4.1 Changes of traffic accessibility and equilibrium level along Beijing-Shanghai HSR
4.11 Changes of traffic accessibility
This paper calculates the economic potential of the cities along the Beijing-Shanghai HSR according to Eq. 3. The results are shown in Table 2. Due to space limitations, this paper only lists the results of 2009 and 2016 for comparison, and only lists the calculation results including self-potential. The first two columns in Table 2 are calculated based on the minimum railway travel time, and the last two columns are calculated based on the average railway travel time. These results all take into account the impact of train frequency.
As can be seen from Table 2, with the continuous construction of HSR, from 2009 to 2016, the economic potential calculated by the minimum travel time increased by 96.4% on average, and the four-hour economic potential increased by 93.7% on average. The economic potential calculated by the average travel time increased by 76.5% and the four-hour economic potential increased by 70.9% on average. On the whole, the accessibility level of the cities along the Beijing-Shanghai HSR has risen dramatically under the impetus of the national railway construction, regardless of the calculation criteria.
Table 2 calculates the economic potential of cities based on the minimum travel time. According to the calculation results in Table 2, the greatest change in economic potential is not the regional core cities, such as Shanghai, Beijing, Tianjin and other cities. On the contrary, economic potential of some medium-sized cities have changed greatly, which are two situations as follow: one is some cities are close to core cities (usually the provincial capital), these cities will benefit from the externalities of convenient transportation in big cities. Therefore the traffic accessibility of these cities has been improved a lot, such as Cangzhou, Suzhou, Wuxi and so on; the other is some medium-sized city located at the middle of the Beijing Shanghai HSR, railway construction has greatly shortened the travel time of these cities to other core cities, which gradually reflects the location advantages of these cities, such as Xuzhou, Dezhou and so on. However, for some small cities far away from the core city, such as Suzhou, Zaozhuang and so on, the improvement of their accessibility is limited. This situation is more obvious, especially for the calculation results based on the average travel time. For example, the economic potential of Suzhou base on the above results increased by only 35.5%, while the economic potential of four hours increased by only 22.3%. In addition, for the megacities such as Beijing and Shanghai, because their original traffic accessibility level is high, it is very difficult to continue to improve significantly. Furthermore, the population size of megacities was strictly controlled in recent years, so their economic potential has not increased significantly.
Table 3
Economic potential of the cities along Beijing-Shanghai HSR based on travel time
| Based on the minimum Travel Time by railway | Based on the average travel time by railway |
| Economic Potential(m1) (Unit: Millions of people /per hour) | Four Hours Economic Potential(fourm1) (Unit: Millions of people /per hour) | Economic Potential(m2) (Unit: Millions of people /per hour) | Four Hours Economic Potential(fourm2) (Unit: Millions of people /per hour) |
cities | 2009 | 2016 | change(%) | 2009 | 2016 | change(%) | 2009 | 2016 | change(%) | 2009 | 2016 | change(%) |
Beijing | 34.7 | 49.5 | 42.5 | 33.8 | 46.5 | 37.5 | 33.9 | 46.4 | 37.0 | 32.8 | 43.3 | 31.8 |
Langfang | 5.1 | 9.4 | 84.6 | 4.7 | 8.5 | 80.6 | 4.9 | 8.0 | 62.8 | 4.6 | 7.2 | 57.3 |
Tianjin | 21.9 | 36.3 | 66.2 | 21.1 | 33.5 | 59.0 | 20.9 | 32.6 | 56.3 | 20.1 | 30.3 | 50.5 |
Cangzhou | 4.9 | 12.2 | 148.7 | 4.4 | 10.9 | 145.0 | 4.3 | 8.7 | 103.6 | 3.9 | 7.4 | 91.1 |
Dezhou | 5.4 | 14.0 | 157.0 | 4.8 | 12.2 | 154.2 | 4.6 | 10.7 | 132.5 | 3.7 | 9.2 | 145.5 |
Jinan | 11.5 | 25.7 | 123.7 | 10.7 | 22.7 | 111.4 | 10.8 | 21.7 | 101.2 | 9.7 | 18.4 | 90.4 |
Tai’an | 7.2 | 12.2 | 69.2 | 6.3 | 10.6 | 68.1 | 6.6 | 9.6 | 44.5 | 5.9 | 7.8 | 30.9 |
Zaozhuang | 7.3 | 11.6 | 58.9 | 6.9 | 10.8 | 58.0 | 7.2 | 10.0 | 40.0 | 6.7 | 8.4 | 24.8 |
Xuzhou | 8.9 | 21.7 | 142.9 | 7.4 | 20.6 | 177.3 | 7.9 | 16.7 | 111.4 | 6.4 | 11.8 | 84.8 |
Suzhou | 7.0 | 10.8 | 54.4 | 6.4 | 9.8 | 53.7 | 6.6 | 9.0 | 35.5 | 6.0 | 7.4 | 22.3 |
Bengpu | 7.0 | 13.8 | 96.2 | 5.6 | 12.4 | 123.4 | 5.9 | 9.7 | 63.6 | 4.7 | 6.8 | 45.6 |
Chuzhou | 4.5 | 9.0 | 97.9 | 4.2 | 7.8 | 85.0 | 4.0 | 6.7 | 68.0 | 3.5 | 4.8 | 37.8 |
Nanjing | 17.5 | 38.6 | 120.1 | 16.8 | 35.9 | 113.5 | 15.8 | 30.5 | 92.8 | 13.7 | 27.7 | 102.9 |
Zhenjiang | 8.8 | 22.2 | 152.7 | 8.4 | 21.1 | 152.8 | 7.2 | 16.1 | 123.7 | 5.7 | 14.9 | 160.4 |
Changzhou | 12.2 | 30.7 | 152.3 | 11.6 | 29.5 | 153.3 | 10.4 | 23.9 | 129.3 | 10.0 | 22.4 | 124.2 |
Wuxi | 13.3 | 34.9 | 161.9 | 12.8 | 33.5 | 161.9 | 11.4 | 27.0 | 136.9 | 10.9 | 25.6 | 134.0 |
Suzhou | 13.1 | 38.3 | 193.1 | 12.7 | 36.7 | 189.8 | 11.3 | 31.4 | 179.1 | 10.9 | 29.8 | 174.3 |
Shanghai | 37.0 | 55.8 | 50.8 | 36.3 | 53.2 | 46.4 | 35.9 | 51.2 | 42.5 | 34.9 | 48.7 | 39.5 |
mean | 12.6 | 24.8 | 96.4 | 11.9 | 23.1 | 93.7 | 11.6 | 20.6 | 76.5 | 10.8 | 18.4 | 70.9 |
According to the calculation results based on the minimum travel time and the average travel time by railway, the change of the two is similar except the increase of the latter is relatively small. However, for some cities, there is a big difference between the two kinds of calculation results. For example, for the four-hour economic potential based on the minimum travel time by railway, the three cities with the greatest increase are Suzhou, Xuzhou and Wuxi, but in the calculation results based on the average travel time by railway, the three cities with the greatest increase are Suzhou, Zhenjiang and Dezhou. Except for Suzhou, the order of other cities has changed. This result shows that it should not to assess the change of traffic accessibility only based on the result by the minimum travel time. For example, the minimum travel time between some cities has been greatly shortened, but the average travel time has not increased much, so the real situation is not as significant as the calculation results of the former.
4.12 Changes in the equilibrium degree of traffic accessibility
This paper use two commonly methods to measure the change of the equilibrium degree of traffic accessibility along the Beijing-Shanghai HSR, namely the coefficient of variation and the Gini coefficient. If the absolute change of two coefficients is negative, the equilibrium degree will increase, and vice versa. In this paper, the corresponding calculation is shown in Table 3 and in Table 4.
According to Table 4, it can be found that from 2009 to 2016, no matter what method is adopted, and the degree of equilibrium increased. On the whole, as the calculation of economic potential is a combination of population size and transportation cost. Because the population size of the core cities is large, so the economic potential of mega-cities is always at the forefront, but the gap of traffic accessibility between large cities and small-medium cities has narrowed.
Table 4
Non-equilibrium index of traffic accessibility along Beijing-Shanghai HSR
| Based on Minimum Railway Travel Time | | Based on Average Railway Travel Time |
| 2009 | 2016 | Change | | 2009 | 2016 | change |
Coefficient of variation | | | | | | | |
Economic Potential | 0.762 | 0.589 | -0.173 | | 0.815 | 0.659 | -0.156 |
Four Hours Economic Potential | 0.803 | 0.609 | -0.194 | | 0.871 | 0.721 | -0.150 |
Gini coefficient | | | | | | | |
Economic Potential | 0.363 | 0.316 | -0.047 | | 0.379 | 0.345 | -0.034 |
Four Hours Economic Potential | 0.382 | 0.326 | -0.056 | | 0.402 | 0.376 | -0.026 |
In addition, according to the calculation results based on average travel time, the equilibrium degree of economic potential is relatively lower, and its decline is limited, which also shows that it is inaccurate to appraise the change of the equilibrium degree of economic potential only by the calculation result based on the minimum travel time. Although the minimum travel time between many cities has been greatly reduced, the decrease of average travel time is relatively small. Meanwhile, there are always great differences in population size between cities. Therefore, the increase of equilibrium degree of economic potential is not as significant as the result calculated by the minimum travel time.
4.2 The impact of the improvement of traffic accessibility on the economic growth of cities along the line
Based on Eq.2, this paper uses panel fixed effect estimation method to analyze the impact of economic potential (representing the level of traffic accessibility) on the economic growth of cities along Beijing-Shanghai HSR. Hsiao C. (2007) pointed out that panel data contains more information, more changes and less collinearity between variables than cross-section data. In addition, the application of panel data leads to greater degrees of freedom, which enhances the efficiency of regression estimation.
According to the common processing of panel data estimation, Hausman test is carried out at first. The results show that the fixed effect should be adopted (see Table 4), which also conforms to the theoretical basis underlying the economic growth model. Second, in order to deal with the common cross-sectional heteroscedasticity problem, robust standard error is used in this paper. Third, as there is certain collinearity between various forms of economic potential, it is substituted into the model one by one for regression (see Table 4). Meanwhile, The cross-product term of travel time and economic potential is also substituted into the regression model. For example, TMP1 is the cross-product of T1ik (the minimum travel time from each city to the nearest core city k by railway, taking natural logarithm) and mpl, TMP2 is the product of T2ik (the average travel time from each city to the nearest core city k by railway, also taking natural logarithm) and mp2, and so on. Furthermore, in order to control for endogenous problem that are common in economic analysis as much as possible, all kinds of economic potentials are regressed using their first order lag form, and all models control the time fixed effect and regional fixed effect. The estimated results are shown in the table 4.
Table 4
Estimated results of the impact of traffic accessibility on economic growth.
variables | (1) | (2) | (3) | (4) | (5) | (6) | (7) | (8) |
K | 0.214*** | 0.213** | 0.202*** | 0.199** | 0.210** | 0.209** | 0.200*** | 0.209** |
| (0.071) | (0.074) | (0.066) | (0.072) | (0.073) | (0.079) | (0.067) | (0.076) |
L | 0.150* | 0.140* | 0.113* | 0.129* | 0.191* | 0.182* | 0.116* | 0.164* |
| (0.117) | (0.119) | (0.109) | (0.113) | (0.116) | (0.122) | (0.111) | (0.119) |
G | 0.064** | 0.061** | 0.052** | 0.060** | 0.079** | 0.075** | 0.058** | 0.082** |
| (0.025) | (0.025) | (0.022) | (0.025) | (0.027) | (0.026) | (0.024) | (0.030) |
MP1 | 0.175 | | | | | | | |
| (0.129) | | | | | | | |
TMP1 | 0.011 | | | | | | | |
| (0.038) | | | | | | | |
MP2 | | 0.196* | | | | | | |
| | (0.103) | | | | | | |
TMP2 | | 0.009 | | | | | | |
| | (0.031) | | | | | | |
M1 | | | 0.525*** | | | | | |
| | | (0.156) | | | | | |
TM1 | | | -0.010 | | | | | |
| | | (0.030) | | | | | |
M2 | | | | 1.009*** | | | | |
| | | | (0.341) | | | | |
TM2 | | | | -0.116 | | | | |
| | | | (0.076) | | | | |
fourMP1 | | | | | 0.202 | | | |
| | | | | (0.146) | | | |
TfourMP1 | | | | | -0.009 | | | |
| | | | | (0.040) | | | |
fourMP2 | | | | | | 0.133* | | |
| | | | | | (0.073) | | |
TfourMP2 | | | | | | 0.009 | | |
| | | | | | (0.022) | | |
fourM1 | | | | | | | 0.484*** | |
| | | | | | | (0.144) | |
TfourM1 | | | | | | | -0.001 | |
| | | | | | | (0.029) | |
fourM2 | | | | | | | | 1.311** |
| | | | | | | | (0.465) |
TfourM2 | | | | | | | | -0.208* |
| | | | | | | | (0.098) |
constant | 11.265*** | 11.427*** | 10.847*** | 10.705*** | 11.093*** | 11.291*** | 10.841*** | 10.336*** |
| (1.172) | (1.254) | (0.881) | (0.908) | (1.211) | (1.381) | (0.895) | (0.946) |
With-in R2 | 0.871 | 0.870 | 0.879 | 0.875 | 0.859 | 0.855 | 0.876 | 0.863 |
F value | 148.218 | 158.254 | 164.509 | 152.553 | 87.081 | 145.943 | 151.611 | 141.925 |
Hausman | 66.10 | 93.85 | 929.80 | 209.87 | 91.62 | 90.74 | 449.62 | 92.47 |
Observations | 144 | 144 | 144 | 144 | 144 | 144 | 144 | 144 |
note: Robust standard errors in parentheses; *** p < 0.01, ** p < 0.05, * p < 0.1 |
It can be seen that the coefficients of fixed assets investment are positive and significant in table 4. The coefficients of labor force and local government expenditure are the same as above, which are consistent with the expectations of common economic theory.
First, according to the estimated results in Table 4, the coefficients of economic potential in various forms are positive and significant, which is consistent with the previous expectations. That is to say, the construction of railways will bring about an increase in the speed of railways, improving the potential market capacity of cities along the line and promoting the economic growth of these cities. From an economic point of view, on the one hand, the development of railway greatly save travel time and improve economic exchanges between regions, flow of labor force, capital and other factors of production have become more frequent, more product choice for consumers, deepening regional economic integration, lower transaction costs, enhanced technological exchanges and so on. On the other hand, the development of railways has also increased the total volume of traffic flow between cities, thus expanding the market size of various cities. All of these mentioned above accelerate economic growth of the cities along the line.
Second, for the economic potential calculated by the minimum travel time and the average travel time, the coefficient and statistical significance of the latter are higher than those of the former, which shows that the economic growth is more sensitive to the economic potential calculated by the average travel time, that can truly represent the potential market capacity of each city. Only the decrease of the average travel time between cities can truly reflect the reduction of transportation costs, instead of the reduction of the minimum railway time.
Third, for two forms of economic potentials, which includes own potentials and excludes own potentials, both the coefficients and statistical significance of the former are higher than those of the latter. That is to say, for measuring the impact of economic potentials on economic growth, we cannot neglect self-potential of the city itself. Because the market demand of a region includes not only external demand, but also internal demand, which is also very important. Although some cities are far away from other cities, resulting in low external potential, if their own population and economic scale is large, then their internal potential will also promote rapid economic development.
Fourth, all the cross-product terms are not significant or negative, which indicates that the economic spillover of core cities shows a declining trend with distance increase. It also shows that although some relatively remote cities (such as Suzhou, Zaozhuang and so on) have access to HSR, so the travel time to other cities is greatly reduced, the increase of their accessibility is limited compared with other cities closer to core cities. Therefore, it can be concluded that for these small cities far from the core cities, the development of railway only leads to a moderate reduction in transport costs. Due to the small economic scale and poor economic attraction of these cities, their economic resources tend to flow to the surrounding large cities, and the promotion of these cities on economic growth brought by the development of railways is relatively limited.