Kiran Sharma

FACTORS EXPLAINING REAL GDP GROWTH IN CHINA Tore Økland Department of Economics Indiana University of Pennsylvania Indiana, PA 15705 -- Kiran Sharma Department of Economics Indiana University of Pennsylvania Indiana, PA 15705 -- ABSTRACT The determinants of the change in real Gross Domestic Product (GDP) per capita of China are analyzed in this paper. Using data from-, the regression equation is estimated using ordinary least squares (OLS). The dependent variable is the change in real GDP per capita. The independent variables fit into the following categories: fertility rate, gross capital formation, logged GDP, and mortality rate, with a trend variable added. The results indicate that the variable of fertility rate had a positive effect on economic growth, logged GDP had a positive impact upon the change in real GDP growth per capita (in line with previous studies), and gross capital formation and mortality rate had negative and positive impacts upon the change in real GDP growth per capita, respectively, albeit insignificant amounts. JEL Codes: E01, E10, F0, H52 Keywords: Real GDP per capita, emerging markets, BRICs 1. INTRODUCTION 1.1 Background One of the most well-known methods of measuring a nation’s economic activity is to calculate its GDP, or gross domestic product. A country’s GDP, simply put, is the value of all goods and services produced within a country during a period; often yearly. The basic formula for calculating GDP is , where represents total consumer spending of a country, represents investments by businesses, represents total government expenditures, and equals the difference between a country’s exports and imports. In this paper, however, real GDP determinants of growth are examined: real GDP is an inflation or deflation-adjusted figure that reflects the value of all goods and services produced by an economy during a period. Real GDP per capita can predict changes in output with greater accuracy than, for example, nominal GDP, and can be used to predict future economic growth of a region. 1.2 Emerging Markets In terms of analysis, real GDP per capita and corresponding growth rates can be used to identify emerging markets, predict either positive or negative rates of change for future periods, and can be used to rank countries based on measured GDP. Emerging markets, for example, are characterized by swift economic growth within short periods of time, increasingly larger per-capita incomes, and strong indications of budding industrialization. With globalization on the heels of the 21st century, economies that were otherwise stagnant or mainly agrarian began to grow due to foreign investments, larger government expenditures on healthcare, education, and other welfare programs, and increased involvement in trade (Yeyati and Williams, 2014). The four largest emerging economies were classified under the acronym BRICs, or “Brazil, Russia, India, and China.” These economies had significant growth in terms of real GDP, trade, and foreign investments, with China emerging as the largest of these four economies, and the second-largest in the world. 1.3 China’s Economic Growth Boasting the largest population in the world, with a current figure of 1.357 billion people, China has not only experienced a strong, constant growth of population from 1960-onwards, but from 1995-onwards, has experienced an almost exponential growth of real GDP per capita (World Bank). As an emerging economy, it has garnered attention, attracted foreign investments, become one of the world’s largest producers of everyday items, ranging from electronics to clothing, and continues to grow rapidly. While it seems obvious that China has a large and ever-growing economy, unfortunately, few pieces of literature that examine multiple and varied determinants behind China’s real GDP growth rate have been published. Research papers regarding growth determinants of other, smaller economies have been published (Pryor, F., 2010), but we believe that no comprehensive work of literature regarding multiple determinants of the real GDP growth of China over a period has been published. Of course, hesitance lies in the widely-held claim that the Chinese government has manipulated their GDP figures and determinants of growth (Fernald, Hsu, Spiegel, 2015), which could explain the lack of articles concerning growth factors of China’s real GDP per capita. Regardless of the lack of research investigating the magnitude of effect of multiple determinants of real GDP growth of China, the goal of this paper is to interpret the effectiveness of specific variables with regards to the change in real GDP per capita. By doing so, this paper can contribute to the existing literature, as well as pique additional interest into the economic workings of China. 1.4 Outline of Paper This paper attempts to determine which determinants significantly affect the change in real GDP per capita growth of China. The second section will review literature relevant to determinants of macroeconomic growth concerning gross domestic product. In the third section, data and variables will be mentioned. The fourth section will discuss the model used for the study. In the fifth section, the regression and results will be presented. In the last section, conclusions drawn from the analysis will be discussed. 2. LITERATURE REVIEW Sala-i-Martin et al. (2004) examined the determinants of long-term economic growth for eighty-eight countries over the period of-. Bayesian Averaging of Classical Estimates (BACE) was used to analyze the cross-sectional data. The dependent variable was the growth rate of GDP, while sixty-seven independent variables were used. Eighteen of the explanatory variables were statistically significant, and the variables of relative price of investment, primary school enrollment, and the initial level of real GDP per capita had the strongest relationship with the dependent variable. Imran et al. (2015), using fixed effect and pooled ordinary least squares methods, estimated the relationship between unemployment and economic growth in twelve Asian countries over the period of-. The authors used both time-series and cross-sectional data, which was combined into panel data. The dependent variable in the regression was GDP per capita, while the independent variables were the unemployment rate, gross capital formation, population growth rate, trade openness, real interest rate, domestic credit, inflation rate, government expenditures, natural resources, savings, and foreign direct investment (FDI). Imran et al. (2015) concluded that high unemployment rates reduce economic growth, while the other independent variables had mixed effects or were not statistically significant. Perez-Rodríguez et al. (2014) investigated the dependence between GDP and tourism. They limited the data sample to two developed economies United Kingdom (non-tourist oriented) and Spain (tourist oriented), and one emerging economy, Croatia (tourist oriented). The GDP data source was the International Monetary Fund while tourism data were collected from the Office for National Statistics for the UK, Bank of Spain, and Croatian National Bank. Perez-Rodríguez et al. (2014) employed a copula-based GARCH technique and a two-step MLE procedure to estimate the parameters in the study. The authors determined that there is a statistically significant constant and positive asymmetric dependence between tourism receipts and GDP growth rates. In other words, countries can stimulate their economies by increasing tourism. Therefore, countries that have suffered from terrorist attacks recently, such as France and Turkey, will witness lower tourism and thus experience adverse effects on their economies. Sunny and Unnikrishnan (2015) focused on the determinants of capital inflows in BRICs economies. The authors used a Vector Autoregression (VAR) technique to identify the determinants and Johansen cointegration and Augmented Dickey Fuller tests to analyze the data. Since the Johansen cointegration test suggested that there was cointegration between the variables, the authors used the Vector Error Correction model (VECM). The key variables included in the study were capital inflows (sum of FDI and portfolio equity) and GDP. Per the authors, capital inflows play a significant role in economic growth and macroeconomic fluctuations. Data was collected from the World Bank for the period- and was analyzed with OLS. Sunny and Unnikrishnan (2015) studied each BRICs country separately and found that GDP is essential in determining the capital inflows to all the BRICs economies. Demographic changes also have played an important role in both GDP growth and capital market returns. Arnott and Chaves (2012) used ordinary least squares regression analysis (OLS), a large sample of countries, and over sixty years of economic and demographic data for each, to determine that demographics do account for changes in GDP growth and capital market returns. The authors “force-fitted” a polynomial to their collected data, resulting in strong statistical significance, and concluded that via demographics, young-adults were the most responsible for driving GDP growth, while middle-aged adults were responsible for driving capital market returns. Barro (1998) examined determinants of economic growth in a panel data study. The paper expanded on an earlier cross-sectional study written by the same author. About 100 countries were examined for the period 1960 to 1990. The dependent variable was economic growth, as measured by real per capita GDP, while the independent variables used were initial level of GDP, initial level of human capital, fertility rate, government consumption, the Rule-of-Law Index, terms of trade, and regional dummy variables. The paper found that economic growth, as measured by real per capita GDP, is increased by maintaining rule of law, lower levels of government spending, and lower inflation. An increase in political rights has a positive effect on the dependent variable, but this becomes less effective once it reaches a certain level. 3. DATA This study uses data that was collected from the World Bank and Bloomberg for the period of-. Our dependent variable is the change in real gross domestic product per capita in 1972 US dollars, while our independent variables are fertility rate in average births per woman, gross capital formation in 1972 US dollars, the natural log of GDP, and infant mortality rate. Further explanations of independent and dependent variables appear in Table 1. 3.1 Expected Signs        The size of the economy (LOGGDP), as measured by the natural log of GDP, is expected to have a negative effect on the dependent variable. It will become increasingly difficult to maintain a high growth rate as the size of the economy becomes large. This can be observed in developed countries in North American and Western Europe. This explanatory variable has been used in other studies (Sala-i-Martin, 2004). The population size (LOGPOP), as measured by the natural log of the total population, is expected to have an ambiguous sign. Strong population growth causes real GDP per capita to stagnate, but a lesser population growth does not have a positive effect on economic growth (Berry, 2014). Fertility rate (FERTRATE) is expected to have a negative effect on the dependent variable; the higher a nation’s real GDP per capita growth and level of education, the fewer births per woman occur. Gross capital formation (GRSSCAPFRM) is expected to have a positive effect on the dependent variable, as an increase in fixed assets within an economy (land, buildings) will net a higher overall GDP, and by extension, a higher change in real GDP per capita. Lastly, mortality rate (MORTRATE) is expected to have a negative change with respect to the dependent variable; industrialized countries are likely to have advanced medical care and a higher real GDP per capita, so mortality rates are expected to drop. 3.2 Descriptive Statistics The mean change in real GDP per capita of China, from-, is 281.33, while the minimum value is 130.14, and the maximum value is 932.99. The sizable difference in the maximum and minimum values of real GDP per capita for China may be attributable to China’s almost exponential GDP growth rate from about 1995-onwards (World Bank). The period from-, on which this study is based, captures the rapidity of economic growth that China has experienced in the past twenty years. The standard deviation is 207.67 of real GDP per capita, indicating a variation of values spread over a wide range. The variables of FERTRATE, GRSSCAPFRM, AND MORTRATE have standard deviations of 0.924, 1.14E+ 12, and 16.25, respectively. These values also indicate a wide range of values, and may also be attributed to the general growth rate of China’s economy, China’s implementation of their “One Child” policy, and a decrease in mortality rate. Other descriptive statistics appear in Table 2. 3.3 Data Limitations The US Economist magazine remains skeptical about the truthfulness of the Chinese reporting of gross domestic product. Per a survey by the Wall Street Journal (Sparshott, 2015), more than 96% of American economists believed that the actual second quarter GDP growth was smaller than the 7% reported by the Chinese government. The accuracy of the model may, therefore, be affected if the reported gross domestic product numbers have been artificially high in certain years during the period we are examining. No studies have previously been published using long-term time-series data on China’s economic growth. There is a significant lack of available data on relevant independent variables, especially over long period such as the one used in this study. This leads to omitted variable bias in the models. 4. ECONOMETRIC MODEL This study utilizes ordinary least squares regression (OLS) to measure the impact of determinants on real GDP per capita growth rates in China from-. The original hypothesized equation was created with explanatory variables that had been proven to be statistically significant in related literature. Both energy consumption and CO2 consumption were added because of the general trend of increased energy consumption with an increase in the change in real GDP per capita. Fertility rate in average births per woman was added due to the implementation of China’s One Child Policy in 1979; from 1979 onwards, there was a steady decline in the country’s birth rate. The natural logs of both GDP and population were included to reduce overall skewness. Urban population from- was added to illustrate China’s increasing rate of residence in more metropolitan areas, while gross capital formation emphasized the increase in additions of fixed assets to China’s economy (increasing in recent years). Lastly, mortality rate was included to demonstrate the decline in China’s overall infant mortality rate, in accordance with China emerging as a heavily-industrialized economy, with presumably better and more accessible healthcare. Thus, the original equation was as follows: (1) Because of the lack of statistical significance regarding the ENERGYCON, CO2, URBAN, AND LOG(POP) variables, we decided to omit them. A trend variable was also added. Our final equation is as follows: (2) 4.1 Econometric Procedures We tested for serial correlation using the Durbin-Watson test. For equation (1), the Durbin-Watson statistic was-. Possible Durbin-Watson statistic values, according to Studenmund, if below 1.12, imply positive serial correlation, values above 1.66 indicate no evidence of positive serial correlation, and values between the two figures are inconclusive. For equation (2), the Durbin-Watson statistic was measured to be-. Therefore, both Durbin-Watson statistics mentioned indicate no evidence of positive serial correlation. Also, a time trend variable was included in the second model to account for multicollinearity, which is common for time-series data. 5. RESULTS The results from models 1 and 2 are shown in tables 3 and 4, respectively. In model 2, which is the final model, the following independent variables are significant at a 1% level: GRSSCAPFRM, LOGGDP, and MORTRATE. Additionally, FERTRATE is significant on a 5% level. The R-squared and adjusted R-squared are 0.5435 and 0.4801, respectively. We do not have any previous time-series studies on economic growth in China to compare with, but we suspect that the fit of the model has room for improvements. ENRGYCON, GRSSCAPFRM, and MORTRATE had unexpected signs in model 1, but each of these independent variables had a very small coefficient. Additionally, FERTRATE also had an unexpected sign. CO2, LOGGDP, and URBAN had the signs that were expected. We only used the independent variables in model that were statistically significant in model 2. A trend variable was implemented in model 2 to adjust for the time trend in economic growth. The trend variable was positive, which is consistent with economic theory, and statistically significant at the 1% level. Surprisingly, neither CO2 nor ENRGYCON were statistically significant. We suspect that we may have a simultaneity problem in the models in addition to omitted variable bias. This could explain the inconsistencies of the statistical significance levels of CO2 and ENRGYCON. As mentioned by Fernald, Hsu, and Spiegel (2015), there exists the possibility that data available about the Chinese economy may have undergone manipulation before being made public. 6. CONCLUSION With an absence of empirical literature regarding the real GDP per capita growth of China, this study sought to highlight several factors and their individual impacts on Chinese real GDP per capita growth. The results show the importance of logging two variables, population and GDP; since both population and GDP are multiplication-based, and in this instance, have grown exponentially, applying logarithms to these two variables can better analyze the growth rates attributed to each, versus using total population and total GDP. China’s current status as an “emerging market” is evident by its real GDP per capita growth over the past two decades. From a largely agriculture-based economy, in a less than a fifth of a century, it has become one of the world’s largest producers of everyday goods, and currently has the second largest gross domestic product in the world. As the largest member of the “BRICs,” it has, as a nation and an economy, gained both attention and notoriety for its almost exponential rate of growth. But despite having such a global presence, few pieces of economic literature about the country’s real GDP per capita growth have been published. It is our intention that this paper has provided a new kind of analysis of the real GDP per capita of China, using both environmental and demographic variables. By identifying the determinants and their effects on real GDP per capita, better economic analyses can be conducted regarding the topic, and more literature can be published with different focuses - demographic, social, economic, environmental, and so on. REFERENCES Berry, Craig. "The Relationship Between Economic Growth And Population Growth". SPERI British Political Economy Brief 7 (2014): n. pag. Print. Arnott, Robert D., and Denis B. Chaves. “Demographic Changes, Financial Markets, and the Economy.” Financial Analysts Journal, vol. 68, no. 1, 2012, pp. 23–46. www.jstor.org/stable/-. Barro, Robert J. "Determinants of Economic Growth: A Cross-Country Empirical Study." NBER Working Paper, 1998. Web. 29 Oct. 2016. Fernald, John, Eric Hsu, Mark M. Spiegel. 2015. “Is China Fudging its Figures? Evidence from Trading Partner Data.” Federal Reserve Bank of San Francisco Working Paper 2015-12. http://www.frbsf.org/economic-research/publications/working-papers/wp2015-12.pdf Group, T. W. B. (2016). GDP per capita (current US$), China. Retrieved December 12, 2016, from World Bank, http://data.worldbank.org/indicator/NY.GDP.PCAP.CD?locations=CN Imran, M., Mughal, K., Salman, A., & Makarevic, N. (2015). Unemployment and Economic Growth of Developing Asian Countries: A Panel Data Analysis. European Journal of Economic Studies, 13(3), 147-160. doi:-/es- Perez-Rodríguez, J. V., Ledesma-Rodríguez, F., & Santana-Gallego, M. (2014, December 13). Testing dependence between GDP and tourism's growth rates. Tourism Management, 48, 268-282. Retrieved September 12, 2016. http://dx.doi.org/10.1016/j.tourman- Pryor, F. (2010). Medium-Term Economic Growth in the Caribbean. Social and Economic Studies, 59(3), 127-140. Retrieved from http://www.jstor.org/stable/- Sala-i-Martin, X., Doppelhofer, G., & Miller, R. I. (2004, September). Determinants of Long-Term Growth: A Bayesian Averaging of Classical Estimates (BACE) Approach. The American Economic Review, 94(4), 813-835. http://dx.doi.org/10.1257/- Sparshott, Jeffrey. "WSJ Survey: China's Growth Statements Make U.S. Economists Skeptical." Wall Street Journal. Wall Street Journal, 11 Sept. 2015. Web. 25 Oct. 2016. Sunny, D., & Unnikrishnan, A. (2015, April 3). Pattern of capital inflows in BRICS: Aspects to Ponder for Policy Implementation. The Journal of Developing Areas, 49(6), 219-234. doi:10.1353/jda- Takamine, T. (2006). The Political Economy of Japanese Foreign Aid: The Role of Yen Loans in China's Economic Growth and Openness. Pacific Affairs, 79(1), 29-48. Retrieved from http://www.jstor.org/stable/- Verme, P. (2010). A Structural Analysis of Growth and Poverty in the Short-Term. The Journal of Developing Areas, 43(2), 19-39. Retrieved from http://www.jstor.org/stable/- Yeyati, Eduardo Levy and Martin, Tomas. “Financial Globalization in Emerging Economies: Much Ado about Nothing?” Economia 14.2 (2014): 91-131. Web. APPENDIX TABLE 1: DEFINITIONS OF VARIABLES Variable Explanation Expected Sign Dependent RGDPCH Change in real GDP per capita N/A Independent ENERGYCON Energy consumption in kilograms of oil and oil equivalents (+) FERTRATE Average births per woman (-) CO2 CO2 emissions in metric tons per capita (+) GRSSCAPFRM Addition to fixed assets in the economy (+) LOGGDP Natural log of GDP (+) LOGPOP Natural log of the population ? URBAN Fraction of people living in urban areas (+) MORTRATE Infant mortality rate (-) TABLE 2: DESCRIPTIVE STATISTICS GDP CO2 ENRGYCON FERTRATE GRSSCAPFRM MORTRATE POP URBAN  Mean  -  -  -  -  7.11E+11  -  1.15E+09  -  Median  -  -  -  -  1.77E+11  -  1.17E+09  -  Maximum  -  -  -  -  4.52E+12  -  1.36E+09  -  Minimum  -  -  -  -  3.52E+10  -  8.62E+08  -  Std. Dev.  -  -  -  -  1.14E+12  -  1.54E+08  -  Skewness  -  -  -  -  -  - -  -  Kurtosis  -  -  -  -  -  -  -  -  Jarque-Bera  -  -  -  -  -  -  -  -  Probability  -  -  -  -  -  -  -  -  Sum  -  -  -  -  2.99E+13  -  4.82E+10  -  Sum Sq. Dev.  -.  -  -.  -  5.29E+25  -  9.74E+17  -  Observations  42  42  42  42  42  42  42  42 TABLE 3: STARTING REGRESSION, EQUATION (1): ALL VARIABLES Dependent Variable: GDPCH Method: Least Squares Date: 12/06/16 Time: 14:14 Sample:- Included observations: 42 Variable Coefficient Std. Error t-Statistic Prob.   C- ENRGYCON - - FERTRATE -**- - CO- GRSSCAPFRM -1.29E-13** 5.27E-14 - LOGGDP-**- LOGPOP - - URBAN- MORTRATE-**- R-squared-     Mean dependent var- Adjusted R-squared-     S.D. dependent var- S.E. of regression-     Akaike info criterion - Sum squared resid-     Schwarz criterion - Log likelihood-     Hannan-Quinn criter. - F-statistic-     Durbin-Watson stat- Prob(F-statistic- *Significant at the 10% level **Significant at the 5% level ***Significant at the 1% level TABLE 4: FINAL REGRESSION, EQUATION (2): ALL SIGNIFICANT VARIABLES *Significant at the 10% level **Significant at the 5% level ***Significant at the 1% level