Jose L. Tongzon, jtongzon@inha.ac.kr Dong Yang, yangdong@nus.edu.sg 3 rd International Workshop on Port Economics and Policy December 9 10, 2013. Singapore
1. Introduction The economic rise of China The growth of Chinese ports The overall average growth of Chinese ports for 1980 2010 29 % South ports (Shenzhen, Guangzhou, Zhongshan, Zhuahai and Zhanjiang) 32 % Middle ports (Shanghai, Ningbo, Xiamen, Liangyungang, Quanzhou and Fuzhou) 29 % North ports (Qingdao, Tianjin, Dalian, Yingkou, Yantai, Rizhao and Qinhuangdao) 27 %
Figure 1: Port throughputs in All China, North, Middle and South China Co ontainer Throughput of Chinese ports (ln TEU) 20 18 16 14 AllChina North ports Middle ports South Ports 12 10 8 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
Period All China North(7) middle(6) South(5) Shanghai Shenzhen Guangzhou Qingdao TianJin Ningbo Xiamen Dalian 1981 1990 38% 39% 36% 46% 33% 68% 40% 32% 38% 83% 53% 35% 1991 1999 31% 25% 31% 37% 29% 73% 30% 32% 20% 46% 42% 23% 2001 2010 21% 21% 21% 19% 19% 20% 23% 19% 20% 31% 19% 18% 85% 88% 103% 80% 160% 93% 83% 78% 161% 113% 77% Period All China North(7) middle(6) South(5) Shanghai Shenzhen Guangzhou Qingdao TianJin Ningbo Xiamen Dalian 1981 1985 54% 60% 52% 63% 48% 62% 31% 67% 142% 35% 1986 1990 21% 18% 20% 30% 19% 68% 17% 33% 10% 83% 17% 36% 1991 1995 30% 25% 32% 36% 27% 73% 35% 35% 20% 49% 55% 23% 1996 2000 31% 24% 30% 39% 30% 73% 26% 29% 20% 43% 29% 23% 2001 2005 28% 25% 29% 29% 27% 31% 31% 24% 23% 42% 25% 22% 2005 2010 13% 17% 14% 9% 11% 8% 14% 14% 16% 21% 12% 15%
Average Growth Rate of Main Chinese Ports 1980 2010 0.6 0.5 0.482038285 0.476997508 0.4 0.343875746 0.361818379 0.3 0.295380962 0.281958215 0.29485261 0.26771866 0.308719917 0.277123058 0.259059526 0.248439605 0.2 0.1 0 All China North(7) middle(6) South(5) Shanghai Shenzhen Guangzhou Qingdao TianJin Ningbo Xiamen Dalian
Table 1: Top 10 container ports in the world in 2010 Ports Container throughput (in million TEUs) 1. Port of Shanghai 29.069 2. Port of Singapore 28.400 3. Port of Hong Kong 23.530 4. Port of Shenzhen 22.510 5. Port of Busan 14.180 6. Port of Ningbo 13.144 7. Port of Guangzhou 12.550 8. Port of Qingdao 12.012 9. Port of Dubai 11.600 10. Port of Rotterdam 11.100 Sources: Websites <http://www.supplychaindigital.com>
TOP 12 WORLD CONTAINER PORTS Rank Port, Country Volume 2012 (Million TEUs) Volume 2011 (Million TEUS) 1 Shanghai, China 32.53 31.74 2 SIngapore,Singapore 31.65 29.94 3 Hong Kong, China 23.10 24.38 4 Shenzhen, China 22.94 22.57 5 Busan, South Korea 17.04 16.18 6 Ningbo-Zhoushan, China 16.83 14.72 7 Guangzhou Harbor, China 14.74 14.42 8 Qingdao, China 14.50 13.02 9 Jebel Ali, Dubai, United Arab Emirates 13.30 13.00 10 Tianjin, China 12.30 11.59 11 Rotterdam, Netherlands 11.87 11.88 12 Port Kelang, Malaysia 10.00 9.60
Did the rise of Chinese ports have a beneficial or negative impact on the major East Asian container ports? Much of this rise was driven by Chinese local cargoes However, with continued investments in Chinese ports in operational efficiency and policy of decentralization (Qiu, 2008), more and more transshipment could have been shifted to China 3
For example, after the reform in 2002 the ports of Shanghai and Shenzhen have had enjoyed two-digit annual growth rates while the ports of Singapore and Hong Kong have had only either low two-digit or one-digit annual growth rates resulting in the port of Shanghai replacing the ports of Singapore and Hong Kong to become the largest container port in the world (Qiu, 2008).
1. Literature Review Seabrooke et al. (2002) Assessed implications of the rise of South Chinese ports for the port of HK and concluded a slower growth for HK Cullinane et al. (2004) More in depth analysis on competition between the ports of Shenzhen and Hong Kong using Robinson s criteria and concluded that port of HK will retain its dominant role. Yap and Lam (2006) Yap et al. (2006) Assessed nature and intensity of inter port competition in East Asia and concluded that inter port competition in the region would intensify but failed to empirically investigate the impact of the rise of Chinese ports on East Asian ports due to unavailability of data. Tongzon and Chang (2007) The implications of the rise of China for the port of Busan Yap and Lam (2011) The relationship between Shenzhen and Hong Kong both competitive and complementary Tongzon (2011) The relationship between Shanghai port and Singapore port
Literature Review There is therefore a need to empirically assess the implications of the growth of Chinese ports based on latest available data Taking into account all major ports in China and East Asia Extending the period of analysis 1980 2010
2. Methodology Container movements through Chinese and major non Chinese ports in East Asia (Singapore, Hong Kong, Kaohsiung, Keelung, Pusan, Yokohama, Tokyo and Nagoya) for 1980 2010 Nature of relationships: substitutes or complements Co-integration analysis is justified and suitable in this study because we want to assess the overall relationships over time for the entire period in consideration. Secondly, we are dealing with time-series data for a considerable period of time, which could lead to spurious or nonsense regressions in time series. This happens because economic time series are dominated by smooth, long term trends. Under this condition it is possible to generate coefficients which make appear to be stationary, although such an empirical result tells us little of the nature of the relationship between variables. Co-integration therefore allows us to examine the average (overall) relationship between ports over time. However, to show the short-term dynamics and speed of adjustment to the long-run relationship between ports, the error correction modeling (ECM) is employed. In cases where two ports are not co-integrated, we use the general-to-specific modeling approach to determine the nature of relationships after making them stationary.
Perform ADF unit root test to check if variables have the same order of integration and ensure the robustness of its results by Phillips Perron Yes Run co integration. No Consult critical values for presence of cointegrating equation. Yes No Variables are not co integrated and therefore no presence of long term relationships. Confirm the presence of long term relationships. Determine if relationship is competitive or complementary based on the signs of co integrating equations. Construct VEC to determine short run dynamics in the relationships. Perform general to specific modeling approach after making all the variables stationary to reveal the relationships between these ports. Figure 4: Diagrammatic representation of the methodology used
If they are co-integrated, then their short-run relationships can be estimated based on error correction modeling (ECM) techniques (Granger, 1988): Port a t = c 1 + α 1 e t-1 + Port b t = c 2 + α 2 e t-1 +
On the other hand, for those variables that are not co-integrated, their short-run relationships can be estimated employing the general-to-specific modelling approach after making them stationary by differencing. Gilbert (1986) outlines the advantages of the general to specific approach originally due to Sargan (1969). This outcome restricts the number of lags included in a regression model from a maximum to those lag terms proving significant in estimation. This method is appropriate here given the lack of co-integration among the variables. Unless the variables are co-integrated, modelling in levels will lead to the problems identified by Granger and Newbold (1974). Hence, the general equation used in this case is follows: Port a =c + αport b + βt + a t-1 + (5)
Since the focus of this paper is on the impact of Chinese ports on non-chinese East Asian ports based on a pair-wise analysis, it may not be able to capture the relationships among Chinese ports. On the other hand, it would be cumbersome if we include all other ports in one co-integration and error correction modeling equation due to many ports covered in this study. To overcome this weakness, we also examine the nature of relationships between Chinese ports to see if they competitive or complementary based on the framework adopted for this paper.
Table 2: Unit root test results Order of I(0) I(1) I(2) difference Ports (1% Significant level) Kobe, Ningbo, Dalian Ports (5% Significant level) Kobe, Ningbo, Dalian, Keelung, North, Tianjing HongKong, Kaohsiung, Keelung*, Nagoya, Yokohama, Pusan, Tokyo, Allchina, North*, Middle, South, Shanghai, Guangzhou, Qingdao, Tiangjin*, Xiamen, Singapore, HongKong, Kaohsiung, Nagoya, Yokohama, Pusan, Tokyo, Allchina, Middle, South, Shanghai, Guangzhou, Qingdao, Xiamen Shenzhen Singapore, Shenzhen Notes: * denotes the test results of ADF test and PP test are different. In this case, the adopted result is subject to the ADF test. Allchina is the sum of container throughput of all the major ports involved.
3. Findings Long term relationships Only ports of Kaohsiung and Keelung benefitted from the rise of Chinese ports Short term dynamics Positive impact Singapore, from all major Chinese ports especially Shanghai except Tianjin Hong Kong, except from port of Xiamen Negative Kaohsiung, Keelung, Pusan and other Japanese ports
Table 3: Cointegrating equations (CE) and their port coefficients: 1980 2010 Port pair Allchina North Middle South Shanghai Singapore -1.61 0.52-0.66-0.64-1.55 Hong Kong -0.76-1.17 none -0.53-0.78 Kaohsiung 0.44 0.72 0.59-0.07 0.40 Keelung 0.02 0.002 0.01 none 0.01 Pusan -0.46-0.99 none -0.37-0.47 Nagoya -0.13-0.22-0.17 none -0.20 Yokohama none -0.12 none none none Tokyo -0.27-0.29 none -0.29 none Port pair Shenzhen Guangzhou Qingdao Tianjin Xiamen Singapore -0.36 none. -0.65-0.06-0.74 Hong Kong n.a none -0.57-1.85-0.21 Kaohsiung n.a. -0.49-0.54 none -0.27 Keelung n.a. none -0.11-0.03-0.005 Pusan n.a. none -0.45-0.80-0.34 Nagoya n.a. none -0.13-0.29-0.14 Yokohama n.a. none -0.28 none none Tokyo n.a. none -0.28 none -0.23 Notes: The relationship is competitive if the sign of the coefficient is negative and complementary if the sign is positive; n.a. = not applicable; none=no cointegration exists.
Table 4: Error correction mechanisms (ECM) between Chinese and major East Asian ports: 1980-2010 ECM AllCN North Middle South SH GZ QD TJ XM SZ DL NB Singapore 0.03-0.03 0.04 0.30 0.30-0.097 0.14-0.06 0.18 0.49 0.190-0.012 Hong Kong 0.10 0.01 0.222 0.37 0.10-0.057 0.14 0.05-0.21 0.063 0.235-0.016 Kaohsiung -0.02-0.02-0.01-0.21-0.01-0.04 0.11-0.040 1.91 0.079-0.035-0.020. Keelung -0.11-0.19-0.05 0.166-0.04 0.166 0.06-0.56 0.20 0.099 0.441 0.105. Yokohama 0.393-0.28 0.420 0.203 0.403 0.102-0.03-0.013 0.188 0.206-0.012-0.010 Pusan -0.13 0.06 0.127-0.50-0.20 0.044-0.23-0.22-0.61 0.138-0.015-0.010 Tokyo -0.57-0.57 0.191-0.09 0.189 0.024-0.85-0.007-0.61 0.122-0.006-0.007 Nagoya -0.14-0.28-0.11 0.171-0.09 0.20-0.19-0.51-0.11 0.135-0.028-0.025 Kobe -0.075-0.098-0.187 0.283-0.217 0.226-0.287 0.276 0.042-0.345 0.448 0.008 Notes: A negative sign denotes that the impact from Chinese ports is negative and a positive sign denotes that the impact is positive. SH is Shanghai, GZ is Guangzhou, QD is Qingdao, TJ is Tianjin, XM is Xiamen, SZ is Shenzhen, DL is Dalian, NB is Ningbo. Figures in bold were estimated based on equation (5).
Table 5: Co-integration and their port coefficients among Chinese ports: 1980-2010 Competition among Chinese ports Table 5: Co integration and their port coefficients among Chinese ports: 1980 2010 Port pair Shanghai Tianjin Guangzhou Qingdao Xiamen Shanghai 1.33 1.00 0.80 None Tianjin 0.78 0.60 None Guangzhou 0.83 None Qingdao 0.82 Xiamen
Competition among Chinese ports Table 6 ECM tests among Chinese ports: 1980 2010 Port pair Shanghai Tianjin Guangzhou Qingdao Xiamen Shanghai to 0.40 0.07 0.25 None to Shanghai 0.58 0.58 0.13 None Tianjin to 0.45 0.37 None to Tianjin 0.42 0.09 None Guangzhou to 0.38 None To Guangzhou 0.04 None Qingdao to 0.25 To Qingdao 0.01 Xiamen to To Xiamen
4. Policy and Research Implications The different implications imply the need to analyze interport relationships on a disaggregated level Chinese ports found to be long run competitors with most of major East Asian ports In some cases the relationships varied depending on the time periods considered For those oseports that have aeo long term competing relationships with Chinese ports Identify factors behind this negative relationship Implement strategies Forge closer links with shipping lines and other major stakeholders in the supply chain Create complementary relationships with neighbouring Chinese ports More emphasis on value added services for high value and time sensitive freight
Policy and Research Implications For those with long run complementary relationships Constant monitoring and analysis of the factors behind the complementary relationships Monitoring of changing market conditions and shipping lines needs
5. Conclusion Significant shift over time in the centre of port and shipping activity away from ports of Singapore, Hong Kong and Pusan towards ports of Kaohsiung, Keelung to major ports in mainland China In the short run, the effects were different in the case of Singapore, Hong Kong, Kaohsiung and Keelung, but similar for other major ports in East Asia Future research: Factors underlying the nature of these relationships Identifications of costs and non cost factors
THANK YOU!