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?

For more details on Kyushu Shinkansen’s wider effects on the Kyushu region, refer to 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” in this volume, which discusses the changes in the number of tourists from the Kansai region and the changes in the people’s flows among three prefectures within the Kyushu region, including the route map and the more detailed description of Kyushu Shinkansen.

When we say “Hakata,” we usually divide the city center of Fukuoka City by the river flowing south to north in two districts, east and west or Tenjin and Hakata, and refer to the opposite side of Tenjin as the Hakata district. However, this sense of the Hakata district is considerably broader than the JR Hakata Station and its neighborhood because of including other shopping core sites, Canal City Hakata, an independent shopping mall, and Hakata Riverain, the commercial redevelopment complex implemented by Fukuoka City Government. Therefore, this chapter uses the term JR Hakata Station district if referring to JR Hakata Station and its neighborhood to distinguish it from the broader Hakata district.

Japan consists of four islands, from north to south, Hokkaido, Honshu, the largest main island, Shikoku, and Kyushu islands. The Kyushu island is the southmost one composed of seven prefectures, roughly from north to south, Fukuoka, Saga, Oita, Kumamoto, Miyazaki, and Kagoshima Prefectures, with a population of 12.8 million people as of 2020, according to census 2020.

Our consumer behavior approach started with forecasting the changes in consumers’ choices of the frequency of visits and their effects on retail sales. Refer to Saito and Yamashiro [3] and Saito et al. [4]. After these studies, the treatment of economic effects by consumer behavior approach has evolved from the effort for measuring consumers’ time value of shopping (refer to Saito and Yamashiro [5], Saito et al. [6]) to the direct estimation of consumers’ utility functions to evaluate the welfare changes. (Refer to Yamashiro and Saito [7], Saito et al. [8].)

The framework of economic effects of Kaiyu developed from the city center one-dollar circuit buses and was extended to the economic effects of city center cafés. Refer to Saito and Yamashiro [9, 10], Saito et al. [11, 12].

For more details on the intervening opportunity, refer to Saito et al. [14], Chapter 9, “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,” in this volume, which has discussed the estimation and prediction of Tenjin and Hakata’s intervening opportunity effects on the destination, Hakata and Tenjin, measured by the number of visitors to the destination. Also see Yamashiro et al. [15], and Saito et al. [16], which dealt with the closely related topics of Hub functions.

As for the weighted Poisson, refer to Saito et al. [17, 18] and for the consistent estimation, see Saito and Nakashima [19], and Saito et al. [20].

Hakata Riverain was developed by the redevelopment project led by the subsidiary of the Fukuoka City Government. Saito et al. [21] discussed the background and bankruptcy of Hakata Riverain’s redevelopment process from the perspective of the risk inherent in the collective social decision-making process.

Canal City Hakata is the first large-scale commercial complex developed by the private sector in the city center of Fukuoka City in 1994. It comprises a large retail shop, a shopping mall, hotels, a theater, a cinema complex, restaurants, and an amusement park, attracting many visitors for shopping and leisure. Saito [22] and Saito et al. [23] discussed its physical design with many curves from the angle of information evolution by town walking.

For more detailed description of Tenjin, refer to Saito et al. [24] which discussed and verified that the connectivity of Tenjin has increased by the extension of the Tenjin underground shopping mall.

Precisely speaking, the predicted number of visitors to Tenjin and Hakata while considering the intervening opportunities are not the entrance visits to Tenjin and Hakata since the question items to ask the respondents about their frequency of visits to Tenjin and Hakata are not referring to the number of joint visits to Tenjin and Hakata but just total frequency of visits to each of Tenjin and Hakata.

Similarly, since we are using the same question items to ask the respondents about their frequency of visits to Tenjin, Hakata, and Canal City, in this logit model, after all, we have hypothesized that the respondents’ frequency of visits to each destination is proportional to their entrance choice probability for each destination.

More precisely, we have expanded by 10,000. There are no sound theoretical foundation to determing the number.

To avoid misunderstandings and multi-understandings about the entire city center of Fukuoka City, we show the respondents its district extent definition in the questionnaire sheet.

We say the Kaiyu behavior proceeds if the Kaiyu step progresses or the Kaiyu movement occurs. We assume that a Kaiyu movement occurs if the place or the purpose of the Kaiyu behavior changes. Thus we can describe the phenomenon that a consumer continues to undertake his/her Kaiyu behavior at the same place with changing his/her purpose of continuing his/her Kaiyu.

We discussed the estimation and forecasting of the intervening opportunity effects that both Tenjin and Hakata Station districts affect each other in a separate chapter in this volume. Refer to Saito et al. [14], Chapter 9, “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,” in this volume.

Saito et al. [14] estimated another formulation of the intervening opportunity, which replaces squ_k with squ_k/t_ik for \( k=j,\overline{j} \) in Chapter 9, “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,” in this volume.

Another reason for adding the visit frequency to Canal City Hakata to the multivariate Poisson model is that the present formulation expresses the intervening opportunity for the destination j as a constant under the dependent variable Y_ij. Thus we must have more than three dependent variables to incorporate two, Tenjin and Hakata’s, intervening opportunity explanatory variables.

In later sections concerning the consistent estimation and the Kaiyu Markov model, we address the problem of how visitors’ entrance visits to the destinations and their Kaiyu behaviors between shopping sites determine the total visits to the destinations. Thus the total visits to Tenjin and Hakata Station districts forecasted here are independent results not related to the forecast in the later Kaiyu Markov model sections because the models they base on are different. Thus, a future challenge remains to formulate how consumers jointly determine the visit frequency to the destinations and the number of Kaiyu steps between them.

Saito et al. [20] first developed the algorithm for consistently estimating the Kaiyu path density over all Kaiyu paths. Then, Saito and Nakashima [19] first applied the consistent estimation method to the estimation for the pedestrian flows or Kaiyu flows within the Daimyo district neighboring Tenjin in the city center of Fukuoka City.

Saito and Ishibashi [31] first introduced this definition to deal with the staying within the same large retail facility; see also Saito and Ishibashi [32, 33], Chapter 5, “Kaiyu Markov Model with Covariates to Forecast the Change of Consumer Kaiyu Behaviors Caused by a Large-Scale City Center Retail Redevelopment,” in this volume.

From Table 10, we obtain the total visit density excluding staying for node six as follows. The column sum for the node six minus the staying density = 0.3224 − 0.1381 = 0.1843. Thus the estimated number of net incoming visitors to the entire city center of Fukuoka City becomes 55,000/0.1843 = 298,417.

Also, refer to Ishibashi and Saito [34] and Saito and Ishibashi [33], respectively, Chapters 7 and 5, “A Disaggregate Kaiyu Markov Model to Forecast the Sales of Retail Establishments Based on the Consumers’ Frequency of Visit” and “Kaiyu Markov Model with Covariates to Forecast the Change of Consumer Kaiyu Behaviors Caused by a Large-Scale City Center Retail Redevelopment,” in this volume. As to the fundamentals for the Kaiyu Markov model estimation, refer to Saito et al. [35], Chapter 6, “Estimation of Disaggregate Huff and Kaiyu Markov Model: A Lecture Note on Conditional Logit Model,” in this volume.

All calculations to estimate and forecast the Kaiyu flows in this section are based on the postulates presented in Chapter 6, “Basics of Kaiyu Markov Models: Reproducibility Theorems—A Validation of Infinite Kaiyu Representation,” in this volume, Saito et al. [36]. For more details on Kaiyu Markov models, readers are requested to refer to Chapter 6, “Basics of Kaiyu Markov Models: Reproducibility Theorems—A Validation of Infinite Kaiyu Representation”. However, for completeness, we list below the formula for calculating the number of Kaiyu flows between the shopping sites or districts. We prepare the notations. Let I and H denote the shopping nodes and exit/entrance, respectively. Let P be the probability matrix of Kaiyu choice probabilities, including exit, comprising P_II and P_IH. Also, we denote by F_HI the number of entrance visitors to shopping nodes, I. Let RM denote the number of Kaiyu flows. We use the convention for the superscript x to denote before and after the opening when x = Before or After, respectively. With these notations, the number of Kaiyu flows between the shopping nodes or districts, I, is calculated by the following formula:

See the matrix formed by the first three rows and columns or three districts in Table 19(A). For example, the number in the cell for the first row and the second column means that the Kaiyu movement from Tenjin to Hakata is 35,287 persons per day.

Note that we excluded the Kaiyu movements within the same retail facility as stated before. Thus, this result does not reflect the Kaiyu movements in JR Hakata City.

On 26 February 2011, 4 days before the pre-opening of JR Hakata City, FQBIC (Fukuoka University Institute of Quantitative Behavioral Informatics for City and Space Economy) publicized the newsrelease [37] about the impacts of JR Hakata City’s opening on the city center of Fukuoka City; Forecasts of the number of visitors, retail sales, and the changes in people’s Kaiyu flows between Tenjin and Hakata. The newsrelease was reported in the articles in the newspapers (cf. [38‐40]).

The abridged versions of some parts of this chapter first appeared in Yamashiro and Saito et al. [47] and Yamashiro, Sato, and Saito et al. [48].