Reasearch on Regional Differences Between China's Economic Growth and Export Trade Based on the Analysis of a Random Effects Model

This paper analyzes the factors that influence the economic growth of the provinces of China by means of a Panel Data Model. Traditional analytical methods of economic growth are compared with a Panel Data Model. The results of empirical research indicate that the changes of fixed assets investment, gross domestic export, and macroeconomic policies will affect China's GDP. It is finally concluded that export is the driving force behind economic growth in China. This conclusion is quite different from traditional analysis.

First of all, a descriptive statistical analysis on time series data of economic variables of more than 30 provinces, municipalities, and autonomous regions throughout the country discovers that there are more unusual value points in the economic variables of the Tibet Autonomous Region and that common insertion and supplementation as well as smoothing cannot reflect the missing data effectively.It is reasonable for us to doubt the authenticity and accountability of such statistical data.As a result, in the positive analysis of the paper, the Tibet Autonomous Region is removed from the samples of the western region and the whole country.
Furthermore, until 1997 the data of Sichuan Province included the data of Chongqing Municipality.For the purpose of accurate measurement, the data of both Sichuan Province and Chongqing Municipality are eliminated.Thus the data of 27 provinces are obtained from such changes.
All the sequences represent the observed data of 27 provinces from 1995 to 2004.The sources of the data are mainly the 1995-2004 China Statistical Yearbook, the State Information Center, and the China Economic Information Network Data Co., Ltd.For the convenience of calculation, current values of different periods are used without considering the impact of the price index.When processing data, the natural logarithm is used to eliminate heteroscedasticity.The exchange rate is the annual average conversion rate of RMB versus key foreign currencies, i.e. the annual middle price (1995)(1996)(1997)(1998)(1999)(2000)(2001)(2002)(2003)(2004).
The study on regional differences in economic growth during a specific period includes cross section data and time series data.Only using time series data or cross section data will not meet the requirements for economic analysis.For example, the growth of the GDP in each province is influenced not only by the economic structure of the region but also by the macroeconomic policies of China during different periods.Cross section data (that is, choosing data from different provinces at the same time as sampled observed values) alone can only support analysis of the impact of different industrial structures in the provinces or municipalities on economic growth; however, time series data (that is, choosing data from the same province or municipality at different time periods as sampled observed values) can only analyze the impact of macroeconomic policies or structures (consumption, savings, investment, and net export) on a country's economic growth.Similarly, such analysis cannot discover the impact of structural differences among the regions' economic growth.As a result, the model adopted for the analysis is the Panel Data Model.This combines cross section data with time series data.
The formula of the general model for economic growth is given in Equation 1.
The logarithm used is given in Equation2.
In the above formulas, GDP represents gross domestic product of the provinces; IN is the data of fixed assets investment; CON means consumption; EX refers to the export activity at the current period; IM means the import data of the current period.In addition, i represents the provinces and t the years.
The introduction of dummy variables (also usually called qualitative interpretation variables) can help find how economic change in each region is influenced not only by quantitative variables (e.g.export and fixed assets investment) at some clear sizes but also by qualitative variables (Asian Financial Crisis).For instance, under the same conditions and in case of the same factors, China's economic growth at the regional level will necessarily suffer negative influence when financial crisis occurs.Meanwhile, China's foreign exports will also be greatly ) ) (3) In accordance with the model, 0 β ) is used to interpret the differences in economic growth among the provinces, municipalities, and autonomous regions.

DATA
The growth rate of the GDP 1 for the provinces of China is used to represent the speed of their economic growth.
For the convenience of analysis, we define the rate of actual economic growth of Province i in the years of 1995-2004 in Equation 4.
Similarly, RIN is defined as the growth rate of fixed assets investment; RCON: the growth rate of household consumption; and REX: the growth rate of export data at the current period.In addition, i represents the provinces and t the years in Equation (5).

;
) ; In order to eliminate the impact of price fluctuations, we divide the above by the 'index of gross domestic market value,' 'price index of fixed assets investment,' and 'price index of household consumption' (calculated by comparable price which is 100 in the previous year) (See the data in Table 1).China (1995China ( -2004) ) Then, the HHC is the representative index of household consumption; FAI is the representative index of fixed assets investment; and GDP is the representative index of the deflating index.
In order to make the dimensions of the data the same, the USD is adopted uniformly as the measurement unit, while the exchange rate is the annual average conversion rate of RMB vs. key foreign currencies, that is, the annual middle price (1995)(1996)(1997)(1998)(1999)(2000)(2001)(2002)(2003)(2004)) (See the data in Table 2).Sichuan, Guizhou, Yunnan, and Tibet (See the data in Tables 3-7).

REPORT AND ANALYSIS OF THE MODELS OF REGIONAL ECONOMIC GROWTH OF CHINA
Data Science Journal, Volume 6, Supplement, 9 June 2007 Many economists use models such as Fixed Effects and Random Effects to evaluate regional economic growth; however, the two models are never used in relevant documents available in China.In this paper, we apply the two models to the evaluation of the 7 regions of China (Hsiao, 2003;Greene, 1997).The specific models are presented as follows: the Fixed Effects Model is given in Equation ( 6).
In accordance with this model, 0 β ) is used to interpret the difference of economic growth in the provinces, municipalities and autonomous regions.The Random Effects Model is given in Equation ( 7).
E-VEWS3.1 statistical software can be used to evaluate the model for the 7 regions of China.The Fixed Effects weighted model is used in Formulas 4 and 5, and the results are shown in Table 8.The regression model indicates that it is helpful for data simulation and that the values of T and F have been tested.Meanwhile, the dummy variables of the year are also tested and in agreement with the actual variables.
This means that there is certain inflation in China in 2003.
When the Random Effects model is used, the results are in table 9.
Data Science Journal, Volume 6, Supplement, 9 June 2007 ε are both disturbance terms.In the quantitative analysis, Hausman's test is often used to decide which is more effective, Fixed Effects or Random Effects (Hadri, & Larsson, 2005;Jung, Shin, & Oh, 2005;Hausman, 1978).The test formula is as follows: As for models 7 and 8, H is subject to the Chi-squared distribution with 2 as the degree of freedom.The critical value of Chi-squared with 2 as the degree of freedom is 5.99.If H>5.99, the Fixed Effects model is accepted; otherwise the Random Effects model is accepted.B is the evaluated coefficient of Fixed Effects, while β is the evaluated coefficient of Random Effects.The model is selected in accordance with the Hausman Test by means of STATA8.0 as given in Table 10.
Data Science Journal, Volume 6, Supplement, 9 June 2007 The results of test cannot reject a null hypothesis, which indicates that the Random Effects model is more suitable for this study.

CONCLUSION AND DISCUSSION
There are some similarities between the conclusions of this research and the evaluated results made by other scholars by means of traditional theories of economic growth, that is, there is a common ground of economic growth inside the economic belts.What differs from those scholars in this work is that there is little difference in Model to interpret the differences of economic growth in China, and using reasonable, scientific methods leads to making correct conclusions; and (2) the simulation results indicate that there are differences in economic growth between economic belts in China.According to the elasticity analysis of the impact of material capital and human capital on economic growth rate in the years 1995-2004, the output elasticity of material capital in the economic regions of China is higher than that of export demand, which indicates that increasing the input of material capital is more effective in the short term.However, the output elasticity of material capital and export demand is relatively advantageous in some regions.
The problems on which this paper needs to make further study are also very important.First, in the years 1994-2003, Chinese export trade was in a period of high-speed development, and spatial change and expansion of the samplings may have had certain effects on the conclusions.This indicates that the export demand had a great impact on economic growth during this period.Second, the last part of the paper examines the Hausman test on the Fixed Effects and Random Effects of the Panel Data Model, which is quite important for the correct selection of models and analysis.However, the analysis on dummy variables including time needs further extension, which will provide more detailed interpretation on finalizing the driving effect of export trade over the economic growth of a country.

Figure 1 .
Figure 1.Different Price Indices of Economic Growth inChina (1995China ( -2004)   ) economic growth between different economic belts.Our work greatly diverges from other works in the definition of model interpreting variables and reasonable interpretations.It does not only merely demonstrate the use of the Panel Data Model but also demonstrates the difference between different economic belts.The findings indicate that China was greatly affected by the Asian Financial Crisis in the years 1997-1999.There was certainly inflation in 2003.This paper tries to make two points clear: (1) it is effective to use the Panel Data

Table 1 .
Different Price Indices of Economic Growth inChina (1995China ( -2004)   ) 1 GDP is usually calculated by its deflating index, with the formula below: GDP deflating index=100*GDP nominal growth rate/ GDP actual growth rate.As a result, the GDP index here refers to GDP deflating index.

Table 3 .
China's seven major economic regions' GDP

Table 8 .
Regressive Results of the Growth Rate of the Seven Regions of China(Fixed Effects)

Table 9 .
Regressive Results of the Growth Rate of the Seven Regions of China (Random Effects)

Table 10 .
Selecting the Model in accordance with the Hausman Test