Neighborhood Spatio-Temporal Impacts of SDG 8.9: The Case of Urban and Rural Exhibition-Driven Tourism by Multiple Methods

Abstract

Rural arts events (triennials/festivals) are mainly aimed at local and regional revitalization. This exhibition-driven tourism (unlike traditional festivals, conferences, and exhibitions) has existed for more than 20 years in Japan. The curators of exhibition-driven tourism hope that these events can promote the economy and stop population decline as a result of the aging population. Therefore, this paper attempts to evaluate the effects of urban and rural arts event tourism in local and neighborhood areas in Niigata, Japan from the perspective of SDG 8.9. The Echigo-Tsumari Art Triennial and Water and Land Niigata Art Festival were chosen as case studies. Panel data (1997–2019) concerning tourists, income, and population in Niigata were evaluated using multiple empirical methods with descriptive correlation statistics (simple linear regression (SLR) and one-way ANOVA) and spatial analysis (Moran’s I). Through multiple-method analysis, the positive impacts of urban and rural arts event tourism in local and neighborhood areas in relation to Sustainable Development Goal 8.9 were evaluated. The findings presented herein have meaningful implications for tourism academia and the industry in general.

1. Introduction

Rural arts events (triennials/festivals) are mainly aimed at local and regional revitalization. The curators of exhibition-driven tourism hope to develop the economy and reduce population decline as a result of the aging population. Moreover, urban arts events (triennials/festivals) mainly focus on cultural development and revitalization [1]. In previous studies, the different economic, policy, human, social culture, and environmental protection scenarios were taken as the principal impact items of festivals [2,3]. Moreover, sustainable development has become the main focus of tourism policymakers and researchers [4]. Sustainable development, combined with mainstreaming tourism, economic, and social responsibility, has become one of the main headings of the World Tourism Organization (UNWTO) targets [5]. The 17 Sustainable Development Goals (SDGs) are an urgent call for action by all countries, with No. 8.9 stating: “By 2030, devise and implement policies to promote sustainable tourism that creates jobs and promotes local culture and products”. Therefore, this paper attempts to evaluate the effects of urban and rural arts event tourism in local and neighborhood areas from the perspective of SDG 8.9.

However, various empirical papers show that tourism is less sustainable than expected (e.g., the adverse reactions to tourism growth in Venice in relation to Venice Biennale, one of the most comprehensive examples of exhibition-driven arts event tourism) [6,7]. However, the SDGs and millennium development goals (MDGs) have become the main focus when studying the contribution of tourism to the sustainability of the entire tourism industry [5,8]. Despite this conflict, more positive samples for empirical analyses focused on sustainable tourism are needed.

As an SDG-response study, using quantitative empirical analysis for arts-event-driven tourism (unlike traditional conferences and exhibitions or festivals), the current paper is a new attempt to study the tourism industry and sustainable cities. After the economic recession in the 1990s, more than 120 art exhibitions aimed at revitalizing the areas in which they took place through art and local resources (attracting tourists) began appearing throughout the Japanese territory. Niigata is one of Japan’s earliest and most important art-exhibition-hosting regions [9,10]. In addition, Niigata is the only prefecture in Japan in which two arts events, a rural event and an urban event, are held simultaneously. Thus, the Echigo-Tsumari Art Triennial (ETAT) (in Tokamachi and Tsunan) and the Water and Land Niigata Art Festival (WLNAF) in Niigata city were selected for the empirical evaluation. The concept of a triennial is that exhibitions are hosted once every 3 years [11]. Figure 1 shows the structure of the current paper. Through the multiple-method analysis, the positive impacts of urban and rural arts event tourism in local and neighborhood areas from the perspective of SDG 8.9 were evaluated. Although this process is controversial, a new evaluation for exhibition-driven tourism must be established. The current paper attempts to fill the gaps in the assessment of spatio-temporal impacts as related to the exhibition-driven tourism industry.

2. Literature Review

2.1. Sustainable Development Goals (SDGs) and Sustainable Tourism

Recently, various empirical papers showed that tourism is “less sustainable” than expected [6,7]. Hall [6] provided an anti-institutional perspective on the tourism sector’s approach to the sustainable development goals and sustainable tourism framework. Ahmad et al. [12] studied the correlations between tourism and lower-middle-income economies. Rutty et al. [7] found that there was less emphasis on the environmental and social consequences than the positive economic impacts of tourism. After destinations such as Venice (e.g., with Venice Biennale representing a key example of exhibition-driven tourism) produced a series of adverse reactions to tourism growth, concerns about the contribution of tourism to sustainable development have also become local-scale issues (e.g., World Travel and Tourism Council (WTTC)).

The World Tourism Organization (UNWTO) refers to sustainable tourism and its economic significance. The SDGs have become the main focus when researching the tourism contribution to the sustainability of the entire tourism industry [5,8]. There are designated journals specifically related to sustainable tourism and many texts and journal articles, which account for as much as 5% of journal output [13]. Sustainable tourism is the concept of visiting places without damaging local communities and nature and positively impacting health, the environment, technological methods, and the economy [14]. Various researchers have studied the relationship between sustainable tourism and the attitudes of tourists/residents [15]. Other scholars have studied the relationship between sustainable tourism with ecotourism [16]. Although this remains controversial, more positive samples for empirical analyses focused on sustainable tourism are needed.

2.2. Urban and Rural Arts Event Tourism

First, events depend on positive perceptions of the destination and tourism products [17], including the art triennial and festival studied herein. Event tourism may be associated with specific spatial resources, e.g., with attracting and planning the event in relation to natural and other tourist values [18]. The first article we reviewed is specifically related to event tourism in JTR and was published by J.R.B. Ritchie and Beliveau in 1974 [18]. Kersulić et al. [19] reviewed the strategic planning sustainability elements in relation to broader sport tourism events. Recently, leading arts festivals have emerged “bottom-up”, developing organically in urban and rural areas [20]. Therefore, certain scholars have studied rural tourism through lifestyle, livelihood, and artistic careers [21]. For example, Wise et al. [22] studied the local tourism economy via a sense of rural community, potential industry opportunities, and social impacts.

Second, the attitude of tourists is seen as the main positive force in promoting the tourism economy [23,24]. Andersen et al. [25] considered the image of Denmark held by the visitors to the art exhibition in their study. Camarero et al. [26] analyzed the four elements of brand equity, brand image/value, loyalty, and perceived quality in the art exhibition. They evaluated the state of art exhibitions in Spain. Chen et al. [27] studied satisfaction and service quality for event promotion. Liu et al. [28] and Ruan et al. [29] examined the relationship between natural capital and tourism image. Fu et al. [26] examined the basic dimensions of place attachment in the exhibition environment and their impact on participant satisfaction.

Third, the relationship between tourism and the economy has been studied in many previous papers [30,31,32]. According to the report of the WTO (World Tourism Organization), tourism consumption was USD 462 billion per day in 2018, while it was only USD 1.3 billion in 2001 [33]. The majority of studies are focused on tourism demand and the impact of exchange rates on tourism income [34]. Others have studied political and economic impacts using time series analyses [35]. However, few previous papers have studied exhibitions’ direct spatio-temporal impact on tourism.

2.3. Art Event Tourism and Economics: Total Income, Tertiary industry Income, and Per capita Income

Art event tourism (exhibition-related triennials and festivals)-related impact on economic growth has been studied in previous papers. One of the earliest studies, by Della et al. [36] in 1977, assessed the economic impact of the Rhode Island high mast sailing ceremony. Kim et al. [37] studied the overall exhibition industry’s economic impact. Rephann [38] assessed the impact of economic activities during the construction and operation of exhibition venues. Nesticò et al. [39] studied the economic evaluation of sustainable indicators. Zhang [40] also used a process–goal method to study tourism development in relation to environmental, economic, and social goals. Chhabra et al. [38] demonstrated that festivals are usually a strategic choice for rural economic growth. However, the economic impact of festivals depends on the characteristics of the festival, such as the number of days the festival is held and the parts of the local economy that are brought together (the different products from rural areas that can be sold during the festival). Hwang et al. [41] studied elderly tourism in relation to promoting South Korea’s economic growth. Vasiliev et al. [42] examined the relationship between sustainable development’s economic, social, and environmental dimensions. Ying et al. [43] studied the correlations between the circular economy and green exhibition. Cai et al. [10] analyzed the influence of the exhibition industry on sustainable local development. Other researchers studied the motivation and purpose behind the consumption of conferences and exhibitions [44,45].

An event may significantly increase local economic activity. However, the net impact in neighboring areas and cities may be more significant than the local (the hosting areas) impact (e.g., the big/national effect often exceeds the small/state effect), with the impact on the local/hosting areas even sometimes being negative [46,47]. However, the impact format of these exhibitions is mainly related to transactions [48,49]. Thus, accurately measuring the economic contribution of art exhibitions or exhibition-driven tourism is challenging. The earliest mention of the concept of “rising tide” in the tourism industry was in 1996 in a study by Mason and Mowforth [50]. They studied a rising tide in relation to codes of conduct in tourism.

The correlations between tourism and total income have been studied in many papers from the 1990s [51,52,53]. Saint Akadiri et al. [54] examined the role of real total income, globalization, and tourism in sustainable targets using autoregressive distributed lag and the vector error correction model with the Granger test. Wagner et al. [55] examined the economic effects of tourism using a social accounting matrix for the evaluation of tourism policies in eight disaggregated levels of income, two governments, and four types of taxes. Louca et al. [56] analyzed the existence and nature of long-term relationships between income derived from the tourism industry and tourist arrivals.

The correlations between tourism and tertiary industry income have been studied in previous papers [57]. Lee and Kang [58] reported that tourism from tertiary industries improves the lower-income class more than the primary- and secondary-income classes. Hung et al. [59] described a positive relationship between household income, household head age, and car ownership. Li et al. [60] demonstrated that on-screen tourism had positive impacts on the tertiary industry but negative impacts on the primary and secondary sectors.

The correlations between tourism and per capita income have been studied in previous papers [61]. Garin-Munoz et al. [62] tested the effects of real per capita income, exchange prices, and rates using tourist services panel data (1985–1995). Brau et al. [63] analyzed the relationship between growth, size, and tourism by controlling for initial per capita income using panel data (1980–2003). Zaman et al. [64] studied the relationship between economic growth, carbon dioxide emissions, and tourism development using hypothesis panel data (2005–2013) to verify per capita income and tourism carbon emissions.

2.4. Art Event Tourism and Population: Total Population, Labor Population, and Household Number

The correlation between population growth and economic development is a constantly changing issue (with different methods being used in different periods) in demographic economics. One of Malthus’s most influential studies is “Population” [65]. You et al. [66] studied the factors of settlement intention in floating populations in Chinese cities. Tamura et al. [67] studied a small Japanese town by comparing spatial population distribution patterns with environment and infrastructure costs. Egidi et al. [68] studied worldwide urban and city size population trends from 1950 to 2030. However, there are few studies on the direct connections between exhibitions and population. Cai et al. [10] found a positive correlation between exhibition-driven tourism and population. Getz [69] studied the long-term impacts of change in tourism and population in the Scottish Highlands. Khalid et al. [70] and Nam et al. [71] used empirical tests to show that the local community in their study wanted a thriving sustainable tourism.

On the other hand, with the aging of Japanese society and the low fertility rate, urban shrinkage has had a negative impact on the sustainable development of Japanese cities. Mallach et al. [72] believe that Japan’s urban shrinkage is due to demographic change. Martinez-Fernandez et al. studied shrinking cities in Australia, Japan, Europe, and the USA in relation to economic development, greening, revitalization, and social inclusion [73]. Japanese population shrinkage moves from urban centers towards the countryside [74,75]. Events or green events are entertainment-focused economic drivers with the local community and cultural identity at their core [76]. Although many scholars have conducted long-term and extensive research on population issues in various fields, this remains a critically underexplored issue.

3. Exhibition-Driven Tourism in Niigata

3.1. Japanese Arts Events: Festival and Exhibition

Rural arts events (Triennials/festivals) are mainly aimed at local and regional revitalization. They aim to develop the economy and reduce population shrinkage as a result of the aging population. Moreover, urban arts events (Triennales/festivals) focus on cultural development and revitalization [1]. From 1961 to 2019, especially after the recession in the 1990s, hundreds of Japanese art exhibitions were established with the purpose of revitalizing sustainable development in the host areas [10]. In this study, exhibitions in Niigata (the Echigo-Tsumari Art Triennial (ETAT) and the Water & Land Niigata Art Festival (WLNAF)) were selected for the empirical analysis. Every three years, artists from all countries are invited to create specific artworks related to the Niigata region’s environmental, social, and cultural background. Since the first ETAT event in 2000, thousands of works of art have been displayed, including sculptures, sound works, theater works, art installations, performances, musical performances, landscape design, urban design projects, and architectural structures.

3.2. Urban and Rural Arts Events: ETAT and WLNAF in Niigata

From 2000 to 2018 (once every three years), more than JPY 65,279.671 million were obtained, and more than 132,577.645 million tourists visited the host areas (Tokamachi (No.5), Tsunan (No.4), and Niigata City (No.21)). The creative city calls on people to engage in imaginative activities to help develop and manage urban life, encouraging the population to plan, think, and creatively solve urban problems [77]. Scholars have attempted to understand the potential of creative art and culture in rural environments [78]. However, more in-depth research is required to enhance our general understanding of the topic as this field is in its infancy [79,80].

Rural arts event tourism: Echigo-Tsumari Art Triennial (ETAT) was hosted in Tokamachi (No.5) and Tsunan (No.4) in 2000, 2003, 2005, 2009, 2012, and 2018. ETAT stemmed from a county-level incentive that encourages regions to overcome the socio-economic recession by profiting from the particularity of their environment [10]. However, there are almost no previous papers that study the quantitative economic impact of ETAT. For example, Ahn [69] studied the ETAT’s cultural and artistic impact on the hosting areas. In addition, Klien [4] and Kitagawa [81,82] explored the correlations between art and nature in this setting. Favell and Boven et al. [83,84] studied abandoned schools and their reuse to promote sustainable goals in ETAT. However, these papers lack empirical data, i.e., they only focus on descriptive statistics. Cai et al. [10] showed that the ETAT positively impacted sustainable tourism in the hosting areas. This paper selected panel data and assessed them from a new perspective to study the impacts of the ETAT. Therefore, exploring the relationship using a quantitative analysis of ETAT is essential.

Urban arts event tourism: the Water and Land Niigata Art Festival (WLNAF) was held in Niigata City (No.21) in 2009, 2012, and 2018. Similarly, since 2009, WLNAF has aimed to explore how Niigata’s local culture is equally affected by the region’s land and water resources. It also requires participants to reflect on the relationship between nature and the humanities. Koizumi [85] examined the “social roles” of WLNAF in a shrinking society with a decreasing population in the hosting areas. However, there are almost no previous papers regarding the quantitative economic impact of WLNAF.

According to previous research [10], exhibitions are rarely used as social forces to comprehensively evaluate and demonstrate their role in exhibition-driven tourism. These two exhibitions have been described by the foreign media as unique in quality and scale and are regarded as a new model for art exhibitions. Community building through art has attracted the attention of curators and people in the art world in the United States, Europe, and Asia and from local government delegations. [86]. Thus, the current research attempts to fill the gap related to the correlation between exhibition-driven impact and sustainable tourism economics via an empirical investigation.

4. Methods

Panel data were selected from the statistical yearbook and county survey of Niigata (1997–2019) and include: (1) tourist number and its growth rate; (2) total income/tertiary industry income/per capita income and their growth rate; and (3) total population/labor population/household number and their growth rate. The panel data (1997–2019) of tourists, income, and population in Niigata were evaluated using multiple empirical methods with descriptive correlation statistics (simple linear regression (SLR) and one-way ANOVA) and spatial analysis (Moran’s I). The positive impacts of urban and rural arts event tourism in local and neighborhood areas were evaluated from the perspective of SDG 8.9 (Figure 1). Panel data are complex and require a multiple methods analysis [10]. Therefore, the current paper selected descriptive statistics, simple linear regression (SLR), one-way ANOVA analysis, and Moran’s I for the assessment. Simple linear regression (SLR) and one-way ANOVA analysis were used for the time series data from the panel data. Moran’s I was used for the spatial sequence data. The positive impacts of urban and rural arts event tourism in local and neighborhood areas were evaluated from the perspective of SDG 8.9 (Figure 1).

4.1. Panel Data

Panel data contain observations of multiple phenomena obtained for the objects over multiple periods. Thus, panel data clustering is an essential part of decision-making and expert analysis [87,88]. Di Lascio et al. [89] used panel data analysis to study the relationship between cultural tourism and temporary art exhibitions. Moreover, panel data are more informative than other types of data because they provide more variability, so the estimation is more efficient [89]. Many studies use panel data to analyze tourism’s impact on the economy. Naudé and Saayman [90] identified five main areas for empirical research in tourism. There are many different estimation methods available [91]. Bhattarai [92] found that fixed and random effects estimates indicate that investment, rather than aid, is a factor that promotes growth when reviewing important applications of panel data models.

In the current study, we collected data on tourist number and its growth rate/tertiary industry/per capita income.

Table 1. Categorical variables.

Abbreviation		Variables	Year	Name	Sources
X	NO	NO	Before 2000	the year before the hosting of the ETAT	ETAT Official website
	Y1	YES1	2000/2003/2006	the hosting year of the ETAT
	B1	BETWEENNESS1	2001/2002/2004/2005/2007/2008	the year between the hosting of the ETAT
	Y2	YES2	2009/2012/2015/2018	the hosting year of the ETAT+WLNAF
	B2	BETWEENNESS2	2010/2011/2013/2014/2016/2017	the year between the hosting of the ETAT+WLNAF

shows the following: (1) categorical data, including the year before the exhibition (hereafter NO)—for example, the years before 2000 are denoted as NO; (2) the hosting year of the ETAT (hereafter Y1)—for example, the years 2000, 2003, and 2006 are denoted as YES; (3) the years between the hosting of the ETAT (hereafter B1)—for example, the years 2001, 2002, 2004, 2005, 2007, and 2008 are denoted as B1; (4) categorical data, including the hosting year of the ETAT and WLNAF (hereafter Y2)—for example, the years 2009, 2012, 2015, and 2018 are denoted as Y2; (5) the years between the hosting of the ETAT and WLNAF (hereafter B2)—for example, the years 2010, 2011, 2013, 2014, 2016, and 2017 are denoted as B2.

Table 2. Continuous variables 1.

NO.	Areas 1	Tourism		Economic
		Tourist Number (TN)	Tourist Number Growth Rate 2 (TNGR)	Total Income (TI)	Total Income Growth Rate (TIGR)	Per capita Income (PCI)	Per capita Income Growth Rate (PCIGR)	Tertiary Industry Income (TII)	Tertiary Industry Income Growth Rate (TIIGR)
0	Niigata Areas 3	TN0	TNGR0	TI0	TIGR0	PCI0	PCIGR0	TII0	TIIGR0
1	Itoigawa City	TN1	TNGR1	TI1	TIGR1	PCI1	PCIGR1	TII1	TIIGR1
2	Myoko	TN2	TNGR2	TI2	TIGR2	PCI2	PCIGR2	TII2	TIIGR2
3	Joetsu City	TN3	TNGR3	TI3	TIGR3	PCI3	PCIGR3	TII3	TIIGR3
4	Tsunan	TN4	TNGR4	TI4	TIGR4	PCI4	PCIGR4	TII4	TIIGR4
5	Tokamachi	TN5	TNGR5	TI5	TIGR5	PCI5	PCIGR5	TII5	TIIGR5
6	Yuzawa Town	TN6	TNGR6	TI6	TIGR6	PCI6	PCIGR6	TII6	TIIGR6
7	Minamiuonuma	TN7	TNGR7	TI7	TIGR7	PCI7	PCIGR7	TII7	TIIGR7
8	Uonuma City	TN8	TNGR8	TI8	TIGR8	PCI8	PCIGR8	TII8	TIIGR8
9	Ojiya City	TN9	TNGR9	TI9	TIGR9	PCI9	PCIGR9	TII9	TIIGR9
10	Nagaoka	TN10	TNGR10	TI10	TIGR10	PCI10	PCIGR10	TII10	TIIGR10
11	Kashniwazaki	TN11	TNGR11	TI11	TIGR11	PCI11	PCIGR11	TII11	TIIGR11
12	Kariwa Village	TN12	TNGR12	TI12	TIGR12	PCI12	PCIGR12	TII12	TIIGR12
13	Izumozaki Town	TN13	TNGR13	TI13	TIGR13	PCI13	PCIGR13	TII13	TIIGR13
14	Mitsuke City	TN14	TNGR14	TI14	TIGR14	PCI14	PCIGR14	TII14	TIIGR14
15	Sanjo City	TN15	TNGR15	TI15	TIGR15	PCI15	PCIGR15	TII15	TIIGR15
16	Tsubame City	TN16	TNGR16	TI16	TIGR16	PCI16	PCIGR16	TII16	TIIGR16
17	Yahniko Village	TN17	TNGR17	TI17	TIGR17	PCI17	PCIGR17	TII17	TIIGR17
18	Kamo City	TN18	TNGR18	TI18	TIGR18	PCI18	PCIGR18	TII18	TIIGR18
19	Tagami Town	TN19	TNGR19	TI19	TIGR19	PCI19	PCIGR19	TII19	TIIGR19
20	Gosen	TN20	TNGR20	TI20	TIGR20	PCI20	PCIGR20	TII20	TIIGR20
21	Niigata City	TN21	TNGR21	TI21	TIGR21	PCI21	PCIGR21	TII21	TIIGR21
22	Agano City	TN22	TNGR22	TI22	TIGR22	PCI22	PCIGR22	TII22	TIIGR22
23	Aga Town	TN23	TNGR23	TI23	TIGR23	PCI23	PCIGR23	TII23	TIIGR23
24	Shibata City	TN24	TNGR24	TI24	TIGR24	PCI24	PCIGR24	TII24	TIIGR24
25	Seiromachni	TN25	TNGR25	TI25	TIGR25	PCI25	PCIGR25	TII25	TIIGR25
26	Wombai city	TN26	TNGR26	TI26	TIGR26	PCI26	PCIGR26	TII26	TIIGR26
27	Sekikawa	TN27	TNGR27	TI27	TIGR27	PCI27	PCIGR27	TII27	TIIGR27
28	Murakami City	TN28	TNGR28	TI28	TIGR28	PCI28	PCIGR28	TII28	TIIGR28
29	Awashnimaura	TN29	TNGR29	TI29	TIGR29	PCI29	PCIGR29	TII29	TIIGR29
30	Sado City	TN30	TNGR30	TI30	TIGR30	PCI30	PCIGR30	TII30	TIIGR30

¹ From the Niigata Statistical Yearbook; ² Compared with the previous year; ³ Niigata area including 30 areas.

shows the future tourist number and the total/tertiary industry/per capita income for the hosting areas and other cities in the Niigata area. Figure 2 shows the locations of the hosting areas and other cities in the Niigata area.

4.2. The Descriptive Statistics

Descriptive statistics were used for data analysis as they are an easy, visual way to understand the data [93,94]. Hwang et al. [95] used descriptive statistics to study elderly tourism wellbeing perception and its outcomes. Various scholars have studied the relationship between tourism and sustainability using descriptive statistics [33,96,97], and others have studied the relationship between economics and tourism using descriptive statistics [29,33,98].

4.3. Simple Linear Regression (SLR)

The paper used the SPSS26 software (IBM, New York, United States). A correlation analysis is commonly used to evaluate the relationship between two variables. A high correlation means the relationship between variables is vital [99,100,101]. Two random variables (X and Y) are usually tested in simple linear regression (SLR) [102]. The p-value helps researchers to reject or fail to reject a hypothesis. If the p-value is < 0.05, the analysis is considered significant for the next step. The least square method was used to calculates simple linear regression and the Pearson’s correlation model (Y = a + bx). The following formula was used for the slope (b) and the Y-intercept (a) (Y = linearly related to x; r² = the proportion of the total variance (s²) of Y that the linear regression of Y can explain on x; 1 − r² = the balance that is not defined by the regression; thus, 1 − r² = s²xY/s²Y):

$$\mathrm{b}=\frac{{\sum }_{\mathrm{i}=1}^{\mathrm{n}}\left({\mathrm{x}}_{\mathrm{i}}-\bar{\mathrm{x}}\right)\left({\mathrm{Y}}_{\mathrm{i}}-\bar{\mathrm{Y}}\right)}{{\sum }_{\mathrm{i}=1}^{\mathrm{n}}{\left({\mathrm{x}}_{\mathrm{i}}-\bar{\mathrm{x}}\right)}^2}$$

(1)

$$\mathrm{a}=\bar{\mathrm{Y}}-\mathrm{b}\bar{\mathrm{x}}$$

(2)

$$\mathrm{b}=\frac{{\sum }_{\mathrm{i}=1}^{\mathrm{n}}\left({\mathrm{x}}_{\mathrm{i}}-\bar{\mathrm{x}}\right)\left({\mathrm{Y}}_{\mathrm{i}}-\bar{\mathrm{Y}}\right)}{\sqrt{{\sum }_{\mathrm{i}=1}^{\mathrm{n}}{\left({\mathrm{x}}_{\mathrm{i}}-\bar{\mathrm{x}}\right)}^2{\sum }_{\mathrm{i}=1}^{\mathrm{n}}{\left({\mathrm{Y}}_{\mathrm{i}}-\bar{\mathrm{Y}}\right)}^2}}$$

(3)

Using Fisher’s z, the transformation was constructed for r using confidence limits. The null hypothesis that r = 0 (i.e., no association) was evaluated using a modified t-test [103,104]. These belts represent the reliability of the regression estimate (the tighter/belt, the more reliable/estimate) [105].

4.4. The One-Way ANOVA Analysis

A one-way analysis of variance (ANOVA) was used to determine whether there were any statistically significant differences between the means of the three or more independent (unrelated) groups. An F distribution can be used to compare this technique. A one-way ANOVA compares the means between related groups and determines whether these means are statistically significantly different from each other [106].

Step 1: ANOVA. Independent elements in the sum of squares are indicated by degrees of freedom (DF). The degrees of freedom for each component of the model are $\mathrm{DF} \left(\mathrm{Factor}\right)=\mathrm{r}-1$, $\mathrm{F} \mathrm{Error} ={\mathrm{n}}_{\mathrm{T}}-\mathrm{r}$, and$\mathrm{Total} ={ \mathrm{n}}_{\mathrm{T}}-1$ (${\mathrm{n}}_{\mathrm{T}}$ = total number of observations; r = number of factor levels). The F-value means that the degrees of freedom for the numerator are $\mathrm{r}-1$. The degrees of freedom for the denominator are ${\mathrm{n}}_{\mathrm{T}}-1$. The mean squares (MS) calculation for the factor/error is as follows (MS = mean square; SS = sum of squares; DF = degrees of freedom):

$$\mathrm{MS} \mathrm{Factor}=\frac{\mathrm{SS} \mathrm{Factor}}{\mathrm{DF} \mathrm{Factor}}$$

(4)

$$\mathrm{MS} \mathrm{Error}=\frac{\mathrm{SS} \mathrm{Error}}{\mathrm{DF} \mathrm{Error}}$$

(5)

Step 2: Post hoc tests—multiple comparisons and LSD. Post hoc tests are used for multiple comparisons with a control. Minitab offers four different confidence interval methods for comparing various factor means in a one-way analysis of variance with equal variances between the groups: Tukey’s, Fisher’s, Dunnett’s, and Hsu’s MCB. Fisher’s least significant difference (LSD) was used in the current paper for the individual error rate and the number of comparisons to calculate the simultaneous confidence level for all confidence intervals ($\overline{{\mathrm{Y}}_{\mathrm{i}}}$ = sample mean for the ${\mathrm{i}}^{\mathrm{th}}$ factor level; ${\mathrm{n}}_{\mathrm{i}}$ = number of observations in level i; r = number of factor levels; s = pooled standard deviation or sqrt (MSE); ${\mathrm{n}}_{\mathrm{T}}$ = total number of observations; α = probability of making α Type I error).

$$\overline{{\mathrm{Y}}_{\mathrm{i}}}-\overline{{\mathrm{Y}}_{\mathrm{ij}}} \pm \mathrm{t} \left(1-\frac{\alpha }{2};{ \mathrm{n}}_{\mathrm{i}}-\mathrm{r}\right)\mathrm{s}\sqrt{\frac{1}{{\mathrm{n}}_{\mathrm{i}}}+\frac{1}{{\mathrm{n}}_{\mathrm{j}}}}$$

(6)

Step 3: Mean plots. Mean: the average of the observations at a given factor level (${\mathrm{n}}_{\mathrm{i}}$ = number of observations at factor level i; ${\mathrm{y}}_{\mathrm{ji}}$ = value of the ${\mathrm{j}}^{\mathrm{th}}$ observation at the ${\mathrm{j}}^{\mathrm{th}}$ factor level).

$${\bar{\mathrm{x}}}_{\mathrm{i}}=\frac{{\sum }_{\mathrm{j}=1}^{{\mathrm{n}}_{\mathrm{i}}}{\mathrm{y}}_{\mathrm{ji}}}{{\mathrm{n}}_{\mathrm{i}}}$$

(7)

Figure 3 shows the sample of mean plots. The X axis represents the categorical variable X (time held), and the Y axis represents the continuous variable.

Step 4: Descriptions. Standard deviation (SD) ($\overline{{\mathrm{Y}}_{\mathrm{i}}}$ = mean of observations at the ${\mathrm{i}}^{\mathrm{th}}$ factor level; ${\mathrm{n}}_{\mathrm{i}}$ = number of observations at the ${\mathrm{i}}^{\mathrm{th}}$ factor level; ${\mathrm{y}}_{\mathrm{ji}}$ = observations at the ${\mathrm{i}}^{\mathrm{th}}$ factor level) was calculated as follows:

$${\mathrm{s}}_{\mathrm{i}}=\sqrt{\frac{{\sum }_{\mathrm{j}=1}^{{\mathrm{n}}_{\mathrm{i}}}{\left({\mathrm{y}}_{\mathrm{ji}}-\overline{{\mathrm{y}}_{\mathrm{i}}}\right)}^2}{{\mathrm{n}}_{\mathrm{i}}-1}}$$

(8)

4.5. Moran’s I

Moran’s I is one of the most frequently used methods for spatial cluster analysis [107,108,109]. Various previous papers studied geographic information systems (GIS) in relation to tourism. Yang et al. [110] studied the spatial tourist flows with cities and its growth rates using Moran’s I in China. Sarrión-Gavilán et al. [111] studied tourism flows using exploratory spatial data analysis (ESDA) and GIS in Andalusia. The range of Moran’s I is between −1 and 1 (1 = perfect positive autocorrelation; −1 = negative autocorrelation; 0 = no autocorrelation). This statistic was calculated for each year as follows (W = spatial weighting matrix; ${\mathrm{W}}_{\mathrm{ii}}$ on the diagonal were set to zero;${ \mathrm{z}}_{\mathrm{t}}$ = (${\mathrm{z}}_{1\mathrm{t}}$, ${\mathrm{z}}_{2\mathrm{t}}$, ${\mathrm{z}}_{3\mathrm{t}}$,…, ${\mathrm{z}}_{\mathrm{nt}}$) = tourist arrivals’ observation vector (from the mean) for the year t to the n cities in deviation; ${\mathrm{W}}_{\mathrm{ij}}$ = the way in which city I spatially connects to city j):

$${\mathrm{I}}_{\mathrm{t}}=\frac{\mathrm{n}}{{\mathrm{S}}_{\mathrm{o}}}·\frac{{\mathrm{z}}_{\mathrm{t}}^{\prime }{\mathrm{Wz}}_{\mathrm{t}}}{{\mathrm{z}}_{\mathrm{t}}^{\prime }{\mathrm{z}}_{\mathrm{t}}}$$

(9)

Moreover, a Moran’s scatterplot was used to measure the local spatial correlation and the local spatial correlation indicator in order to establish the importance of local spatial autocorrelation and hot spots. The local spatial association was studied by means of Moran’s scatterplot [112]. In addition, local indicators of spatial association (LISA) were used to test the hot spots’ significance via Moran’s scatterplot. Local Moran’s I was chosen as the LISA (${\mathrm{u}}_{\mathrm{t}}$ = the mean value across cities (year t); ${\mathrm{w}}_{\mathrm{ij}}$ = the spatial weighting matrix W factor; ${\mathrm{z}}_{\mathrm{it}}$ = the number of tourist arrivals (year t) to city i; ${\mathrm{I}}_{\mathrm{it}}$ (positive value) = a city and its neighbors for the spatial clustering of similar values):

$${\mathrm{I}}_{\mathrm{it}}=\frac{\left({\mathrm{z}}_{\mathrm{it}}-{\mathrm{u}}_{\mathrm{t}}\right)}{{\mathrm{m}}_{\mathrm{o}}}\sum _{\mathrm{j}}{\mathrm{w}}_{\mathrm{ij}}\left({\mathrm{z}}_{\mathrm{it}}-{\mathrm{u}}_{\mathrm{t}}\right)$$

(10)

$${\mathrm{m}}_{\mathrm{o}}=\sum _{\mathrm{j}}{\left({\mathrm{z}}_{\mathrm{it}}-{\mathrm{u}}_{\mathrm{t}}\right)}^2/\mathrm{n}$$

(11)

Negative values indicate the spatial clustering of different values. In this article, the I value of the local Moran is used as LISA statistics. To further visualize the tourist flow in hotspots (HH clusters), Moran importance maps were used. These contain information from Moran scatter plots and LISA, showing cities with important LISA statistics and indicating hotspot areas with color-coded quadrants in the Moran scatter plots to which these cities belong [113].

5. Results

5.1. The Descriptive Statistics

Figure 4 shows the changes in the number of tourists and the growth rate of tourists. These changes can be divided into three stages: (1) the fixed period before the exhibition; the number of tourists declined before 2001, and there was a slight fluctuation in the growth rate of tourists. (2) The fluctuation period of a single exhibition; the number of tourists fluctuated from 2001 to 2008, while the growth rate of tourists also began to fluctuate. (3) A double exhibition growth period; since 2009, the number of tourists has been increasing throughout the Niigata area; at the same time, the growth rate was the most significant fluctuation in these three stages. This shows the following points: (1) the exhibition has a “rising tide” impact on tourism in the area and (2) the double exhibition period has a more significant “rising tide” effect on the growth rate of tourists than the single exhibition period. This shows that different exhibitions complement each other and can better stimulate and drive tourism.

Figure 5 shows the economic ripple effect of ETAT and WLNAF from 2000 to 2018. Overall, the financial spread between 2000 and 2003 was the largest from the years studied. The ripple effect often colloquially denotes what would be called a multiplier in macroeconomics [114]. Infrastructure can influence the event tourism economics [115]. Because most of the ETAT hosting areas are located in rural areas, much of the infrastructure is initially required. Tamura et al. [67] studied a small Japanese town with spatial population distribution patterns in relation to environment and infrastructure costs. However, the infrastructure costs are not a SDGs engine. On the other hand, the WLNAF was hosted in Niigata (urban areas) with no need for infrastructure. The polynomial regression (divided into six stages: 2000–2003, 2003–2006, 2006–2009, 2009–2012, 2012–2015, and 2015–2018) shows that after experiencing the tourism-based economic growth driven by the investment in exhibition infrastructure in the early stages, the tourism economy entered a sustainable growth state in 2009 (two exhibitions were held at the same time, which no longer relied purely on exhibition infrastructure).

5.2. Simple Linear Regression (SLR) between Hosting Areas and Niigata Areas

The SLR was used for the next analysis (

Table 3. SLR: Model Summary.

Model a	R b	R Square	Adjusted R Square	Std. Error of the Estimate	Change Statistics
					R Square Change	F Change	df1	df2	Sig.
TN	0.824	0.679	0.615	0.529	0.679	10.593	3	15	0.001
TI	0.624	0.390	0.275	37,318	0.390	3.405	3	16	0.043
TII	0.989	0.978	0.972	99,244	0.978	151.364	3	10	0.000
PCI	0.979	0.959	0.951	22	0.959	123.518	3	16	0.000
TP	0.993	0.986	0.984	11,328	0.986	447.687	3	19	0.000
LP	0.999	0.999	0.999	4093	0.999	5218.00	3	19	0.000
HN	0.972	0.945	0.937	10,070	0.945	109.318	3	19	0.000

a. Predictors: (Constant), TN4/5/21, TI 4/5/21, TII 4/5/21, PCI 4/5/21, TP 4/5/21, LP4/5/21, HN4/5/21; b. Dependent Variable: TN/TI/TII/PCI 0.

, Figure 6). The p value is <0.05, which shows that the results are significant.

First, the adjusted R square of TN was 0.615 (>0.5). This shows that the hosting areas (TN21, TN5, and TN4) had a high positive impact on the Niigata area (TN0). Second, the adjusted R square of TI was 0275 (<0.3). This shows that the hosting areas (TI 21, TI 5, and TI 4) had an insignificant positive impact on the Niigata area (TI 0). Third, the adjusted R square of TII/PCI/TP/LP/HN was >0.8. This shows that the hosting areas (TII/PCI/TP/LP/HN 21, TII/PCI/TP/LP/HN 5, and TII/PCI/TP/LP/HN 4) had a very strong positive impact on the Niigata area (TII/PCI/TP/LP/HN 0). That is to say that (1) tourist number, (2) total income/tertiary industry income/per capita income, and (3) total population/labor population/household number in the hosting areas had a positive impact on the Niigata area. The impacts of arts event tourism go beyond the hosting areas. The next step was to test their impact on their local and neighborhood areas.

5.3. One-Way ANOVA: Total Income

5.3.1. Total Income

First,

Table 4. ANOVA (Total Income).

	df		MS	F	Sig.		MS	F	Sig.		MS	F	Sig.
a	4	TI0	656,415,244,879	17.516	0.000
b	18		37,475,971,024
a	4	TI1	1,210,085,983	8.171	0.001	TI11	28,895,330,302	11.841	0.000	TI21	2,946,305,756	1.182	0.377
b	18		148,099,707				2,440,224,174				2,580,796,734
a	4	TI2	8,273,877,931	15.434	0.000	TI12	1,357,891,695	7.824	0.002	TI22	5,031,478	0.111	0.977
b	18		536,079,989				173,544,468				45,372,332
a	4	TI3	2,544,811,438	2.428	0.097	TI13	16,303,468	16.769	0.000	TI23	334,226,911	14.022	0.000
b	18		1,047,993,256				972,247				23,835,344
a	4	TI4	101,059,294	10.033	0.000	TI14	32,066,970	0.502	0.735	TI24	226,598,415	2.351	0.104
b	18		10,073,133				63,873,350				96,376,084
a	4	TI5	690,228,981	12.893	0.000	TI15	615,334,483	3.489	0.035	TI25	201,890,786	0.436	0.781
b	18		53,533,123				176,357,993				463,442,867
a	4	TI6	928,871,622	7.313	0.002	TI16	1,338,954,477	13.590	0.000	TI26	167,400,453	1.707	0.204
b	18		127,010,165				98,527,695				98,063,293
a	4	TI7	427,438,014	3.734	0.029	TI17	28,914,694	13.248	0.000	TI27	35,925,215	6.622	0.003
b	18		114,486,226				2,182,521				5,425,128
a	4	TI8	1,507,644,034	16.934	0.000	TI18	209,312,770	13.320	0.000	TI28	5,461,628,820	3.147	0.048
b	18		89,030,682				15,713,655				1,735,666,800
a	4	TI9	2,553,681,975	34.620	0.000	TI19	14,349,980	8.350	0.001	TI29	1,769,860	5.319	0.008
b	18		73,763,483				1,718,472				332,763
a	4	TI10	2,828,820,421	1.680	0.210	TI20	165,474,567	5.073	0.010	TI30	4,258,672,838	13.087	0.000
b	18		1,683,574,885				32,620,003				325,405,173

a = between groups; b = within groups. Note: MS = mean square.

shows that the p-value (>0.5) of TI3/10/14/21/22/24/25/26 was not a significant fit. An event may significantly increase local economic activity, but the net impact within neighboring areas and cities may be more significant than the local (hosting areas) impact (e.g., the big/national effect often exceeds the small/state effect). In addition, the impact on the local/hosting areas can even be negative [46]. That is why the total income of the hosting area in Niigata City (T21) was negative. However, 73% of the cities (22 cities and the Niigata area) in Niigata passed the one-way ANOVA analysis. Second, the 22 cities and the Niigata area were selected for multiple comparisons testing using LSD (

Table A1. Total income—post hoc tests—multiple comparisons using LSD.

I	J		MD (I–J)	Sig.		MD (I–J)	Sig.		MD (I–J)	Sig.		MD (I–J)	Sig.		MD (I–J)	Sig.		MD (I–J)	Sig.
NO	Y1	TI0	368,697	0.035	TI5	8963	0.156	TI9	4579	0.524	TI15	8638	0.439	TI19	1265	0.257	TI28	−27,532	0.432
	B1		499,675	0.003		16,832	0.006		18,499	0.009		18,076	0.075		1700	0.088		−26,258	0.388
	Y2		1,032,318	0.000		33,075	0.000		55,381	0.000		29,131	0.018		4487	0.001		−88,540	0.021
	B2		1,037,094	0.000		32,726	0.000		57,300	0.000		32,666	0.006		4651	0.000		−87,016	0.016
Y1	B1		130,979	0.355		7869	0.151		13,920	0.038		9437	0.332		434	0.647		1273	0.966
	Y2		663,621	0.001		24,112	0.001		50,802	0.000		20,493	0.080		3222	0.009		−61,009	0.095
	B2		668,398	0.000		23,763	0.001		52,720	0.000		24,028	0.033		3385	0.004		−59,484	0.083
B1	Y2		532,643	0.002		16,244	0.007		36,882	0.000		11,056	0.259		2787	0.009		−62,282	0.053
	B2		537,419	0.001		15,895	0.005		38,801	0.000		14,591	0.111		2951	0.004		−60,758	0.040
Y2	B2		4777	0.975		−349	0.951		1919	0.774		3535	0.733		164	0.873		1524	0.962
NO	Y1	TI1	31,152	0.007	TI6	20,488	0.043	TI11	87,731	0.047	TI16	9601	0.256	TI20	11,730	0.025	TI29	649	0.190
	B1		38,015	0.001		24,254	0.009		98,712	0.013		15,898	0.040		17,046	0.001		620	0.151
	Y2		52,347	0.000		40,538	0.001		222,228	0.000		42,672	0.000		10,628	0.039		1663	0.003
	B2		41,769	0.001		41,363	0.000		214,491	0.000		44,040	0.000		16,286	0.002		1649	0.002
Y1	B1		6863	0.438		3766	0.644		10,981	0.758		6297	0.385		5317	0.209		−30	0.943
	Y2		21,196	0.051		20,049	0.047		134,497	0.005		33,070	0.001		−1102	0.817		1014	0.049
	B2		10,618	0.272		20,874	0.029		126,761	0.005		34,438	0.000		4556	0.314		1000	0.040
B1	Y2		14,333	0.118		16,283	0.060		123,516	0.003		26,773	0.002		−6418	0.134		1044	0.023
	B2		3755	0.640		17,108	0.034		115,779	0.003		28,141	0.001		−761	0.840		1029	0.015
Y2	B2		−10,578	0.274		825	0.925		−7737	0.840		1368	0.859		5658	0.216		−14	0.975
NO	Y1	TI2	−3636	0.850	TI7	17,011	0.072	TI12	21,189	0.069	TI17	−2321	0.075	TI23	15,027	0.002	TI30	28,740	0.071
	B1		−613	0.971		19,581	0.021		23,113	0.026		−70	0.948		16,305	0.000		42,350	0.005
	Y2		76,648	0.001		30,317	0.004		49,754	0.000		4488	0.002		24,038	0.000		84,034	0.000
	B2		90,940	0.000		26,580	0.006		46,569	0.000		4031	0.003		25,775	0.000		83,113	0.000
Y1	B1		3023	0.856		2570	0.739		1924	0.839		2251	0.049		1278	0.717		13,610	0.304
	Y2		80,284	0.001		13,307	0.150		28,565	0.019		6809	0.000		9011	0.040		55,295	0.002
	B2		94,576	0.000		9569	0.261		25,380	0.024		6352	0.000		10,747	0.012		54,373	0.001
B1	Y2		77,261	0.000		10,736	0.178		26,641	0.013		4558	0.001		7733	0.042		41,685	0.006
	B2		91,553	0.000		6999	0.328		23,456	0.015		4101	0.001		9469	0.009		40,763	0.004
Y2	B2		14,292	0.433		−3738	0.654		−3185	0.756		−456	0.692		1736	0.649		−922	0.948
NO	Y1	TI4	6930	0.018	TI8	22,350	0.012	TI13	2633	0.006	TI18	8694	0.018	TI27	2704	0.177
	B1		8792	0.002		29,952	0.001		2152	0.008		13,969	0.000		4016	0.029
	Y2		13,769	0.000		50,716	0.000		5494	0.000		18,667	0.000		7425	0.002
	B2		13,477	0.000		52,020	0.000		4926	0.000		19,577	0.000		7899	0.001
Y1	B1		1862	0.421		7601	0.274		−481	0.502		5275	0.081		1311	0.439
	Y2		6839	0.019		28,366	0.002		2861	0.003		9974	0.008		4721	0.026
	B2		6546	0.017		29,670	0.001		2293	0.009		10,884	0.003		5195	0.011
B1	Y2		4977	0.044		20,764	0.008		3342	0.000		4699	0.116		3410	0.057
	B2		4684	0.038		22,069	0.003		2774	0.001		5609	0.046		3883	0.022
Y2	B2		−292	0.906		1304	0.859		−567	0.464		910	0.768		474	0.794

) and mean plots (Figure 7). Table A1 shows the descriptive statistics. Here, we can see that the hosting of double exhibitions positively impacted the neighborhood areas and the whole Niigata area. In addition, the rising tide process for 95% of the fitting areas was the same (excepting TI28). Parts of the mean total income plots exhibited a positive growth trend, but the majority exhibited a negative growth trend.

5.3.2. Tertiary Industry Income

First,

Table 5. ANOVA (Tertiary Industry Income).

	df		MS	F	Sig.		MS	F	Sig.		MS	F	Sig.
a	4	TII0	1,342,753,512,113	83.599	0.000
b	18		16,061,887,716
a	4	TII1	604,895,005	33.164	0.000	TII11	8,893,164,576	2.451	0.114	TII21	192,877,319,781	9.653	0.002
b	18		18,239,409				3,627,656,709				19,980,050,156
a	4	TII2	356,764,030	63.426	0.000	TII12	497,009,534	2.488	0.110	TII22	247,337,672	87.687	0.000
b	18		5,624,929				199,722,627				2,820,700
a	4	TII3	10,625,978,362	59.464	0.000	TII13	3,489,956	16.659	0.000	TII23	88,216,224	15.081	0.000
b	18		178,697,467				209,492				5,849,310
a	4	TII4	83,230,682	5.158	0.016	TII14	404,983,310	149.886	0.000	TII24	1,948,667,310	361.804	0.000
b	18		16,136,347				2,701,935				5,385,971
a	4	TII5	123,196,734	0.130	0.968	TII15	2,531,481,017	82.470	0.000	TII25	1,537,895,068	15.237	0.000
b	18		946,397,616				30,695,961				100,933,115
a	4	TII6	298,309,518	6.329	0.008	TII16	1,761,355,407	331.565	0.000	TII26	113,212,883	32.277	0.000
b	18		47,132,916				5,312,249				3,507,562
a	4	TII7	1,599,530,277	137.053	0.000	TII17	15,585,059	102.551	0.000	TII27	5,898,076	10.819	0.001
b	18		11,670,906				151,974				545,143
a	4	TII8	633,321,292	71.193	0.000	TII18	141,213,761	62.679	0.000	TII28	1,102,284,497	0.774	0.567
b	18		8,895,780				2,252,971				1,424,808,776
a	4	TII9	375,855,603	96.340	0.000	TII19	20,039,870	112.971	0.000	TII29	1,179,623	55.907	0.000
b	18		3,901,342				177,390				21,100
a	4	TII10	30,587,378,331	194.814	0.000	TII20	405,722,020	55.925	0.000	TII30	1,020,896,820	43.165	0.000
b	18		157,008,447				7,254,778				23,651,042

a = between groups; b = within groups. Note: MS = mean square.

shows that the p-value (>0.5) of TII5/11/12/28 was not a significant fit. However, 87% (26 cities and the Niigata area) of the cities in Niigata passed the one-way ANOVA analysis. Second, the 26 cities and the Niigata area were selected for multiple comparisons testing using LSD (

Table A2. Tertiary industry income—post hoc tests—multiple comparisons using LSD.

I	J		MD (I–J)	Sig.		MD (I–J)	Sig.		MD (I–J)	Sig.		MD (I–J)	Sig.		MD (I–J)	Sig.		MD (I–J)	Sig.
NO	Y1	TII0	−1,387,250	0.000	TII4	−12,260	0.003	TII9	−26,337	0.000	TII15	−60,416	0.000	TII19	−5622	0.000	TII23	−12,459	0.000
	B1		−1,473,547	0.000		−10,549	0.004		−22,391	0.000		−59,950	0.000		−5404	0.000		−9798	0.000
	Y2		−1,104,793	0.000		−8083	0.043		−15,628	0.000		−53,109	0.000		−4254	0.000		−2905	0.196
	B2		−1,102,233	0.000		−8520	0.034		−15,805	0.000		−58,273	0.000		−4737	0.000		−4146	0.076
Y1	B1		−86297	0.394		1711	0.589		3946	0.026		466	0.915		218	0.513		2661	0.180
	Y2		282,457	0.035		4177	0.281		10,709	0.000		7307	0.179		1368	0.005		9554	0.001
	B2		285,017	0.033		3740	0.332		10,532	0.000		2143	0.681		885	0.044		8313	0.004
B1	Y2		368,754	0.007		2466	0.495		6763	0.003		6841	0.184		1150	0.010		6893	0.008
	B2		371,314	0.007		2030	0.573		6586	0.003		1677	0.734		667	0.097		5652	0.022
Y2	B2		2560	0.984		−437	0.916		−178	0.930		−5164	0.373		−483	0.279		−1241	0.619
NO	Y1	TII1	−31,655	0.000	TII6	−24,428	0.001	TII10	−198,932	0.000	TII16	−51,674	0.000	TII20	−25,494	0.000	TII24	−52,867	0.000
	B1		−29,380	0.000		−16,389	0.007		−216,664	0.000		−51,694	0.000		−25,021	0.000		−53,023	0.000
	Y2		−15,695	0.002		−5919	0.343		−186,900	0.000		−39,084	0.000		−17,284	0.000		−43,875	0.000
	B2		−13,776	0.004		−9879	0.128		−198,395	0.000		−44,050	0.000		−16,726	0.000		−51,922	0.000
Y1	B1		2275	0.501		8039	0.156		−17,732	0.094		−20	0.991		473	0.823		−156	0.932
	Y2		15,961	0.002		18,509	0.014		12,032	0.318		12,590	0.000		8210	0.008		8992	0.002
	B2		17,879	0.001		14,549	0.043		537	0.963		7624	0.005		8768	0.005		945	0.665
B1	Y2		13,686	0.004		10,470	0.109		29,764	0.021		12,610	0.000		7737	0.008		9149	0.001
	B2		15,604	0.002		6510	0.299		18,269	0.123		7645	0.003		8295	0.005		1101	0.596
Y2	B2		1919	0.663		−3960	0.577		−11,495	0.381		−4966	0.057		558	0.840		−8048	0.006
NO	Y1	TII2	−23,899	0.000	TII7	−48,607	0.000	TII13	−2579	0.000	TII17	−5170	0.000	TII21	−360,834	0.007	TII25	−57,498	0.000
	B1		−21,321	0.000		−49,990	0.000		−2074	0.000		−4697	0.000		−577,842	0.000		−31,931	0.001
	Y2		−6733	0.008		−38,916	0.000		−1380	0.006		−2411	0.000		−482,311	0.003		−41,433	0.001
	B2		−8345	0.002		−39,750	0.000		−1145	0.016		−3056	0.000		−489,660	0.003		−31,937	0.004
Y1	B1		2578	0.185		−1382	0.608		506	0.179		473	0.143		−217,009	0.072		25,566	0.008
	Y2		17,165	0.000		9692	0.011		1199	0.017		2759	0.000		−121,478	0.369		16,065	0.110
	B2		15,554	0.000		8857	0.018		1435	0.006		2114	0.000		−128,827	0.342		25,561	0.019
B1	Y2		14,588	0.000		11,074	0.004		694	0.111		2286	0.000		95,531	0.453		−9501	0.300
	B2		12,976	0.000		10,240	0.006		929	0.041		1641	0.001		88,182	0.488		−5	1.000
Y2	B2		−1612	0.512		−835	0.812		236	0.618		−645	0.129		−7349	0.960		9496	0.367
NO	Y1	TII3	−114,203	0.000	TII8	−30,797	0.000	TII14	−21,259	0.000	TII18	−15,040	0.000	TII22	−19,366	0.000	TII26	−14,246	0.000
	B1		−112,225	0.000		−31,882	0.000		−24,050	0.000		−14,735	0.000		−19,525	0.000		−12,136	0.000
	Y2		−115,113	0.000		−18,074	0.000		−24,652	0.000		−9703	0.000		−15,332	0.000		−10,555	0.000
	B2		−141,230	0.000		−17,005	0.000		−23,850	0.000		−8402	0.000		−15,353	0.000		−10,116	0.000
Y1	B1		1978	0.850		−1085	0.644		−2792	0.050		305	0.796		−159	0.904		2110	0.171
	Y2		−910	0.942		12,723	0.001		−3394	0.047		5336	0.003		4034	0.025		3691	0.056
	B2		−27,027	0.051		13,792	0.000		−2592	0.115		6637	0.001		4013	0.026		4131	0.036
B1	Y2		−2888	0.808		13,809	0.000		−602	0.681		5032	0.003		4193	0.016		1581	0.353
	B2		−29,005	0.031		14,877	0.000		200	0.891		6333	0.001		4172	0.017		2020	0.241
Y2	B2		−26,117	0.079		1069	0.728		802	0.636		1301	0.406		−21	0.990		440	0.819

) and mean plots (Figure 8). Table A2 shows the descriptive statistics. This shows that the hosting of double exhibitions positively impacted the tertiary industry income (TII) in the neighborhood areas and the whole Niigata area. In addition, the rising tide process for 73% of the fitting areas (excepting TII1/6/9/16/22/23/25) was the same. Parts of the mean tertiary industry income plots exhibited a negative growth trend. In total, 100% of the local areas (the hosting areas) and 52 of neighborhood areas exhibited a positive growth trend in relation to tertiary industry income. The Niigata area also exhibited a positive growth trend.

5.3.3. Per Capita Income

First,

Table 6. ANOVA (Per Capita Income).

	df		MS	F	Sig.		MS	F	Sig.		MS	F	Sig.
a	4	PCI0	28,133	3.541	0.03
b	18		7946
a	4	PCI1	57,229	4.298	0.015	PCI11	36,112	3.273	0.039	PCI21	24,928	3.587	0.029
b	18		13,316				11,034				6949
a	4	PCI2	197,425	9.466	0.000	PCI12	125,264	4.354	0.014	PCI22	11,229	1.321	0.305
b	18		20,856				28,767				8499
a	4	PCI3	43,692	3.733	0.025	PCI13	27,183	4.145	0.017	PCI23	188,472	14.439	0.000
b	18		11,704				6559				13,053
a	4	PCI4	102,938	5.764	0.005	PCI14	18,779	0.742	0.577	PCI24	51,513	6.853	0.002
b	18		17,858				25,307				7517
a	4	PCI5	66,873	8.407	0.001	PCI15	11,398	1.145	0.371	PCI25	153,565	3.782	0.024
b	18		7954				9953				40,604
a	4	PCI6	404,902	2.427	0.091	PCI16	14,456	0.842	0.519	PCI26	26,376	1.332	0.301
b	18		166,819				17,175				19,805
a	4	PCI7	26,872	4.892	0.009	PCI17	81,863	5.88	0.004	PCI27	79,889	5.555	0.005
b	18		5493				13,921				14,381
a	4	PCI8	79,044	7.936	0.001	PCI18	16,463	2.754	0.064	PCI28	145,313	10.623	0.000
b	18		9960				5977				13,679
a	4	PCI9	59,046	5.935	0.004	PCI19	47,729	10.816	0.000	PCI29	336,389	4.496	0.013
b	18		9948				4413				74,825
a	4	PCI10	18,019	1.599	0.223	PCI20	24,966	2.89	0.056	PCI30	255,331	12.197	0.000
b	18		11,267				8639				20,933

a = between groups; b = within groups. Note: MS = mean square.

shows that the p-value (<0.5) of PCI6/10/14/15/16/18/20/22/26 was not a significant fit. However, 70% (21 cities and the Niigata area) of the cities in Niigata passed the one-way ANOVA analysis. Second, the 21 cities and the Niigata area were selected for multiple comparisons testing using LSD (

Table A3. Per capita income—post hoc tests—multiple comparisons using LSD.

I	J		MD (I–J)	Sig.		MD (I–J)	Sig.		MD (I–J)	Sig.		MD (I–J)	Sig.		MD (I–J)	Sig.		MD (I–J)	Sig.
NO	Y1	PCI0	128	0.078	PCI4	202.333	0.065	PCI9	38	0.628	PCI17	148	0.119	PCI24	155	0.033	PCI29	338.3	0.125
	B1		165	0.011		281.5	0.005		154	0.029		154	0.060		193	0.003		561.4	0.006
	Y2		228	0.004		399.667	0.001		296	0.001		365	0.001		299	0.000		721.3	0.003
	B2		176	0.009		378.6	0.001		244	0.002		315	0.001		259	0.000		654.7	0.003
Y1	B1		37	0.564		79.167	0.414		116	0.119		6	0.947		38	0.549		223.2	0.266
	Y2		100	0.188		197.333	0.089		259	0.006		216	0.039		144	0.059		383.0	0.106
	B2		48	0.469		176.267	0.09		206	0.012		167	0.070		104	0.120		316.4	0.133
B1	Y2		63	0.334		118.167	0.229		142	0.061		211	0.023		107	0.102		159.8	0.421
	B2		11	0.840		97.1	0.248		90	0.156		162	0.038		67	0.223		93.2	0.581
Y2	B2		−52	0.438		−21.067	0.832		−53	0.481		−49	0.576		−40	0.536		−66.6	0.743
NO	Y1	PCI1	192	0.044	PCI5	176.333	0.02	PCI11	−1	0.994	PCI19	109	0.047	PCI25	194	0.225	PCI30	295.3	0.017
	B1		206	0.014		226.667	0.001		38	0.587		147	0.004		229	0.098		393.8	0.001
	Y2		303	0.003		324	0		203	0.022		268	0.000		494	0.006		621.7	0.000
	B2		286	0.002		307	0		173	0.026		255	0.000		441	0.005		602.2	0.000
Y1	B1		14	0.864		50.333	0.436		38	0.614		37	0.438		35	0.812		98.5	0.350
	Y2		110	0.259		147.667	0.06		203	0.031		159	0.010		299	0.088		326.3	0.014
	B2		94	0.282		130.667	0.062		174	0.038		146	0.008		246	0.114		306.9	0.010
B1	Y2		96	0.256		97.333	0.142		165	0.041		122	0.020		265	0.082		227.8	0.041
	B2		80	0.271		80.333	0.156		136	0.049		108	0.016		212	0.102		208.4	0.030
Y2	B2		−17	0.847		−17	0.797		−30	0.705		−13	0.785		−53	0.723		−19.5	0.856
NO	Y1	PCI2	192	0.101	PCI7	124.167	0.043	PCI12	−192	0.158	PCI21	142	0.041	PCI27	222	0.027
	B1		245	0.018		162.833	0.004		−53	0.638		173	0.005		262	0.004
	Y2		537	0.000		214.167	0.002		187	0.168		206	0.005		379	0.001
	B2		504	0.000		181.9	0.002		249	0.044		159	0.012		314	0.001
Y1	B1		53	0.609		38.667	0.471		139	0.262		32	0.597		40	0.644
	Y2		346	0.010		90	0.156		379	0.015		64	0.359		157	0.128
	B2		312	0.009		57.733	0.302		440	0.003		18	0.776		92	0.311
B1	Y2		292	0.011		51.333	0.342		240	0.063		33	0.589		117	0.187
	B2		259	0.009		19.067	0.677		301	0.010		−14	0.782		52	0.487
Y2	B2		−33	0.756		−32.267	0.559		61	0.627		−47	0.454		−65	0.466
NO	Y1	PCI3	82	0.336	PCI8	221.167	0.01	PCI13	181	0.01	PCI23	347	0.001	PCI28	292	0.005
	B1		171	0.026		274.5	0.001		184	0.003		400	0.000		346	0.000
	Y2		294	0.003		365.167	0		208	0.004		542	0.000		488	0.000
	B2		185	0.021		318.3	0		143	0.018		519	0.000		447	0.000
Y1	B1		89	0.261		53.333	0.461		3	0.961		54	0.517		55	0.519
	Y2		212	0.029		144	0.096		27	0.685		195	0.053		197	0.056
	B2		104	0.209		97.133	0.201		−38	0.526		172	0.055		155	0.088
B1	Y2		123	0.128		90.667	0.217		25	0.674		142	0.099		142	0.105
	B2		14	0.829		43.8	0.479		−41	0.414		119	0.105		101	0.174
Y2	B2		−108	0.189		−46.867	0.529		−66	0.283		−23	0.790		−41	0.634

) and mean plots (Figure 9). Table A3 shows the descriptive statistics. This shows that the hosting of double exhibitions positively impacted tertiary industry income (TII) in the neighborhood areas and the whole Niigata area. In addition, the rising tide process for 73% of the fitting areas (excepting PCI12) was the same. Only one of the mean per capita income plots exhibited a positive growth trend. The majority exhibited a negative growth trend.

5.4. One-Way ANOVA: Population

5.4.1. Total Population

First,

Table 7. ANOVA (Total Population).

	df		MS	F	Sig.		MS	F	Sig.		MS	F	Sig.
a	4	TP0	38,242,857,927	13.606	0.000
b	19		2,810,822,930
a	4	TP1	64,469,522	16.328	0.000	TP11	120,552,719	14.499	0.000	TP21	181,562,716	7.099	0.001
b	19		3,948,516				8,314,793				25,575,465
a	4	TP2	37,456,965	18.530	0.000	TP12	309,797	11.366	0.000	TP22	24,720,539	14.885	0.000
b	19		2,021,369				27,256				1,660,783
a	4	TP3	197,172,852	12.527	0.000	TP13	1,497,464	19.904	0.000	TP23	15,459,175	17.031	0.000
b	19		15,740,076				75,233				907,733
a	4	TP4	5,440,750	18.302	0.000	TP14	7,858,503	15.275	0.000	TP24	53,353,977	14.714	0.000
b	19		297,279				514,469				3,625,992
a	4	TP5	96,658,471	15.440	0.000	TP15	72,987,681	17.580	0.000	TP25	476,898	27.672	0.000
b	19		6,260,177				4,151,862				17,234
a	4	TP6	942,279	21.408	0.000	TP16	17,428,149	9.776	0.000	TP26	16,200,960	20.658	0.000
b	19		44,016				1,782,738				784,254
a	4	TP7	40,717,060	13.293	0.000	TP17	93,722	1.903	0.151	TP27	2,469,633	17.676	0.000
b	19		3,063,006				49,257				139,720
a	4	TP8	58,453,611	16.084	0.000	TP18	24,626,076	15.977	0.000	TP28	2,802,885,271	13.572	0.000
b	19		3,634,193				1,541,380				206,519,346
a	4	TP9	22,732,489	13.807	0.000	TP19	2,410,253	15.201	0.000	TP29	5045	14.150	0.000
b	19		1,646,498				158,557				357
a	4	TP10	23,219,140	1.377	0.279	TP20	52,097,693	14.473	0.000	TP30	192,168,084	18.898	0.000
b	19		16,857,229				3,599,749				10,168,506

a = between groups; b = within groups. Note: MS = mean square.

shows that the p-value (>0.5) of TP10/17 was not a significant fit. However, 93% (28 cities and the Niigata area) of the cities in Niigata passed the one-way ANOVA analysis. Second, the 28 cities and the Niigata area were selected for multiple comparisons testing using LSD (

Table A4. Total population—post hoc tests—multiple comparisons using LSD.

I	J		MD (I–J)	Sig.		MD (I–J)	Sig.		MD (I–J)	Sig.		MD (I–J)	Sig.		MD (I–J)	Sig.		MD (I–J)	Sig.
NO	Y1	TP0	4060	0.926	TP5	708	0.733	TP11	915	0.702	TP16	−625	0.573	TP22	68	0.949	TP27	142	0.648
	B1		27,691	0.469		1758	0.333		2125	0.310		−478	0.618		620	0.505		290	0.286
	Y2		158,290	0.001		8122	0.000		9175	0.001		2448	0.027		3878	0.001		1316	0.000
	B2		180,083	0.000		9392	0.000		10,587	0.000		3214	0.002		4569	0.000		1511	0.000
Y1	B1		23,631	0.536		1050	0.560		1210	0.560		147	0.878		552	0.552		149	0.580
	Y2		154,230	0.001		7414	0.001		8261	0.001		3073	0.007		3810	0.001		1174	0.001
	B2		176,023	0.000		8684	0.000		9672	0.000		3839	0.000		4501	0.000		1369	0.000
B1	Y2		130,599	0.001		6364	0.001		7050	0.001		2926	0.003		3258	0.001		1025	0.000
	B2		152,392	0.000		7634	0.000		8462	0.000		3692	0.000		3949	0.000		1220	0.000
Y2	B2		21,794	0.510		1270	0.418		1412	0.434		766	0.361		691	0.392		195	0.405
NO	Y1	TP1	614	0.709	TP6	78	0.654	TP12	290	0.045	TP18	198	0.847	TP23	341	0.666	TP28	11,893	0.324
	B1		1565	0.279		239	0.123		419	0.002		664	0.459		712	0.304		15,663	0.140
	Y2		6693	0.000		847	0.000		645	0.000		3973	0.000		3275	0.000		52,188	0.000
	B2		7734	0.000		949	0.000		677	0.000		4599	0.000		3773	0.000		53,492	0.000
Y1	B1		951	0.507		161	0.290		129	0.283		466	0.602		371	0.589		3769	0.715
	Y2		6079	0.001		769	0.000		355	0.011		3775	0.001		2934	0.001		40,295	0.002
	B2		7120	0.000		871	0.000		387	0.003		4401	0.000		3431	0.000		41,598	0.000
B1	Y2		5128	0.001		607	0.000		226	0.047		3310	0.001		2563	0.001		36,526	0.001
	B2		6169	0.000		710	0.000		258	0.009		3936	0.000		3061	0.000		37,829	0.000
Y2	B2		1041	0.403		103	0.434		32	0.757		626	0.421		498	0.404		1303	0.884
NO	Y1	TP2	131	0.911	TP7	371	0.798	TP13	175	0.444	TP19	−21	0.949	TP24	−477	0.762	TP29	−3	0.831
	B1		719	0.483		992	0.433		331	0.104		152	0.596		90	0.947		−3	0.834
	Y2		4856	0.000		5091	0.001		1110	0.000		1171	0.001		5192	0.002		51	0.002
	B2		5588	0.000		6042	0.000		1225	0.000		1400	0.000		6171	0.000		58	0.000
Y1	B1		588	0.566		621	0.622		156	0.431		173	0.547		568	0.678		1	0.971
	Y2		4725	0.000		4721	0.002		935	0.000		1192	0.001		5670	0.001		54	0.001
	B2		5456	0.000		5671	0.000		1050	0.000		1421	0.000		6648	0.000		61	0.000
B1	Y2		4137	0.000		4100	0.002		779	0.000		1019	0.001		5102	0.001		53	0.000
	B2		4868	0.000		5050	0.000		894	0.000		1248	0.000		6081	0.000		61	0.000
Y2	B2		731	0.411		951	0.386		115	0.504		229	0.359		979	0.412		8	0.524
NO	Y1	TP3	−1840	0.577	TP8	408	0.796	TP14	−211	0.722	TP20	167	0.915	TP25	−294	0.013	TP30	982	0.710
	B1		−605	0.832		1031	0.454		84	0.871		906	0.508		−322	0.003		2364	0.308
	Y2		9189	0.007		6235	0.000		2011	0.002		5700	0.001		−712	0.000		11,543	0.000
	B2		11,274	0.000		7087	0.000		2382	0.000		6638	0.000		−788	0.000		13,140	0.000
Y1	B1		1235	0.665		624	0.649		295	0.568		738	0.589		−28	0.769		1382	0.547
	Y2		11,029	0.002		5827	0.001		2222	0.001		5533	0.001		−417	0.001		10,561	0.000
	B2		13,114	0.000		6680	0.000		2593	0.000		6470	0.000		−493	0.000		12,158	0.000
B1	Y2		9793	0.001		5204	0.000		1927	0.001		4795	0.001		−390	0.000		9178	0.000
	B2		11,879	0.000		6056	0.000		2298	0.000		5732	0.000		−466	0.000		10,775	0.000
Y2	B2		2085	0.401		852	0.474		371	0.409		937	0.430		−76	0.357		1597	0.424
NO	Y1	TP4	257	0.570	TP9	355	0.739	TP15	−176	0.917	TP21	−11337	0.013	TP26	−106	0.885
	B1		528	0.187		764	0.410		966	0.511		−11991	0.003		348	0.585
	Y2		2021	0.000		3936	0.001		6657	0.000		−1714	0.662		3131	0.000
	B2		2288	0.000		4510	0.000		7682	0.000		−46	0.989		3548	0.000
Y1	B1		270	0.492		410	0.657		1142	0.438		−653	0.857		454	0.477
	Y2		1763	0.000		3582	0.002		6833	0.000		9623	0.022		3237	0.000
	B2		2030	0.000		4155	0.000		7858	0.000		11,291	0.004		3654	0.000
B1	Y2		1493	0.000		3172	0.001		5691	0.000		10,276	0.005		2783	0.000
	B2		1760	0.000		3746	0.000		6716	0.000		11,944	0.000		3200	0.000
Y2	B2		267	0.434		574	0.474		1025	0.421		1668	0.596		417	0.451

) and mean plots (Figure 10). Table A1 shows the descriptive statistics. Only two of the total population plots exhibited a positive growth trend. The majority exhibited a negative growth trend.

5.4.2. Labor Population

First,

Table 8. ANOVA (Labor Population).

	df		MS	F	Sig.		MS	F	Sig.		MS	F	Sig.
a	4	LP0	59,308,416,492	15.111	0.000
b	19		3,924,942,746
a	4	LP1	65,920,441	18.533	0.000	LP11	142,602,155	14.431	0.000	LP21	4,074,518,569	15.284	0.000
b	19		3,556,895				9,881,935				266,587,308
a	4	LP2	31,229,521	17.364	0.000	LP12	320,266	13.180	0.000	LP22	24,725,157	11.645	0.000
b	19		1,798,499				24,300				2,123,244
a	4	LP3	467,260,853	18.088	0.000	LP13	676,597	16.151	0.000	LP23	9,497,889	19.553	0.000
b	19		25,832,996				41,893				485,752
a	4	LP4	2,728,448	17.360	0.000	LP14	19,647,008	16.526	0.000	LP24	102,413,697	13.960	0.000
b	19		157,169				1,188,826				7,336,351
a	4	LP5	98,798,611	17.548	0.000	LP15	160,522,118	20.311	0.000	LP25	5052	0.208	0.931
b	19		5,630,303				7,903,370				24,311
a	4	LP6	1,526,743	22.534	0.000	LP16	74,452,415	17.526	0.000	LP26	23,013,872	20.652	0.000
b	19		67,754				4,248,176				1,114,374
a	4	LP7	36,435,378	11.114	0.000	LP17	499,814	6.391	0.002	LP27	1,392,328	16.224	0.000
b	19		3,278,454				78,207				85,821
a	4	LP8	42,832,982	16.146	0.000	LP18	25,547,365	17.914	0.000	LP28	10,939,589	2.144	0.115
b	19		2,652,920				1,426,080				5,102,115
a	4	LP9	29,300,613	15.746	0.000	LP19	3,668,188	13.684	0.000	LP29	4420	22.696	0.000
b	19		1,860,885				268,060				195
a	4	LP10	422,125,699	10.047	0.000	LP20	63,616,834	15.299	0.000	LP30	121,186,584	21.249	0.000
b	19		42,016,840				4,158,202				5,703,272

a = between groups; b = within groups. Note: MS = mean square.

shows that the p-value (>0.5) of LP25/28 was not a significant fit. However, 93% (28 cities and the Niigata area) of the cities in Niigata passed the one-way ANOVA analysis. Second, the 28 cities and the Niigata area were selected for multiple comparisons testing using LSD (

Table A5. Labor population post hoc tests—multiple comparisons using LSD.

I	J		MD (I–J)	Sig.		MD (I–J)	Sig.		MD (I–J)	Sig.		MD (I–J)	Sig.		MD (I–J)	Sig.		MD (I–J)	Sig.
NO	Y1	LP0	52,509	0.318	LP5	2342	0.242	LP10	4385	0.418	LP15	2894	0.223	LP20	1747	0.307	LP26	761	0.389
	B1		76,059	0.102		3474	0.052		6289	0.186		4298	0.044		2533	0.095		1221	0.118
	Y2		223,150	0.000		9341	0.000		18,643	0.001		11,930	0.000		7391	0.000		4269	0.000
	B2		254,377	0.000		10,547	0.000		21,445	0.000		13,342	0.000		8332	0.000		4819	0.000
Y1	B1		23,550	0.601		1133	0.508		1904	0.682		1404	0.489		787	0.592		460	0.545
	Y2		170,641	0.002		7000	0.001		14,258	0.010		9036	0.000		5645	0.002		3508	0.000
	B2		201,867	0.000		8206	0.000		17,061	0.001		10,448	0.000		6585	0.000		4059	0.000
B1	Y2		147,091	0.002		5867	0.001		12,354	0.008		7632	0.000		4858	0.002		3049	0.000
	B2		178,317	0.000		7073	0.000		15,156	0.000		9044	0.000		5798	0.000		3599	0.000
Y2	B2		31,227	0.426		1206	0.417		2803	0.489		1411	0.423		940	0.461		550	0.405
NO	Y1	LP1	1927	0.226	LP6	337	0.129	LP11	3131	0.237	LP16	2283	0.191	LP21	13,764	0.315	LP27	307	0.214
	B1		2903	0.042		531	0.009		4376	0.064		3129	0.045		18,539	0.125		426	0.054
	Y2		7711	0.000		1225	0.000		11,322	0.000		8254	0.000		58,084	0.000		1121	0.000
	B2		8622	0.000		1349	0.000		12,838	0.000		9241	0.000		66,103	0.000		1264	0.000
Y1	B1		975	0.473		194	0.304		1244	0.582		846	0.568		4775	0.684		118	0.575
	Y2		5783	0.001		888	0.000		8191	0.003		5972	0.001		44,319	0.002		814	0.002
	B2		6695	0.000		1012	0.000		9706	0.000		6958	0.000		52,339	0.000		957	0.000
B1	Y2		4808	0.001		694	0.001		6947	0.003		5126	0.001		39,544	0.001		695	0.002
	B2		5719	0.000		817	0.000		8462	0.000		6112	0.000		47,564	0.000		838	0.000
Y2	B2		911	0.440		124	0.447		1515	0.441		986	0.444		8020	0.432		143	0.436
NO	Y1	LP2	1011	0.367	LP7	1554	0.306	LP12	341	0.015	LP17	166	0.476	LP22	989	0.416	LP29	10	0.376
	B1		1541	0.121		1945	0.145		438	0.001		208	0.306		1434	0.180		19	0.067
	Y2		5049	0.000		5479	0.001		642	0.000		627	0.008		4371	0.001		62	0.000
	B2		5692	0.000		6433	0.000		710	0.000		743	0.001		5169	0.000		67	0.000
Y1	B1		530	0.583		391	0.763		97	0.391		42	0.835		445	0.671		9	0.382
	Y2		4038	0.001		3924	0.011		301	0.021		461	0.044		3382	0.007		51	0.000
	B2		4681	0.000		4879	0.001		368	0.002		577	0.007		4180	0.000		57	0.000
B1	Y2		3508	0.001		3533	0.007		204	0.057		420	0.031		2938	0.006		42	0.000
	B2		4151	0.000		4488	0.000		272	0.004		535	0.002		3735	0.000		48	0.000
Y2	B2		643	0.443		955	0.400		68	0.488		116	0.507		798	0.383		5	0.537
NO	Y1	LP3	3823	0.368	LP8	1346	0.324	LP13	218	0.208	LP18	1005	0.316	LP23	856	0.149	LP30	2097	0.296
	B1		6057	0.108		1894	0.117		311	0.045		1504	0.091		1229	0.022		3341	0.063
	Y2		19,427	0.000		5958	0.000		792	0.000		4599	0.000		3014	0.000		10,141	0.000
	B2		22,076	0.000		6752	0.000		885	0.000		5225	0.000		3342	0.000		11,335	0.000
Y1	B1		2234	0.542		547	0.640		93	0.529		499	0.561		374	0.458		1244	0.470
	Y2		15,604	0.001		4611	0.001		574	0.002		3594	0.001		2158	0.001		8045	0.000
	B2		18,253	0.000		5405	0.000		667	0.000		4220	0.000		2486	0.000		9238	0.000
B1	Y2		13,370	0.001		4064	0.001		481	0.002		3095	0.001		1784	0.001		6801	0.000
	B2		16,019	0.000		4858	0.000		574	0.000		3721	0.000		2113	0.000		7994	0.000
Y2	B2		2649	0.405		794	0.436		93	0.466		626	0.403		328	0.451		1194	0.424
NO	Y1	LP4	433	0.197	LP9	1329	0.247	LP14	996	0.277	LP19	325	0.451	LP24	2030	0.370
	B1		660	0.030		1807	0.077		1368	0.092		497	0.191		2816	0.158
	Y2		1601	0.000		5080	0.000		4070	0.000		1674	0.000		9014	0.000
	B2		1788	0.000		5723	0.000		4632	0.000		1947	0.000		10,435	0.000
Y1	B1		227	0.428		478	0.626		372	0.635		172	0.644		786	0.686
	Y2		1168	0.001		3751	0.002		3074	0.002		1349	0.003		6984	0.003
	B2		1355	0.000		4394	0.000		3636	0.000		1622	0.000		8405	0.000
B1	Y2		942	0.002		3273	0.001		2702	0.001		1177	0.002		6197	0.002
	B2		1128	0.000		3916	0.000		3264	0.000		1450	0.000		7618	0.000
Y2	B2		187	0.451		642	0.451		562	0.411		273	0.399		1421	0.402

) and mean plots (Figure 11). Table A1 shows the descriptive statistics. All parts of the mean labor population plots exhibited a negative growth trend.

5.4.3. Household Number

First,

Table 9. ANOVA (Household Number).

	df		MS	F	Sig.		MS	F	Sig.		MS	F	Sig.
a	4	HN0	8,065,587,979	22.458	0.000
b	18		359,133,120
a	4	HN1	463,929	20.976	0.000	HN11	7,713,658	24.666	0.000	HN21	2,281,557,957	18.864	0.000
b	18		22,117				312,722				120,946,518
a	4	HN2	462,189	21.303	0.000	HN12	8621	19.845	0.000	HN22	1,852,761	20.861	0.000
b	18		21,696				434				88,817
a	4	HN3	68,240,679	25.286	0.000	HN13	13,958	12.588	0.000	HN23	232,285	8.512	0.000
b	18		2,698,732				1109				27,290
a	4	HN4	19,928	9.251	0.000	HN14	3,547,127	19.697	0.000	HN24	15,509,385	19.284	0.000
b	18		2154				180,080				804,268
a	4	HN5	252,088	7.932	0.001	HN15	14,477,563	21.691	0.000	HN25	862,771	16.388	0.000
b	18		31,781				667,450				52,648
a	4	HN6	186,192	6.872	0.002	HN16	14,773,235	20.624	0.000	HN26	960,011	22.296	0.000
b	18		27,093				716,328				43,058
a	4	HN7	3,278,312	18.224	0.000	HN17	112,066	46.289	0.000	HN27	16,720	9.89	0.000
b	18		179,885				2421				1691
a	4	HN8	377,603	15.39	0.000	HN18	431,793	25.545	0.000	HN28	121,389,622	2.71	0.063
b	18		24,535				16,903				44,786,751
a	4	HN9	513,973	24.351	0.000	HN19	104,182	20.434	0.000	HN29	257	2.493	0.080
b	18		21,107				5098				103
a	4	HN10	263,777,661	24.402	0.000	HN20	2,964,408	25.616	0.000	HN30	1,192,474	9.675	0.000
b	18		10,809,744				115,724	,			123,259

a = between groups; b = within groups. Note: MS = mean square.

shows that the p-value (>0.5) of HN28/29 was not a significant fit. However, 93% (28 cities and the Niigata area) of the cities in Niigata passed the one-way ANOVA analysis. Second, the 28 cities and the Niigata area were selected for multiple comparisons testing using LSD (

Table A6. Household number—post hoc tests—multiple comparisons using LSD.

I	J		MD (I−J)	Sig.		MD (I−J)	Sig.		MD (I−J)	Sig.		MD (I−J)	Sig.		MD (I−J)	Sig.		MD (I−J)	Sig.
NO	Y1	HN0	−63,893	0.002	HN5	−440	0.014	HN10	−6453	0.045	HN15	−2271	0.007	HN20	−1211	0.001	HN25	−363	0.101
	B1		−51,060	0.004		−431	0.008		−8073	0.008		−2527	0.001		−1339	0.000		−471	0.022
	Y2		−107,511	0.000		−707	0.000		−17,946	0.000		−4657	0.000		−2246	0.000		−1017	0.000
	B2		−116,484	0.000		−703	0.000		−19,571	0.000		−4993	0.000		−2351	0.000		−1125	0.000
Y1	B1		12,834	0.351		9	0.943		−1620	0.495		−255	0.664		−128	0.601		−109	0.512
	Y2		−43,618	0.007		−267	0.066		−11,492	0.000		−2386	0.001		−1035	0.001		−655	0.002
	B2		−52,591	0.001		−263	0.043		−13,117	0.000		−2722	0.000		−1140	0.000		−763	0.000
B1	Y2		−56,452	0.000		−276	0.028		−9873	0.000		−2131	0.001		−907	0.001		−546	0.002
	B2		−65,424	0.000		−272	0.011		−11,498	0.000		−2467	0.000		−1012	0.000		−654	0.000
Y2	B2		−8973	0.449		4	0.972		−1625	0.430		−336	0.510		−105	0.621		−108	0.451
NO	Y1	HN1	−1039	0.000	HN6	−147	0.340	HN11	−1943	0.001	HN16	−1941	0.022	HN21	−22,450	0.038	HN26	−684	0.002
	B1		−972	0.000		−138	0.317		−2118	0.000		−2261	0.004		−25,634	0.011		−748	0.000
	Y2		−1033	0.000		−409	0.010		−3603	0.000		−4479	0.000		−54,355	0.000		−1274	0.000
	B2		−1002	0.000		−491	0.001		−3779	0.000		−4871	0.000		−59,178	0.000		−1332	0.000
Y1	B1		66	0.536		9	0.939		−175	0.663		−320	0.599		−3184	0.687		−64	0.670
	Y2		6	0.957		−262	0.052		−1661	0.001		−2538	0.001		−31,906	0.001		−590	0.002
	B2		37	0.720		−343	0.006		−1836	0.000		−2930	0.000		−36,728	0.000		−648	0.000
B1	Y2		−60	0.539		−271	0.020		−1486	0.001		−2218	0.001		−28,722	0.001		−527	0.001
	B2		−30	0.716		−352	0.001		−1661	0.000		−2610	0.000		−33,544	0.000		−584	0.000
Y2	B2		31	0.742		−82	0.429		−175	0.615		−392	0.460		−4823	0.483		−58	0.656
NO	Y1	HN2	−790	0.000	HN7	−1190	0.007	HN12	−4	0.822	HN17	−171	0.001	HN22	−939	0.003	HN27	−21	0.580
	B1		−795	0.000		−1223	0.002		12	0.508		−210	0.000		−991	0.001		−4	0.918
	Y2		−1009	0.000		−2250	0.000		−65	0.002		−410	0.000		−1717	0.000		88	0.024
	B2		−1035	0.000		−2405	0.000		−75	0.000		−423	0.000		−1846	0.000		105	0.005
Y1	B1		−5	0.961		−33	0.913		16	0.297		−39	0.283		−52	0.810		18	0.551
	Y2		−219	0.067		−1061	0.004		−61	0.001		−239	0.000		−778	0.003		109	0.003
	B2		−245	0.024		−1215	0.001		−70	0.000		−252	0.000		−906	0.000		126	0.000
B1	Y2		−214	0.037		−1027	0.001		−77	0.000		−201	0.000		−727	0.001		91	0.003
	B2		−240	0.007		−1182	0.000		−86	0.000		−214	0.000		−855	0.000		108	0.000
Y2	B2		−26	0.773		−155	0.559		−10	0.455		−13	0.668		−128	0.491		17	0.502
NO	Y1	HN3	−4934	0.004	HN8	−877	0.000	HN13	−73	0.028	HN18	−570	0.000	HN23	−271	0.089	HN30	−1179	0.002
	B1		−5679	0.000		−883	0.000		−38	0.179		−574	0.000		−207	0.142		−1143	0.001
	Y2		−10,292	0.000		−939	0.000		45	0.140		−902	0.000		131	0.371		−383	0.224
	B2		−10,863	0.000		−918	0.000		56	0.048		−934	0.000		219	0.110		−232	0.414
Y1	B1		−745	0.530		−6	0.957		35	0.160		−5	0.960		64	0.591		36	0.887
	Y2		−5358	0.000		−62	0.611		117	0.000		−332	0.004		402	0.005		796	0.008
	B2		−5929	0.000		−41	0.704		128	0.000		−365	0.001		490	0.000		947	0.001
B1	Y2		−4613	0.000		−56	0.586		83	0.001		−327	0.001		338	0.005		761	0.004
	B2		−5184	0.000		−35	0.685		94	0.000		−360	0.000		426	0.000		911	0.000
Y2	B2		−571	0.577		21	0.828		11	0.588		−33	0.685		88	0.396		151	0.492
NO	Y1	HN4	−38	0.386	HN9	−687	0.000	HN14	−962	0.023	HN19	−293	0.000	HN24	−2245	0.013
	B1		8	0.842		−657	0.000		−1071	0.006		−300	0.000		−2442	0.004
	Y2		94	0.031		−972	0.000		−2160	0.000		−442	0.000		−4730	0.000
	B2		115	0.006		−1049	0.000		−2384	0.000		−469	0.000		−5083	0.000
Y1	B1		45	0.184		30	0.777		−108	0.722		−7	0.886		−198	0.759
	Y2		132	0.002		−285	0.019		−1198	0.002		−149	0.014		−2486	0.002
	B2		152	0.000		−363	0.002		−1421	0.000		−177	0.002		−2839	0.000
B1	Y2		86	0.010		−315	0.004		−1090	0.001		−142	0.006		−2288	0.001
	B2		107	0.000		−392	0.000		−1313	0.000		−169	0.000		−2641	0.000
Y2	B2		21	0.477		−78	0.393		−224	0.401		−27	0.539		−353	0.529

) and mean plots (Figure 12). Table A1 shows the descriptive statistics. Only one of the mean household number plots exhibited a negative growth trend. Approximately 67% of local areas (the hosting areas) and 100% of neighborhood areas exhibited a positive growth trend in relation to household number (HN). The Niigata area also exhibited a positive growth trend.

5.5. Spatial Impact

5.5.1. Moran’s I Test of Tourism Indicators

The current paper continues to study the spatial autocorrelation of these 30 areas in the Niigata area. Some of the statistics are as follows: a z-Value > 1.65 (positive) or a z-Value < 1.65 (negative) indicate the significance of spatial autocorrelation (10%, 5%, 1%) (

Table 10. Different standards of Moran’s I value.

z-Value	p-Value	p-Value
<−1.65 or >+1.65	<0.10	90%
<−1.96 or >+1.96	<0.05	95%
<−2.58 or >+2.58	<0.01	99%

). The four quadrants of the scatter plot indicate the local spatial (between the city and its neighbors) associations: HH (a high-value city surrounded by high-value cities); LH (a low-value city surrounded by low-value cities); LL (a low-value city surrounded by low-value cities); a HL (a high-value city surrounded by low-value cities). Moran’s scatterplot can be used to determine the correlations between cities and neighbors.

5.5.2. Local Indicators of Spatial Association (LISA)

On the basis of ANOVA, the majority of tertiary industry income (PCI) and household number (HN) areas exhibited a positive growth trend. Thus, LISA statistics were used to further examine arts event tourism flows (PCI and HN) in meaningful clusters and hot spots in the Niigata area. A significant collection of hosting year data was also utilized. Local indicators of spatial association (LISA) show the spatial cluster effect in 30 Niigata areas in the hosting year of 2000, 2003, 2006, 2009, 2012, 2015, and 2018. It was divided into two periods: (1) only rural arts event (ETAT) tourism was hosted in 2000, 2003, and 2006; (2) rural/urban arts event (ETAT+WLNAF) tourism was hosted at the same time in 2009, 2012, 2015, and 2018.

• Tertiary Industry Income

First, in the first period (only rural arts events (ETAT) tourism), a spatial cluster began to form in the local/hosting areas and the neighborhood areas (Figure 13 and Figure 14). However, the scale of the spatial set was small and low in 2000, the first hosting year. After that, the scale of the spatial collection became more significant. In 2006 (the third hosting year) in particular, the most significant spatial clusters were both in the local/hosting areas and the neighborhood areas. This shows that hosting rural arts event (ETAT) tourism had a positive spatial impact on both the local/hosting areas and the neighborhood areas.

Second, in the period (rural/urban arts event (ETAT+WLNAF) tourism), the most significant spatial clusters were surrounded by the urban local/hosting areas and the neighborhood areas. This shows that hosting urban arts event (WLNAF) tourism had a more substantial positive spatial impact on both the local/hosting areas and the neighborhood areas than rural arts event (ETAT) tourism. Moreover, in 2018, there were no spatial clusters in the rural and urban arts event (ETAT+WLNAF)-tourism-hosting areas. An event can significantly increase local economic activity, yet the net impact within the neighboring areas and cities may be more significant than the local (hosting areas) impact (e.g., the big/national effect often exceeds the small/state effect). In addition, the impact on the local/hosting areas can even be negative [46,47]. This shows that after encouraging rural and urban arts event (ETAT+WLNAF) tourism over these years, the surrounding neighborhood areas also demonstrated solid economic development potential. This finding can be used to drive positive change in the future.

• Household Number

First, in the first period (only rural arts event (ETAT) tourism), the spatial cluster began to form in the local/hosting areas’ neighborhood areas (Figure 15 and Figure 16). However, the scale of the spatial collection was not significant in 2000, the first hosting year. After 2003, the scale of the spatial cluster became more significant in the local/hosting areas and the neighborhood areas. This shows that hosting rural arts event (ETAT) tourism had a positive spatial impact on both the local/hosting areas and the neighborhood areas. On the other hand, from 2003, the areas hosting urban arts event (WLNAF) tourism had a spatial cluster that was more significant than the areas hosting rural arts event (ETAT) tourism even though the urban arts events (WLNAF) had not been held yet.

Second, during the period (rural/urban arts events (ETAT+WLNAF) tourism), the most significant spatial clusters were surrounded by the urban local/hosting areas and their neighborhood areas. This shows that hosting urban arts event (WLNAF) tourism had a more substantial spatial impact on both local/hosting areas and the neighborhood areas as compared to rural arts event (ETAT) tourism. Moreover, this shows that before or after the rural and urban arts events (ETAT+WLNAF) tourism drive over these years, people remained more attracted to urban areas than to rural areas.

6. Discussion and Conclusions

As a result of population shrinkage in both rural and urban areas throughout Japan and the loss of population in the Niigata area, total income and per capita income began to decline after reaching their highest values in the 1980s [116,117]. However, this decline started to slow after these areas began hosting two arts events, with certain economic and population aspects beginning to show positive growth.

6.1. Implications for Theory

First, an event may significantly increase local economic activity, yet the net impact within the neighboring areas and cities may be more significant than the local (hosting areas) impact (e.g., the big/national effect often exceeds the small/state effect). In addition, the impact on the local/hosting areas can even be negative [46,47]. This shows that after the rural and urban arts event (ETAT+WLNAF) tourism drive over these years, the surrounding neighborhood areas also demonstrated solid economic development potential. This finding can be used to drive positive change in the future. It shows that hosting urban arts event (WLNAF) tourism has a more substantial positive spatial impact on both local/hosting areas and its neighborhood areas than rural arts event (ETAT) tourism. Moreover, it shows that before and after the rural and urban arts event (ETAT+WLNAF) tourism drive over these years, people remained more attracted to urban areas than to rural areas.

Second, the economic results from spatial Moran’s I show that after the rural and urban arts events (ETAT+WLNAF) tourism drive over these years, the surrounding neighborhood areas also demonstrated solid economic development potential. This finding can also be used to drive positive change in the future. The population results from spatial Moran’s I show that hosting urban arts event (WLNAF) tourism had a more substantial positive spatial impact on both local/hosting areas and its neighborhood areas than hosting rural arts event (ETAT) tourism. Moreover, it shows that before or after the rural and urban arts event (ETAT+WLNAF) tourism drive over these years, people remained more attracted to urban areas than to rural areas.

Third, from the perspectives of tourism, economics, and the population from SLR, ANOVA, and spatial Moran’s I, hosting rural and urban arts events was shown to positively impact local and neighborhood areas. This shows that holding an exhibition positively increased the local population’s income. This achieves the goal of “promoting local by sustainable tourism” in the SDGs. Moreover, the panel data were studied using multiple-method analysis, which brings a certain validity to the paper, making its analysis of the spatio-temporal impacts more valuable from a scientific point of view.

6.2. Implications for Practitioners and Policy Makers

The current paper shows the positive impacts of exhibition-driven tourism using quantitative analysis. The following conclusions were drawn: (1) the rural arts events were mainly aimed at local and regional revitalization and developing the economies there, which can have a positive effect on population shrinkage; (2) urban arts events have more potential to affect population shrinkage. SDG 8.9 was also empirically confirmed in the results. The changes in the world have exceeded our expectations. Therefore, a new evaluation of exhibition-driven tourism must be established. Although this process may be controversial, this study adds to our knowledge regarding the spatio-temporal impacts of the exhibition-driven tourism industry. The findings in this paper will help to guide operators/practitioners in the tourism industry to obtain market research support aimed at improvement measures. Moreover, these findings also is a policy support role for governmental or non-governmental policymakers in the tourism industry.

6.3. Limitations and Future Research Directions

It is well known that the factors affecting the economy and population are very complex. Thus, the current study has certain limitations. For example, in this study, we only conducted empirical research on two arts events from three perspectives. The scope must be expanded in further investigations. Moreover, similar and different impacts related to rural arts event and urban arts event tourism are an essential research direction for the future with sustainable development goals in mind.