Do remittances worsen export diversification?

Vardanyan, Erik Working Paper Do remittances worsen export diversification? Economics Discussion Papers, No. 2019-46 Provided in Cooperation with: Kiel Institute for the World Economy – Leibniz Center for Research on Global Economic Challenges Suggested Citation: Vardanyan, Erik (2019) : Do remittances worsen export diversification?, Economics Discussion Papers, No. 2019-46, Kiel Institute for the World Economy (IfW), Kiel This Version is available at: https://hdl.handle.net/10419/201801 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/4.0/ Discussion Paper No. 2019-46 | August 12, 2019 | http://www.economics-ejournal.org/economics/discussionpapers/2019-46 Do remittances worsen export diversification? Erik Vardanyan Abstract The paper explores the impact of workers’ remittances on the level of export diversification. The hypothesis is that significant inflow of remittances causes overvaluation of real exchange rate, which in turn deteriorates diversity of export. The theoretical base is in line with the Dutch disease phenomenon. The paper uses annual cross-national panel data over 2000–2016 period and System GMM methodology. The evidence suggests that indeed large inflow of remittances is associated with less diversified export. The economic intuition behind is that remittance-caused real exchange rate appreciation unevenly suppresses export of goods: some goods “suffer” more than others do. In terms of the number of product-names, a percentage point increase in remittances to GDP sent home “reduces” variety of export by approximately five active lines. There are other interesting findings as well. An improvement of government effectiveness facilitates overall export diversification; terms of trade improvement and rise of real exchange rate volatility mostly increase export concentration rather than alter number of exported product-names. JEL F14 F24 F31 Keywords Remittances; export diversification; export concentration; export variety; real exchange rate; System GMM Authors Erik Vardanyan, Central Bank of Armenia, Yerevan, Armenia, [email protected] The author would like to thank Armen Nurbenkyan, Xavier Raurich, Hayk Avetisyan, Vahe Movsisyan, Vahe Charchyan, Vahagn Jerbashian, Mattias Polborn, Karen Poghosyan, Laura Aydinyan, Aleksandr Shirkhanyan and Vardan Adibekyan for useful comments and suggestions. The views expressed in the paper are those of the author and do not necessarily represent the views of the Central Bank of Armenia. All errors are the author's responsibility. Citation Erik Vardanyan (2019). Do remittances worsen export diversification? Economics Discussion Papers, No 2019-46, Kiel Institute for the World Economy. http://www.economics-ejournal.org/economics/discussionpapers/2019-46 Received July 26, 2019 Accepted as Economics Discussion Paper August 8, 2019 Published August 12, 2019 © Author(s) 2019. Licensed under the Creative Commons License - Attribution 4.0 International (CC BY 4.0) Introduction Remittances have risen spectacularly in recent decades, fueling research interest about their pros and cons for recipient economies. For some developing economies, these kinds of financial inflows constitute the main source of foreign exchange, even exceeding received official development assistance and FDI 1 . However, economic implications of remittances are diverse. They can improve the well-being of families left behind, especially in terms of reduction of poverty, improvement of health, education and income distribution (Koczan and Loyola 2018). However, sizable amount of remittances may pose a challenge for the long run economic growth by creating culture of dependence, lowering labor force participation, harming country competitiveness, promoting conspicuous consumption etc. (Amuedo-Dorantes 2014). Academic and policy-oriented research has not come to a consensus, whether remittances contribute to longer-term growth by building human and financial capital or degrade growth by creating public moral hazard problem (Ebeke 2012) and harming economy’s competitiveness. Although, several studies state that there is no example of a country with remittances driven economic growth (A. Barajas, et al. 2009), (Clemens and McKenzie), the complete answer to the question “why this is the case?” is yet to be found. This necessitates conducting research like the actual one. Which aims to contribute to filling this gap. Namely, I investigate possible negative consequences of remittances on the quality of external sector, which is an important determinant of long-run growth (Hesse 2008). The paper claims that significant inflow of remittances eventually leads to the real exchange rate appreciation, which in turn worsen economy’s export diversification. The underlying notion is in spirit of Dutch disease phenomenon. On top of that, here I impose a realistic assumption that elasticities of export of goods with respect to real exchange rate fluctuations are uneven. Meaning, that some types of products are relatively immune to an appreciation shock while others are not. After occurrence of such a shock, some goods may even drop out from the export basket of a country or they may not occur in the first place. Eventually, economies associated with sizable inflow of remittances may end up having poorer external sector. Large data available for many countries allow me to use dynamic panel technique to test this hypothesis. The dynamic panel information helps to isolate unobserved time-invariant country-specific characteristics. Moreover, usage of the System GMM methodology (Arellano and Bond 1991) helps to deal with endogeneity of explanatory variables. The estimation results show robust evidence across specifications that sizable inflow of remittances deteriorate diversification of export. This is true for both variety and concentration of the export. Additionally, the government effectiveness is 1 During the last decades remittances sent to developing countries by migrants has increased exponentially, peaking to 441 billion US Dollars in 2015. As a share of GDP, Tajikistan (42 percent), the Kyrgyz Republic (30 percent), Nepal (29 percent), Tonga (28 percent), and Moldova (26 percent) were among the largest recipients of remittances (World Bank, 2016) one of the strong determinants of export diversification. Terms of trade improvement negatively affects export concentration, while variety dimension of export is insensitive to this movement. Background The literature on remittance caused real exchange rate appreciation is extensive, e.g. (AmuedoDorantes and Pozo 2004), (Acosta, Mandelman and Lartey 2012), (Acosta, Lartey and K.K.Lartey 2009), (A. Barajas, et al. 2010), (López, Molina and Bussolo 2008), etc. On top of that, economic literature has documented the adverse impact of the real exchange rate appreciation on the export diversification, e.g. (Goya 2014), (Bahar and Santos 2018). The actual study lies at the juncture of these two strands of economic literature, as it explores remittances impact on export diversification. The economic logic behind this connection is in the spirit of the Dutch disease phenomenon: significant inflow of foreign exchange increases aggregate consumption and prices of non-tradables. As a result, appreciated real exchange rate harms the country’s competitiveness and induces recourse movement from tradable to non-tradable sector. Consequently, as it was described above, the remittance caused real exchange rate appreciation may adversely affect the structure of the export. The Figure 1 illustrates positive relationship between remittances and export diversification for period of 2015. Importantly, there is no regional pattern of this relationship. Figure 1. Received Remittances as percent of GDP (>5%) vs. Theil index. 2015 2 2 The Theil Index is one of indices used in this research to capture dynamics of export diversification. Higher is the index less diversified is country’s export. Different branches of economic literature prove the beneficial impact of export diversification on the economic activity (Cadot, Carre`re and Strauss-Kahn 2011), (Klinger and Lederman 2004), (Hausmann, Hwang and Rodrik 2005), (Koren and Tenreyro 2007). Academics come to the consensus that export diversification is of the factors promoting high economic growth. First, it helps to moderate negative impact of external instability by diversifying corresponding risks. Moreover, diversification of production, and in turn the export, helps developing economies to undergo structural transformation. It is the process of movement from production of “poorcountry goods” to “rich-country goods” (Hesse 2008). Therefore, it is vital to explore factors, which could delay this transformation. To analyze external sector in depth, I distinguish two dimensions of the export diversification: its variety and concentration. By variety, I refer to the number of different product-names exported. The concentration, on the other hand, is about the distribution of shares of different products in the export. Moreover, to see the picture broadly and to ensure robustness of findings, I use several indices of export diversification (Gini index, Herfindahl-Hirschman Index (HHI), Theil Index, Number of Active Lines). Data To calculate export diversification indices, I use export data (in current USD) from UN COMTRADE database. I use cross-national annual panel data for 135 economies over the 2000-2016 period. The database includes 17 low income, 34 lower middle income, 42 upper middle income and 42 high income countries 3 . For the research, I consider only export of goods and neglect services due to inherent complication of calculating diversification indices for them. 4-digit disaggregated data of export is used (Standard Industry Trade Classification (revision 2), with 786 possible product names); higher level of disaggregation may be noisier and with such data, there may be the risk of taking minor variations as a sign of diversification change. The Gini index, the Herfindahl–Hirschman Index (HHI), and the within component of the Theil Index (Theil Within) are considered as concentration indices. On the other hand, the between component of the Theil Index (Theil Between) and the number of product-names in export of a country (Active Lines) are considered as variety indices. The usage of broad range of indicators allows understanding 3 The full list of countries used in the research are displayed in the Table 3 of the Appendix Variety Concentration Export Diversification Real Exchange Rate Output Growth Remittances underlying mechanisms of diversification 4 . Increase of the Gini Index, HHI and both Theil indices indicate decline in diversification. At the same time, literally increase of the Active Lines reflects improvement of variety component of the export diversification. Other variables used in the research are natural logarithm of GDP per capita (PPP constant 2011 international), Real Effective Exchange Rate (REER) (as a proxy of real exchange rate), volatility of real exchange rate (Standard deviation of the last 5 observations of the REER), Terms of Trade index, the Government Effectiveness Index and the growth rate of population. The absence of some observations is common practice in empirical research. The actual one is not an exception. To fill these gaps, I use Kalman filtering technique, which based on likelihood function replaces empty cells with probabilistic values of the corresponding variable. The sources and summary description of data used in the research are placed in the Table 1 and Table 2 of the Appendix of the paper. The Methodological Procedure and the Estimation There are several channels through which remittances affect real exchange rate (Fajnzylber and López 2008). The first possibility is associated with the external equilibrium (remittances affecting net financial position of a country against the rest of the world). The second one is linked to the internal equilibrium (remittances stimulating domestic demand and pushing prices of nontradables up, plus it increases reservation wage, therefore it is appreciating the real exchange rate). Finally, remittances may affect real exchange rate through the GDP growth. The GDP growth increases domestic demand with consequent implications for real exchange rate described in the preceding point. Naturally, overall macroeconomic implication of remittances mostly depends on the households’ spending preferences. If hand-to-mouth wage earners receive these financial means, then the impact would be contractionary. A positive remittances (demand) shock would create upward pressure on the relative prices of non-tradables to tradables, and therefore appreciate real exchange rate. The opposite would be true when credit-constrained entrepreneurs are main receivers of remittances. Here the overall effect can be expansionary, since received financial means may be utilized as productive investment (Bahadir, Chatterjee and Lebesmuehlbacher 2018). This case is particularly relevant for financially less developed economies (Giuliano and Ruiz-Arranz 2009). Conventional wisdom says that addressees of remittances are comparably poorer households. Naturally, one can expect their propensity to consume to be relatively high: so, as it follows remittances would rather have appreciation effect on real exchange rate. Anyway, in practice, evidence is mixed. For instance, (Brahim, Nefzi and Sambo 2017), (Mongardini and Rayner 2009) conclude that remittances have negative or no impact on real exchange rate. They claim that 4 Technical notes on the calculation of diversification indices are in the Appendix grants and remittances are easing supply constraints and boosting productivity in non-tradable sector in the recipient economy leading to domestic price downward adjustment and so to depreciation of real exchange rate. However, as a matter of speculation, one may link this to the characteristics of the regions those papers are referencing. Here the economies may have significant space of resource utilization and therefore, remittances could have facilitating role in managing it. On the other hand, other research tackling the remittances inflow/real exchange rate connection ( (Lopez, Molina and Bussolo 2007), (Acosta, Lartey and K.K.Lartey 2009), (Amuedo-Dorantes and Pozo 2004)) are in favor of the appreciation hypothesis. For some economies, even visual examination reveals vivid positive correlation between remittances inflow and real exchange rate. The Figure 2 displays some of those cases. At the end of the day, the interplay of various macroeconomic and country specific factors may change the direction of the remittances inflow vs. real exchange rate relationship. Therefore, the hypothesis about the nature of this relationship is an empirical question rather than a given fact (A. Barajas, et al. 2010). Hence, for soundness of further analysis I need to test this hypothesis using our data. Figure 2. The positive association of remittances and real effective exchange rate (increase indicates appreciation) is straightforward in several economies (the data are scaled) For the test, I employ the methodology applied in the (Acosta, Mandelman and Lartey 2012). Usage of System GMM estimation technique allows addressing reverse causality problem. On one hand, if remittance senders are altruistically driven, the inflow of foreign exchange will increase during downturns (remittances are countercyclical with respect to income in migrant’s home economy). On the other hand, irrespective of wellbeing status of those left behind, migrants may define size of remittances to be sent just by looking at the exchange rate level or conditioning it with their own earnings (in other words, here remittances are procyclical with respect to income in migrant’s host economy) (Frankel 2009). Therefore, to deal with possible endogeneity problem, I employ System GMM regression technique, which is superior to OLS. Since the latter would provide biased and inconsistent estimates. The equation (1) is the considered econometric model to be estimated. Here, aside from lag of real exchange rate and remittances, I include a vector 𝑿𝒊𝒕 of explanatory variables as determinants of real exchange rate. The vector includes GDP per capita, broad money (M2 as a percentage of GDP), terms of trade index, trade openness (sum of export and import as a percentage of GDP), annual growth rate of real GDP (%), a dummy variable Fixed Exchange Rate (FER) indicating a type of conducted exchange rate policy (1 if the fixed exchange rate regime is in place, 0 otherwise) and Foreign Direct Investments FDI (% of GDP). As in the reference paper, the first lagged differences and the second lag level of all explanatory variables used as standard “IV-style” instruments. The estimation outcome of different specifications of the econometric model (1) is represented in the Table 1. 𝑹𝑬𝑬𝑹𝒊𝒕 = 𝜷𝟏𝑹𝑬𝑬𝑹𝒊𝒕−𝟏 + 𝜷𝟐𝑹𝒆𝒎𝒊𝒕𝒕𝒂𝒏𝒄𝒆𝒔𝒊𝒕+∑ 𝜷𝒏𝑿𝒊𝒕 𝒑 𝒏=𝟑 +𝝁𝒊+ 𝒗𝒊𝒕 (1) For i = 1…N, t= 1…, – country and time indicators. 𝛽Coefficients to be estimated 𝑅𝐸𝐸𝑅𝑡 - Real Effective Exchange Rate 𝑅𝑒𝑚𝑖𝑡𝑡𝑎𝑛𝑐𝑒𝑠 - Net Received Remittances (% GDP) 𝑋𝑖𝑡 - Vector of explanatory variables 𝜇𝑖Time-period dummy variable 𝑣𝑖𝑡An error term. Table 1: System GMM estimation. The dependent variable is the REER, 2000-2016 (1) (2) (3) (4) (5) (6) (7) Real GDP Growth Rate 1.336** 1.243** 1.247** 0.592 0.583 0.631 0.593 (0.624) (0.615) (0.620) (0.415) (0.410) (0.427) (0.414) Remittances 0.306 0.485** 0.437** 0.578** 0.535** 0.804** 0.578** (0.236) (0.236) (0.213) (0.263) (0.248) (0.378) (0.258) Terms of Trade 0.032 0.031 0.042** 0.042** 0.030 0.042** (0.021) (0.022) (0.021) (0.020) (0.022) (0.021) Trade Openness 0.006 -0.013 -0.003 -0.009 -0.013 (0.019) (0.020) (0.022) (0.024) (0.020) GDP per capita 1.788* 1.673* 3.278* 1.759* (1.036) (0.996) (1.900) (1.025) FDI -0.055 (0.057) M2 -0.067 (0.046) FER 0.373 (1.669) Number of observations 1740 1740 1740 1740 1740 1740 1740 Number of groups 116 116 116 116 116 116 116 Number of instruments 105 107 109 111 113 113 112 AR (2) p-value 0.903 0.912 0.912 0.976 0.966 0.959 0.975 Hansen test of overidentification restrictions 0.097 0.097 0.122 0.091 0.098 0.090 0.092 Difference-in-Hansen tests of exogeneity of instrument subsets IV instruments for first differences Hansen test excluding group (p-value) 0.088 0.093 0.099 0.077 0.078 0.072 0.079 Difference (null H = exogenous) (p-value) 0.583 0.396 0.684 0.548 0.611 0.596 0.519 Note: Significance levels: .01 - ***; .05 - **; .1 - *, Standard Errors (SE) are in parentheses below the corresponding coefficients, SE are Windmeijer corrected and clustered by countries. For these estimations, advanced economies are excluded from the sample (Acosta, Mandelman and Lartey 2012). In the estimation, first two lags of the dependent variable and Remittances are used as “GMM-style” instruments. First lagged differences and second lag levels are specified as “IV-style” instruments. For conventional purposes, the outcomes for the Difference-inHansen tests are reported only for the lagged differences. Corresponding estimates for second lag levels and GMM instruments indicate results similar to those reported, and are available upon request. The Fixed Exchange Rate (FER) is a dummy variable with value 1 if the fixed exchange rate regime policy is conducted and 0 otherwise (De-facto IMF classification); The M2 is broad money (% of GDP); FDI stands for net inflow of Foreign Direct Investments (% of GDP); Remittances represent inflow of remittances (% of GDP); GDP per capita is PPP adjusted (constant 2011 international $) natural logarithm of per capita output; Real GDP Growth Rate is in local currency units and at constant prices. Other variables are self-explanatory. Columns labeled from (1) to (7) represent different specifications of the econometric model. The outcome of the econometric model (1) estimation supports the hypothesis of remittance based real exchange rate appreciation. The result is immune to specification changes. A percentage point increase in inflow of remittances, appreciates the real exchange rate by approximately 0.5 units. This finding is in line with the one described in (Acosta, Mandelman and Lartey 2012). On top of that, GDP per capita is another significant determinant of real exchange rate. The significance of coefficients of this two factors will play an interesting role in the further analysis illustrated in the second part of the paper. From the policymaking perspective, development of proper policy response is rather difficult; this is due to the nature of these financial inflows. The first option coming to mind could be the taxation. However, taxing remittances directly is not a viable policy. Since the income generating the remittances has already been taxed at the origin. On top of that, generally, the beneficiaries of remittances are comparable poor population of an economy; therefore, the taxation could increase the vulnerability of households to income shocks and worsen the income inequality within the society. Moreover, an introduction of remittances taxation would drive workers to transfer money through the black market. To cope with negative consequences of remittances, some research advocate a switch from direct to indirect taxation: decreasing payroll taxes and increasing sales taxes (López, Molina and Bussolo 2008). Encouragement of saving may be considered as another possible policy option. Nevertheless, (Maimbo and Ratha 2005) show that forcing remittance recipients to save more and consume less, as Lesotho, Turkey, Mexico and others have done in the past, reduces consumer welfare. Nevertheless, there may be ways to indirectly increase the development impact of remittances Encouraging account-to-account remittance flows instead of cash transfers would result in increased saving by recipients (and senders) and better matching (by banks) of available saving and investment demand. In this context, the incentive to invest and its subsequent productivity will depend on the policy environment. Good policy environment will increase the return on investment and hence will raise the opportunity cost of consumption (Burnside and Dollar 2004). Conclusion Using large dataset of countries for the period of seventeen years, I examine the possible impact of workers’ remittances on export diversification of a receiver economy. The main hypothesis is that sizable amount of remittances inflow causes real exchange rate appreciation, which worsen diversity of export. Corresponding estimations support this hypothesis. It appears that indeed remittance-driven overvaluation of real exchange rate unequally affects export of products. Shares of some productions shrinks, and/or the products drop out of the export basket, being unable to overcome the negative shock, and/or production of some potentially feasible products are not profitable to export in an environment of overvaulted real exchange rate. As a result, remittances dependent economies end up having relatively poorer external sector. The analysis reveal that the outcome is valid for both variety of products and concentration of the export. Further research needed to address questions originated from above mentioned outcome. For instance, is there a hysteresis problem: do dropped active lines recover after an adverse shock? What are the most vulnerable sectors? Appendix Table 1. Description and Sources of data Variable Description Source Export data 4-digit disaggregated data (Standard Industry Trade Classification (revision 2), with 786 possible product names), in USD UN COMTRADE database Remittances Personal remittances, received (% of GDP)5 World Bank REER Real Effective Exchange Rate, CPI based. Increase indicates appreciation. For more complete data (Darvas 2012a), (Darvas 2012b), (Darvas 2012c) REER Volatility Standard deviation of the last 5 observations of the REER GDP per capita PPP (constant 2011 international $) World Bank GDP Real Growth Rate Annual GDP, local currency unit at constant prices. World Bank Terms of Trade Net barter terms of trade index (2000=100) World Bank M2 Broad Money World Bank FDI Foreign direct investment, net inflow (% of GDP) World Bank FER Fixed Exchange Rate. A dummy variable; 1 if the fixed exchange rate regime is conducted. De-facto classification Annual Report on Exchange Arrangements and Exchange Restrictions, IMF Trade Openness The sum of exports and imports of goods and services (% GDP) World Bank Population Growth Annual growth of population (%) World Bank Government Effectiveness Is a country score ranging within -2.5 the worst governance) to 2.5 (the best) Worldwide Governance Indicators. World Bank Inflation Inflation, consumer price (annual %) World Bank Unemployment Unemployment, total (% of total labor force, modeled ILO estimate) World Bank Nominal exchange rate Official exchange rate (LCU per US$, period average) World Bank Human capital School enrollment, primary (% gross) World Bank 5 From the perspective of economic relevancy, usage of net received remittances would be more appropriate. Anyway, I substitute it with just received remittances because of scarcity of data. In fact, for remittances dependent economies those two variables pretty much resemble each other. Table 2. Summary statistics Number of observations Mean Standard Deviation Min Max HHI 2380 0.16 0.20 0.01 0.99 Gini Index 2380 0.91 0.06 0.49 0.998 Theil Index 2380 3.79 1.32 1.69 7.11 Theil Within Index 2380 2.83 1.01 0.47 5.81 Theil Between Index 2380 0.96 0.66 0.48 6.03 Active lines 2380 547.9 200.4 3 771 Remittances 2380 4.21 6.35 0.00003 53.83 REER 2380 102.99 30.55 14.50 1282.61 REER Volatility 2380 6.36 7.57 0.27 172.94 GDP per capita 2380 16687.69 17318.05 561.61 129349.90 GDP Real Growth Rate 2380 3.83 4.32 -37.26 54.19 Terms of Trade 2380 111.49 46.20 21.39 810.65 M2 2380 61.04 37.35 -8.92 267.39 FDI 2380 5.51 16.45 -43.46 451.72 Trade Openness 2380 84.82 38.88 19.79 325.86 Government Effectiveness 2380 0.12 0.89 -2.27 2.35 Population Growth 2380 1.33 1.44 -2.85 16.33 Calculation of Export Diversification Indices Gini Index G=2 𝑛 (∑𝑘𝑅𝑘 𝑛 𝑘=1 ) ∑𝑅𝑘 𝑛 𝑘=1 −𝑛−1 𝑛 k – index of a product 𝑅𝑘 - export revenue of product k. Such that 𝑅𝑘<𝑅𝑘+1 n – number of exported products The Gini coefficient lies between zero (perfect equality) and one (complete inequality). Herfindahl-Hirschman Index (HHI) HHI=∑𝑆𝑖 2 𝑛 𝑖=1 −(1 𝑛) 1−(1 𝑛) n – number of exported products (active lines) 𝑆𝑖 - share of every product in the export basket The HHI is normalized as to lie within zero and one. Theil Index T=1 𝑁∑𝑥𝑘 𝜇 𝑁 𝑘=1 𝑙𝑛(𝑥𝑘 𝜇) Where 𝜇= 1 𝑁∑𝑥𝑘 𝑁 𝑘=1 𝑥𝑘– amount of export of product k 𝜇 - mean of 𝑥𝑘 N - number of all possible products in the world (for 4-digit disaggregated data it is 786) Theil’s Index: this index can be separated into two components • Between or Extensive margin (Tb) • Within or Intensive margin (Tw) Such that T= Tb+Tw 𝑇𝑤=1 𝑛∑𝑥𝑘 𝜇𝑒 𝑛 𝑘=1 ln(𝑥𝑘 𝜇𝑒) 𝑇𝑏=ln (𝑁 𝑛) 𝑥𝑘– amount of export of product k 𝜇𝑒mean of 𝑥𝑘 (only active lines) n - number of exported products In in the case of perfect diversification the Theil index is taking value 0, while the maximum value of the Theil index (perfect concentration) is ln(n). Active Lines The number of active lines at 4-digit disaggregation SITC rev. 2 (out of 786). Table 3. The list of countries used in the research Albania Cabo Verde Finland Japan Mozambique Russian Federation Togo Algeria Cambodia France Jordan Namibia Rwanda Tonga Argentina Cameroon Gambia Kazakhstan Netherlands Saint Lucia Trinidad and Tobago Armenia Canada Georgia Kenya New Zealand Saint Vincent and the Grenadines Tunisia Australia Chile Germany Kuwait Nicaragua Samoa Turkey Austria China Ghana Kyrgyzstan Niger Sao Tome and Principe Uganda Azerbaijan Colombia Greece Latvia Nigeria Saudi Arabia Ukraine Bangladesh Comoros Guatemala Lebanon Norway Senegal United Kingdom Barbados Costa Rica Guinea Lesotho Oman Serbia United Rep. of Tanzania Belarus Croatia Guyana Lithuania Pakistan Slovakia Uruguay Belgium Cyprus Honduras Madagascar Panama Slovenia USA Belize Czech Republic Hungary Malawi Paraguay Solomon Is. Venezuela Benin Denmark Iceland Malaysia Peru South Africa Viet Nam Bolivia Dominican Rep. India Maldives Philippines Spain Yemen Botswana Ecuador Indonesia Mali Poland Sri Lanka Zambia Brazil Egypt Iraq Malta Portugal Suriname Bulgaria El Salvador Ireland Mauritius Qatar Sweden Burkina Faso Estonia Israel Mexico Rep. of Korea Switzerland Burundi Ethiopia Italy Mongolia Rep. of Moldova TFYR of Macedonia Côte d'Ivoire Fiji Jamaica Morocco Romania Thailand Please note: You are most sincerely encouraged to participate in the open assessment of this discussion paper. You can do so by either recommending the paper or by posting your comments. Please go to: http://www.economics-ejournal.org/economics/discussionpapers/2019-46 The Editor © Author(s) 2019. Licensed under the Creative Commons License - Attribution 4.0 International (CC BY 4.0).