3 Data Description

All data are collected from the China Statistical Yearbook 1990 and 2012. The dependent variables were GDP per capita in 2010 and the average GDP growth rate from 2004 to 2010, respectively. To avoid the endogeneity problem, all the data for explanatory variables are in 2004 which is the initial year of the period studied. Also, for Tibet has some explanatory variables with extreme values, to remove the effect of these extreme values, the observation of Tibet is deleted when we estimate the regression models, which leaves the sample size as 30 provinces. Table 1 shows the meanings for each variable and the summaries of descriptive statistics.

Table 1 Descriptive Statistics

Table 1 Descriptive Statistics-Continued

Due to no agreement on the measurement standard of natural resource abundance, such as the share of resource-based exports to GDP, the amount of cultivated land per capita, the proportion of primary sector employment in total employment, energy reserves, the ratio of resources rent to GDP, and the proportion of primary sector output in GDP, it is the first big obstacle for our research.

Here, we take three variables, water, land and coal, to reflect natural resources that are commonly thought of as being primary factors of production. Water is considered to be one of the crucial resource constraints for some provinces, especially those in the Northwest of China. It also has more significant influence in agriculture provinces than in industrialized provinces. Usable land can be the resource constraints for provinces with complicated geological features, such as deserts, mountains, and primitive forests, etc. Coal is the primary energy source for production in China and produced in almost every province, and it is used here just as the proxies of mineral resources.

Based on the quartile plot of the indicator of the natural source and regional economy of China’s province, we find that the western region of China has a higher volume of coal per capita and water, while the eastern region of China has more useable land, GDP and GDP per capita, and the growth rate of real GDP is different. Due to Stijns’s research (2005), an abundance of a mineral resource is apt to the resource curse, while abundance of a primary product such as agricultural product, will not be affected by resource curse. Therefore, the indicator, coal per capita, is a crucial variable to represent the natural resource. Based on the quartile plot of the coal per capita, we find that the eight natural resource-rich provinces include Inner Mongolia (NMG), Xinjiang (XJ), Shaanxi (SAX), Shanxi (SX), Qinghai (QH), Yunnan (YN), Guizhou (GZ), Ningxia (NX). Except SAX, other provinces are located in the western region of China.

From Table 2, we can find some fluctuations in the economic growth rate of China’s provinces. In 1990, the ranks of the growth rate of the resource-rich provinces’ economy were distributed with equal distance. In 1995, all of them were located in the last half. In 2000, some of them entered into the first half, such as NMG, SX, which are nearer to the metropolitan, economic center such as Beijing. In 2005, due to the Western Development Policy of the central government since 1999, the last province (YN) had better economic performance than before, with a growth rate of 8.9%. But the gap between them is huge. We think a possible reason is geography. Taking NMG and SX as examples, we find they are neighbor to developed region among the natural resource-rich provinces, and reach faster economic growth rate by exporting natural resources to nearby developed regions. In 2010, except XJ, others performed well. In my views, it is partly due to the declining growth rate of the eastern region caused by the external economic sector, partly due to the supporting policies of the central government.

Table 2 Ranking of the Growth Rate of Real GDP of China’s Province

Table 2 Ranking of the Growth Rate of Real GDP of China’s Province-Continued

According to the comparative analysis with the data of different provinces, we can draw a simple conclusion that natural resources can to some degree affect economic growth. Compared to many other factors that can influence economic growth, natural resources may not be the most important, but it has wildly positive effect, especially in the developing stage.

I use transportation and education level to represent the hardware and software infrastructure. The meaning of all these variables is very straightforward. Yet, one caveat should be pointed out. The variable, river density, which measures the density of inland river waterway, can reflect both the transportation infrastructure level and the topological feature of one province. The same may be true for road density and rail density but to a lesser extent.

Finally, Openness, FDI and privatization are used to represent the extent of market reform and opening, which are all commonly used in development literature as to a lesser determinants of development level.