1
University of Amsterdam
Amsterdam School of Economics Economics – Monetary Policy & Banking
Master Thesis
Empirical Research on the Impact of Foreign Direct Investment on Urban-rural Income Inequality: A Panel Analysis of China’s Provinces.
Student name: Yuqi Sun Student number: 11252162
E-mail address: yuqi.sun@student.uva.nl Supervisor: Alex J. Clymo
Date: 11/7/2017 Word count: 7,627
2 Abstract
China is the world’s biggest recipient of foreign direct investment (in overall volume) since 2002. On one hand, FDI brings China new technology and management method. On the other, it also brings about some problems. This paper is going to examine the effects of FDI on China’s urban-rural income inequality. After setting up an empirical regression model, I use the panel data of China’s 31 provinces from 2002 to 2012 to solve the research question. The results of empirical model indicate that FDI
narrowed the urban-rural income gap in China. What’s more, the impact of FDI on income inequality depends on the development level of a province.
3 Table of Contents 0. List of definitions. ... 5 1. Introduction ... 6 1.1 Research Question... 7 1.2 Composition ... 7 2. Theory Review ... 8 2.1 Modernization Theory ... 8 2.2 World-systems Theory ... 9 3. Literature Review ... 10
3.1 FDI decreases income inequality ... 10
3.2 FDI increases income inequality ... 11
3.3 FDI’s impact on income inequality varies... 13
4. Data and Methodology ... 15
4.1 Source Data ... 16 4.2 Methodology ... 17 Time / Frequency ... 17 Provinces ... 17 Models ... 18 4.3 Variables ... 19
Dependent variable – urban-rural income gap ... 21
Independent variable – FDI inflows ... 21
Independent control variable – GDP per capita ... 22
Independent control variable –investment into fixed assets ... 22
Independent control variable –trade ... 22
Independent control variable –education ... 23
Independent control variable –urban population ... 23
Independent control variable –transportation infrastructure ... 24
Independent control variable –CPI ... 24
Independent control variable –provincial dummy and time dummy ... 24
5. Results and Discussion ... 25
5.1 Estimation Results... 25 5.2 Robustness Checks ... 30 5.3 Region-related Analysis ... 32 6. Conclusion ... 33 6.1 Conclusion Remarks ... 33 6.2 Limitation of Paper ... 34 6.3 Future Agenda ... 35 7. Reference: ... 37
4 List of tables and figures:
1 Figures...
Fig.1 Kuznets Curve. ... 9
Fig.2 Map of China. ... 18
Fig.3 Urban-rural income gap and FDI inflow as percentage of GDP (national). ... 27
2 Tables ... Table 1. Explanation of Variables... 21
Table 2. Summary statistics. ... 22
Table 3. Effect of FDI on urban-rural income gap (Baseline OLS - national). ... 27
Table 4. Effect of FDI on urban-rural income gap (model 1 - national). ... 28
Table 5. Effect of FDI on urban-rural income gap (model 2 - national). ... 30
Table 6. Robustness check- Effect of FDI on urban-rural income gap (national) ... 32
5 0. List of definitions.
1. FDI (Foreign direct investment): refers to the investment, including capital assets or other production factors, made by foreign companies or individuals.
2. Inward FDI/ FDI inflows: refers to the (net) value of FDI flows into the receive country.
3. Host country: refers to the country which receives the FDI.
4. Gini coefficient: refers to a ratio measuring the income distribution, higher the Gini coefficient, bigger the income gap.
5. Absorptive capacity: refers to the ability of maximizing the utilization of FDI in this paper.
6. Total value of imports and exports of operating units: refers to value of imports and exports carried out by companies which registered by the local Customs house and are vested with right to run import export business1.
7. Total value of imports and exports of destinations and catchments: refers to the value of export goods of the places of their origin or the places of the goods dispatched2.
8. Two way FE/RE model: refers to the fixed effects/random effects estimation
model with time fixed effect.
1
http://data.stats.gov.cn/english/
2
6 1. Introduction
With the rapid development of economic globalization, international capital flows have increased a lot. And as the main component of international capital flows, foreign direct investment is playing more and more crucial role in the world economy. Since China’s “open door policy” started in 1978, China also joined the session of economic globalization and attracted huge amount of foreign direct investment. The data of value of FDI actually utilized in China is available from 19833, which is only $0.92 billion at that year, grew up very quickly and reached $52.743 billion in 2002 after China joining the WTO, and China became the largest host country of FDI all over the world. Then, the FDI inflow4 in China continued growing after that and until 2015, the FDI actually utilized was as much as $126.3 billion.
The huge amount of FDI has great impact on China’s economic growth. It brings new technology and management method to China, and also helps a lot in capital accumulation, structure change, industrial restricting, technology upgrading, etc. At the same time, the foreign invested firms provide great number of job spaces, which reliefs the employment pressure in China.
However, FDI also brings some problems, as the distribution of FDI inflows is extreme uneven. For one thing, FDI mainly targets at China’s eastern (coastal) provinces, which have access to cheap shipment, rather than the central and western (inland) provinces. It indeed promotes the economic growth in eastern provinces, but at the same time, exacerbates the inter-regional income gap in China. For the other, FDI only concentrates in the manufacturing sector but shows no interest in agriculture sector. With the support of FDI, wage in manufacturing sector grows rapidly, while without FDI, wage in agriculture sector increases relatively very slow. Therefore, the income gap between manufacturing workers and agricultural workers is bigger and bigger. In China, most of manufacturing workers live in urban area, while the
3
Data available at the website of National Bureau of Statistics of China.
4
7
agricultural workers live in rural area, so the income gap between these two kinds of workers becomes the urban-rural income inequality. The rising of urban-rural gap reduces the enthusiasm of rural agricultural workers to increase production, and also drives them to move to cities due to higher wages. With too many labor forces move to manufacturing sector, there is huge area of idle lands in rural area, which is a severe waste of land resource. The increasing urban-rural income gap can impede the further economic growth and social development, so narrowing the urban-rural income inequality becomes one of the top issues needed to be solved. As FDI is one of the main factors contributed to the inequality, investigating the impact of FDI on urban-rural income gap also becomes one of the current concerns.
1.1 Research Question
So the intent of this paper is to analyze the relationship between FDI and urban-rural income inequality in China, my research question is
how foreign direct investment affects the urban-rural income gap in China’?
In this paper, to investigate the relationship between FDI and urban-rural income inequality in China’s provinces, I will use a panel data covers 31 provinces, spanning over the period 2002 through 2012.
1.2 Composition
The rest of paper is organized as follows. Section 2 is a theory review of how FDI affecting a country’s income inequality. Section 3 is a review of existing literatures on the impact of FDI on income inequality. Section 4 is the empirical model analysis, including source data, data description, and set-up of regression model. Section 5 is the results interpretation and discussion. Section 6 is the conclusions, limitation of this paper, and future research agenda.
8 2. Theory Review
Here I use mainly two theories to analyze how FDI affecting a country’s income inequality: 1) modernization theory; 2) world-systems theory.
2.1 Modernization Theory
The core of modernization theory is the Kuznets “inverted-U” Curve (Kuznets, 1955), which points out that as a country develops, its economic inequality first increases and reaches a peak, then decreases when the GDP per capita is relatively high (see Fig.1).
Fig 1. Kuznets Curve.
To be specific, Kuznets divided the development of an economy into two stages, primary rural society and industrialized urban society. He hypothesized that at the early stage of economic development, people with some capital have more chances to invest and make even more money, while the poor don’t. Therefore, the income gap
9
between relatively rich and poor people widens. Moreover, the non-agricultural sectors in cities can offer higher wages than the agricultural sector in rural area, so rural labors are attracted and move to cities. But the influx of rural labors makes supply exceed the demand in labor market, thus wages in cities are kept down and income gap between working class and capital owners is widening. So overall, at the early years of development of primary rural society, the income inequality increases.
However, Kuznets believes that with time going by, the country becomes more industrialized and urbanized, when it reaches a certain level of income per capita, it will benefit from trickle-down effect and the income inequality will decrease.
According to this theory, China is still a middle-income developing country, so with the economic growth, the income gap in China will be bigger and bigger in recent years, as the FDI firstly affects the industries which absorb more of it. But many years later if China becomes a developed country, FDI will have deeper effect on the whole economy, and help to narrow the income gap.
2.2 World-systems Theory
World system theory, which was first put forward by Wallerstein in 1970s, refers that there exists a world economic system in which countries are divided into two categories, some of them benefit while others are exploited. The dominant capitalist countries are called core countries, and they exploit the peripheral countries, which have undeveloped industries, for labor and raw materials.
According to Girling (1973), FDI inflows will increase the income inequality of the host countries. It is because that a country’s income distribution is determined by the relative position of this country in world economic system, but is not relevant to GDP or other factors. Core countries have more severe inequality problem than the peripheral and semi-peripheral countries. That is to say, the more advanced a country is, the more FDI it uses, the bigger income gap in this country.
Girling (1973) explained the mechanics by analyzing the differences in wages between elite and ordinary workers. With more and more FDI inflows to the host
10
countries, there forms a new class of elite in the multinational firms in recipients. Compared to the ordinary workers who work for other domestic industries, these elites have as much as 3 to 9 times more wages. At the same time, the wage in traditional sectors also goes up due to the development of the industry. The higher wage in traditional sectors makes the employers decrease the labor demand, thus the unemployment in traditional sectors rises. In all, the income disparities are exacerbated. What’s worse, these elite will go along with the foreign investors to edge local firms out of the market to safeguard their own interest, which continues to worse the income distribution.
China now can be considered as peripheral or semi-peripheral country in the world economic system, and recent decades have witnessed its rapid economic growth. There also occurs an elite class who try to harm the market competition, and the income gap between ordinary workers and them is widening now. Most of foreign invested firms or multinational firms in China are located in cities, and most of FDI concentrates at urban area. So from this respective, urban-rural income gap in China will increase with the inflowing FDI.
To sum up, modernization theory and world-systems theory have different predictions, which are also shown in empirical papers. In next part, I’m going to summarize these papers.
3. Literature Review
Till now, many economists have empirically studied the impact of FDI on income inequality, due to the differences in theoretical models, empirical models, and data selections, their results and conclusions are not the same and even contradict.
3.1 FDI decreases income inequality
Only small portion of economists support the view that FDI inflows can narrow the income gap and improve inequality. Milanovic (2003) used the approach of
11
household surveys to find that at low average income level, rich people benefit a lot from the openness. But when the income increased to around $5-7,000 per capita at international prices, the situation changed and it is mainly the lower and middle income level households who benefits from openness. So in the long run, FDI makes the income distribution better.
3.2 FDI increases income inequality
There are plenty of researchers who in favor of the view FDI inflows exacerbate the income inequality. Most of them use a panel data, but some of them employ a state/province-level panel of a sore country, while others use a cross-country dataset.
To begin with, I will summarize some previous literature using state/province-level panel data. Some of economists investigate the relationship between FDI and income inequality in developed countries. For example, Chase-Dunn (1975) used a panel analysis of America’s 1950 and 1970 data to test the relationship between FDI inflows and income inequality (in Gini Coefficient). The estimation results reported a positive relationship, that is to say, FDI increased income inequality in America at the selected two years. Chintrakarn, Herzer, and Nunnenkamp (2011) also used a state-level panel data of U.S and applied panel cointegration techniques, which allow for cross-sectional heterogeneity and cross-sectional dependence, they found that in the long run, FDI has a positive relationship with income inequality in some states (21 out of 48 cases). Taylor and Driffield (2005) used UK panel data and controlled technology and trade, which are two most common explanations of income inequality, to investigate if FDI has contributed to the rising wage inequality in UK. And they found inward flows of FDI
5
have significant effect on wage inequality in UK and even contributed 11% of the increased inequality.
Also, there are some papers investigate developing countries using
5
12
state/province-level data. For instance, Feenstra and Hanson (1997) used a penal data of Mexico from 1975 to 1988 to study the effect of FDI on the wages of skilled workers. And their regression results showed that FDI inflows to Mexico increased the wages of skilled workers a lot but didn’t have that much impact on unskilled workers’ wages. Therefore, FDI to Mexico worsened the income distribution there. Using the data of Bolivia, Nunnenkamp, Schweickert, and Wiebelt (2006) carried out a computable general equilibrium (CGE) analysis of the country. The simulation results they got indicates the conclusion that even though FDI inflow enhances Bolivia’s economic growth and reduces poverty, it makes the income distribution very unequal, especially widens the income disparities between urban and rural areas. Sumie and Fukushige (2007) used the 1975-1995 South Korea data to test the effect of economic openness of South Korea on its income inequality. Their methodology is non-linear regression model with some economic factors as control variables. Their research shows that FDI increases income inequality both in the short and long run, while opening of good markets reduced inequality.
Moreover, there is a published paper dealing with FDI’s impact on regional inequality in China. Wei, Yao and Liu (2009) employed a Cobb-Douglas production function with a large panel dataset covering all China’s provinces from 1979 to 2003. Based on the results of model, they pointed out that FDI is an important factor contributed to regional inequality in China from 1979, but it is the uneven distribution of FDI rather than FDI itself that caused the inequality. My research question may seem similar to that of Wei et al. (2009), but actually I will concentrate on income inequality between cities and rural area, while Wei et al. (2009) investigated the income disparities among different provinces/regions in China. And the models we use are different, too.
Furthermore, there are also some existing literature use cross-country dataset to test FDI’s effect on income disparities. For example, Mahutga and Bandelj (2008) used a panel data of Central and Eastern European countries, and set up a series of fixed effects regression models relate income inequality to FDI. And their estimation
13
results show that FDI has positive effects on income disparities in some Central and Eastern Europe countries in short term.
Moreover, there are also articles using an even broader world-wide data. Choi (2006) based his research on the World Bank’s World Development Indicators and used Gini coefficient6 of 119 countries from 1993 to 2002 in the database. Then, he ran a linear regression model with time dummy and regional dummies of Asia and Latin America, and found that Gini coefficient has a positive relationship with FDI-GDP ratio. What’s more, Latin American ones show a less equal income distribution among all the 119 countries. Based on the panel data of 119 developing countries, Basu and Guariglia (2007) set up a growth model of a dual economy in which the agricultural sector has diminishing returns, while the industrial sector has a rapid growth and becomes the core of whole economy with the help of FDI. The model predicts that FDI contributes to both inequality and growth, and at the same time reduced the share of agricultural sector to total GDP in these developing countries. Castro (2011) used the same panel dataset of Galbraith and Kum (2003)7 and applied their suggestion of using a new variable called “pay inequality” to measure income inequality instead of Gini coefficient. The paper concluded that FDI is but not the only factor increasing the income inequality; education and geography also affect inequality.
3.3 FDI’s impact on income inequality varies
There are also some researches showing that FDI’s impact is different among countries/ regions.
Some of them use a dataset of selected typical countries. For example, Tsai (1995) ran a linear regression model with a panel data of less-developed countries, and the results indicated that FDI has negative impact on income inequality in some
6Gini coefficient is a ratio measuring the income distribution, higher the Gini coefficient, bigger the income gap.
7
Their dataset is based on Deininger and Squire (1996), which includes a “high-quality” panel of income and expenditure inequality in 693 countries starting from 1947.
14
developing countries in Asia, but has no significant effect on Latin American countries. Avalos and Savvides (2003) also tested the relationship between economic openness and income inequality using the panel data of countries in Latin America and East Asia during the last thirty years. They set up a labor supply and demand model and the findings show that overall the trade openness is negatively related to wage gaps, but the wage gaps in two regions responded differently to various determinants. Velde and Morrissey (2004) used panel data of five East Asian countries, which includes Korea, Singapore, Hong Kong, Philippines, and Thailad, and found that overall there is no evidence of FDI reducing wage inequality. But after controlling the domestic influences, they found FDI has increased the wage gap in Thailand, which is because that the education system in Thailand cannot offer sufficient skilled human resources to maximize the benefits from FDI.
There are also other researchers using a wider range of observations. For instance, using a panel of more than 100 countries over 1980 to 2002, Figini and Gorg (2011) did a non-linear regression for OECD and non-OECD countries respectively, and found two different patterns: for non-OECD countries, income inequality was correlated positively with FDI inward stock, while for OECD countries it was negatively correlated with FDI inward stock. Wu and Hsu (2012) tested the correlation of FDI and income inequality and asked whether this relationship depends on a country’s absorptive capacity, which refers to the ability of maximizing the utilization of FDI. They used a cross-section data of 54 countries from 1980 to 2005 and adopted an endogenous threshold regression model. Finally, their research shows that for the host country (refers to country who receives FDI) with lower absorptive capacity, FDI increases income inequality; for the host country with higher absorptive capacity, FDI has little impact on inequality.
Lessmann (2013) used a panel data of 55 countries at different development stages and calculated three different indexes of income inequality, the coefficient of variation (CV), the adjusted Gini coefficient (GINI), and the population-weighted coefficient of variation (WCV), as dependent variables of his regression. Then, he
15
selected some of the economic factors as control variables, such as GDP per capita, unemployment rate, trade to GDP ratio, urbanization ratio, log of population, agricultural employment, etc. He also included country/ region fixed effects and time fixed effects in the regressions. The results of his paper indicate that FDI inflows increase income gap in low and middle income countries, while there is no significant effect on high income countries. I found Lessmann (2013)’s methodology is straightforward and very suitable for my research question as well, so it becomes the main reference of my regression model of this paper.
However, some researchers find that overall the FDI inward stock has no impact on inequality. Mah (2003) use 20-year panel data of Korea and the estimation results of her non-linear regression model show that there is weak evidence of the Kuznets hypothesis. During the period 1975-1995, economic globalization does not affect income distribution in Korea. According to Sylwester (2005), a panel of less developed countries over 1970-1989 was used, and the paper concludes that FDI even has no impact at all on income gap in the less developed countries. Bhandari (2007) used the data of transitional countries in Eastern Europe and Central Asia (1990-2002) to see if FDI could affect income inequality. By using fixed effects, she found that overall the FDI inflows don’t affect the income inequality, but if breaking the effect into its components, the results suggested that FDI inflows increased wage income inequality, while decrease capital income inequality.
4. Data and Methodology
In this section, I’m going to first present and discuss the data source. Then, I will describe the methodology of empirical analysis, including time, frequency, provinces, and format of models. At least, discuss specifically each variable I’m going to use in the empirical model.
16 4.1 Source Data
Almost all the data for this paper are downloaded from the website of National Bureau of Statistics of the People’s Republic of China (NBSC). NBSC provides very broad range of economic and demographic data in China for the public, and the data provided also covers different levels, from national data to provincial data. The national data usually available from 1970s, but the provincial ones have a much shorter observations and most of them are only available from 2000s. In this paper, in order to set up a provincial panel, I use the annually observed provincial data from this website.
There is a long-time debate on the reliability and consistency of the NBSC database. Firstly, some economists suspect the data falsification in China. According to Cai (2000), in order to make outstanding achievement for their promotion, some officials in local government can report higher growth rates. But some efficient reforms have been done to make sure the data collection is under supervision. Secondly, there are also some economists (Keidel, 2001) suspecting that NBSC itself may purposely increase the output data in order to meet the economic growth target. But actually the officials in NBSC are obligated by laws to present the public reliable data (Chow, 2006), and Keidel’s (2001) method was based on a biased dataset and even proved not correct by Holz (2003). Thirdly, there are also some papers (Chen, Jefferson, Rawski, Wang, & Zheng, 1988; Koch-Weser, 2013) thought that the NBSC data lacks adjustment for inflation, which was neglected at the 70s when NBSC first made the data available to public. But NBSC has recognized and actually already adjusted for inflation in recent years. And the data I’m going to use in regression covers 2002 to 2012, which is relatively new and already adjusted for inflation. So it is not a problem for my dataset. The development of statistics is also a part of modernization in China, and the government has already and is still making effort to improve the quality and quantity of statistics. As a result, the data now are more reliable and consistent. Therefore, though the NBSC data has its shortcomings, it is
17
still the most used and the best data source on China. So in my paper, I use the data provided by NBSC to run my regression models.
4.2 Methodology
Time / Frequency
Most of provincial data in NBSC dataset are available until 2012 annually. And the most important variable, urban-rural income gap is only available from 2002 to 2012 annually, the time period in my panel data will only cover the 11 years.
Provinces
Fig 2. Map of China8.
There are total 34 provinces in China, but the data of Taiwan is not available due to some political reasons, and Hong Kong and Macao are relatively very small and usually not be considered when doing this kind of economic research. So in this paper,
8
18
I only focus on the 31 provinces and in order to make comparison. Furthermore, I use a very common method to divide them into 3 regions, eastern, central, and western.
Fig.2 shows a map of China and how people usually divide the eastern, central, and western regions. The reason behind it is that there are huge economic as well as demographic differences among the three regions, especially in terms of FDI and international trade. Actually, around 90% of the total FDI in recent years is located in eastern provinces rather than the central and western ones, so I think it is also essential to run the regression not only based on the whole country, but also on the three regions separately.
Models
To exam the effect of FDI on income disparity, I’m going to use a regression approach, of which the models are
based on Yang (2014). is the urban-rural income gap for province i at time t, is the FDI total net inflow in province i and is the idiosyncratic disturbance. I will describe other control variables in next part.
However, according to Lessmann (2013), the effect of FDI on income inequality can also depends on the development level of a country. Therefore, I found it was also interesting to see if the interaction between FDI inflows and GDP per capita also has influences on urban-rural income disparities in China. So I generated a new variable , which is the multiple of FDI inflows and income per capita, and added
it into the model 2.
19
Then, I’m going to discuss what kind of estimation method will be used in empirical models. First method I will use is the ordinary least squares (OLS) with time fixed effect, which can provide the preliminary study of the effects. Then I will apply fixed effects (FE) model to have better estimations. Moreover, I will use two-way FE and RE (random effects) with not only provincial fixed effect but also time fixed effect as well. To determine which of them perform better, I will use a Hausman test. According to Hausman (1978), the null hypothesis of the test is the difference in coefficients ( ) is not systematic. If the difference is large enough, I can reject the null and use the FE model. Otherwise, RE model performs better than the FE.
For national data (panel of 31 provinces), both of models will be applied. But for regional data, in which I divide the whole national data into eastern, central, and western regions, I will only use model 2 with RE method. It is because model 2 is the main model of my paper, while model 1 is just for comparison. For regional regressions, I will drop the regional dummies. The regional dummies I use in the regression of whole country can only capture the average differences in FDI inflows, but keep the other coefficients the same, but by running three regional regressions respectively, the other coefficients vary as well. So even though I have regional dummies in the national regressions, it is still essential for me to run three regional regressions.
What’s more, in order to make sure that the regression results are robust and valid under different conditions, I will conduct several robustness checks after running a regression of model 2 with national data. In model 3, 4, and 5, I will replace with , , and respectively. In model 6, I will replace the variable with another variable .
4.3 Variables
Table 1 shows the simple explanation of variables, and Table 2 shows the observation number, mean, standard deviation, minimum, and maximum of each
20
variable. In the following of this section, I’m going to explain them in detail.
Table 1. Explanation of Variables.
Name of Variables Meaning
Logarithm of urban-rural income gap, which is urban
income per capita divided by rural income per capita.
Logarithm of net inflow of foreign investment in yuan in
province i.
Logarithm of Gross Domestic Product (GDP) per capita.
Logarithm of total investment in fixed assets in province i.
Logarithm of total import and export volume (by the
location of operating units) in yuan.
Logarithm of total import and export volume (by
destinations and catchments) in yuan.
Logarithm of total export volume (by the location of
operating units) in yuan.
Logarithm of total export volume (by destinations and catchments) in yuan.
Logarithm of average years of schooling.
Logarithm of total educational expenditure in yuan.
Logarithm of urban population.
Logarithm of total length of transportation (in 1000 km). Logarithm of Consumer Price Index (preceding year =
100).
Dummy variable, equals to 1 if province i located in eastern region.
Dummy variable, equals to 1 if province i located in central region.
Dummy variable, equals to 1 if province i located in western region.
Time dummy.
.
Table 2. Summary statistics.
Variable Obs Mean Std. Dev Min Max
21 341 21.887 2.272908 17.12592 26.14261 341 9.84514 0.6954927 8.088562 11.44221 341 28.77564 1.136776 25.39216 31.07323 341 25.4361 1.803404 20.79937 29.45745 341 25.47152 1.773294 20.75898 29.5827 341 24.81033 1.769476 20.32492 28.91852 341 24.78554 1.771402 20.15085 29.02135 341 2.092126 0.1558346 1.318662 2.471173 341 23.90303 0.9448595 20.75326 25.96217 341 16.47879 0.9402877 13.17927 18.08386 341 4.491142 0.8279911 2.272229 5.767655 341 4.631902 0.0222331 4.581902 4.701389 341 0.3870968 0.4878019 0 1 341 0.2580645 0.4382127 0 1 341 0.3548387 0.4791675 0 1 341 2007 3.166925 2002 2012 341 216.4932 34.52705 149.7798 291.0288
Dependent variable – urban-rural income gap
(measured in
yuan) is the dependent variable of this model, which refers to urban-rural income gap.
Both the nominator and denominator can be downloaded directly from the NBSC website.
Independent variable – FDI inflows
is total foreign direct investment net inflows (measured in yuan) of
22
exchange rate between USD and CNY in that year to calculated the FDI in yuan. And then take logs of it.
Independent control variable – GDP per capita
is logarithm of Gross Domestic Product (GDP) per capita measured
in yuan, which is also a very common measure used for investigating income gap among different regions.
Independent control variable –investment into fixed assets
is total government investment into fixed assets (measured in yuan)
of province i in year t. According to previous research, the level of investment into fixed assets also affects the income inequality. So I also include this variable into the empirical model. Ideally, the total domestic investment is one of the best control variables for this model, but the data of it is unfortunately unavailable. So I use the investment into fixed assets instead.
Independent control variable –trade
is the total import and export volume (in yuan, measured by the
location of operating units9). Trade is an important part of economic globalization, which should have influence on China’s income disparity (Rodriguez-Pose & Gill, 2006). And the same with FDI, trade also varies a lot across the country. So I include trade as one of the control variables as well.
Actually, NBSC offers another way of calculating it, which is the provincial total value of imports and exports of destinations and catchments10. So in the robustness check, I will replace the variable Trade with Trade2 (measured by destinations and
9
Total value of imports and exports of operating units refers to value of imports and exports carried out by companies which registered by the local Customs house and are vested with right to run import export business.
10
Total value of imports and exports of destinations and catchments refers to the value of export goods of the places of their origin or the places of the goods dispatched.
23
catchments), and use the same method as I did in the main model. What’s more, according to OECD (2005), when measuring globalization, we should only apply exports instead of total volume of exports and imports. So I also use the total volume of exports to check the robustness. is the exports calculated by location of operating units, and is the exports calculated by destinations and catchments.
Independent control variable –education
stands for average years of schooling in province i at year t, which
is not available from NBSC database. According to UNESCO (2013), , in which HS stands for share of population attained the highest level of education l, and YS stands for the schooling years to attain the diploma of education l. From China Statistical Yearbook published by NBSC, I found data of total population aged 6 years old and above in each province, and they are categorized by the highest level of diploma attained: no schooling (0 years of schooling), primary school (6 years of schooling), junior secondary school (9 years of schooling), senior secondary school (12 years of schooling), and college and higher level (16 years of schooling). Then I did the calculation for average years of schooling in 31 provinces from 2002 to 2012 according to the method mentioned above. After that, I took log of it and got the variable I need.
In order to check the robustness, I also use the variable , which is the total government educational expenditure (in yuan) in a province. It can be downloaded from NBSC database directly.
Independent control variable –urban population
is the logarithm of urban population lives in province i at the end of year t. As the urban population in eastern China is much bigger than that in western provinces, I also include the population as one of the control variables.
24
Independent control variable –transportation infrastructure
is log of transportation length (in 1000 km). There are many previous papers (e.g. He & Duchin, 2009; Chen & Groenewold, 2010, Fan et al., 2011) pointed out that transportation infrastructure plays a very important role in economic growth and has significant impact on inequality. So I also include this variable in my regression. But it is also essential to pay attention that the three main transportation modes in China, railway, highway, and waterway, should have different weighted ratios when added together. It is because that they have different transportation capacities. According to Yao and Wei (2007) and Wei et al. (2009), total length of transportation infrastructure = 4.27 * railways + 1.0 * highways + 1.06 * waterways. The length of three ways are available in NBSC website, I downloaded them and used the equation above to calculate the total length of transportation infrastructure in provinces.
Independent control variable –CPI
is the logarithm of Consumer Price Index (year 2002 = 100). According to Desai et al. (2005), there is a positive correlation between inflation and inequality in the U.S. And I wonder if the inflation also has impact on income gap in China, so I include CPI as a measure of inflation into my model. It is also available in the website of NBSC that there is data of CPI in 31 provinces during 2002 to 2012.
Independent control variable –provincial dummy and time dummy
As I mentioned in the introduction, 90% of FDI flows to eastern provinces and there are also regional disparities among China’s provinces, so I create three provincial dummy variables , , and , which equal to 1 if the province located in eastern, central, or western regions, otherwise the value of dummy is zero. In order to control the time related variant, I also create a time dummy variable and use it in some of the regressions.
25
To sum up, in Section 4, I first illustrate the source of data, then the set up of regression model as well as estimation methodology, and at last I describe each variable used in regression.
5. Results and Discussion
In this part of paper, I’m going to interpret the regression results of model 1 & 2 first, and then to do robustness check to make sure that the relationship between FDI and income inequality is not a casual one. After that, I will apply the estimation method on regional data respectively.
5.1 Estimation Results
Fig.3 shows the correlation between log of FDI inflow and log of urban-rural income gap. It can be observed that FDI is negatively correlated with income disparity.
26
Table 3. Effect of FDI on urban-rural income gap (Baseline OLS - national).
(1) (2)
VARIABLES lnURIG lnURIG
lnFDI -0.0612*** -0.0428*** (0.00321) (0.00787) lnGDPPC -0.0607* (0.0314) lninvest -0.0648** (0.0255) lntrade -0.0253** (0.0123) lnschooling -0.380*** (0.0809) lnurbanpop 0.144*** (0.0329) lntransportation -0.00534 (0.0131) lnCPI -0.924 (0.717) eastern 0.180*** (0.0225) western 0.207*** (0.0182) Time-fixed effect Constant Yes 2.392*** Yes 7.514** (0.0767) (3.364) Observations 341 341 R-squared 0.520 0.794
Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
Table 3 shows the regression results of the baseline OLS model, in which I used the panel data of all the 31 provinces in China. Using White test, I found there is heteroskedasticity problem. So I use the robust standard errors to correct it. Column (1) is the unconditional regression in which I only included the dependent variable lnURIG and independent variable lnFDI. In column (2), I used a complete model with control variables as I mentioned above. In the second column, it can be seen that the
27
coefficient of FDI is negative and significant at 1% confidence level, which means FDI inflows can decrease the urban-rural income inequality in China. The coefficients of lnGDPPC is also negative and significant at 1% confidence level, that is to say, provinces with higher GDP per capita and economic significance have less income inequality between urban and rural area.
Table 4. Effect of FDI on urban-rural income gap (model 1 - national).
(1) OLS (2) FE (3) Two-way FE (4) Two-way RE
VARIABLES lnURIG lnURIG lnURIG lnURIG
lnFDI -0.0428*** 0.00180 -0.00776 -0.0119 (0.00787) (0.0231) (0.0211) (0.0204) lnGDPPC -0.0607* -0.0515 0.0753 0.0480 (0.0314) (0.0516) (0.0612) (0.0597) lninvest -0.0648** 0.0257 0.00505 -0.00214 (0.0255) (0.0366) (0.0349) (0.0332) lntrade -0.0253** -0.0627* -0.0774** -0.0785*** (0.0123) (0.0333) (0.0288) (0.0264) lnschooling -0.380*** -0.328** -0.187 -0.249* (0.0809) (0.124) (0.179) (0.140) lnurbanpop 0.144*** 0.107 0.238* 0.0861* (0.0329) (0.135) (0.136) (0.0463) lntransportation -0.00534 0.0835*** 0.0392 0.0405 (0.0131) (0.0264) (0.0521) (0.0336) lnCPI -0.924 0.478*** 0.0671 -0.00582 (0.717) (0.161) (0.353) (0.340) Time-fixed effect Constant Yes 7.514** No -1.240 Yes -1.655 Yes 1.708 (3.364) (1.780) (2.727) (1.710) Observations 341 341 341 341 R-squared 0.794 0.229 0.398 Number of province 31 31 31
Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
28
Table 4 shows the regression results of model 1 using different estimation methods, OLS, FE, two-way FE, and two-way RE. The column (1) is exactly the same as the second column in Table 3, which are the results of OLS method with time fixed effect. In column (2) and (3) I applied the fixed effects and two-way fixed effects method, the main difference between the two methods is the year dummy variable. Column (4) presents the results of two-way random effects method. It can be seen from Table 4 that only when using OLS method, the coefficient of lnFDI is significant. In column (2), (3), and (4), the coefficients are not significant and even have different signs. What’s more, the estimation results of lnGDPPC are not significant in last three columns as well. Therefore, using model 1, the estimation results are not as good as I expected. One of the main reasons behind it should be the omitted variables. According to Lessman (2013), the effect of FDI on income inequality can also depends on the development level of a country. So I think the interaction term of lnFDI and lnGDPPC can be very important variable in my regression, and model 1 without interaction term shows a biased result.
Table 5. Effect of FDI on urban-rural income gap (model 2 - national).
(1) OLS (2) FE (3) Two-way FE (4) Two-way RE
VARIABLES lnURIG lnURIG lnURIG lnURIG
lnFDI -0.194*** -0.169*** -0.240*** -0.224*** (0.0397) (0.0543) (0.0512) (0.0527) lnGDPPC -0.410*** -0.403*** -0.313*** -0.350*** (0.0883) (0.133) (0.111) (0.112) interaction 0.0156*** 0.0176*** 0.0236*** 0.0214*** (0.00405) (0.00546) (0.00540) (0.00569) lninvest -0.0542** 0.0137 -0.00538 0.00438 (0.0250) (0.0359) (0.0315) (0.0299) lntrade -0.0317*** -0.0504* -0.0481** -0.0638*** (0.0120) (0.0259) (0.0181) (0.0165) lnschooling -0.289*** -0.337*** -0.137 -0.184** (0.0730) (0.0884) (0.108) (0.0890)
29 lnurbanpop 0.121*** -0.0205 0.182* 0.0472 (0.0316) (0.156) (0.102) (0.0532) lntransportation 0.0154 0.0935*** 0.0549 0.0730* (0.0143) (0.0246) (0.0462) (0.0379) lnCPI -0.715 0.414*** 0.457 0.386 (0.697) (0.130) (0.280) (0.274) Time-fixed effect Constant Yes 9.862*** No 4.568* Yes 0.767 Yes 3.734** (3.218) (2.575) (2.355) (1.617) Observations 341 341 341 341 R-squared 0.804 0.355 0.592 Number of province 31 31 31
Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
Then, I ran the regression of model 2 with interaction term, and the results are shown in Table 5. As I did in model 1, this time I also applied 4 exactly the same estimation methods for national data of 31 provinces from 2002 to 2012. And it is clear from Table 5 that this time all the coefficients of lnFDI are significant at 1% confidence level with a negative sign. Moreover, the lnGDPPC and interaction term are also significant at 1% confidence level. As I mentioned before, to see whether the FE or RE is more suitable for my model, I conducted a Hausman test, which null hypothesis is that RE is better than FE. The p-value of Hausman test here is 0.9945, which is bigger than 0.05, so I cannot reject the null hypothesis, that is to say, RE method performs better than FE in model 2. Therefore, my conclusion of model 2 will mainly base on the results of RE model shown in column (4).
So I can conclude from the results that FDI inflow has decreased the urban-rural income gap in China during the period 2002-2012, and to be specific, 1 unit more of FDI inflows can decrease urban-rural income gap by 0.174 unit. And this impact was also influenced by income level of the provinces. Provinces with higher per capita income have lower urban-rural income disparities. 1 unit increase of GDP per capita contributes to 0.4 unit decrease of urban-rural income gap.
30
negative and significant at 5% confidence level, that is to say, bigger volume of international trade help to decrease the urban-rural gap. lnschooling is negatively correlated with income inequality, which indicates that education can narrow down the urban-rural gap. The coefficients of CPI are positive and significant, indicating that higher inflation increases the income disparities.
To sum up, Table 5, especially the column (4) shows the results of my main model, and it can be concluded from it that from 2002 to 2012, the inward FDI has decreased the urban-rural income gap in China. And the impact of FDI on urban-rural income gap depends on the development level of a province.
5.2 Robustness Checks
To a wide range of estimators, Table 5 shows robust results. But it is also important to test if the results are still robust when different choices of proxies are applied. Till now, the variable Trade I used in all the regressions is provincial total value of imports and exports of operating units. Actually, NBSC offers another way of calculating it, which is the provincial total value of imports and exports of destinations and catchments. So in the robustness check, I replace the variable Trade with Trade2, and use the same method as I did in Table 5. The new regression results of two-way RE method are shown in the first column in Table 6. Then, I also use the volume of export only instead of volume of import and export. NBSC also offers two way of calculating export volume, one is by locations of operating units, and the other is by destinations and catchments of commodities. So I name the two variables Export and Export2 respectively and replace the variable Trade with them. The results are shown in the column (2) and (3) in Table 6. Then I replace average years of schooling with EDU, which stands for government educational expenditure. The results of the forth robustness check are in last column of Table 6.
It can be seen that the all the coefficients of FDI inflows and GDP per capita are still negative and significant, which are the same with the results in Table 5. Therefore, from the results of robustness check in Table 6, it can still be concluded that the FDI
31
inflows decreased the urban-rural income gap in China, and provinces with higher GDP per capita had lower income inequality during the period 2002 to 2012.
Table 6. Robustness check- Effect of FDI on urban-rural income gap (national)
Two-way RE (1) (2) (3) (4)
VARIABLES lnURIG lnURIG lnURIG lnURIG
lnFDI -0.245*** -0.227*** -0.249*** -0.237*** (0.0573) (0.0484) (0.0493) (0.0451) lnGDPPC -0.390*** -0.361*** -0.388*** -0.415*** (0.122) (0.106) (0.111) (0.0966) interaction 0.0229*** 0.0217*** 0.0237*** 0.0229*** (0.00613) (0.00529) (0.00546) (0.00491) lninvest 0.00483 -0.00187 -0.00680 0.00766 (0.0330) (0.0299) (0.0322) (0.0316) lntrade2 -0.0385** (0.0172) lnschooling -0.237** -0.171** -0.210** (0.0946) (0.0854) (0.0872) lnurbanpop 0.0366 0.0455 0.0587 0.00618 (0.0586) (0.0511) (0.0552) (0.0537) lntransportation 0.0659 0.0757** 0.0709* 0.0838** (0.0412) (0.0374) (0.0391) (0.0384) lnCPI 0.530** 0.176 0.301 0.387 (0.260) (0.277) (0.272) (0.295) lnexport -0.0578*** (0.0144) lnexport2 -0.0530*** (0.0149) lntrade -0.0670*** (0.0185) lnedu 0.0317 (0.0614) Constant 3.293** 4.779*** 4.466** 3.801** (1.603) (1.705) (1.758) (1.561) Observations 341 341 341 341 Number of province 31 31 31 31
Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
32 5.3 Region-related Analysis
As China is a country with very broad area, the development level varies among provinces. So it is also interesting to see if impact of FDI on urban-rural disparities performs distinctively among different regions.
Table 7. Effect of FDI on urban-rural income gap (model 2 - regional).
Two-way RE (1) Eastern (2) Central (3) Western
VARIABLES lnURIG lnURIG lnURIG
lnFDI -0.174*** -1.521* -0.345** (0.0610) (0.875) (0.144) lnGDPPC -0.343** -3.392* -0.676** (0.138) (1.836) (0.344) interaction 0.0128** 0.156* 0.0338** (0.00523) (0.0912) (0.0154) lninvest 0.0793** -0.0484 0.0206 (0.0355) (0.0707) (0.0678) lntrade 0.00593 -0.184** -0.0940*** (0.0316) (0.0808) (0.0203) lnschooling -0.263 -0.318 -0.0464 (0.218) (0.389) (0.160) lnurbanpop -0.0753* 0.328* 0.0149 (0.0385) (0.172) (0.0863) lntransportation 0.0357 -0.222 0.135* (0.0259) (0.159) (0.0710) lnCPI 0.100 1.858 0.604 (0.377) (1.398) (0.537) Time-fixed effect Constant Yes 4.246* Yes 27.25* Yes 6.307 (2.325) (16.52) (4.234) Observations 132 88 121 Number of provinces 12 8 11
Robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
33
respectively using RE method with fixed time effect. The reason of choosing two-way RE rather than two-way FE is quite similar to the one in national model 2. I use the Hausman test again, and find that the p-value of it is 0.1361, which is bigger than 0.05. So I accept the null hypothesis that RE performs better than FE in regional regressions. The way of categorizing provinces is shown in Fig.2. The FDI inflow and GDP per capita are still negatively correlated with urban-rural income gap in eastern, central, and western regions. However, in central provinces, the magnitudes of coefficients become much bigger. One of the reasons behind this can be the fewer observations and less fluctuations of FDI inflow in central provinces. But it can still be concluded that FDI inflows are negatively correlated to urban-rural income gap in different regions of China.
To sum up, in part 5, I explained the results of all the models, especially the model 2, using national data. And I also mentioned the estimation results of model 2 using regional data.
6. Conclusion
Part 6 is the conclusion of the whole paper, where I will first summarize the main steps to reach the final conclusions of this paper. And then I will discuss some of the shortcomings and limitation of this paper. At last, I’m going to discuss the direction of future related research.
6.1 Conclusion Remarks
In this paper, I first put forward the research question how foreign direct
investment affects the urban-rural income gap in China and motivation of it. Then I
did a review of existing literature and related theories. After that, I set up my own regression model and collected data as well as calculated variables for it. Then, I chose the proper estimation methods and ran the regression in STATA software.
34
Finally, I use economic knowledge I learnt to interpret the estimation results of regression model and came to the conclusion that FDI narrows the urban-rural income gap in China’s provinces, but its impact also depends on the development level of a province. Provinces with higher GDP per capita always have lower urban-rural income gap.
6.2 Limitation of Paper
This paper also has its limitations. In this part I’m going to discuss them and illustrate the reasons why I could not overcome these limitations.
First of all, there is problem of variable selections. In the regression model, I used the variable to control the domestic investment factor. Ideally, I should use the total domestic investment as one of the control variables in my paper. However, I could not find it anywhere in online database. So at last I use another variable, which is total government investment into fixed assets (measured in yuan), to control the domestic investment factor in my regression model.
Moreover, it would be better if I could have a larger observation numbers. The panel I used in this paper includes data of 31 provinces from 2002 to 2012 annually, which means that the overall observations for each variable is only (31*11= ) 341. If I could get data with longer time period, perhaps I could have better estimation results. But the economic data of China’s provinces are only available from 2002 and those of recent few years are not available on the website yet. Therefore, I can only have access to the provincial data from 2002 to 2012.
Last but not least, the regression model I used in this paper only includes some of the important economic control variables, but in reality, there can be more other economic factors which also influence the relationship between urban-rural income gap and FDI inflows in China. It is impossible to include all the potential influential factors into the regression, so in this paper I take all the other economic factors, which are not included in the model, as exogenous and have no effect on the outcome of the regression results.
35 6.3 Future Agenda
In this paper, I investigated the impact of FDI inflows on urban-rural income inequality in China’s provinces, and found that FDI can narrow down this inequality depending on the development level of the province. But the conclusions I made are all based on the previous data and economic environment, I could not provide evidence that FDI will always benefit income distribution in future. As China is still a developing country and its economic policies and environment are always on changing, it is essential to keep an eye on this research question.
Even though overall FDI inflows has narrowed down the urban-rural income gap in China, and from Fig 4 it is clear that the urban-rural inequality has decreased during the past 11 years, the disparities between urban citizens and rural residents in China are still very big. Fig 4 shows us that in 2012, urban citizens earn three times as much as people living in rural area, so it is also essential to find out solutions to decrease this gap. Here are two possible approaches.
For one thing, make adjustments on hukou system. Hukou system is one of the unique social and geographic control systems in China, which ties people’s access services to their hukou categories and thus restricts rural workers to move from their hometown to prosperous urban area. Knowing that hukou system is one of the main explanations for the urban-rural income gap, China’s national government has already carried out revolutions in 2014 to remove the distinction between urban and rural residents (Branigan, 2014). However, the adjustments of policy and revolutions of system are on a very slow process and don’t perform very well till now. Therefore, I could do some related researches to see how the government can better reform hukou system and further narrow down the urban-rural income gap.
For another, change the current development policies. There is a long-standing policy in China that government favors urban people over the rural ones, and it is also one of the potential explanations of large urban-rural income inequality in China (Xie & Zhou, 2014). I’m also interested in this research question, as there is no existing
36
literature investigating the impact of government development policies on urban-rural gap in China quantitatively.
To sum up, even though FDI inflows has helped to narrow down urban-rural income gap in China during 2002 to 2012, it is still uncertain that the impact will continue in the future. So it is essential to keep investigating this research question. What’s more, there are some other economic factors also affecting the urban-rural disparities. Therefore, it is also important to conduct related researches to help further decrease the gap.
37 7. Reference:
Avalos, A., & Savvides, F, A. 2003. On the Determinants of the Wage Differential in
Latin America and East Asia: Openness, Technology Transfer and Labor Supply. LAEBA working paper No. 19.
Basu, P., & Guariglia, A. 2007. Foreign Direct Investment, Inequality, and Growth.
Journal of Macroeconomics, 29 (4): 824-839.
Bhandari, B. 2007. Effect of Inward Foreign Direct Investment on Income Inequality in Transition Countries. Journal of Economic Integration, 22(4): 888-928. Branigan, T. 2014. China Reforms Hukou System to Improve Migrant Workers’
Rights. The Guardians, retrieved at
https://www.theguardian.com/world/2014/jul/31/china-reform-hukou-migrant-wo rkers.
Cai, Y. 2000. Between State and Peasant: Local Cadres and Statistical Reporting in Rural China. The China Quarterly, 163: 783-805.
Castro, G. A. 2011. The Effect of Trade and Foreign Direct Investment on Inequality: Do Governance and Macroeconomic Stability Matter? Economic Mexicana, 20 (1): 181-219.
Chase-Dunn, C. 1975. The Effects of International Economic Dependence on Development and Inequality: A Cross-national Study. American Sociological
Review, 40 (Dec): 720-738.
Chen, A. &f Groenewold, N. 2010. Reducing Regional Disparities in China: An Evaluation of Alternative Policies. Journal of Comparative Economics, 38: 189-198.
Chen, K., Jefferson, G., Rawski, T., Wang, H., & Zheng, Y. 1988. New Estimates of Fixed Investment and Capital Stock for Chinese State Industry. The China
Quarterly, 114: 243-266.
Chintrakarn, R., Herzer, D., & Nunnenkamp, P. 2011. FDI and Income Inequality: Evidence from a Panel of U.S States. Economic Inquiry, 50 (3): 788-801.
38
Choi, C. 2006. Does Foreign Direct Investment Affect Domestic Income Inequality?
Applied Economic Letters, 2006 (13): 811-814.
Chow, G. 2006. Are Chinese Official Statistics Reliable? CESifo Economic Studie,. 52 (2): 396-414.
Deiniger, K., & Squire, L. 1996. A New Data Set Measuring Income Inequality. World
Bank Economic Review, 10: 565-591.
Desai, R, M., Olofsgard, A., & Yousef, T, M. 2005. Inflation and Inequality: Does Political Structure Matter? Economics Letters, 87(1): 41-46.
Fan, S., Kanbur, R., & Zhang, X. 2011. China’s Regional Disparities: Experience and Policy. Review of Development Finance, 1: 47-56.
Feenstra, R. C., & Hanson, G. H. 1997. Foreign Direct Investment and Relative Wages: Evidence from Mexico’s Maquiladoras. Journal of International
Economics, 42 (1997): 371-393.
Figini, P., & Gory, H. 2011. Does Foreign Direct Investment Affect Wage Inequality?
An Empirical Investigation. Retrieved at
https://www.ifw-members.ifw-kiel.de/publications/does-foreign-direct-investmen t-affect-wage-inequality-an-empirical-investigation/working-paper.
Galbraith, J., & Kum, H. 2003. Inequality and Economic Growth: A Global View Based on Measures of Pay. CESifo Economic Studies, 49(4): 527-556.
Girling, R. 1973. Dependency and persistent income inequality. Stanford, CA: Institute of Political Studies.
Hausman, J. A. 1978. Specification Tests in Econometrics. Econometrica, 46 (6): 1251-1271.
He, L. & Duchin, F. 2009. Regional Development in China: Interregional Transportation Infrastructure and Regional Comparative Advantage. Economic
Systems Research, 21: 3-22.
Holz, C., A. 2003. “Fast, Clear and Accurate”: How Reliable Are Chinese Output and Economic Growth Statistics? The China Quarterly, 173: 122-163.
39
(4): 355-367.
Koch-Weser, I., N. 2013. The Reliability of China’s Economic Data: An Analysis of
National Output. U.S. – China Economic and Security Review Commission Staff
Research Project, January 2013, U.S – China Economic and Security Review Commission, retrieved at
http://www.uscc.gov/sites/default/files/Research/TheReliabilityofChina'sEconomicDa ta.pdf.
Kuznets, S. 1955. Economic Growth and Income Inequality. The American
Economic Review, 1995 (3): 1-28.
Lessmann, C. 2013. Foreign Direct Investment and Regional Inequality: A Panel Data Analysis. China Economic Review, 24 (2013): 129-149.
Mah, Jai S. 2003. A Note on Globalization and Income Distribution – The Case of Korea, 1975 – 1995. Journal of Asian Economics, 14 (1): 157-164.
Mahutga, M. C., & Bandelj, N. 2008. Foreign Investment and Income Inequality: The Natural Experiment of Central and Eastern Europe. International Journal of
Comparative Sociology, 2008, 49 (6): 429-454.
Milanovic, B. 2003. Can We Discern the Effect of Globalization on Income
Distribution? Evidence from Household Surveys. Retrieved from https://www.cgdev.org/doc/event%20docs/MADS/Milanovic.pdf
Nunnenkamp, P., Schweickert, R., & Wiebelt, M. 2006. Distributional Effects of
FDI: How the Interaction of FDI and Economic Policy Affects Poor Households in Bolivia. Kiel Working Paper No. 1281.
OECD. 2005. Measuring Globalization: OECD Handbook on Economic
Globalization Indicators 2005. Retrieved at:
http://www.oecd.org/sti/sci-tech/oecdhandbookoneconomicglobalisationindicator s.htm
Rodriguez-Pose, A., & Gill, N. 2006. How Does Trade Affect Regional Disparities?
World Development, 34 (7): 1201-1222.
40
and Long Run: The Case of South Korea 1975-1995. Journal of Asian
Economics, 20 (1): 62-68.
Sylwerster, K. 2005. Foreign Direct Investment, Growth and Income Inequality in Less Developed Countries. International Review of Applied Economics, 19 (3): 289-300.
Taylor, K., & Driffield, N. 2005. Wage Inequality and the Role of Multinationals: Evidence from UK Panel Data. Labor Economics, 12 (2): 223-249.
Tsai, P. 1995. Foreign Direct Investment and Income Inequality: Further Evidence.
World Development, 23 (3): 469-483.
Velde, D. T., & Morrissey, O. 2004. Foreign Direct Investment, Skills and Wage Inequality in East Asia. Journal of the Asia Pacific Economy, 9 (3): 348-369. Wei, K., Yao, S., & Liu, A. 2009. Foreign Direct Investment and Regional Inequality
in China. Review of Development Economics, 13 (4): 778-791.
Wu, J., & Hsu, C. 2012. Foreign Direct Investment and Income Inequality: Does the Relationship Vary with Absorptive Capacity? Economic Modeling, 29 (6): 2183-2189.
Xie, Y., & Zhou, X. 2014. Income Inequality in Today’s China. Proceedings of the
National Academy of Sciences of the United States of America, 111(19):
6928-6933.
Yang, Xu. 2014. Foreign Direct Investment and Urban-Rural Income Inequality:
the Case of China. Master thesis, the Graduate School of Clemson University.
Yao, S. & Wei, K. 2007. Economic Growth in the Presence of FDI: The Perspective of Newly Industrializing Economies. Journal of Comparative Economics, 25: 211-234.
