba Zeydanlı, Çalışma Tebliği No. 13/22, Mayıs 2013) Türkiye İşgücü Piyasasında Mesleklerin Önemi: Hizmetler Sektörü İstihdamı, İşgücü ve Ücret Kutuplaşması (Semih Akçomak, H. Burcu Gürcihan Çalışma Tebliği No. 13/21, Nisan 2013) Gecelik Kur Takası Faizleri ve İMKB Gecelik Repo Faizleri (Doruk Küçüksaraç, Özgür Özel Çalışma Tebliği No. 13/20, Nisan 2013) Undocumented Workers’ Employment Across US Business Cycles (David Brown,Şerife Genç, Julie Hothckiss, Myriam Quispe-Agnoli Working Paper No. 13/19, March 2013) Systemic Risk Contribution of Individual Banks (Hüseyin Çağrı Akkoyun,Ramazan Karaşahin,Gürsu Keleş Working Paper No. 13/18, March 2013) External Financial Stress and External Financing Vulnerability in Turkey: Some Policy Implications for Financial Stability (Etkin Özen, Cem Şahin, İbrahim Ünalmış Working Paper No. 13/17, March 2013) End-Point Bias in Trend-Cyle Decompositions: An Application to the Real Exchange Rates of Turkey (Fatih Ekinci, Gazi Kabaş, Enes Sunel Working Paper No. 13/16, March 2013) Did the Rising Importance of Services Decelerate Overall Productivity Improvement of Turkey during 2002-2007? (Murat Üngör Working Paper No. 13/15, March 2013) Some Thought Experiments on the Changes in Labor Supply in Turkey (Murat Üngör Working Paper No. 13/14, March 2013)