Chapter 2. On Trade Surplus from Property Rights 35
3.6 Estimate of the Proportion of the Service
Compared with agriculture and manufacturing, the service industry has two characteristics. First, the vast majority of services are sup- plied and consumed in the same place and at the same time. The spatial distribution of the demands is very important to the service industry. Services gather in the city and create the scale economy.
Usually, restaurants, bookstores, and banks are opened together to gain a stream of prosperous people, thus causing businesses to boom.
Therefore, the market scale has a greater influence on the service industry than on the manufacturing industry. Second, among the element inputs in the service industry, labor accounts for a larger
Interpreting the Economic Scale of China 79
proportion than capital. The workers in the service industry and their face-to-face interaction with the customers promote the development of the service industry. Moreover, the population density and the concentration of labors and enterprises (the level of urbanization) decide the development of the service industry to a great extent.6 To estimate the proportion of the service industry in the GDP, three related variables are selected: Per capita GDP level, urbanization rate, and population density.
According to the world development index (WDI) of the World Bank, the panel data of 112 countries from 1980 to 2008 are selected for the regression analysis. Through the Hausman Specification Test, the hypothesis of setting the regression equation as the fixed effect random model cannot be refused. It is set to be a simple binary regression model (Equation (1)). The proportions of the service industry in the GDPs of different countries serve as the dependent variable, and the per capita GDP, urbanization rate, and population density are the independent variables. Here, the natural logarithms of the dependent and independent variables are taken to reduce the random error terms caused by the lack of homoscedasticity of the model and the resulting heteroscedasticity.
lnSerit=α0+β1lnGDP pc pppit+β2lnUrbanit
+β3lnP OP denit+γit∗ +εi, (1) (i= 1,2, . . . ,112, t= 1980,1981, . . . ,2008),
where,Seris the proportion of the value-added of the service industry in the total output;GDPpc ppp is the per capita GDP calculated by the PPP;Urbanis the urbanization rate, that is, the proportion of the urban people in the total population; andPOPdenis the population density, that is, the population per square kilometer.
Before regressing using Equation (1), the co-integration test is conducted on the variables to decide whether the linear combina- tion of the non-stationary series has the co-integration relationship, thus judging whether the linear regression equation is set reasonably
6The Scientific Research Institute of the National Bureau of Statistics, China’s Service Industry in Perspective of New Economic Geography, the Research Topic Group of the Wage Difference in China’s Service Industry, Aug. 22, 2011.
80 From Trade Surplus to the Dispute over the Exchange Rate
Table 3.7: Co-integration test of the proportion of the service industry.
Kao test H0:ρ= 1 ADF −1.799 (0.0036)∗
Pedroni test H0:ρ= 1; Panel v-statistic −1.540 (0.0382)∗ H1:(ρi=ρ)<1 Panel rho-statistic 2.639 (0.0580)∗ Panel PP-statistic −4.087 (0.0000)∗ Panel ADF-statistic −2.493 (0.0000)∗ H0:ρ= 1; Group-rho-statistic 6.316 (0.3843) H1:(ρi=ρ)<1 Group-PP-statistic −5.508 (0.0000)∗ Group-ADF-statistic −4.167 (0.0000)∗
and whether the stable equilibrium relationship exists between the dependent variables and the explanatory variables of the regression equation. The co-integration test method proposed by Engle and Granger (1987)7 is applied to perform the ADF test on the residual of Equation (1). The results are shown in Table 3.7.
According to the report of the World Bank, the countries in the world are classified into different groups according to the per capita income, namely, the low-income, the lower middle-income, the upper middle-income, and the high-income countries. The pooled least square method is applied to perform the regression of difficult groups of Equation (1). The results are shown in Table 3.8.
The coefficient estimates in Table 3.8 are placed in Equation (1) to obtain the regression equations of the different income groups as follows.
The low-income countries:
lnSerit= 1.954∗∗∗+ 0.217∗∗∗lnGDP pc pppit+ 0.117∗∗∗lnUrbanit
−0.022 lnP OP denit+γit∗ +εi (2) The lower middle-income countries:
lnSerit= 2.418∗∗∗+ 0.221∗∗∗lnGDP pc pppit
+ 0.013 lnUrbanit−0.076∗∗∗lnP OP denit+γit∗ +εi (3)
7Robert F. Engle and C. W. J. Granger (1987). Co-integration and error correction:
Representation, estimation and testing.Econometrica, 55, 251–276.
FromTradeSurplustotheDisputeovertheExchangeRate9inx6inb2268-ch03page81
InterpretingtheEconomicScaleofChina81
Table 3.8: Pooled least square regression results of the proportion of the service industry.
Constant Urbanization Population Per capita Number
Different groups term rate density GDP of sample R2 F statistics
All samples 1.918∗∗∗ 0.053∗∗∗ 0.036∗∗ 0.198∗∗∗ 3244 0.762 92.390 (0.104) (0.014) (0.015) (0.011)
Low-income countries 1.954∗∗∗ 0.117∗∗∗ −0.022 0.217∗∗∗ 551 0.499 26.710 (0.186) (0.035) (0.039) (0.026)
Lower middle-income countries 2.418∗∗∗ 0.013 −0.076∗∗∗ 0.221∗∗∗ 1189 0.746 82.388 (0.154) (0.022) (0.021) (0.017)
Upper middle-income countries 1.084∗∗∗ 0.118∗∗∗ 0.095∗ 0.229∗∗∗ 812 0.671 56.013 (0.268) (0.035) (0.041) (0.027)
High-income countries 2.079∗∗∗ 0.035 0.328∗∗∗ 0.055∗ 663 0.809 113.37 (0.308) (0.026) (0.029) (0.022)
Explication: The per capita GDP level here is converted using the PPP. The figures in the brackets represent the standard deviation of the regression coefficient; ∗means that the original hypothesis is accepted at the 1% significant level;∗∗means that the original hypothesis is accepted at the 5% significant level; ∗∗∗means that the original hypothesis is accepted at the 10% significant level.
Low-income, countries refer to those whose per capita income is below $1,005; lower middle-income countries refer to those whose per capita income is between $1,006 and $3,975; upper middle-income countries refer to those whose per capita income is between $3,976 and $12,275; and high-income countries refer to those whose per capita income is more than $12,276.
82 From Trade Surplus to the Dispute over the Exchange Rate
The upper middle-income countries:
lnSerit= 1.084∗∗∗+ 0.229∗∗∗lnGDP pc pppit
+ 0.118∗∗∗lnUrbanit+ 0.095∗lnP OP denit+γ∗it+εi (4) The high-income countries:
lnSerit= 2.079∗∗∗+ 0.055∗lnGDP pc pppit+ 0.035 lnUrbanit + 0.328∗∗∗lnP OP denit+γit∗ +εi (5) The coefficients of the regression model pass the significance test, and the estimates and the direction of change are in accordance with the theoretical assumptions and experience judgment of the new eco- nomic geography.
First, the growth of the per capita income level and the devel- opment of urbanization will promote the development of the service industry to different extents. For the low-income and lower middle- income countries, the increase in population density is not conducive to the development of the service industry. In the take-off stage, an excessively dense population will produce many social problems, such as public security, environmental pollution, and employment pressure, which hinder the construction of the social service system to a certain extent.
Second, in any group, the per capita GDP growth and the elastic- ity coefficient of the development of the service industry are signifi- cantly positive. In other words, the rising income level will increase the demands for various services, but the correlation coefficients of the different groups are different. In low-income countries, an increase of 1% in the per capita income pulls the proportion of the service industry to increase by 0.217%. In the lower middle-income coun- tries, an increase of 1% in the per capita income pulls the proportion of the service industry to increase by 0.221%. The most significant pulling effect is in the upper middle-income countries (up to 0.229%), whereas the least significant is in the high-income countries (0.055%).
As the service industry has been developing maturely in high-income countries, and its proportion in the GDP has been more than 70%, the growth of the per capita income will not significantly drive the increase in the proportion of the service industry.
Interpreting the Economic Scale of China 83
According to the WDI database of the World Bank, if the cal- culation was based on the constant price in 2000, the per capita GDP level of China would have increased significantly from $186 in 1980 to $2,206 in 2009. Statistically, within 20 years, the proportion of the service industry in the GDP of China would have increased from 22% to 43% in 2009, which is an average annual growth rate of only 0.046%. This rate is significantly lower than the per capita GDP change of 0.54%. Based on the changes in the time series of the development of the service industry in China, the proportion of the service industry in the GDP had two periods of stagnation, that is, from 1996 to 1998 and from 2002 to 2008. During these two periods, the per capita GDP maintained a high growth momentum, but the proportion of the service industry in the GDP remained unchanged.
This situation deviates from the general law of economy. The only explanation for this is that the statistics of the service industry during these two periods were inaccurate.
China is classified as a lower middle-income country, and the elas- ticity coefficient of its per capita GDP growth rate and the change in the proportion of the service industry should be 0.221, accord- ing to the estimate of Equation (3). In other words, the average annual change rate of the proportion of the service industry should be 0.119% and not 0.046% (0.221×0.54 = 0.119).
With 1980 as the starting point, adjustments are made on the proportion of the service industry in China over the years according to the elasticity estimated above. The results and the change trend are shown in Table 3.9 and Fig. 3.2, respectively.
Based on Table 3.9, the lag of the development of the service industry occurred in the 1990s. From 1990 to 2009, the per capita GDP of China (calculated based on the constant price in 2000) increased from $391 to $2,206. In theory, the proportion of the service industry in the GDP of China should increase from 31.5%
to 52.8%.
Considering the relationship between the time series of the service industry development in China and the RMB exchange rate in 2011, the development of the service industry has two stagnations (i.e., from 1992 to 1996 and from 2002 to 2008). If the service industry
84 From Trade Surplus to the Dispute over the Exchange Rate
Table 3.9: Comparison of the proportion of the service industry before and after adjustment (%).
Before adjustment After adjustment
1980 22.0 22.0
1981 22.0 22.5
1982 21.9 23.4
1983 22.4 24.5
1984 24.8 26.1
1985 28.7 27.5
1986 29.1 28.4
1987 29.6 29.5
1988 30.5 30.7
1989 32.1 31.0
1990 31.5 31.2
1991 33.7 32.2
1992 34.8 33.7
1993 33.7 35.2
1994 33.6 36.6
1995 32.9 37.8
1996 32.8 38.8
1997 34.2 39.8
1998 36.2 40.6
1999 37.8 41.4
2000 39.0 42.3
2001 40.5 43.2
2002 41.5 44.2
2003 41.2 45.3
2004 40.4 46.4
2005 40.5 47.7
2006 40.9 49.1
2007 41.9 50.7
2008 41.8 51.8
2009 43.4 52.8
Source: The data before the adjustment are from theChina Statistical Yearbook2011, and those after the adjustment are from the results of the analysis above.
developed normally during these two periods, the proportion of the service industry in the GDP of China would have increased to 58%, which is an increase of 16% compared with that in 2009. This rate is close to the level of the upper middle-income countries rather than
Interpreting the Economic Scale of China 85
0 10 20 30 40 50 60
80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07
Before Adjustment After Adjustment
Figure 3.2: Adjustment of the proportion of the service industry in the GDP.
GDP per capita Service % GDP
62.4%
55.2%
43%
47.1%
48.9%
A0 A1 A2 A3
The rational area of China’s service sector
Figure 3.3: Reasonable proportion of the service industry based on the per capita GDP level.
Definition of indicator: A0is the per capita GDP level of least developed countries, A1is the per capita GDP level of low-income countries, A2is the per capita GDP level of lower middle-income countries, and A3 is the per capita GDP level of upper middle-income countries.
Source: WDI, the World Bank, 2011.
86 From Trade Surplus to the Dispute over the Exchange Rate
to the level even lower than that of low-income countries, as shown by the current statistics shown (Xu, 2011).8
When the Atlas method is applied, the per capita GDP of China is equivalent to that of low-income countries, but when the PPP is used, the figure will be equivalent to that of the upper middle-income countries. Therefore, the proportion of the tertiary industry in China should range from 55.2% to 62.4% (Fig. 3.3).
If the proportion of the service industry in China in 2009 increased from 43.4% to 52.8% (according to the simulation results in Table 3.9), the GDP should increase from $8.3 trillion to $9.9 trillion based on the constant price in 2005; this is an adjustment range of approximately 20%. If it was raised to 55.2% (the average level of the lower middle-income countries), the GDP of China should be adjusted to $10.4 trillion, and its economic scale should increase by 26.4%. If it was raised to 62.4% (the average level of upper middle- income countries), the GDP of China should reach $12.4 trillion.9
If the proportion of the service industry in the GDP of China was increased to 52.8%, (Scenario 1), the economic scale of Japan would only be equivalent to 38.2% of that of China; if it was increased to 55.2% (Scenario 2), the economic scale of Japan would be equiv- alent to 36.3% of that of China; and if it was increased to 62.4%
(Scenario 3), the economic scale of Japan would be equivalent to 30.5% of that of China. Under different assumptions of the propor- tion of the service industry in China, the GDP of China is 2.17 to 3.28 times higher than that of Japan according to the PPP (Table 3.10).
8Xu Jianguo (2011). RMB depreciation and stagnant development in China’s service sector.The World Economy, 3, 3–20. (in Chinese).
9The adjustment of the GDP is conducted in two steps. First, 100 minus the proportion of the service industry estimated above is the relative scale of the primary and secondary industries. Assuming that the relative proportion of the primary and the secondary industries is unchanged, the proportion of the primary industry after the adjustment of the service industry can be obtained. Second, the current output of the primary industry divided by its adjusted proportion in the GDP is the current adjusted GDP. Here, the agricultural output, instead of the industrial output, is taken as a frame of reference to estimate the GDP because of the incomplete tax system of China and the loss of statistical information. In other words, the statistics of the service industry coincides with that of the industry in different vocations. Compared with that of industry, the service degree of agriculture is low. Thus, the adjustment of the GDP with agriculture as the frame of reference can reduce the repeated calculation.
Interpreting the Economic Scale of China 87
Table 3.10: GDPs of China in the different service industries (2009).
Proportion of the Service Industry (%)
GDP with the PPP (Based on the constant price in 2005, 100 million $)
Before adjustment 43.4 82553.3
After adjustment 52.8 99064.8
55.2 104371.9
62.4 124358.1
If the proportion of the service industry in China was increased to the level of upper middle-income countries (i.e., up to 62.4% (Sce- nario 3)), in 2009, the U.S. economic scale would only be 103.1% of that of China. That is, the GDPs of these two countries would have been very close. Clearly, such an estimate is difficult to accept. We do not insist on it either. However, this possibility cannot be excluded in the research (Table 3.11).
In recent years, many studies at home and abroad have discussed the position of the GDP scale of China in the world. China has surpassed Japan in the economic scale and has become the second- largest economy in the world, which is non-controversial. People have begun discussing when China would overtake U.S. to become the largest economy in the world.
Prof. Yao Yang of the China Center for Economic Research of Peking University pointed out that a series of assumptions is required to estimate when China will overtake the U.S. in terms of the GDP.
If the economic growth rate of China is maintained at 8% while that of the U.S. is maintained at 3%, the inflation rate of the former will be 3.6% and that of the latter will be 2%; the annual appreciation of RMB against the dollar? will be 3%; and the GDPs of both countries will be $24 trillion in 2021. If the economic growth rate of China was maintained at 9% to 10% five years ago and is 6% to 7% after five years, China would also catch up with the U.S. in 2021.10
10See Yao (2012). When will China’s economy overtake America’s?China Daily.
FromTradeSurplustotheDisputeovertheExchangeRate9inx6inb2268-ch03page88
TradeSurplustotheDisputeovertheExchangeRate
1 The U.S. 128209 155.3 128209 129.4 128209 122.8 128209 103.1
2 China 82553 100.0 99064 100.0 104372 100.0 124358 100.0
3 Japan 37876 45.9 37876 38.2 37876 36.3 37876 30.5
4 India 34308 41.6 34308 34.6 34308 32.9 34308 27.6
5 Germany 26410 32.0 26410 26.7 26410 25.3 26410 21.2
6 The UK 19879 24.1 19879 20.1 19879 19.0 19879 16.0
7 Russia 19308 23.4 19308 19.5 19308 18.5 19308 15.5
8 France 19089 23.1 19089 19.3 19089 18.3 19089 15.4
9 Brazil 18317 22.2 18317 18.5 18317 17.5 18317 14.7
10 Italy 16005 19.4 16005 16.2 16005 15.3 16005 12.9
11 South Korea 12427 15.1 12427 12.5 12427 11.9 12427 10.0
12 Canada 11663 14.1 11663 11.8 11663 11.2 11663 9.4
13 Indonesia 8768 10.6 8768 8.9 8768 8.4 8768 7.1
14 Turkey 8386 10.2 8386 8.5 8386 8.0 8386 6.7
15 Poland 6373 7.7 6373 6.4 6373 6.1 6373 5.1
16 Thailand 4920 6.0 4920 5.0 4920 4.7 4920 4.0
17 Pakistan 4020 4.9 4020 4.1 4020 3.9 4020 3.2
18 Malaysia 3495 4.2 3495 3.5 3495 3.3 3495 2.8
19 Philippines 2958 3.6 2958 3.0 2958 2.8 2958 2.4
20 Romania 2319 2.8 2319 2.3 2319 2.2 2319 1.9
21 Czech 2318 2.8 2318 2.3 2318 2.2 2318 1.9
22 Bulgaria 2086 2.5 2086 2.1 2086 2.0 2086 1.7
23 Hungary 1693 2.1 1693 1.7 1693 1.6 1693 1.4
24 Zimbabwe 869 1.1 869 0.9 869 0.8 869 0.7
Source: The original data are from the WDI, the World Bank, 2011. The data after adjustment are based on the simulation results above.
Interpreting the Economic Scale of China 89
The IMF predicted in the World Economic Outlook released in April 2011 that China would overtake the U.S. in the economic scale to become the largest economy in the world in 2016. At that time, the GDP of China will reach $18.7 trillion, whereas that of the U.S.
will be $18.3 trillion.
Prof. Robert Feenstra of UC Davis calculated the GDP using the revenue method. According to him, the World Bank used the price indexes of the towns and the surrounding areas in estimating the real GDP so the result could be underestimated by 50%. Correcting this indicator means that the real GDP of China will overtake that of the U.S. in a shorter period of time, that is, in 2012 or 2013, instead of 2016 as predicted by the IMF (Table 3.12).11
Another American professor Arvind Subramanian estimated that China had become the largest economy in the world as early as 2010.
According to the argument above, if the Atlas method was used for calculation, the GDP of China would be $5.9 trillion in 2010 based on the data of the World Economic Outlook of the IMF, and it would overtake Japan ($5.5 trillion) to become the second-largest economy in the world for the first time. If the proportion of the service industry in China was adjusted, the GDP of China would have been
$5.6 trillion in 2007, which is higher than that of Japan ($4.4 trillion) using the Atlas method. China’s GDP was approximately $3.3 trillion in 2001, which is higher than that of Japan ($3.3 trillion) using the PPP. According to the data after the adjustment of the proportion
Table 3.12: Comparison of the GDPs between China and the U.S. calculated by the American professor.
2005 2008 2011 2012 2013
The U.S. 12364 12716 13078 13201 13325
China 6863 8916 11583 12639 13791
Source: Robert Feenstra (2012), How Big is China? Economics (Quarterly Publication), 11(2), p. 367.
11See Robert Feenstra (2012). How big is China?Quarterly Journal of Economics, No. 1.
(in Chinese)
90 From Trade Surplus to the Dispute over the Exchange Rate
0 2000 4000 6000 8000 10000 12000 14000 16000 18000 20000
1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016
U.S. China
China(adjustment) Japan
Figure 3.4: Proportions of the service industry in China, Japan, and the U.S.
using the Atlas method and the GDP scale of China after adjustment.
Definition of Data: The data during 2011 and 2016 are the predictions of the IMF.
Source: World Economic Outlook, World Economic Outlook, 2011 ed., IMF.
of the service industry, the GDP of China ($3.3 trillion) would be higher than that of Japan ($3.1 trillion) (Fig. 3.4).
By comparing the GDPs of China and the U.S., the IMF predicted that China would overtake the U.S. to become the largest economy in the world in GDP for the first time in 2016 using the PPP. If the GDP is measured after the adjustment, the GDP scale of China would have reached $14 trillion in 2008, which is higher than that of the U.S. ($14 trillion), and thus it would have been the largest economy in the world (Fig. 3.5).
The above estimate has a very strict premise, that is, China will not suffer from a financial crisis in the future. The economy of Japan grew rapidly in the 1980s; thus, many people speculated when Japan would overtake the U.S. However, when the bubble economy of Japan collapsed in 1990, Japan remained in trouble, and its economic growth rate was nearly zero. As a result, it became lost for 20 years.
At present, no one can say that Japan is on top. If China falls into the same trap, all the previous estimates will be meaningless.