Structural change, trade, and inequality: Some cross-country evidence

Roy, Rudra Prosad; Roy, Saikat Sinha Working Paper Structural change, trade, and inequality: Some crosscountry evidence ADBI Working Paper, No. 763 Provided in Cooperation with: Asian Development Bank Institute (ADBI), Tokyo Suggested Citation: Roy, Rudra Prosad; Roy, Saikat Sinha (2017) : Structural change, trade, and inequality: Some cross-country evidence, ADBI Working Paper, No. 763, Asian Development Bank Institute (ADBI), Tokyo This Version is available at: https://hdl.handle.net/10419/179219 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-nc-nd/3.0/igo/ ADBI Working Paper Series STRUCTURAL CHANGE, TRADE, AND INEQUALITY: SOME CROSS-COUNTRY EVIDENCE Rudra Prosad Roy and Saikat Sinha Roy No. 763 July 2017 Asian Development Bank Institute The Working Paper series is a continuation of the formerly named Discussion Paper series; the numbering of the papers continued without interruption or change. ADBI’s working papers reflect initial ideas on a topic and are posted online for discussion. ADBI encourages readers to post their comments on the main page for each working paper (given in the citation below). Some working papers may develop into other forms of publication. ADB recognizes “China” as the People’s Republic of China; “Hong Kong” as Hong Kong, China; and “Korea” as the Republic of Korea. Suggested citation: Roy, R. P. and S. S. Roy. 2017. Structural Change, Trade, and Inequality: Some Cross-Country Evidence. ADBI Working Paper 763. Tokyo: Asian Development Bank Institute. Available: https://www.adb.org/publications/structural-change-trade-and-inequality-some-cross-countryevidence Please contact the authors for information about this paper. Email: [email protected], [email protected] An earlier version of the paper was presented at the workshop on Structural Change and Inclusive Growth organized by ADB Institute, Tokyo, on 20–21 September 2016. The paper was also presented at the 9th International Conference on Empirical Issues in International Trade and Finance organized by the Indian Institute of Foreign Trade (IIFT), Kolkata, on 16–17 December 2016. The authors thank Yoshino Naoyuki, Wan Guanghua, Saumik Paul, Zhang Yuan, S. P. Jayasooriya, Kireyev Alexei, Siddhartha Mitra, Arindam Mandal, and other participants in these two conferences for their invaluable comments. The authors also thank Sachin Chaturvedi, N. R. Bhanumurthy, Achin Chakraborty and Indrani Chakraborty for their help at different stages of research. The usual disclaimer applies, however. Rudra Prosad Roy is research scholar and Saikat Sinha Roy (corresponding author) is professor, both at the Department of Economics, Jadavpur University, Kolkata, India. The views expressed in this paper are the views of the author and do not necessarily reflect the views or policies of ADBI, ADB, its Board of Directors, or the governments they represent. ADBI does not guarantee the accuracy of the data included in this paper and accepts no responsibility for any consequences of their use. Terminology used may not necessarily be consistent with ADB official terms. Working papers are subject to formal revision and correction before they are finalized and considered published. Asian Development Bank Institute Kasumigaseki Building, 8th Floor 3-2-5 Kasumigaseki, Chiyoda-ku Tokyo 100-6008, Japan Tel: +81-3-3593-5500 Fax: +81-3-3593-5571 URL: www.adbi.org E-mail: [email protected]g © 2017 Asian Development Bank Institute ADBI Working Paper 763 Roy and Roy Abstract The process of transition from a low-income to a high-income country involves a structural transformation of the economy along with a change in the distribution of income and wealth in the economy. This study examines how this process of structural change impacts on inequality for a sample of advanced, emerging, and transition economies. Trade liberalization, through a reduction in tariff and removal of nontariff barriers, aids the process of structural transformation and simultaneously changes the income distribution in an economy. This study investigates whether structural change impacts inequality. Using a panel of 217 countries during the period 1991–2014 and the System GMM method of dynamic panel data analysis, it is found that the process of structural change increases income inequality, while trade liberalization and FDI inflows help to reduce it. Income distribution is found to be more equal to infrastructure development. The econometric results are robust and have important policy implications. Keywords: inequality, growth and structural change, trade, System GMM JEL Classification: D63, F1, O18, O40, C23 ADBI Working Paper 763 Roy and Roy Contents 1. INTRODUCTION ......................................................................................................... 1 2. LITERATURE REVIEW ............................................................................................... 5 3. DATA, EMPIRICAL MODEL, AND ESTIMATION METHOD ....................................... 9 4. EMPIRICAL RESULTS AND DISCUSSIONS ........................................................... 11 5. CONCLUSIONS ........................................................................................................ 17 BIBLIOGRAPHY ................................................................................................................... 18 APPENDIX ............................................................................................................................ 25 ADBI Working Paper 763 Roy and Roy 1. INTRODUCTION Reduction in inequality within and across countries is one of the main targets of the recent Sustainable Development Goals (SDGs). Even though global inequality is found to have remained stable or, at best, declined1, there has been a rising incidence of income inequality in many countries since the 1980s. There is great heterogeneity in within-country inequality across countries and regions (Klasen et al. 2016). Inequality is seen to increase in developing countries, transition economies, and emerging market economies; these are the economies that have undergone structural transformation in the recent past. Within-country inequality is associated with different drivers, which vary across countries. Structural change is one such driver. This study aims to examine the causal relationship between income inequality and structural transformation while considering the role of international trade and foreign direct investment (FDI), as it is widely believed that globalization is one of the key drivers of increasing inequality both in advanced and developing countries. Structural change at a narrow level refers to changes in the structure of the economy, while at a broader level, it refers to social, political, cultural, societal, and other changes (Aizenman, Lee and Park 2012). Although there are many definitions of structural change, the most common meaning refers to long-term and continual shifts in the sectoral composition of economic systems (Chenery, Robinson and Syrquin 1986; Syrquin 2007; UNIDO 2009). According to Machlup (1991), structural change is “the different arrangements of productive activity in the economy and different distributions of productive factors among various sectors of the economy, various occupations, geographic regions, types of product, etc…..” Thus, in the process of structural change, a gradual shift of resources is observed from traditional to more sophisticated sectors. A rise in the relative share of the manufacturing sector is seen to occur, followed by a rise in the relative share of the service sector2. Before discussing the association between structural change and income inequality, it is important to study the pattern of both. Since the literature largely discusses the link between structural transformation and wage inequality, it is important to understand the change in wage gap between skilled and unskilled labor during the period1995–20093. Figure 1 shows that in major advanced countries, such as the United States, Japan, Canada, Australia, Germany, and some emerging countries, such as the People’s Republic of China (PRC), Brazil, Mexico, and Indonesia, the wage gap increased between 1995 and 2009. On the other hand, in other developing and emerging countries such as India; the Russian Federation; Taipei,China; and the Republic of Korea, the wage gap decreased between these two time periods. A discussion on the trend in inequality and the pattern of structural change across geographic regions can give some insights into the relationship between the two. 1 Nino-Zarazua, Roope and Tarp (2016) show that while relative global inequality declined substantially during the period 1975–2010, global inequality measured using ‘absolute’ and ‘centrist’ measures registered a pronounced increase during this period of time. 2 See Johnston (1970) for some other definitions of structural change. 3 Wage gap is calculated using the Socio-Economic Accounts (SEA) database of the World Input Output Database (WIOD). First, for all industries, the total hours worked by high-skilled and low-skilled employees and the total compensation for high-skilled and low-skilled labor are calculated. Then hourly compensation for high-skilled and low-skilled labor and the gap between them have been estimated for 1995 and 2009. The wage gap is the ratio of hourly compensation for high-skilled labor to that for low-skilled labor. 1 ADBI Working Paper 763 Roy and Roy Figure 1: Change in Wage Gap between High-skilled and Low-skilled Labor between 1995 and 2009 Source: Authors’ calculation on the basis of Socio-Economic Accounts (SEA) of the World Input Output Database (WIOD). Heterogeneity in income inequality, measured in terms of the Gini coefficient, can be found across countries in different geographical regions. Cross-country comparison of inequality is difficult on account of the lack of coverage and inconsistent data and methodology. In this exercise the World Bank database on inequality is used for purposes of comparison. Three indicators are considered, namely the difference between income shares of the top 20% and bottom 20% of the population, the difference between income shares of the top 10% and bottom 10% of the population, and the Gini index; these indicators can, however, be used interchangeably. From the yearly data on different indicators of inequality, the average indicators are calculated for the periods 1991–2000 and 2001–2010, as given in Table A2.The highest level of inequality is found in African countries, followed by South American and North American countries. Inequality is the lowest in European countries. In what follows, details can be found on each region. 2 ADBI Working Paper 763 Roy and Roy In Africa, very high inequality is seen in countries like Botswana, the Central African Republic, Namibia, and South Africa. In Botswana, Kenya, Ethiopia, Nigeria, and Cameroon, all three indicators show a downward trend from the 1990s to the 2000s, whereas in countries like Egypt, Morocco, and South Africa they show an upsurge. As an emerging market economy, South Africa showed high economic growth in the 2000s, and also experienced an increase in inequality, as the Gini coefficient is found to increase from 57.96 to 63.33. Inequality in Asian countries is not as severe as in African countries. From the 1990s to the 2000s, when the Chinese economy showed a huge increase in inequality, the Indian and Indonesian economies experienced a moderate increase. Small countries like Jordan, Kazakhstan, Malaysia, Pakistan, and the Philippines, and large countries like the Russian Federation and Thailand, showed a decline in inequality, whereas Bangladesh and Sri Lanka experienced an increase in inequality. The inequality measured in terms of the Gini coefficient for Bangladesh increased from 30.5 to 32.9. The Chinese economy showed almost a 15% increase in inequality in the last two decades. In the case of India, the Gini coefficient increased from 30.8 to 33.6. Among ASEAN countries, in Indonesia, Lao PDR, and Viet Nam, the Gini coefficient increased from 29.37 to 34.3, from 32.7 to 34.7, and from 35.6 to 36.8, respectively. Some other ASEAN countries like the Philippines and Thailand showed a downward trend in inequality. In Australia, inequality increased slightly4. In the 1990s, the average inequality measured in terms of the Gini coefficient was 33.7, and in the 2000s it increased to 34.1. In general, inequalities across countries in Europe are lower than those among Asian and African countries. In the 21st century, European countries show a mixed trend in terms of decline in inequality. Inequality has declined in countries such as Austria, France, Greece, Ireland, Moldova, the Netherlands, Spain, and Ukraine and increased in all other countries. At the same time, Switzerland has successfully reduced its level of inequality from 37.10 to 32.70 (in terms of the Gini coefficient); and in counties like Belgium it has gone up from a level of 26.75 to 33.14 (in terms of the Gini coefficient). Inequality in North American countries is higher than that in Asian and European countries but lower than in African countries. Inequality has increased in the 21st century in almost all major countries in this continent, though the magnitude varies across countries. In the United States, inequality measured in terms of all three indicators has increased marginally. Some countries, such as Guatemala, Mexico, Nicaragua, and Panama, however, have shown a marginal decline in inequality. In all South American countries, inequality is very severe. A high level of income and consumption inequality persists in countries like Bolivia, Brazil, Chile, Colombia, Paraguay etc. From the last decade of the 20th century to the beginning of the 21st century, inequality increased in all countries except Brazil, Chile, and Ecuador. In Paraguay and Peru it increased marginally. On the whole, inequality is highest in the Latin American countries followed by the Caribbean and Sub-Saharan African countries, while it is lowest in countries in Europe, and Central and South Asia (see Table 1). On the other hand, inequality across countries in Europe and North America is lower than that in Asian and African countries, in general. Inequality in all countries across regions is seen to have increased in 2000 and to have decreased thereafter, with high inequality prevailing in some African countries. Since 2000, the largest decline in the level of inequality can be seen among countries in East Asia and the Pacific (25.51%) followed 4 Data for New Zealand are not available. 3 ADBI Working Paper 763 Roy and Roy by countries in Latin America and the Caribbean (8.61%), and the Middle East and North Africa (8.12%). Table 1: Trend in Income Inequality across Regions Region Time Period 1990 1995 2000 2005 2010 East Asia and Pacific 35.63 38.48 50.06 40.86 37.29 Europe and Central Asia NA 33.19 33.34 32.15 31.25 Latin America and Caribbean 49.20 51.38 53.42 51.61 48.82 Middle East and North Africa 41.01 38.50 40.73 37.71 37.42 North America NA NA 37.06 NA 37.07 Sub-Saharan Africa NA 46.46 45.69 44.26 45.32 South Asia 32.85 34.52 33.06 32.71 31.44 World 42.67 43.25 43.57 38.40 36.41 Note: NA refers to nonavailability of data. Source: Authors’ own calculation on the basis of data obtained from WDI. Although structural change can only be observed over the long term, countries across geographical regions are found to have undergone structural transformation over a period of 25 years (1990–2014). By the early 1990s, most countries had started moving away from the agricultural sector towards the manufacturing and services sectors. Table 2 provides a snapshot of structural change across regions. Across geographical regions, the shares of the agriculture and manufacturing sectors are found to have decreased over time and that of services has increased. For Latin American and Caribbean countries, where inequality is the greatest, the share of the agricultural sector fell from 8.77% to 5.50% and that of the service sector increased from 53.66% to 65.53% between 1990 and 2014. In Sub-Saharan Africa, where inequality is also very high, the share of the agricultural sector decreased from 23.62% to 17.09%, and that of the manufacturing sector also decreased from 13.62% to 10.61%. Interestingly, in South Asian countries the share of the manufacturing sector remained more or less unchanged over the period. A shift is found to occur from agriculture to the services sector. Another interesting fact that can be observed is that the share of agriculture in North American countries is increasing marginally along with a shift from manufacturing to the services sector. On the whole, it can be seen that structural change is widespread in regions where inequality is high. Thus a relationship between the two is expected to exist. The paper thus investigates whether structural change determines inequality in countries across regions during globalization. The rest of the paper is structured as follows. This short introduction is followed by a review of existing literature in Section 2. In Section 3, data, the empirical model, and empirical methodological issues are discussed. Empirical results are discussed in Section 4, and in Section 5, a summary of the findings is presented. 4 ADBI Working Paper 763 Roy and Roy yit =δyi,t−1 + xit ′β+ uit i = 1, 2,.….., N; t = 1,2,……., T (2) where δ is a scalar, xit ′ is a 1 x K vector of strictly exogenous regressors, and β is a K x 1 vector of coefficients. The uit is assumed to follow a one-way error component model uit = µi+ vit (3) where µi and vit are independent of each other and IID with a mean of 0 and variance ofσµ 2 and σv, 2 respectively. The ineluctable correlation between yi,t−j, i.e., the lagged dependent variables, andui, i.e., the unobserved panel level effects, makes the OLS estimator biased and inconsistent even though vit is not serially correlated. Anderson and Hsiao (1981) show that first differencing of the model gives a consistent estimator. But this does not necessarily produce an efficient estimator. A generalized method of moments (GMM) procedure suggested by Arellano and Bond (1991) gives us a consistent estimator that is certainly more efficient than Anderson and Hsiao’s (1981) estimator. Before using GMM, the Arellano-Bond (1991) technique transforms all regressors by taking the first difference, and hence the technique is popularly known as the “difference GMM” technique (Hansen 1982). However, in the presence of too large autoregressive parameters, or if the ratio of the panel-level effect to the variance of idiosyncratic error is too large, this estimator can perform poorly. Based on the study of Arellano and Bover (1995), Blundell and Bond (1998) developed an estimator assuming the absence of autocorrelation in the idiosyncratic errors and no correlation between panel-level effects and the first difference of the dependent variable. The first difference GMM model is found to have very poor finite sample properties in terms of biasness and precision, especially when the series is persistent as the instruments are then weak predictors of endogenous changes. As a remedy, the level restrictions and the use of extra moment conditions that depend on certain stationarity conditions of the initial observation suggested by Arellano and Bover (1995) are factual and also augmented by Blundell and Bond (1998) by making an additional assumption of no correlation between the first difference of instrument variables and fixed effects. In doing so, one can increase efficiency by introducing more instruments. This method is called “System GMM” as it deals with a system of two equations – the original equation and the transformed equation. This System GMM estimator not only improvises precision but also reduces finite sample bias even when covariates are weakly exogenous. With a large sample of individuals or cross section of units observed for a small number of time periods, difference GMM estimators have been found to produce unsatisfactory results (Mairesse and Hall 1996). However, with large T first difference GMM estimator performs relatively well. Blundell and Bond (1998) suggest use of extra moment conditions with small T. In this study, since we have considered many panels with few time periods, we consider a system estimator as suggested by Blundell and Bond (1998). 4. EMPIRICAL RESULTS AND DISCUSSIONS Table 3 shows summary statistics of the Gini index and some other important determinants of income inequality. It can be seen that the average level of inequality is highest in South American countries. Among African countries, the average level of income (measured by average per capita GDP) is the lowest and at the same time the average inequality is very high. Interestingly, among North American countries, both the average inequality and average income are very high in contrast to European countries, where the average income is very high and average inequality is the lowest. 11 ADBI Working Paper 763 Roy and Roy Table 3: Summary Statistics Continent Variable Statistics Gini PCGDP TO FDI Infrastructure Manu Serv Africa Mean 44.89 3,545.89 0.76 7.04E+08 43,323.08 11.01 48.06 S.D 8.46 10,936.72 0.46 2.01E+09 108,750.90 7.09 13.43 Min 29.81 115.44 0.05 1.00E+02 1,682.33 0.24 12.87 Max 65.76 94,903.20 3.38 2.37E+10 591,906.50 45.67 93.22 Asia Mean 36.50 10,422.48 0.88 5.88E+09 89,230.85 15.55 51.64 S.D 6.38 14,744.27 0.60 2.25E+10 175,048.70 7.56 13.56 Min 19.49 314.88 0.09 1.00E+01 1,682.33 0.86 16.56 Max 69.47 74,632.24 4.00 2.91E+11 1525,740.00 40.45 83.70 Europe Mean 31.65 26,171.71 0.94 1.34E+10 49,724.03 16.95 62.82 S.D 4.27 22,939.06 0.48 4.16E+10 78,080.52 7.95 15.08 Min 16.23 690.92 0.17 1.00E+03 1,925.26 0.69 2.43 Max 44.42 145,221.20 3.61 7.34E+11 591,906.50 47.34 93.76 North America Mean 49.23 10,982.34 0.95 1.34E+10 34,612.17 12.52 65.57 S.D 6.61 12,630.64 0.74 4.61E+10 40,693.69 6.91 11.00 Min 31.15 662.28 0.16 3.00E+05 1,682.33 1.28 33.40 Max 60.91 50,662.41 4.48 3.50E+11 220,406.50 29.01 92.98 Oceania Mean 41.13 8,596.69 0.58 2.20E+09 33,510.30 7.29 63.56 S.D 8.05 13,273.78 0.28 8.80E+09 38,402.92 5.19 12.44 Min 33.72 1,047.45 0.24 1.00E+01 1,682.33 0.38 22.81 Max 61.18 54,232.66 1.32 6.56E+10 200,919.90 19.93 88.02 South America Mean 51.54 6,624.60 0.53 7.49E+09 104,135.30 15.84 55.13 S.D 5.03 3,728.78 0.26 1.66E+10 171,701.40 4.11 8.69 Min 40.20 1,397.18 0.10 7.30E+06 1,925.26 3.68 26.12 Max 63.00 14,687.98 1.31 1.12E+11 591,906.50 28.31 72.85 Across countries, variation in inequality can also be understood from the standard deviation of the Gini index. The highest variability is observed among African countries and the least variability is found to exist among European countries. On the other hand, the size of the manufacturing sector and service sector (measured in terms of average GDP share of manufacturing and service sectors, respectively) is largest in European countries and smallest in African countries. Before the empirical estimation, it is important to check the possibility of the presence of multicollinearity. Table A3 (see Appendix 2) presents the correlation matrix of the explanatory variables. It can be seen that the GDP share of the service sector, FDI, and the quality of the infrastructure have high correlations with PCGDP. Thus, while estimating the empirical model, PCGDP has been considered an endogenous variable. The absence of a high correlation among any other pairs of explanatory variables is evidence in favor of the absence of a multicollinearity problem. Table 4 presents the results of the first set of estimations. In each model, the Sargan test has been carried out to check the validity of the overidentifying restriction. The Sargan test 11 accepts the null hypothesis of valid overidentifying restrictions. Instrumental variables must be uncorrelated with the structural error term and correlated with the endogenous regressors. Here all models are overidentified or the number of additional instruments used in each model exceeds the number of 11 See Sargan (1958). 12 ADBI Working Paper 763 Roy and Roy endogenous regressors, and instruments are uncorrelated with the error term. In Table 4, four different specifications of equation (1), with changes only in the measure of inequality, have been shown. Regression equations using four different dependent variables, namely the Gini index, income share of the top quintile, income share of the bottom quintile and the ratio of the two quintiles, and controlling for a group of basic variables (PCGDP, TO, FDI, quantity and quality of infrastructure, urbanization), as well as the two variables of interest, share of manufacturing sector and share of service sector, are estimated. Table 4: Estimation Result (Overall) Model 1 Model 2 Dependent Variable lnGini ln Q1 Coef. SE Coef. SE lnInequalityt-1 0.7663 (0.0321)*** 0.8328 (0.0367)*** lnInequalityt-2 0.1795 (0.0326)*** 0.0981 (0.0355)*** lnPCGDPt –0.0267 (0.0107)** –0.0185 (0.0086)** lnManufacturing_Sharet 0.0366 (0.0156)** 0.0228 (0.0127)* lnServices_Sharet 0.1031 (0.0296)*** 0.0540 (0.0214)** lnTOt –0.0227 (0.0122)* –0.0236 (0.0101)** lnFDIt –0.0047 (0.0027)* –0.0038 (0.0021)* lnInfrastructure_Quantityt –0.0135 (0.0053)** –0.0013 (0.0037) lnInfrastructure_Qualityt –0.0222 (0.0116)* –0.0057 (0.0090) lnUrbanizationt 0.0185 (0.0108)* 0.0008 (0.0080) Constant 0.1785 (0.1757) 0.2367 (0.1475) Sargan test p value 0.5654 0.9719 Model 3 Model 4 Dependent Variable ln Q5 ln (Q1/Q5) Coef. SE Coef. SE lnInequalityt-1 0.7029 (0.0361)*** 0.7408 (0.0358)*** lnInequalityt-2 0.1837 (0.0338)*** 0.1629 (0.0338)*** lnPCGDPt 0.0179 (0.0265) –0.0406 (0.0332) lnManufacturing_Sharet –0.0667 (0.0388)* 0.0942 (0.0486)* lnServices_Sharet –0.1192 (0.0682)* 0.1716 (0.0843)** lnTOt 0.0527 (0.0300)* –0.0735 (0.0378)* lnFDIt 0.0100 (0.0068) –0.0125 (0.0084) lnInfrastructure_Quantityt 0.0083 (0.0117) –0.0100 (0.0145) lnInfrastructure_Qualityt –0.0250 (0.0270) 0.0142 (0.0341) lnUrbanizationt –0.0053 (0.0247) 0.0079 (0.0310) Constant 0.5139 (0.3526) –0.1259 (0.4522) Sargan test p value 0.2332 0.4514 Note: (a) Standard errors are given in parentheses. (b) An * implies significance at the 10% level; ** implies significance at the 5% level; and *** implies significance at the 1% level. (c) PCGDP, Manufacturing Share and Services Share are considered to be endogenous. 13 ADBI Working Paper 763 Roy and Roy In the first model, log of Gini index has been used as the dependent variable. Clear evidence of path dependence can be seen from the result as the lagged dependent variable is found to be positive and significant. So it is likely that if inequality exists in the present period, it will prevail in the future period as well, if not controlled. Per capita GDP is found to be negative and significant, and thus there is evidence of a trickledown effect. Trade openness – as measured by the ratio of exports and imports to GDP – tends to make income distribution more equal. This clearly confirms the findings of White and Anderson (2001), Dollar and Kraay (2002), Edwards (1997), and Higgins and Williamson (1999); however, it contradicts the finding of Barro (2000), Calderon and Serven (2004, 2008), and Wan, Lu and Chen (2006a). The coefficient of FDI is significant and negative, suggesting that FDI reduces income inequality. This is consistent with Markusen and Venables (1997), Blonigen and Slaughter (2001), and Aghion and Howitt (1998); however, it contradicts the finding of Wan et al. (2006b). A negative and significant relationship between infrastructure stock and income inequality is found. That is, the larger stock of infrastructure, the more equal the distribution of income. This result is consistent with the findings of Calderon and Chong (2004) and Seneviratne and Sun (2013). Similarly, there is a negative and significant relationship between the quality of infrastructure and income inequality. In short, the better the quality of infrastructure, the more equal the distribution of income. This confirms the findings of Seneviratne and Sun (2013); however, it contradicts the findings of Calderon and Chong (2004). Urbanization is found to have a positive significant relationship with income inequality. This is consistent with the finding of Wan et al. (2006b) but at the same time contradicts the result of Wan et al. (2006a). The two variables of interest, namely the share of the manufacturing sector and that of the service sector, are found to be positive and significant. This implies that the process of structural transformation results in a more unequal distribution of income. A 1% increase in the share of the manufacturing sector in GDP results in a 3% increase in income inequality. On the other hand, a 1% increase in the GDP share of the service sector increases income inequality by 0.10%. To confirm this, three models have been estimated considering three other dependent variables. In the second model, where the income share of the top 20% of the population has been used as the dependent variable, GDP shares of the manufacturing and service sectors are found to be positive and significant. A 1% increase in the GDP share of the manufacturing and service sectors increases the income share of the top 20%of the population by 0.02% and 0.05%, respectively, and thus increases income inequality. On the other hand, when the income share of the bottom 20% of the population has been considered as the dependent variable, the two coefficients have been found to be negative and significant. It can be seen that a 1% increase in the GDP share of the manufacturing and service sectors decreases the income share of the bottom 20%of the population by 0.06% and 0.11%, respectively, and thus makes the income distribution more unequal. The result is the same even when the ratio of the income shares of the two groups or the difference between the income groups has been considered as the dependent variable. The gap in the income shares between the two income groups increases by 0.09% and 0.17% when the share of the manufacturing sector and that of the service sector, respectively, increase by 1%. This clearly proves the robustness of the results. Now, to check the heterogeneity across regions, instead of GDP shares of the manufacturing and service sector as a whole, the interaction of sectoral shares with region dummies has been considered. For each region, a high correlation has been found between the share of the manufacturing sector and that of the service sector (see Table A4 in Appendix 2). Thus Model 5-8 in Table 5 considers interaction dummies only with the manufacturing share, and Model 9–12 in Table 6 considers interaction dummies only with the service share. Due to insufficient data on inequality 14 ADBI Working Paper 763 Roy and Roy measures, estimation for two regions, namely Africa and Oceania, has not been done. It can be seen that the expansion of the manufacturing sector is significantly increasing inequality in two regions, namely North America and South America (see Model 5 in Table 5). On the other hand, in all four regions, income distribution is found to become more unequal due to the expansion of the service sector (see Model 9 in Table 6). Table 5: Region-Specific Estimation Result (Manufacturing) Model 5 Model 6 Dependent Variable lnGini ln Q1 Coef. SE Coef. SE lnInequalityt-1 0.7432 (0.0327)*** 0.7915 (0.0393)*** lnInequalityt-2 0.0984 (0.0354)*** 0.0439 (0.0369) lnPCGDPt 0.0120 (0.0106) 0.0019 (0.0082) D_Europe×lnManufacturing_Sharet 0.0207 (0.0138) 0.0135 (0.0114) D_NorthAmerica×lnManufacturing_Sharet 0.0483 (0.0146)*** 0.0305 (0.0117)*** D_SouthAmerica×lnManufacturing_Sharet 0.0473 (0.0170)*** 0.0319 (0.0131)** D_Asia×lnManufacturing_Sharet 0.0226 (0.0147) 0.0136 (0.0118) lnTOt –0.0270 (0.0143)* –0.0199 (0.0106)* lnFDIt –0.0050 (0.0028)* –0.0026 (0.0021) lnInfrastructure_Quantityt –0.0065 (0.0049) –0.0016 (0.0034) lnInfrastructure_Qualityt 0.0077 (0.0121) 0.0096 (0.0091) lnUrbanizationt –0.0138 (0.0122) –0.0156 (0.0092)* Constant 0.5356 (0.1659)*** 0.6018 (0.1604)*** Sargan test p value 0.8600 0.9198 Model 7 Model 8 Dependent Variable ln Q5 ln (Q1/Q5) Coef. SE Coef. SE lnInequalityt-1 0.6028 (0.0373)*** 0.6498 (0.0375)*** lnInequalityt-2 0.1064 (0.0343)*** 0.0883 (0.0345)** lnPCGDPt –0.0419 (0.0243)* 0.0450 (0.0305) D_Europe×lnManufacturing_Sharet –0.0460 (0.0341) 0.0642 (0.0432) D_NorthAmerica×lnManufacturing_Sharet –0.1267 (0.0344)*** 0.1658 (0.0438)*** D_SouthAmerica×lnManufacturing_Sharet –0.1525 (0.0402)*** 0.1911 (0.0509)*** D_Asia×lnManufacturing_Sharet –0.0444 (0.0363) 0.0685 (0.0457) lnTOt 0.0269 (0.0322) –0.0475 (0.0406) lnFDIt 0.0017 (0.0063) –0.0042 (0.0081) lnInfrastructure_Quantityt 0.0143 (0.0106) –0.0153 (0.0131) lnInfrastructure_Qualityt –0.0653 (0.0252)*** 0.0752 (0.0323)** lnUrbanizationt 0.0783 (0.0292)*** –0.0921 (0.0367)** Constant 1.0554 (0.2976)*** –0.0714 (0.3711) Sargan test p value 0.2989 0.5016 Note: (a) Standard errors are given in parentheses. (b) An * implies significance at the 10% level; ** implies significance at the 5% level; and *** implies significance at the 1% level. (c) PCGDP and Manufacturing Share are considered to be endogenous. 15 ADBI Working Paper 763 Roy and Roy Table 6: Region-Specific Estimation Result (Service) Model 9 Model 10 Dependent Variable lnGini ln Q1 Coef. SE Coef. SE lnInequalityt-1 0.6909 (0.0327)*** 0.7524 (0.0404)*** lnInequalityt-2 0.0801 (0.0332)** 0.0211 (0.0367) lnPCGDPt –0.0111 (0.0100) –0.0038 (0.0080) D_Europe×lnServices_Sharet 0.0813 (0.0288)*** 0.0398 (0.0210)* D_NorthAmerica×lnServices_Sharet 0.1066 (0.0293)*** 0.0567 (0.0212)*** D_SouthAmerica×lnServices_Sharet 0.1087 (0.0300)*** 0.0573 (0.0217)*** D_Asia×lnServices_Sharet 0.0853 (0.0296)*** 0.0438 (0.0216)** lnTOt –0.0024 (0.0115) –0.0165 (0.0099)* lnFDIt –0.0042 (0.0027) –0.0016 (0.0021) lnInfrastructure_Quantityt –0.0022 (0.0045) 0.0004 (0.0032) lnInfrastructure_Qualityt 0.0046 (0.0105) 0.0127 (0.0078) lnUrbanizationt –0.0006 (0.0120) –0.0132 (0.0088) Constant 0.6529 (0.1609)*** 0.7033 (0.1611)*** Sargan test p value 0.4712 0.8877 Model 11 Model 12 Dependent Variable ln Q5 ln (Q1/Q5) Coef. SE Coef. SE lnInequalityt-1 0.6146 (0.0377)*** 0.6501 (0.0382)*** lnInequalityt-2 0.1012 (0.0343)*** 0.0810 (0.0345)** lnPCGDPt –0.0048 (0.0238) –0.0019 (0.0299) D_Europe×lnServices_Sharet –0.0831 (0.0630) 0.1328 (0.0791)* D_NorthAmerica×lnServices_Sharet –0.1419 (0.0629)** 0.2090 (0.0793)*** D_SouthAmerica×lnServices_Sharet –0.1438 (0.0645)** 0.2109 (0.0813)** D_Asia×lnServices_Sharet –0.0924 (0.0645) 0.1473 (0.0810)* lnTOt 0.0233 (0.0293) –0.0367 (0.0370) lnFDIt 0.0019 (0.0065) –0.0035 (0.0082) lnInfrastructure_Quantityt 0.0094 (0.0101) –0.0110 (0.0126) lnInfrastructure_Qualityt –0.0592 (0.0227)*** 0.0694 (0.0287)** lnUrbanizationt 0.0390 (0.0270) –0.0470 (0.0338) Constant 0.9834 (0.2882)*** –0.0818 (0.3489) Sargan test p value 0.2217 0.3912 Note: (a) Standard errors are given in parentheses. (b) An * implies significance at the 10% level; ** implies significance at the 5% level; and *** implies significance at the 1% level. (c) PCGDP and Manufacturing Share are considered to be endogenous. Furthermore, these interactive dummies are found to have a positive and statistically significant association with respect to the top 20% income share (see Model 6 in Table 5 and Model 10 in Table 6), and a negative and statistically significant relation with respect to the bottom 20% share of income (see Model 7 in Table 5 and Model 11 in Table 6). A positive relationship is found even when the ratio of the two income groups or the gap between the two income groups has been considered (see Model 8 in Table 5 and Model 12 in Table 6). All these econometric results with variants of 16 ADBI Working Paper 763 Roy and Roy income inequality measure are therefore found to be robust, and thus structural change is found to be associated with an increase in income inequality. 5. CONCLUSIONS In the literature on economic development, one of the earliest and most central themes is structural change. The countries that developed in the last few centuries are those that are able to diversify away from the production and consumption of traditional goods to modern sectors. Since the early 1990s, the developing countries have experienced rapid structural change and at the same time become more integrated with the world economy. The reduction of import tariffs and nontariff barriers (through infrastructure development), FDI flows, and thus globalization facilitated technology transfers to these countries. This reduction in trade barriers, FDI flows, and technology transfers not only promotes growth, but also leads to structural change. In the process, the demand for skilled labor increases, leading to a wage gap, and thus inequality increases. This study empirically shows the positive impact of structural change on income inequality, that is, how structural change results in a more unequal distribution of income. While all previous studies have shown impacts of structural change on wage inequality, this study is the first to show the impact of structural transformation on overall income inequality. The data include a panel of a large number of countries from all income groups and all regions. To check the robustness of the results, different indicators of inequality have been considered. Analysis considering regional interactive dummies shows that among North and South American countries, both expansion of manufacturing and expansion of services are found to increase income inequality. On the other hand, in Asia and Europe, the problem of inequality has worsened with expansion of the service sector only. The study also shows the strong negative impact of trade liberalization on income inequality and weak negative impact of FDI inflow on the same in the long run. The study thus contributes to the literature by raising many important dimensions for policy analysis. The results are of particular importance with regard to Sustainable Development Goal 10 on Reduced Inequalities within and between countries. The widening disparity requires the adoption of sound policies to empower the bottom deciles of income earners through structural transformation, infrastructure development, and focusing on those groups of people where it is most required. Trade liberalization and FDI can be chosen as policy instruments to reduce inequality. This study, however, does not take into account the role of migration and development assistance in bridging the inequalities. 17 ADBI Working Paper 763 Roy and Roy BIBLIOGRAPHY Acharyya, R. (2008). International Trade, Poverty, and Income Inequality: The Indian Experience during the Reform Period 1985–2000. In: Asian Development Bank Trade Policy, Industrial Performance, and Private Sector Development in India. New Delhi: Oxford University Press. 97–140. ———. (2010). Successive Trade Liberalization and Wage Inequality. Keio Economic Studies. 26, 29–42. ———. (2016). International Trade and Widening Wage Gap: A General Equilibrium Approach. In: Bandyopadhyay S., Torre A., Casaca P. and Dentinho T. P. Regional Cooperation in South Asia: Socio-economic, Spatial, Ecological and Institutional Aspects. New Delhi: Springer (Forthcoming). Aghion, P. and Howitt P. (1998). Endogenous Growth Theory. Cambridge, MA: MIT Press. Aghion, P., Caroli, E. and Garcia-Penalosa, C. (1999). Inequality and Economic Growth: The Perspective of the New Growth Theories. Journal of Economic Literature. 37(4), 1615–1660. Aitken, B., Harrison, A. and Lipsey, R.E. (1996). Wages and Foreign Ownership: A Comparative Study of Mexico, Venezuela, and the United States. Journal of International Economics. 40(3–4), 345–371. Aizenman, J., Lee, M. and Park, D. (2012). The Relationship Between Structural Change and Inequality: A Conceptual Overview with Special Reference to Developing Asia. ADBI Working Paper Series No. 396. Alesina, A. and Perotti, R. (1996). Income Distribution, Political Instability and Investment. European Economic Review. 40(6), 1203–1228. Alesina, A. and Rodrik, D. (1994). Distributive Politics and Economic Growth. Quarterly Journal of Economics.108, 465–90. Ali, I., and Son, H. (2007). Defining and Measuring Inclusive Growth: Application to the Philippines. ERD Working Paper Series No. 98.ADB. Manila. Anderson, E. (2005). Openness and Inequality in Developing Countries: A Review of Theory and Recent Evidence. World Development. 33(7), 1045–1063. Anderson, T. W. and Hsiao, C. (1981). Estimation of Dynamic Models with Error Components. Journal of the American Statistical Association. 76(375), 598–606. Arellano, M., and Bond, S. (1991). Some Tests of Specification for Panel Data: Monte Carlo Evidence and an Application to Employment Equations. Review of Economic Studies. 58(2), 277–297. Arellano, M., and Bover, O. (1995). Another Look at the Instrumental Variable Estimation of Error-Components Models. Journal of Econometrics. 68(1), 29–51. Attanasio, O., Goldberg, P. K. and Pavcnik, N. (2004). Trade Reforms and Wage Inequality in Colombia. Journal of Development Economics. 74(2), 331–366. Avalos, A. and Savvides, A. (2006). The Manufacturing Wage Inequality in Latin America and East Asia: Openness, Technology Transfer, and Labor Supply. Review of Development Economics. 10(4), 553–576. Banerjee, A. V. and Newman A. F. (1993). Occupational Choice and the Process of Development. Journal of Political Economy. 101(2), 274–298. 18 ADBI Working Paper 763 Roy and Roy Barro, J and Lee, J.W. (2001). International Data on Educational Attainment: Updates and Implications. Oxford Economic Papers. Oxford University Press. 53(3), 541–563. Barro, R. (2000). Inequality and Growth in a Panel of Countries. Journal of Economic Growth. 5(1), 05–32. Benabou, R. (2000). Unequal Societies: Income Distribution and Social Contract. The American Economic Review. 90(1), 96–129. ———. (2002). Tax and Education Policy in a Heterogeneous-Agent Economy: What Levels of Redistribution Maximize Growth and Efficiency? Econometrica. 70(2), 481–517. Berg, A. and Ostry, J.D. (2011). Inequality and Unsustainable Growth: Two Sides of the Same Coin? IMF Staff Discussion Note 11/08. Berg, A., Ostry, J.D. and Zettelmeyer, J. (2012). What Makes Growth Sustained? Journal of Development Economics. 98(2), 149–166. Bleaney, M., Gemmell, N. and Kneller, R. (2001). Testing the Endogenous Growth Model: Public Expenditure, Taxation, and Growth over the Long Run. Canadian Journal of Economics. 34(1), 36–57. Blonigen, B.A. and Slaughter, M.J. (2001). Foreign-Affiliate Activity and U.S. Skill Upgrading. Review of Economics and Statistics. 83(2), 362–376. Blum, B. (2008). Trade, Technology and the Rise of the Service Sector: The Effects on US Wage Inequality. Journal of International Economics. 4(2), 441–458. Blundell, R. and Bond, S. (1998). Initial Conditions and Moment Restrictions in Dynamic Panel Data Models. Journal of Econometrics. 87(1), 115–143. Bornschier, V. and Chase-Dunn, C. (1985). Transnational Corporations and Underdevelopment. New York: Praeger Press. Calderon, C. and Chong, A. (2004). Volume and Quality of Infrastructure and the Distribution of Income: An Empirical Investigation. Review of Income and Wealth. 50(1), 87–106. Calderon, C. and Serven, L. (2004). The Effects of Infrastructure Development on Growth and Income Distribution. World Bank Policy Research Working Papers No. 3400. ———. (2008). Infrastructure and Economic Development in Sub-Saharan Africa. World Bank Policy Research Working Papers No. 4712. Chandrasekhar C.P. and Ghosh J. (2015). Growth, Industrialisation and Inequality in India. Journal of the Asia Pacific Economy. 20(1), 42–56. Chari, A., Henry, P. B. and Sasson, D. (2012). Capital Market Integration and Wages. American Economic Journal: Macroeconomics. 4(2), 102–32. Chatterjee, S. (1991). The Effect of Transitional Dynamics on the Distribution of Wealth in a Neoclassical Capital Accumulation Model. Federal Reserve Bank of Philadelphia, Working Paper No. 91–22. Chenery H., Robinson S. and Syrquin, M. (1986). Industrialization and Growth, a Comparative Study. Published for The World Bank, Oxford University Press. Choi, C. (2006). Does Foreign Direct Investment Affect Domestic Income Inequality? Applied Economics Letters. 13(12), 811–814. 19 ADBI Working Paper 763 Roy and Roy Chong, A. and Calderon, C. (2000). Institutional Quality and Income Distribution. Economic Development and Cultural Change. 48(4), 761–786. ———. (2001). Volume and Quality of Infrastructure and the Distribution of Income: An Empirical Investigation. Inter-American Development Bank Working Paper No: 450. ———. (2001). Volume and Quality of Infrastructure and the Distribution of Income: An Empirical Investigation. Inter-American Development Bank Working Paper No: 450. Cornia, G. A. (2005). Policy Reform and Income Distribution. DESA Working Paper No. 3. Dabla-Norris, E., Kochhar, K., Suphaphiphat, N., Ricka, F. and Tsounta, E. (2015). Causes and Consequences of Income Inequality: A Global Perspective. IMF Discussion Note No. SDN/15/13. Dastidar A. G. (2012). Income Distribution and Structural Transformation: Empirical Evidence from Developed and Developing Countries. Seol Journal of Economics. 25(1), 25–56. ———. (2004). Structural Change and Income Distribution in Developing Economies: Evidence from a Group of Asian and Latin American Countries. Centre for Developing Economics Working Paper No. 121. Delhi: Delhi School of Economics. Deardorff, A. and Stern, R. (1994). The Stolper–Samuelson Theorem: A Golden Jubilee. Ann Arbor, NI: University of Michigan Press. Deaton, A. and Dreze, J. (2002). Poverty and Inequality in India: A Re-examination. Economic and Political Weekly. 37(36), 3729–3748. Deininger, K. and Squire, L. (1996). A new dataset measuring income inequality. World Bank Economic Review. 10(3), 565-591. Dollar D. and Kraay, A. (2002). Growth is Good for the Poor. Journal of Economic Growth. 7(3), 195–225. Easterly, W. (2007). Inequality Does Cause Underdevelopment: Insights from a New Instrument. Journal of Development Economics. 84(2), 755–776. ECLAC (2012). Structural Change for Equality: An Integrated Approach to Development. Thirty-fourth Session of ECLAC. San Salvador. Edwards, S. (1997). Trade Policy, Growth and Income Distribution. The American Economic Review. 87(2), 205–210. Estache, A. (2003). On Latin America’s Infrastructure Privatization and its Distributional Effects. Washington, DC: The World Bank, Mimeo. Estache, A., Foster, V. and Wodon, Q. (2002). Accounting for Poverty in Infrastructure Reform: Learning from Latin America’s Experience. WBI Development Studies. Washington, DC: The World Bank. Figini, P. and Görg.H. (1999). Multinational Companies and Wage Inequality in the Host Country: The Case of Ireland. Review of World Economics (Weltwirtschaftliches Archiv). 135(4), 594–612. Freenstra R.C. and Hansen, G. (1997). Foreign Direct Investment and Relative Wages: Evidence from Mexico’s Maquiladoras. Journal of International Economics. 42(3–4), 371–393. 20 ADBI Working Paper 763 Roy and Roy Table A2 continued 1990s Income Share Income Share Continent Country Highest 20 % Lowest 20% Difference between Quintiles Highest 10p% Lowest 10% Difference between Deciles Gini Asia Jordan 47.2 6.78 40.42 32.4 2.9 29.51 39.9 Asia Kazakhstan 41.4 7.17 34.24 25.7 2.9 22.73 34 Asia Kyrgyz Republic 50.3 4.84 45.45 34.1 1.9 32.2 44.8 Asia Lao PDR 41.7 8.65 33.03 27.4 3.8 23.61 32.7 Asia Malaysia 53.8 4.51 49.25 37.8 1.8 35.96 48.4 Asia Maldives 65.7 1.41 64.33 48.1 0.4 47.75 62.7 Asia Mongolia 39.5 7.55 31.91 23.9 3.1 20.89 31.7 Asia Nepal 43.5 7.87 35.65 29.1 3.4 25.69 35.2 Asia Pakistan 40.9 8.93 31.99 26.9 3.9 22.95 31.6 Asia Philippines 50.8 5.73 45.05 35 2.5 32.5 44.3 Asia Russian Federation 49.4 5.01 44.39 33.8 1.8 32 44 Asia Slovak Republic 33.1 10.3 22.81 19.5 4.1 15.41 22.7 Asia Sri Lanka 42.7 8.37 34.34 28.3 3.7 24.62 34 Asia Tajikistan 38.1 8.34 29.77 23.3 3.3 20.05 29.5 Asia Thailand 50.9 6.01 44.91 35.1 2.5 32.59 44 Asia Turkey 47.7 5.8 41.88 32.3 2.3 29.99 41.5 Asia Uzbekistan 49.6 3.91 45.65 33.4 1.1 32.27 45.3 Asia Viet Nam 44 7.92 36.08 29.2 3.5 25.64 35.6 Asia Yemen, Rep. 41.2 7.41 33.75 25.9 3 22.88 33.4 Europe Armenia 47.3 6.57 40.73 32.4 2.7 29.69 40.2 Europe Austria 38.6 7.64 31 23.5 2.8 20.76 31 Europe Azerbaijan 42.3 6.94 35.31 27 2.8 24.29 35 Europe Belarus 36 9.4 26.62 21.7 3.9 17.79 26.5 Europe Belgium 36 9.03 26.92 21.5 3.5 18.04 26.8 Europe Bulgaria 37.9 9.09 28.81 23.6 3.8 19.85 28.5 Europe Croatia 37.1 9 28.13 22.5 3.7 18.87 28.1 Europe Czech Republic 36.7 10.3 26.41 23.2 4.5 18.68 26.2 Europe Denmark 34.2 9.93 24.29 20.2 3.8 16.42 24.3 Europe Estonia 43.2 7.16 36.05 27.9 3 24.99 35.7 Europe Finland 34.1 10.7 23.32 20.1 4.6 15.54 23.2 Europe France 40.5 7.92 32.62 25.7 3.2 22.47 32.4 Europe Georgia 45.8 5.44 40.4 30.1 1.9 28.14 40.1 Europe Germany 38.4 8.31 30.13 23.7 3.3 20.37 30 Europe Greece 43.3 5.78 37.5 27.4 1.9 25.49 37.2 Europe Hungary 37.1 9.58 27.55 23.2 4 19.13 27.4 Europe Ireland 44.1 6.96 37.09 28.4 2.8 25.62 36.5 Europe Italy 41.6 6.36 35.23 26.3 2.2 24.18 35.1 continued on next page 27 ADBI Working Paper 763 Roy and Roy Table A2 continued 1990s Income Share Income Share Continent Country Highest 20 % Lowest 20% Difference between Quintiles Highest 10p% Lowest 10% Difference between Deciles Gini Europe Latvia 39.2 8.04 31.2 24.7 3 21.68 31 Europe Lithuania 40.8 7.87 32.94 26.2 3.1 23.05 32.7 Europe Macedonia, FYR 36.7 8.48 28.2 22.1 3.3 18.88 28.1 Europe Moldova 45 6.38 38.61 29.4 2.5 26.93 38.1 Europe Netherlands 38.8 7.8 30.96 23.2 2.5 20.68 30.7 Europe Norway 35.8 9.44 26.32 21.4 3.8 17.63 26.4 Europe Poland 39.6 8.4 31.22 24.8 3.5 21.26 31.1 Europe Romania 37.1 8.82 28.24 22.5 3.6 18.86 28.1 Europe Slovenia 38.2 9.19 28.99 23.8 4 19.81 28.8 Europe Spain 41.8 6.78 35.05 26.4 2.4 24.02 34.7 Europe Sweden 34.6 9.23 25.36 20.1 3.4 16.72 25.5 Europe Switzerland 42.5 5.32 37.2 27.2 0.8 26.4 37.1 Europe Ukraine 40.4 7.91 32.46 25.5 3.3 22.27 32.3 Europe United Kingdom 43.5 6.32 37.17 27.9 2.2 25.66 36.9 North America Canada 39.5 7.32 32.21 24.2 2.7 21.5 32 North America Costa Rica 50.9 3.94 46.91 34.2 1.1 33.07 46.2 North America Ivory Coast 45.3 6.51 38.81 29.6 2.7 26.92 38.4 North America Dominican Republic 54.3 4.19 50.1 38.8 1.5 37.31 49.2 North America El Salvador 56.2 2.84 53.39 39.8 0.7 39.11 52.4 North America Guatemala 59.7 3.14 56.53 44.8 1 43.81 55.8 North America Honduras 58.9 3.09 55.77 42.9 1 41.93 54.6 North America Jamaica 47.3 6.16 41.14 32 2.5 29.48 40.6 North America Mexico 55.1 4.19 50.93 39.4 1.7 37.72 50.1 North America Nicaragua 55.9 3.74 52.12 40.1 1.3 38.81 51.3 North America Panama 60.5 1.55 58.99 43.1 0.2 42.94 57.6 North America United States 44.6 5.28 39.28 28.3 1.8 26.52 39.1 Oceania Australia 40.8 6.8 33.98 24.9 2.1 22.78 33.7 South America Argentina 52.8 4 48.76 36 1.3 34.68 47.9 South America Bolivia 57.2 3.15 54.06 40.7 1.1 39.64 53 South America Brazil 63.1 2.42 60.65 46.6 0.7 45.9 59 South America Chile 61 3.56 57.4 45.5 1.3 44.2 55.8 South America Colombia 58.9 2.94 55.97 43.1 0.8 42.32 54.6 South America Ecuador 58 3.27 54.69 42.2 0.9 41.25 53.4 South America Paraguay 56.5 3.38 53.09 40.1 1.1 38.98 52.1 South America Peru 53.6 4.49 49.08 37.8 1.7 36.1 48.1 South America Uruguay 47.9 5.1 42.84 31.6 1.8 29.75 42.3 South America Venezuela, RB 51.1 4.14 47 34.7 1.3 33.48 46.3 continued on next page 28 ADBI Working Paper 763 Roy and Roy Table A2 continued 2000s Income Share Income Share Continent Country Highest 20p% Lowest 20% Difference between Quintiles Highest 10p% Lowest 10% Difference between Deciles Gini Africa Botswana 67.3 2.56 64.7 51.4 0.9 50.41 62.6 Africa Burkina Faso 48.4 6.27 42.1 33.1 2.7 30.4 41.5 Africa Burundi 42.8 8.96 33.79 28 4.1 23.9 33.3 Africa Cameroon 48.3 6.26 42.04 32.8 2.7 30.09 41.4 Africa Central African Republic 55 4.29 50.71 39.6 1.7 37.92 49.9 Africa Egypt, Arab Rep. 41.3 9.05 32.24 27.5 3.9 23.61 31.9 Africa Ethiopia 40.6 8.61 32.02 26.6 3.6 22.95 31.7 Africa Gambia, The 52.8 4.79 48.05 36.9 2 34.99 47.3 Africa Ghana 48.6 5.24 43.31 32.8 2 30.72 42.8 Africa Guinea 45 6.77 38.2 29.7 2.8 26.91 37.8 Africa Guinea-Bissau 43.2 7.28 35.93 28.1 3.1 25.08 35.5 Africa Kenya 53.2 4.84 48.36 38 2 36.03 47.7 Africa Lesotho 56.7 2.94 53.78 39.7 1 38.66 52.9 Africa Madagascar 49.3 6.14 43.18 33.9 2.5 31.45 42.3 Africa Malawi 49.8 6.16 43.6 35.1 2.5 32.57 43.1 Africa Mali 44.7 6.87 37.78 28.9 2.9 26 37.3 Africa Mauritania 46.9 6.17 40.72 31.5 2.5 28.95 40.3 Africa Morocco 47.8 6.5 41.3 32.8 2.7 30.04 40.8 Africa Mozambique 52.4 5.33 47.05 38 2 35.94 46.4 Africa Namibia 67.4 3.26 64.13 53.3 1.4 51.83 62.6 Africa Niger 45.2 7.38 37.77 30.6 3.1 27.47 37.3 Africa Nigeria 47.5 5.51 41.99 31.4 2.2 29.2 41.5 Africa Senegal 47.1 6.28 40.77 31.6 2.6 28.97 40.3 Africa Seychelles 69.6 3.71 65.92 60.2 1.6 58.52 65.8 Africa South Africa 68.3 2.67 65.58 52 1.1 50.84 63.3 Africa Swaziland 57.9 4.35 53.54 42.2 1.9 40.37 52.4 Africa Tanzania 44.3 7.17 37.11 29.2 3 26.2 36.7 Africa Tunisia 44.9 6.39 38.54 29.3 2.6 26.7 38.1 Africa Uganda 50.8 5.85 44.96 35.9 2.4 33.44 44.3 Africa Zambia 56.4 4.24 52.15 40.8 1.7 39.11 51.2 Asia Bangladesh 42.2 8.78 33.37 27.8 4 23.79 32.9 Asia Cambodia 43.9 8.05 35.81 29.1 3.6 25.54 35.3 Asia PRC 47.9 4.98 42.91 31.2 1.9 29.29 41.4 Asia India 42.6 8.59 34 28.5 3.7 24.81 33.6 Asia Indonesia 42.2 8.4 33.77 27.5 3.7 23.81 34.3 Asia Iran, Islamic Rep. 45.2 6.43 38.73 29.6 2.6 27.01 38.3 Asia Israel 46.3 4.99 41.29 29.9 1.8 28.09 41.3 continued on next page 29 ADBI Working Paper 763 Roy and Roy Table A2 continued 2000s Income Share Income Share Continent Country Highest 20p% Lowest 20% Difference between Quintiles Highest 10p% Lowest 10% Difference between Deciles Gini Asia Jordan 43.1 7.87 35.2 28.2 3.4 24.79 34.8 Asia Kazakhstan 39.1 8.65 30.41 24.2 3.6 20.54 30.3 Asia Kyrgyz Republic 41.9 7.75 34.14 26.4 3.2 23.22 33.8 Asia Lao PDR 43.2 8.06 35.15 28.7 3.5 25.16 34.7 Asia Malaysia 49.2 5.23 43.98 32.7 2.1 30.62 43.4 Asia Maldives 44.2 6.51 37.73 28 2.7 25.32 37.4 Asia Mongolia 42.3 7.28 34.99 26.6 3.1 23.53 34.7 Asia Nepal 46.2 7.4 38.82 31.6 3.3 28.34 38.3 Asia Pakistan 40.5 9.35 31.12 26.6 4.2 22.42 30.8 Asia Philippines 50.6 5.66 44.89 34.3 2.4 31.91 44.1 Asia Russian Federation 45.5 6.51 38.97 29.5 2.6 26.93 38.4 Asia Slovak Republic 37.1 9.25 27.84 23 3.8 19.24 27.6 Asia Sri Lanka 46.9 7.14 39.8 32.2 3.1 29.07 39.2 Asia Tajikistan 40.4 7.87 32.57 25.4 3.1 22.25 32.3 Asia Thailand 48.5 6.4 42.08 32.7 2.7 29.99 41.4 Asia Turkey 46.3 5.66 40.59 30.2 2.1 28.09 40.1 Asia Uzbekistan 42.7 7.79 34.86 27.8 3.1 24.74 34.2 Asia Viet Nam 44.3 7.13 37.21 29 3 25.98 36.8 Asia Yemen, Rep. 44.2 7.84 36.31 29.9 3.3 26.61 35.9 Europe Armenia 41.6 8.49 33.07 27.3 3.6 23.69 32.7 Europe Austria 38.1 8.51 29.61 23.4 3.3 20.03 29.5 Europe Azerbaijan 34.5 11.2 23.29 21.1 5 16.08 23.1 Europe Belarus 36.9 8.94 27.99 22.3 3.7 18.66 27.9 Europe Belgium 41.7 8.35 33.34 28.3 3.3 25.03 33.1 Europe Bulgaria 40 7.23 32.74 25 2.6 22.36 32.4 Europe Croatia 39.9 8.36 31.54 25.1 3.5 21.59 31.2 Europe Czech Republic 36.4 9.45 26.99 22.6 3.8 18.81 26.5 Europe Denmark 35.1 9.68 25.45 21 3.7 17.3 25.4 Europe Estonia 41.3 7.27 34.01 25.9 2.7 23.26 33.6 Europe Finland 37.3 9.32 27.99 23 3.8 19.18 27.9 Europe France 39.6 7.94 31.69 24.6 3.2 21.4 31.5 Europe Georgia 46.5 5.46 41.02 30.5 1.9 28.51 40.6 Europe Germany 39.4 8.38 31.05 24.7 3.4 21.34 30.9 Europe Greece 41.1 6.7 34.38 25.8 2.3 23.51 34.2 Europe Hungary 37.4 8.78 28.66 23 3.6 19.4 28.5 Europe Ireland 40.7 7.76 32.93 25.7 3.1 22.65 32.7 Europe Italy 42.4 6.2 36.2 27.2 2.1 25.03 36.1 continued on next page 30 ADBI Working Paper 763 Roy and Roy Table A2 continued 2000s Income Share Income Share Continent Country Highest 20p% Lowest 20% Difference between Quintiles Highest 10p% Lowest 10% Difference between Deciles Gini Europe Latvia 42.5 6.66 35.85 27 2.4 24.61 35.5 Europe Lithuania 41.5 7.14 34.32 26.2 2.7 23.5 34 Europe Macedonia, FYR 45.9 5.88 40 30 2.3 27.7 39.6 Europe Moldova 42.2 7.43 34.78 27 3 23.97 34.5 Europe Netherlands 38.5 8.2 30.3 23.9 3 20.91 30.1 Europe Norway 37.1 9.2 27.87 23.1 3.5 19.58 27.8 Europe Poland 41.6 7.7 33.89 26.4 3.2 23.27 33.7 Europe Romania 37.9 8.48 29.37 23 3.5 19.56 29.3 Europe Slovenia 36.1 9.3 26.82 21.8 3.8 17.97 26.7 Europe Spain 40.8 6.48 34.29 25.2 2.1 23.08 34.1 Europe Sweden 36.2 9.32 26.84 21.8 3.7 18.07 26.8 Europe Switzerland 40.3 7.67 32.65 24.8 2.9 21.89 32.7 Europe Ukraine 37.4 9.16 28.27 22.8 3.9 18.97 28.1 Europe United Kingdom 44.2 5.98 38.17 28.7 2 26.7 37.9 North America Canada 41 7.02 33.99 25.8 2.6 23.17 33.8 North America Costa Rica 54.3 3.92 50.35 37.7 1.3 36.44 49.3 North America Ivory Coast 48.6 5.69 42.89 33 2.3 30.73 42.3 North America Dominican Republic 54.9 4.24 50.68 39.1 1.6 37.54 49.6 North America El Salvador 52.4 4.11 48.32 36.1 1.4 34.73 47.5 North America Guatemala 58.3 3.13 55.17 42.3 1 41.25 54 North America Honduras 60.1 2.52 57.61 43.7 0.8 42.94 56.5 North America Jamaica 58.5 3.39 55.12 41.9 1.4 40.43 54.3 North America Mexico 54.1 4.49 49.58 38.8 1.7 37.03 48.8 North America Nicaragua 49.2 5.5 43.69 33.5 2.2 31.35 43.1 North America Panama 58 2.83 55.22 41.3 0.9 40.44 54 North America United States 46.2 4.95 41.2 30 1.5 28.53 40.9 Oceania Australia 41.1 6.99 34.15 25.2 2.4 22.89 34.1 South America Argentina 53.1 3.51 49.57 35.9 1.1 34.85 48.9 South America Bolivia 58.3 2.41 55.87 41.9 0.6 41.29 54.7 South America Brazil 60.3 2.93 57.32 44.3 0.9 43.4 55.9 South America Chile 58.6 4.12 54.44 43.4 1.5 41.86 52.9 South America Colombia 60.7 2.93 57.74 45.1 0.9 44.26 56.3 South America Ecuador 56.4 3.61 52.83 40.6 1.1 39.49 51.7 South America Paraguay 57 3.53 53.47 41.6 1.2 40.37 52.6 South America Peru 53.6 3.93 49.68 37.5 1.4 36.12 49 South America Uruguay 51 4.68 46.34 34.4 1.8 32.61 45.7 South America Venezuela, RB 51.9 3.48 48.4 35.2 0.9 34.29 47.7 31 ADBI Working Paper 763 Roy and Roy Table A3: Correlation Coefficients among Explanatory Variables ln PCGDP lnManufacturing_Share lnServices_Share ln TO ln FDI lnInfrastructure_Quantity lnInfrastructure_Quality ln Urbanization ln PCGDP 1 lnManufacturing_Share 0.11 1 lnServices_Share 0.52 0.27 1 ln TO 0.19 0.01 0.06 1 ln FDI 0.62 0.10 0.40 0.00 1 lnInfrastructure_Quantity –0.01 0.03 –0.06 –0.19 0.18 1 lnInfrastructure_Quality –0.55 –0.24 –0.35 –0.06 –0.36 0.06 1 ln Urbanization 0.74 0.14 0.35 0.12 0.47 0.06 –0.32 1 Table A4: Correlation Coefficients among Interaction Dummies D_AfricaXlnManufacturing _Share D_AsiaXlnManufacturing _Share D_EuropeXlnManufacturing _Share D_NorthAmericaXlnManufacturing _Share D_SouthAmericaXlnManufacturing _Share D_PacificXlnManufacturing _Share D_AfricaXlnManufacturing_Share 1 D_AsiaXlnManufacturing_Share –0.32 1 D_EuropeXlnManufacturing_Share –0.32 –0.30 1 D_NorthAmericaXlnManufacturing_Share –0.20 –0.18 –0.19 1 D_SouthAmericaXlnManufacturing_Share –0.16 –0.15 –0.15 –0.09 1 D_PacificXlnManufacturing_Share –0.12 –0.11 –0.12 –0.07 –0.06 1 D_AfricaXlnServices_Share 0.95 –0.33 –0.34 –0.21 –0.17 –0.13 D_AsiaXlnServices_Share –0.32 0.96 –0.31 –0.19 –0.15 –0.12 D_EuropeXlnServices_Share –0.33 –0.31 0.96 –0.19 –0.15 –0.12 D_NorthAmericaXlnServices_Share –0.21 –0.19 –0.20 0.94 –0.10 –0.07 D_SouthAmericaXlnServices_Share –0.16 –0.15 –0.15 –0.09 0.99 –0.06 D_PacificXlnServices_Share –0.15 –0.14 –0.14 –0.09 –0.07 0.82 continued on next page 32 ADBI Working Paper 763 Roy and Roy Table A4 continued D_AfricaXlnServices _Share D_AsiaXlnServices _Share D_EuropeXlnServices _Share D_NorthAmericaXlnServices _Share D_SouthAmericaXlnServices _Share D_PacificXlnServices _Share D_AfricaXlnManufacturing_Share D_AsiaXlnManufacturing_Share D_EuropeXlnManufacturing_Share D_NorthAmericaXlnManufacturing_Share D_SouthAmericaXlnManufacturing_Share D_PacificXlnManufacturing_Share D_AfricaXlnServices_Share 1 D_AsiaXlnServices_Share –0.34 1 D_EuropeXlnServices_Share –0.35 –0.32 1 D_NorthAmericaXlnServices_Share –0.22 –0.20 –0.20 1 D_SouthAmericaXlnServices_Share –0.17 –0.15 –0.16 –0.10 1 D_PacificXlnServices_Share –0.16 –0.14 –0.14 –0.09 –0.07 1 33