Path dependence in port choice: some quasi-experimental evidence

ABSTRACT

This study analyzes the port choice dynamics of Japanese container shippers before and after the 1995 Great Hanshin earthquake, which destroyed Japan’s largest port. We find that the earthquake disintegrated some transportation networks that had initially been maintained because of path dependence. At the same time, it led to establishing a new path dependence. They jointly explain why the negative impact of the earthquake on the damaged port is persistent rather than temporary.

KEYWORDS:

• Port
• path dependence
• container transportation
• Japan

JEL CLASSIFICATION:

• R40
• L92

Acknowledgments

The authors would like to thank Mark P. Taylor (the editor) and an anonymous referee for useful comments.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1 Coordination problem and network externalities are also discussed in their papers to explain the observed path dependence in industry location.

² Osaka port is much smaller in scale than Kobe port. In 1994, the trade container freight handled by Osaka is only one quarter of that by Kobe. Unlike Kobe port, Osaka port was slightly damaged in the 1995 earthquake; however, Osaka was not able to perform as a main substituting port for Kobe due to its limited scale.

³ As discussed in Xu and Itoh (2018), the international transshipment cargoes of Hanshin were largely diverted to Busan port in South Korea.

⁴ Note that container shippers usually sign an annual contract for generalized shipping costs with the container shipping companies or forwarders, rather than a case-to-case freight contract.

⁵ Specifically, the data are available for 1985, 1988, 1993, 1998, 2003, 2008, and 2013; see more details about the CCFS in Xu and Itoh (2018).

⁶ Note that port-level freight volumes were recorded based on the location of customs clearance. That is, if a container was firstly transported from a small-sized regional port in prefecture $i$ to Hanshin, and then exported to the U.S. from Hanshin, this container cargo was recorded as exported by prefecture $i$, and handled by Hanshin rather than the small-sized port, as customs clearance took place in Hanshin.

⁷ Although some prefectures are distant from Hanshin and closer to an alternative major container port (i.e. they are not the hinterlands of Hanshin), a part of the shippers there still use Hanshin, as it was the largest port in the area before the earthquake; container transportation is only available in Hanshin for some intercontinental shipping routes.

⁸ See also a review of the related literature in Ducruet (2017). Actually, the port system in Japan also shows a pattern of de-concentration, as a relative convergence seems to have appeared between Hanshin and Keihin after 2010, and Nagoya has grown substantially in the recent decades and become a third pole of the country’s top port hierarchy (see Figure 1).

⁹ Evidence regarding the impacts of historical events on the shipping sector and the port system can be found in, e.g. Gras (2010) and Marnot (2015).

¹⁰ The container handling volume at the port of Kobe reached its pre-earthquake level for the first time in 2017. However, Kobe’s market share was still significantly lower than before the earthquake, because other ports were growing faster.

Additional information

Funding

This work was supported by the Japan Society for the Promotion of Science [JP16K03684; 17KK0058]; National Natural Science Foundation of China [71703034];