A Micro Behavior Approach to Estimating and Forecasting the Intervening Opportunity Effects with a Multivariate Poisson Model: A Case for the New Terminal Complex of Kyushu Super-Express Railway, JR Hakata City

The research framework described here is what we called the ‘consumer behavior approach.’ We have carried out many studies to evaluate urban development policies based on the consumer behavior approach (cf. [1‐9]).

See, for example, the recent literature [11, 12] and the traditional research [13].

The 37 municipalities in Fukuoka Prefecture as of 2005 are as follows: 7 wards in Fukuoka City, Higashi-ku, Hakata-ku, Chuo-ku, Minami-ku, Nishi-ku, Jonan-ku, and Sawara-ku; 15 cities, Kurume-shi, Iizuka-shi, Chikugo-shi, Ogori-shi, Chikushino-shi, Kasuga-shi, Onojyo-shi, Munakata-shi, Dazaifu-shi, Koga-shi, Fukutsu-shi, Miyawaka-shi, Kama-shi, Asakura-shi, and Itoshima-shi; 15 towns, cho or machi in Japanese, Nakagawa-machi, Umi-machi, Sasaguri-machi, Shime-machi, Sue-machi, Shingu-machi, Hisayama-machi, Kasuya-machi, Okagaki-machi, Onga-cho, Kotake-machi, Keisen-machi, Chikuzen-machi, Tachiarai-machi, and Oki-machi; 2 municipalities in Saga Prefecture as of 2005 are 1 city and 1 town, Tosu-shi, and Kiyama-cho.

Sometimes we define smaller Fukuoka Metropolitan Area, which we define consists of the municipalities satisfying the condition that the percentage of residents who commute to Fukuoka City for work or school is greater than 15%. We used this definition in the chapter “How Would the Kyushu Super-Express Railway Opening Change the Flow of Tourists from the Kansai Region within the Kyushu Wide Area, Japan?: A Micro Behavior Analysis of the Destination’s Hub Function,” Chap. 8 in this volume.

The weighted Poisson model gives each sample a weight expressed as the inverse of his/her visit frequency to the destination. For more details, refer to Saito et al. [22, 23].

This formulation was first given in [24, 25].

We here referred to the positive intervening opportunity effect contributed by Hakata to the number of visitors to Tenjin, which seems strange but is a consequence that we defined the intervening opportunity’s effect on the destination as the difference in the visit frequency to the destination between with and without the intervening opportunity. In the present case, Hakata’s intervening opportunity effect is the difference in the visit frequency to Tenjin between people who visit Tenjin via Hakata and those who visit Tenjin not via Hakata. However, there is a subtle problem with distinguishing those who originally intended to visit Hakata but changed their destination to Tenjin and pass through Hakata, not stopping there from those who have decided to visit Tenjin from the start and visit Tenjin via Hakata. Since the problem is related to people’s behavior changes, to explore the problem further, we need panel data that record the changes in individuals’ behavior, which is not available in this study.

For consumers, the required travel time is the cost for them (cf. [26]).

As for the notion of the intervening opportunity, there are several closely related, similar, but different concepts in the literature. One of them is the competing destination due to Fotheringham [27], which presumes a situation where several destinations are located neighboring each other. Fotheringham diagrammatically showed that the absolute value of the exponent of distance resistance could decrease in competing destinations. While many situations exist showing applicable to competing destinations, if we apply the concept to the case of Hakata and Tenjin, the competing destinations should hypothesize to increase the visitors at both Hakata and Tenjin since one prefers to visit Hakata because one can visit Tenjin too if going further a little bit, and vice versa. Furthermore, in the competing destinations, people visit Tenjin and Hakata through Kaiyu, which we ignore in this study. Hence, we can regard the competing destinations in this case as the hub function, which we dealt with in the chapter, “How Would the Kyushu Super-Express Railway Opening Change the Flow of Tourists from the Kansai Region within the Kyushu Wide Area, Japan?: A Micro Behavior Analysis of the Destination’s Hub Function,” Chap. 8 in this volume [28]. In contrast, in this study, people pass through Hakata and visit Tenjin, not stopping at Hakata. Therefore, the intervening opportunity contributed by Hakata might seem to be close to the concept of the straw effect in the transportation literature, which describes the phenomenon that people visit the remote big city or retail establishments passing through nearby core centers, not visiting or stopping at them. Finally, ignoring Kaiyu, the districts Tenjin and Hakata are both competitive destinations, which we dealt with in the chapter, “How Would the Opening of JR Hakata City, a New Terminal Complex of the Kyushu Super-Express Railway, Change the Number of Visitors, Retail Sales, and Consumers’ Kaiyu Flows in the City Center Commercial District of Fukuoka City?,” Chap. 10 in this volume [29]. Also, refer to Chap. 11 in this volume [34], where we dealt with competitive destinations.

For prediction in this study, we use only the estimated results of the first formulation of the multivariate Poisson model with intervening opportunities.

This study treats the intervening opportunity variable as an exogenous variable. Thus, we need to give the values of intervening opportunities for the representative individuals representing municipalities for forecasting. In particular, we must decide how to set the dummy variable indicating whether they experience the intervening opportunities or not, more specifically, whether or not they go to Tenjin via Hakata and vice versa. We employ the following method. As for the municipality where samples who responded that they experienced the intervening opportunity exist, we use the percentage of these samples in all samples from that municipality for the expansion when estimating the present magnitude and forecasting the magnitude after the JR Hakata City’s opening contributed by the intervening opportunity. When some municipalities without respondents exist that lie along the same railway as those municipalities that have the sample experiencing the intervening opportunity and are sandwiched between them, we set the probability of the intervening opportunity to one for these municipalities. How we dealt with the intervening opportunities perceived by the representative agents for 39 municipalities are shown in the two columns related to Hakata’s and Tenjin’s intervening opportunities to the destinations, Tenjin and Hakata in Table 4, in which the ratio given in each cell implies the share of the persons who experience the intervening opportunity among the whole population in that municipality.

Refer to the literature such as [30‐33].

The original version of this chapter appeared in Yamashiro and Saito et al. [24].