Tourism development and housing price: An interplay

Abstract

This study underscores the significance of the nexus between tourism and housing price. Using panel data from 35 major cities in China, it employed the panel smooth transition regression approach to explore the interplay between tourism development and housing prices. Our findings showed a nonlinear tourism–housing price nexus. Specifically, tourism development can raise housing prices in a nonlinear way, indicating that this positive impact varies at different levels of tourism specialisation. Housing prices had an inverted U-shaped effect on tourism development. Several explanations for these empirical results are provided in the Discussion section, along with policy suggestions.

Keywords

tourism development housing prices China panel smooth transition approach

Introduction

With significant improvement in Chinese people’s living standards and increase in their leisure time, travelling and vacations have become important daily lifestyles. China’s tourism economy has experienced substantial growth over the past few decades. As per the official statistics of China, the country witnessed a significant influx of tourists with a total of 6.3 billion tourist arrivals in 2019, with an annual increase of 8.4%. The number of per capita travels was 3.7. Moreover, the money spent on tourism in China has rapidly increased, as many Chinese people now have opportunities to travel internationally and domestically for leisure (Lin et al., 2015). In 2019, 5.73 trillion yuan was spent on domestic tourism, increasing by 11% annually. Tourism consumption accounted for 18.8% of the total consumption expenditure.

With this rapid increase in tourism, housing price (HP) has risen at the surprising speed. The average housing price increased from 3343 yuan in 2006 to 15,192 yuan in 2020, which grew at 23.6% annually. The real estate exerted a great impetus for boosting China’s national economy. However, the negative externality of an increase in housing price cannot be neglected. Housing costs have already become a large part of household expenditure, which implies that an overly high housing price will curb the consumption of other residents. Although the development of tourism industry has brought considerable economic benefits, concerns about its negative externality, especially housing price, have mounted (Churchill et al., 2021). In some cities, excessive tourism development has adversely affected housing price (Hu and Huang, 2020; Ma et al., 2019). Therefore, the tourism–housing price nexus is a meaningful topic of concern for tourism scholars.

This study investigated the interdependent association between the growth of tourism industry and the fluctuation of housing price across 35 prominent urban centres in China, spanning the time period from 2006 to 2020. It attempts to address the literature gap in nonlinear tourism–housing price nexus research using the frontier approach of panel smooth transition regression (PSTR). In order to attain this objective, the following enquiries are being addressed: Does tourism development increase housing price? Does soaring housing price restrain tourism development? Do their effects vary across 35 cities with different levels of tourism specialisation and housing price? If so, what drives them? Answers to these questions will provide a far-reaching comprehension of the linkage between tourism development and housing price, which will guarantee sustainable development of China’s tourism and real estate industries.

This study enhances the theoretical underpinnings of tourism literature by scrutinising the correlation between the growth of tourism and the prices of housing. First, given the weak theoretical support in tourism research, it endeavours to clarify the debate on the tourism–housing price nexus by establishing a connection between the supply and demand theory and the permanent income hypothesis (PIH). This study shows that both approaches can effectively explain the nonlinear tourism–housing price nexus mechanism and can provide a theoretical underpinning for future studies. Second, the predominant focus of extant literature revolves around the effects of tourism development on housing price, with little attention given to the reciprocal effect. This study is a ground-breaking work to investigate the tourism–housing price nexus in China, the world’s largest tourism market (Zhang and Xiang, 2022). As heterogeneous characteristics vary across cities in China, it is necessary to conduct case-by-case analyses at the city level. The novel finding that the relationship between tourism and housing price is intricate and varied represents a noteworthy addition to the extant body of literature. Finally, this study represents a novel contribution to the field of research methods, as it employs an advanced PSTR approach to analyse the nonlinear tourism–housing price nexus. Numerous studies on the tourism–housing price nexus concern linear effects by applying traditional econometric techniques such as the baseline regression model and panel VAR models (Campbell and Cocco, 2007). This study centred on the nonlinear tourism–housing price nexus that previous studies have generally overlooked.

Literature review and theoretical hypothesis

Theoretical setting

Research on the tourism–housing price nexus has theoretical support. The tourism literature provides a theoretical framework of supply and demand to explain the mechanism by which tourism impacts housing price (Balli et al., 2019; Schäfer and Hirsch, 2017). On the supply side, tourism has traditionally been a significant domain for real estate investment and development, resulting in the construction of numerous tourism accommodations and public service facilities that occupy substantial tracts of farmland. This has contributed to a scarcity of land, thereby driving up land transfer and housing costs (Tsai et al., 2016; Yıldırım and Karul, 2021). On the demand side, tourism development can increase job opportunities, enhance household income at a destination and create extra demand for housing, thereby affecting land and housing price (Churchill et al., 2021; Tsui et al., 2016). Tourists and tourism-related workers in some tourist cities have high demand for housing and compete directly with residents for land and housing (Churchill et al., 2021).

Well-established theoretical explanations of the influence of housing price on tourism are based on the theoretical frameworks of PIH (Friedman, 1957). Scholars suggest that when real estate becomes an integral part of household wealth, rising housing price may support consumption growth because it can increase households’ wealth or ease their credit constraints, namely, the wealth effect (Su et al., 2018). However, according to consumer behaviour theory, the excessive rise in housing price has suppressed household consumption, leading to the ‘crowding-out effect’ (Li et al., 2014). The theoretical rationale behind this is that the rising costs associated with the acquisition and leasing of residential properties have resulted in individuals allocating a greater proportion of their income towards savings, thereby reducing their consumption patterns in order to meet the financial obligations associated with down payments and mortgage repayments (Dong et al., 2017). In tourism literature, some studies have verified the validity of the above theories in estimating the tourism–housing price nexus. However, the empirical results are inconsistent. For example, Zuo and Lai’s (2020) investigation demonstrates a significant influence of housing assets on tourism spending within China. Liu et al. (2016) discovered that in China, the impact of house prices on tourism consumption was more driven by the wealth effect than the crowding-out effect, as indicated by the PIH. Pang (2014) suggested that rural residents’ tourism consumption is in accordance with the PIH, while the tourism behaviours of urban residents do not satisfy it. Tourism consumption, as a typical non-durable luxury good, has a higher income elasticity; thus, tourism consumption may be more sensitive to housing price than the consumption of daily necessities.

The influence of tourism development on housing price

Over the course of recent decades, there has been a growing body of literature examining the impact of tourism on housing price. In tourism research, tourism is increasingly being considered an important driver of housing price. The existing literature shows that the effects of tourism on house prices vary due to tourists’ stay time, housing demand types and leisure activities (Churchill et al., 2021). Most studies have provided substantial evidence that tourism development positively affects housing price (Biagi et al., 2012; Biagi and Faggian, 2004; Kavarnou and Nanda, 2018). Liu (2013) investigated how tourism development increases China’s housing price at the provincial level and found evidence that tourism development had a noteworthy and favourable impact on housing price. However, it was observed that this effect was not contingent on the level of regional economic development. Biagi et al. (2015) found that tourism activities significantly increased housing price in 103 Italian cities during 1996–2007. However, Mikulić et al. (2021) revealed that tourism intensification deteriorates residents’ ability to afford housing.

Numerous academic studies have investigated the correlation between tourism and housing costs, with a particular emphasis on the influence of tourism-related properties on the latter (Sheppard and Udell, 2016; Thackway et al., 2022). For example, Sheppard and Udell (2016) conducted a study on the influence of Airbnb properties on housing price in New York City. Their findings indicated that a greater density of Airbnb properties in a given locality was associated with a rise in the prices of local housing.

Owing to the heterogeneity in cities and tourism activities, a related strand of the literature has shed light on the nonlinear effects of tourism activities on housing price. The impact of tourism flows on housing price in Germany was investigated by Churchill et al. (2021), who discovered that the effects on housing price were both positive and negative. According to Biagi et al. (2016), the proliferation of tourism-related endeavours resulted in a rise in residential property values in approximately 21%–48% of urban areas, while causing a decline in values in 8%–17% of other locations. Tourism development had an insignificant impact on housing price in 50% of the cities. Considering city–tourism heterogeneity, this study develops the following hypothesis:

Hypothesis 1

Tourism development nonlinearly affects housing price.

The influence of housing price on tourism development

The existing economic literature has presented evidence that establishes a close relationship between housing price and consumption. However, the direction of influence between the two variables necessitates further deliberation. Overall, the literature pertaining to the influence of housing price on consumption can be categorised into three distinct strands. According to the first strand of thought, there exists a positive correlation between an upsurge in residential property values and the level of consumption. The impact of fluctuations in house prices on consumption patterns in South Korea was investigated by Kim et al. (2021). The study revealed that alterations in house prices account for a significant proportion of the variation in total consumption, specifically 25%. Furthermore, the quantitative variation in the effects of changes in house prices on consumption categories was observed to range from 0.15% to 46.08%. Simo-Kengne et al. (2013) analysed whether housing price had an impact on determining consumption behaviour in South Africa and found that housing price growth significantly stimulated consumption, while the decrease of housing price insignificantly reduced consumption. The second strand of research finds that its effect is significantly negative. The study conducted by Lin et al. (2019) employs Taiwan as a case study and concludes that the escalation of housing price has a detrimental effect on consumption, thereby leading to a deceleration in economic growth. Cheng and Fung (2008) found that housing price had both a wealth effect and crowding-out effect on household consumption. The third perspective posits that its impact is nonlinear. Dong et al. (2017) conducted a study on the non-symmetrical impact of housing price on consumption across 35 prominent urban areas in China. The study revealed a positive correlation between house prices and cities with a housing price-to-income ratio below 5.088. Conversely, cities that exhibited a housing price-to-income ratio ranging from 5.088 to 5.963 experienced a detrimental impact.

Within the realm of tourism studies, scholarly investigations into the impact of housing price on tourism development are predominantly approached through the lens of consumer behaviour, encompassing both empirical and theoretical research. Several studies have verified the presence of the wealth effect of housing price on tourism expenditure. For example, based on consumption theory, Kim et al. (2012) found that housing wealth expanded outbound travel demand market in South Korea. Fereidouni et al. (2017) provided evidence that housing assets positively and significantly impact Malaysia’s outbound tourism. However, housing price may affect tourism development owing to the heterogeneity of different regions. Liu et al. (2016) suggested that in China, housing price yielded a U-shaped impact on tourist consumption; thus, housing price has both a wealth effect and a collateral effect on tourist consumption. Churchill et al. (2021) found that tourism negatively affected housing price in the early stage but had a strong positive impact on house prices after 2000. Thus, based on the PIH, this study proposes the following hypothesis:

Hypothesis 2

Housing price is nonlinearly correlated to tourist development.

Method and data

Methodology

PSTR model

As mentioned in the literature review, there may be heterogeneity and nonlinearity in tourism development and housing price. Thus, the results based on the linear regression model may ignore the nonlinear and heterogeneous path of tourism and housing price among cities and over the years, which may lead to a biased estimation. Given the circumstances, it is imperative to capture the potential nonlinear correlation between the advancement of tourism and the cost of housing through the utilisation of a nonlinear panel regression model. The research employed a PSTR methodology, which is a fixed effects model incorporating exogenous regressors. This approach is essentially an expansion of the panel threshold regression (PTR) model, enabling the estimated coefficients to gradually alter when transitioning from one regime to another (Gonzalez et al., 2005). The advantageous features of the PSTR model can be articulated as follows: First, it permits the incorporation of the cross-sectional heterogeneity of panel data (Li et al., 2020). Second, it allows smooth transitions between extreme regimes (Duarte et al., 2013). Finally, it can effectively solve the possible endogenous problem between variables (Zhang and Lu, 2022).

As per the findings of Gonzalez et al. (2005), the fundamental PSTR model’s function can be expressed in the subsequent manner

y_{i t} = μ_{i t} + φ x_{i t} + \sum_{j = 1}^{γ} β_{j} x_{i t} g_{j} ({q_{i t}}^{(j)}; γ_{j}; c_{j}) + α z_{i t} + ε_{i t}

(1)

where the variable i ranges from 1 to N, where N signifies the count of cities. Similarly, the variable t ranges from 1 to T, where T denotes the year of the panel. The dependent variable is denoted by y, while the independent variables are represented by x.

φ

and

β

are the linear and nonlinear coefficients, respectively, of the explanatory variables.

μ_{i}

represents the constant individual effect, z_it is the control variable, and α denotes the estimated coefficients.

ε_{i t}

represents an error term.

g_{i t} ({q_{i t}}^{(j)}; γ_{j}; c_{j})

represents the transition function. According to González et al. (2005),

g_{i t} ({q_{i t}}^{(j)}; γ_{j}; c_{j})

function is

g_{i t} ({q_{i t}}^{(j)}; γ_{j}; c_{j}) = (1 + \exp (- γ \prod_{j = 1}^{m} ({q_{i t}}^{(j)} - c_{j})))

(2)

$γ$ >0, c₁ $\leq$ c₂ $\leq$ c₃… $\leq$ c_m

Where the transition variable is represented by $q_{i t}$ , and the parameter of the transition slope γ_j outlines the smoothness of the transition. The variable m denotes the quantity of $g_{i t} ({q_{i t}}^{(j)}; γ_{j}; c_{j})$ , and $c_{j}$ is the location parameter. As highlighted by Gonzalez et al. (2005), it would be adequate to examine the scenarios where m equals 1 or 2, as they encompass frequently encountered forms of parameter variation. The scenario where m equals one showcases two polar regimes that are linked to low and high values of ${q_{i t}}^{(j)}$ in the PSTR model. When m is equal to 2, the transition function attains its minimum value of (c1 + c2)/2. This value is observed to be 1 at both low and high values of ${q_{i t}}^{(j)}$ .

Linearity and non-remaining nonlinearity tests

Prior to performing the regression analysis of the PSTR model, it is imperative to conduct a linearity test to determine the statistical significance of the switching effect. Referring to the work of González et al. (2005), it is advisable to substitute ${q_{i t}}^{(j)}$ with its first-order Taylor expansion centred at γ = 0 and assess the nonlinearity by means of the subsequent auxiliary regression.

y_{i t} = μ_{i t} + β_{0}^{'} x_{i t} + β_{1}^{'} x_{i t} q_{i t} + \dots + β_{m}^{'} x_{i t} q_{i t}^{m} + u_{i t}^{*}

(3)

where the parameter

β_{1}^{'}, β_{2}^{'}, \dots, β_{m}^{'}

can be expressed as multiples of γ and

u_{i t}^{*}

u_{i t}

+ R_m

β_{1}^{'}

x_it, where R_m refers to the remainder of the Taylor expansion. The act of testing the null hypothesis H₀: γ = 0 can be considered analogous to testing the null hypothesis H₀^*:

β_{1}^{'} = β_{2}^{'} = \dots = β_{m}^{'}

in equation (3). The statistical values for the LM, LMF and LR tests are presented in the following manner

L M = \frac{T N ({S S R}_{0} - {S S R}_{1})}{{S S R}_{0}}

(4)

{L M}_{F} = \frac{T N ({S S R}_{0} - {S S R}_{1}) / k}{{S S R}_{0} / (T N - N - K)}

(5)

L R = - 2 [\log ({S S R}_{1}) - \log ({S S R}_{0})]

(6)

where k denotes the quantity of explanatory variables, and SSR₀ refers to the panel sum of squared residuals under the linear panel model with the null hypothesis (H₀). SSR₁ represents the sum of squared residuals under the PSTR model with two regimes and the alternative hypothesis (H₁). The LM and LR statistics conform to

χ^{2} (k)

distributions, and the LM_F statistics adhere to F(k, TN-N-K) distribution. If the null hypothesis is accepted by the significance test, it can be concluded that the PSTR model exhibits linearity. In the event that the null hypothesis is rejected by the significance test, it can be inferred that the model is comprised of at least two regimes. Additionally, a non-remaining nonlinearity test can aid in identifying the quantity of transition functions and regimes (Li et al., 2020).

Finally, an appropriate value for m was selected by judging the values of RSS, AIC and BIC (Zhang and Lu, 2022; Liu et al., 2019). If the RSS, AIC and BIC values for m = 1 are smaller than those for m = 2, m = 1 should be selected; otherwise, m = 2 should be selected.

Variables and data sources

The study focused on two core variables to elucidate the relationship between tourism development and housing price: (1) The level of tourism development was assessed by utilising the tourist flow metric, which was represented by the quotient of tourist arrivals to the population of registered residents (TA) (Cheng and Zhang, 2020). (2) Housing price (HP) was measured as the average housing transaction price in most studies (Campbell and Cocco, 2007; Dong et al., 2017; Ludwig and Slok, 2002). This study used TA and HP as threshold variables in the PSTR model.

The tourism economic literature identifies several control variables that are crucial in driving tourism growth. These variables include the economic development level, proxied by per capita Gross Domestic Product (GDP) (PGDP), population density measured as the proportion of local inhabitants to the city area (DENS), fixed capital investment proxied by the proportion of fixed capital investment to GDP (INVEST), transport facility level proxied by the proportion of total passenger transport by road, railway and civil aviation to local residents (TRAFF), number of A-AAAAA-class scenic spots (Scenic) and number of star hotels (Hotel) (Liu et al., 2016; Zhang, 2023; Zhang and Xiang, 2022; Zuo and Huang, 2018). This study entered the regression model as a control variable.

The dataset of this study for 35 major cities¹ in mainland China from 2006 to 2020 was obtained from the China Real Estate Statistics Yearbook (2007–2021) and the Official Statistical Reports of 35 major cities (2006–2020). All variables were logarithm-sized to eliminate heteroscedasticity in the regression model. Table 1 describes the variables.

Table 1.

Description for variables.

Variable	Definition	Obs	Mean	SD	Min	Max
lnTA	Logarithm of the ratio of tourist arrival to registered residence population	525	2.145	0.730	−0.316	3.977
lnHP	Logarithm of average housing price	525	8.919	0.615	7.612	10.948
lnPGDP	Logarithm of per capital GDP	525	11.065	0.559	9.430	13.056
lnDENS	Logarithm of the proportion of local inhabitants to the city area	525	−0.716	0.8149	−3.971	1.015
lnINVEST	Logarithm of the proportion of fixed capital investment to GDP	525	−0.560	0.604	−3.655	0.626
lnTRAFF	Logarithm of the proportion of total passenger transport by road, railway and civil aviation to local residents	525	2.914	0.895	0.817	6.385
.lnSCENIC	Logarithm of the number of A-AAAAA class scenic spots	525	3.514	0.792	1.609	5.568
lnHOTEL	Logarithm of the number of star hotels	525	4.422	0.682	2.398	6.729

Empirical findings and analyses

The change of housing price and tourism development

This study echoes previous literature by arguing that regional heterogeneities may exist in tourism development and housing price in China. Figure 1(a) and (b) show the changes in average housing price and average tourist arrivals among the three regions consisting of 35 cities during 2006–2020. The average housing price pertains to the aggregate housing price across a specific region, divided by the total count of cities within said region. In Figure 1(a), the average housing prices in the three regions show an upward trend, but with obvious differences. The increase in average housing price in megalopolises has gradually widened the gap between megacities and large cities. The average housing price in megalopolises increases from 5521 Yuan in 2006 to 26,477 Yuan in 2020, higher than those in megacities and large cities increasing from 3797 Yuan and 2450 Yuan in 2006 to 14,711 Yuan and 9920 Yuan in 2020, respectively.

Figure 1.

Change of housing price and tourist arrivals in 2006–2020. (a) Averaging housing price and (b) average tourist arrivals.

In Figure 1(b), the average tourist arrivals continued to rise in 2006–2019, but dramatically declined in 2020 due to the influence of COVID-19 among the three regions. The differences between their growth levels were significant. The average tourist arrivals in megalopolises were higher than those in other regions.

Multicollinearity test

Before estimating the PSTR model, the present study conducted an assessment of multicollinearity in the regression analysis through the utilisation of Pearson correlation and variance inflation factor (VIF) tests. Table 2 displays correlation coefficients for all variables, which are found to be less than 0.5. Additionally, the VIF values for all variables are below 3, indicating the absence of multicollinearity in the regression.

Table 2.

Multicollinearity test.

	VIF	lnHP	lnPGDP	lnDENS	lnINVEST	lnTRAFF	lnSCENIC	lnHOTEL
lnHP	3.57	1.000
lnPGDP	3.77	0.218^***	1.000
lnDENS	1.42	0.274	0.209^***	1.000
lnINVEST	1.30	−0.375^***	−0.279	−0.247^***	1.000
lnTRAFF	1.19	−0.025	−0.074^***	0.259^***	−0.080^*	1.000
lnSCENIC	2.05	0.406^***	0.462^***	0.012	−0.069	−0.122^***	1.000
lnHOTEL	2.20	0.316^***	0.151^**	0.407^***	−0.320^***	0.255^***	0.481^***	1.000

Notes: *p < 0.1, **p < 0.05 and ***p < 0.01.

Panel unit root tests

In order to mitigate the occurrence of erroneous regression outcomes, the current investigation utilised three discrete unit root assessments, specifically the LLC, Fisher-ADF and Fisher-PP tests, to scrutinise the stationarity of the sequence of each variable within the regression model. In Table 3, the findings suggest that the null hypothesis of stationarity for both lnINVEST and lnHOTEL variables has been refuted at a 1% significance level, in both levels and first differences. This implies that these variables are considered to be stable. The majority of the variables (lnTA, lnHP, lnPGDP, lnDENS, lnTRAFF and lnSCENIC) demonstrated acceptance of the null hypothesis of stationarity in their levels, with a significance level of 1%, but passed the stationary test in the first-order difference at the 10% significance level, indicating that they were found to possess an integration order of one, denoted as I(1). Furthermore, based on the panel Kao residual co-integration test, it can be concluded that there exists a long-term equilibrium relationship among the variables. Thus, the utilisation of PSTR analysis in the present investigation was deemed suitable.

Table 3.

Stationarity test.

	LLC	Fisher-ADF	Fisher-PP	Conclusion
lnTA	9.040	8.507	8.230	Nonstationary
D.lnTA	−9.704^***	206.618^***	203.622^***	Stationary
lnHP	20.473	0.162	0.080	Nonstationary
D.lnHP	12.951^***	261.649^***	257.771^***	Stationary
lnPGDP	13.762	3.191	1.119	Nonstationary
D.lnPGDP	−17.726^***	411.710^*	414.190^***	Stationary
lnDENS	2.909	40.023	44.490	Nonstationary
D.lnDENS	−5.182^***	111.701^***	127.694^***	Stationary
lnINVEST	−0.928	147.111^***	167.034^***	Stationary
D.lnINVEST	−23.697^***	523.933^***	544.467^***	Stationary
lnTRAFF	−1.667*	62.861	65.338	Nonstationary
D.lnTRAFF	−16.444^***	330.466^***	329.295^***	Stationary
lnSCENIC	7.575	12.246	9.866	Nonstationary
D.lnSCENIC	−22.499^***	443.899^***	446.443^***	Stationary
lnHOTEL	−8.328^***	130.823^***	173.920^***	Stationary
D.lnHOTEL	−22.803^***	464.682^***	467.881^***	Stationary
Panel Kao residual co-integration test			−3.431^***

Notes: *p < 0.1, **p < 0.05 and ***p < 0.01.

The effect of tourism development on housing price

An equation for the PSTR model is constructed in order to capture the impact of tourism development on housing price, with lnTA serving as the transition variable

{l n H P}_{i t} = μ_{i t} + φ_{01} {l n T A}_{i t} + φ_{02} {l n P G D P}_{i t} + φ_{03} {l n D E N S}_{i t} + φ_{04} {l n I N V E S T}_{i t} + φ_{05} {l n T R A F F}_{i t} + φ_{06} {l n S C E N I C}_{i t} + φ_{07} {l n H O T E L}_{i t} + (β_{01} {l n T A}_{i t} + β_{02} {l n P G D P}_{i t} + β_{03} {l n D E N S}_{i t} + β_{04} {l n I N V E S T}_{i t} + β_{05} {l n T R A F F}_{i t} + β_{06} {l n S C E N I C}_{i t} + β_{07} {l n H O T E L}_{i t}) \times g ({l n T A}_{i t}; γ; c) + ε_{i t}

(7)

In order to establish a basis for comparison with the outcomes of the PSTR model, this study employed both the panel fixed effects and PSTR models to authenticate the impact of tourism development and housing price. Table 4 displays the outcomes of the panel FE and PSTR models, showcasing the linear and nonlinear impacts of tourism on housing price. Based on the results of the LM, LMF and LR tests, it was determined that the null hypothesis of linearity (H0: r = 0) versus the alternative hypothesis of perfect linearity (H1: r = 1) was rejected. However, the null hypothesis of non-remaining nonlinearity (H0: r = 1) versus the alternative hypothesis of two transition functions (H1: r = 2) was accepted, suggesting that the model contained only one transition function. Additionally, the RSS, AIC and BIC were employed for the purpose of determining the optimal number of location parameters. In general, it is preferable to have lower values for the RSS, AIC and BIC. The RSS, AIC and BIC values observed at m = 1 were 16.122, −3.389 and −3.259, respectively. These values were comparatively smaller than the corresponding values of 16.519, −3.359 and −3.221 observed at m = 2. This suggests that the ideal quantity of location parameters was established as one. Consequently, the results underscore the statistical dominance of PSTR models featuring a single transition function (r = 1) and a solitary location parameter (m = 1). This suggests that the examination of the nonlinear correlation between tourist flows and housing price can be efficiently conducted through employment of the PSTR model with dual regimes and a single transition function.

Table 4.

Effect of tourism development on housing price (core variable: lnTA).

	FE model	PSTR model
		lnTA≤3.201	lnTA>3.201
lnTA	0.254^*** (0.030)	0.239^*** (0.055)	0.419^*** (0.254)
lnPGDP	0.462^*** (0.045)	0.440^*** (0.079)	−0.152 (0.091)
lnDENS	−0.145^*** (0.019)	−0.201^*** (0.026)	0.267^*** (0.061)
lnINVEST	−0.015 (0.023)	0.036 (0.032)	−0.532^*** (0.121)
lnTRAFF	−0.032 (0.015)	0.027 (0.015)	−0.022 (0.068)
lnSCENIC	0.018^*** (0.025)	0.010 (0.032)	−0.043 (0.111)
lnHOTEL	−0.033^* (0.032)	−0.038 (0.051)	0.175 (0.143)
_Cons	3.326*** (0.477)
Hausman test	80.91^***
γ			5.078
c			3.201
N	525		525
		H0: r = 0 vs H1: r ≥ 1	H0: r = 1 vs H1: r ≥ 2
	LM	83.250^***	24.789^*
	LM_F	6.407^***	1.635
	LRT	90.644^***	25.393^*
		m = 1	m = 2
	RSS	16.122	16.519
	AIC	−3.389	−3.359
	BIC	−3.259	−3.221

Notes: standard error are in parentheses, *p < 0.1, **p < 0.05 and ***p < 0.01.

The results obtained from the panel fixed effects model are presented in Column 2 of Table 4. The regression analysis indicates that the estimated coefficient of lnTA is statistically significant at the 5% level and positively related to the dependent variable. The findings indicate that the growth of tourism industry has a significant impact on the escalation of residential property values. Nonetheless, it is worth noting that the utilisation of the panel fixed effects method may potentially result in biased outcomes due to its inability to address the issue of endogeneity. The verification of these findings necessitates the utilisation of the frontier PSTR model.

The findings of the PSTR model with a single transition function and location are illustrated in columns 3 and 4, depicting nonlinear outcomes. The values of the slope parameter γ exhibited a considerable magnitude, specifically 5.078. The aforementioned statement suggests that the shift from one regime to another is sudden, indicating that the housing market is highly responsive to a rise in tourist flow. Furthermore, it is observed that the correlation between tourist flow and housing price undergoes a marked alteration as the transition between the two regimes occurs. The location parameter c is 3.201, which divides the sample cities into two tourist groups. When the threshold variable lnTA is less than 3.201, the estimated coefficient of lnTA is 0.239 and exhibits statistical significance at the 1% level. This suggests that a marginal increase of 1% in tourist arrivals results in a corresponding increase of 0.239% in housing price within the low lnTA regime. When the natural logarithm of tourist arrivals (lnTA) exceeds 3.201, the estimated coefficient of lnTA is 0.419 and statistically significant at the 1% level. This indicates that the impact of tourist flow on housing price in the high-lnTA regime is amplified. Consequently, the findings verify that tourist flows affect housing price nonlinearly, supporting Hypothesis 1. Meanwhile, Table 5 illustrates that the growth of the tourism sector in numerous mega- and large-scale urban areas has substantially contributed to the escalation of housing costs. Notably, the count of cities in the high-lnTA regime has surged from 1 in 2014 to 10 in 2019. In the high-lnTA regime, the cities of Xiamen, Wuhan, Guiyang, Hangzhou, Nanchang and Urumqi exhibit a tourism-oriented focus, with a tourism revenue to GDP ratio exceeding 20%.

Table 5.

Number of cities in different regime of tourist flow.

Year	Low lnTA regime (lower than 3.201)	High lnTA regime (higher than 3.201)
2006–2013	35	0
2014	34	1 (Xiamen)
2015	33	2 (Xiamen and Wuhan)
2016–2017	32	3 (Xiamen, Wuhan and Guiyang)
2018	30	5 (Xiamen, Wuhan, Guiyang, Kunming and Xi’an)
2019	24	10 (Xiamen, Wuhan, Kunming, Xi’an, Nanchang, Urumqi, Hangzhou, Shanghai, Taiyuan and Lanzhou)
2020	30	5 (Xiamen, Wuhan, Guiyang, Nanchang and Urumqi)

This study demonstrates that the expected impacts of various control variables, such as economic development (lnPGDP), population density (lnDENS), capital investment (lnINVEST), transportation infrastructure (lnTRAFF), scenic attractions (lnSCENIC) and hotels (lnHOTEL), on housing price are observable. For example, the study finds that there is a significant and positive correlation between economic development, as measured by lnPGDP, and housing price, as measured by lnHP, in the low lnTA range. However, in the high lnTA range, the impact of economic development on housing price becomes statistically insignificant and negative. Similarly, the impact of the remaining control variables on housing price exhibits variability across distinct lnTA regimes. The results obtained are in line with the outcomes reported in prior research (Biagi et al., 2015; Churchill et al., 2021; Garza and Ovalle, 2019).

Effects of housing price on tourism development

To explore whether and how housing price influence tourism development, this study uses housing price (lnHP) as a transition variable. The equation for the PSTR model is as follows:

{l n T A}_{i t} = μ_{i t} + φ_{11} {l n H P}_{i t} + φ_{12} {l n P G D P}_{i t} + φ_{13} {l n D E N S}_{i t} + φ_{14} {l n I N V E S T}_{i t} + φ_{15} {l n T R A F F}_{i t} + φ_{16} {l n S C E N I C}_{i t} + φ_{17} {l n H O T E L}_{i t} + (β_{11} {l n H P}_{i t} + β_{12} {l n P G D P}_{i t} + β_{13} {l n D E N S}_{i t} + β_{14} {l n I N V E S T}_{i t} + β_{15} {l n T R A F F}_{i t} + β_{16} {l n S C E N I C}_{i t} + β_{17} {l n H O T E L}_{i t}) \times g ({l n H P}_{i t}; γ; c) + ε_{i t}

(8)

Based on the diagnostic outcomes of the PSTR model, it has been determined that the optimal approach is to apply a single transition function and a solitary transition location when implementing the PSTR. Table 6 shows the nonlinear effects of housing price on tourism flows, thus verifying Hypothesis 2. In cases where the natural logarithm of housing price (lnHP) is less than 9.681 (HP <16,011), the estimated coefficient of lnHP is 0.624 and statistically significant at the 1% level. This suggests that within the lower range of housing price, there exists a positive correlation between housing price and tourism flow, with a 1% increase in lnHP resulting in a 0.624% decrease in lnTA due to the wealth effect. The study classified a majority of cities, primarily situated in the central and western regions of China, as regions with low-regime characteristics. If the value of lnHP exceeds 9.681, the coefficient estimate of lnHP is −0.257, which is statistically significant at the 1% level. This implies that there exists a negative correlation between rising house prices and tourist flow. This can be interpreted as an increase of 1% in lnHP, decreasing lnTA by 0.257%, showing that in the high-HP regime (HP ≥16,011), the escalation of housing price leads to a phenomenon of crowding-out in the domain of tourism development. In summary, the findings of this study support the notion of a nonlinear relationship between housing price and tourism development, as evidenced by the inverted U-shaped curve observed in the impact of lnHP on lnTA.

Table 6.

Effect of housing price on tourist development (core variable: lnHP).

	FE model	PSTR model
		lnHP≤9.681	lnHP>9.681
lnHP	0.502^*** (0.060)	0.624^*** (0.102)	−0.257^** (0.120)
lnPGDP	0.754^*** (0.060)	0.602^*** (0.101)	0.056 (0.097)
lnDENS	0.109^*** (0.023)	0.120^*** (0.038)	0.030^*** (0.066)
lnINVEST	0.023 (0.032)	0.108^** (0.053)	−0.218^*** (0.076)
lnTRAFF	−0.010 (0.021)	−0.069^*** (0.022)	0.106^** (0.045)
lnSCENIC	0.095*** (0.035)	0.192^** (0.041)	−0.230^*** (0.055)
lnHOTEL	0.015* (0.045)	−0.220^*** (0.052)	0.378^*** (0.064)
_Cons	−10.954^*** (0.497)
Hausman test	75.97^***
γ			5.257
c			9.681
N	525		525
		H0: r = 0 vs H1: r ≥ 1	H0: r = 1 vs H1: r ≥ 2
	LM	114.871^***	31.480^***
	LM_F	9.523^***	2.105^*
	LRT	129.637^***	32.463^***
		m = 1	m = 2
	RSS	29.375	30.477
	AIC	−2.784	−2.753
	BIC	−2.646	−2.623

Notes: standard error are in parentheses, *p < 0.1, **p < 0.05, ***p < 0.01.

Table 7 lists the number of cities with different regimes of house prices. It can be seen that an increasing number of large and medium cities entered the high lnHP regime from 2006 to 2020, indicating a crowding-out effect. In 2020, 10 cities are included in the high level of housing price; most of them located in the eastern region of China are megalopolis and megacities with high level of economic development and high population density. Specifically, these cities in the high lnHP regime include Beijing, Shenzhen, Shanghai, Xiamen, Nanjing, Guangzhou, Hangzhou, Ningbo, Haikou and Tianjin.

Table 7.

Number of cities in different regime of housing price.

Year	Wealth-effected cities (low lnHP regime)	Crowding-out effected cities (high lnHP regime)
2006–2009	35	0
2000–2012	33	2 (Beijing and Shenzhen)
2013	32	3 (Beijing, Shenzhen and Shanghai)
2014–2015	31	4 (Beijing, Shenzhen, Shanghai and Xiamen)
2016	28	7 (Beijing, Shenzhen, Shanghai, Guangzhou, Xiamen, Nanjing and Hangzhou)
2017	29	6 (Beijing, Shenzhen, Shanghai, Guangzhou, Hangzhou and Xiamen)
2018–2019	27	8 (Beijing, Shenzhen, Shanghai, Guangzhou Xiamen, Nanjing, Hangzhou and Ningbo)
2020	25	10 (Beijing, Shenzhen, Shanghai, Guangzhou, Xiamen, Nanjing, Hangzhou, Ningbo, Haikou and Tianjin)

Robust test

To assess the strength and validity of the aforementioned analysis, this study utilises an instrumental variable approach through a two-stage least squares (2SLS) regression to confirm the correlation between the development of tourism and housing price. The findings of the IV-2SLS model regarding the bidirectional association between housing price and tourist flow are presented in Table 8. The utilisation of lnTA as the independent variable reveals a favourable correlation between tourism and housing price, with a heightened effect observed at elevated levels of lnTA. The utilisation of lnHP as the fundamental variable reveals a curvilinear association between housing price and tourism revenue, characterised by an inverted U-shaped pattern. The results obtained through the PSTR model are in agreement with these findings, indicating the reliability of the analysis.

Table 8.

Robustness check.

	lnHP			lnTA
	FE	FE	IV-FE	FE	FE	IV-FE
lnTA	0.361^***	0.101	0.078
lnTA	(0.069)	(0.063)	(0.070)
lnTA^2	0.055^***	0.040^***	0.051^***
lnTA^2	(0.017)	(0.014)	(0.016)
lnHP				6.884^***	4.234^***	5.024^***
lnHP				(0.511)	(0.524)	(0.576)
LnHP^2				−0.317^***	−0.200^***	−0.239^***
LnHP^2				(0.028)	(0.028)	(0.031)
lnPGDP		0.458^***	0.366^***		0.588^***	0.580^***
lnPGDP		(0.044)	(0.048)		(0.062)	(0.063)
lnDENS		−0.148^***	−0.264^***		0.126^***	0.134^***
lnDENS		(0.019)	(0.037)		(0.026)	(0.026)
lnINVEST		−0.008	0.012		0.012	0.012
lnINVEST		(0.023)	(0.024)		(0.031)	(0.031)
lnTRAFF		−0.029^*	−0.032^**		−0.043^**	−0.035
lnTRAFF		(0.015)	(0.016)		(0.020)	(0.021)
lnSCENIC		0.016	0.011		0.073^**	0.067^*
lnSCENIC		(0.025)	(0.028)		(0.033)	(0.035)
lnHOTEL		−0.021	0.025		−0.029	−0.012
lnHOTEL		(0.032)	(0.037)		(0.044)	(0.046)
_Cons	7.862^***	3.444^***	4.209^***	−33.930^***	−26.048^***	−29.942^***
_Cons	(0.069)	(0.475)	(0.527)	(2.297)	(2.160)	(2.431)
N	525	525	490	525	525	490
R ²	0.701	0.801		0.757	0.824

Notes: Standard error are in parentheses, ^*p < 0.1, ^**p < 0.05 and ^***p < 0.01.

Discussions and conclusions

Discussions

In recent decades, China has experienced a significant increase in both tourism influx and housing costs. The correlation between the development of tourism and housing price has been substantiated in academic studies on tourism (Liu et al., 2016; Mikulić et al., 2021; Zhang and Yang, 2021; Zuo and Lai, 2020). However, the majority of prior research has concentrated solely on the linear association between these variables, with minimal exploration of potential nonlinear causal pathways. The present research endeavours to examine the correlation between the growth of tourism industry and the housing price, with a particular emphasis on nonlinear effects, drawing on the principles of supply and demand and the Permanent Income Hypothesis (PIH). This study employs panel data from 35 major cities in mainland China to reveal the intricate nature of the tourism–housing price relationship, a phenomenon that has been frequently disregarded in previous scholarly works. The novel findings are discussed below.

With respect to the impact of tourism development on housing price, this study has discovered a positive correlation between tourism development and housing price, which is consistent with the conclusions drawn by previous studies conducted by Forsyth and Dwyer (1991), Biagi et al. (2015), Yıldırım and Karul (2021), Zhang and Yang (2021) and Mikulić et al. (2021), all of which have demonstrated that the growth of tourism industry contributes to the escalation of housing price. There are several potential reasons for this positive effect from the perspective of supply and demand. On the supply side, the impact of tourism on housing price is a result of tourism-related activities. Therefore, in order to establish a positive reputation as a tourist destination, numerous cities tend to actively construct tourism-related real estate, encompassing scenic attractions, real estate properties situated in scenic spots (such as Expo Park), business real estate properties catering to tourism (such as shopping centres and restaurants), accommodations for tourists (such as hotels, golf resorts and property hotels) and residential real estate properties catering to tourism to facilitate the growth of the tourism industry. To a certain extent, this aggravates the contradiction between land supply and land use and raises land transfer and housing price (Liu, 2013: p33; Tsui et al., 2016). On the demand side, tourism is a comprehensive industry closely related to agriculture, manufacturing, finance, transport and retail trade (Rosalina et al., 2015; Zhang and Xiang, 2022). The growth of the tourism sector leads to an increase in the number of tourists, which in turn stimulates the development of various industries such as accommodation, food and beverage, transportation and retail. This also results in a rise in demand for commercial real estate. Moreover, tourism development increases residents’ employment and income levels, increasing their rigid demand for the housing market, and together with insufficient accommodation facilities, increases housing price. Finally, both artificial and natural tourist attractions contribute to the appreciation of properties located in their proximity, leading to a rise in housing price. (Cebula, 2009; Churchill et al., 2021). For example, according to the research conducted by Schäfer and Hirsch (2017), the rental market in Berlin experiences a favourable impact due to the increasing urban tourism. The study also revealed a positive correlation between housing price and the level of tourist attraction in a given area. The swift progress of tourism can have a substantial impact on the cost of housing, both in a direct and indirect manner.

The study’s empirical results demonstrate a clear deviation from the findings of Biagi et al. (2016) and Sheng (2011), which revealed a detrimental impact of tourism development on housing price. This may be related to the city scale and style of tourism specialisation. For example, Biagi et al. (2016) suggested that tourism exerts a favourable influence on housing price in mountainous regions, while it has an unfavourable effect in marine areas. According to Sheng's (2011) research, small tourism cities may experience negative externalities due to the tourism-oriented mode of operation. This can result in the Dutch disease problem, which reduces the demand for purchasing houses. This study encompassed exclusively urban areas of significant magnitude, characterised by a substantial populace, expansive market reach and robust industrial infrastructure. The tourism sector’s contribution to the GDP in numerous urban areas is notably inferior to that of secondary industries. The present state of tourism development is not conducive to the occurrence of Dutch disease, thereby impeding the escalation of housing price.

For the influence of housing price on tourism development, this study demonstrates a curvilinear relationship, specifically an inverted ‘U-shaped’ pattern, between housing price and tourism development. The empirical findings presented here differ from the theoretical results reported by Liu et al. (2016), where in a nonlinear association between housing price and tourism consumption was observed. In this study, housing price positively affected tourism development in the low-HP regime; thus, with the rise in housing price, the number of tourist arrivals increased accordingly, and the wealth effect of housing price dominated, which is supported by the studies of Zuo and Lai (2020), Chen et al. (2021) and Kim et al. (2012). There are alternative explanations for these positive effects based on the PIH. First, the prominence of the wealth effect of houses is heightened due to the substantial alleviation of household credit constraints as a result of the escalation of housing price (Dong et al., 2017). Second, rising housing price has the potential to stimulate growth in the real estate sector and its associated industries, thereby generating employment opportunities, augmenting the income levels of local inhabitants and bolstering their capacity and inclination to engage in tourism-related activities. Finally, the rise in housing price increases the local government’s income from land transfers and taxes related to real estate, which can be used to construct public infrastructure and tourist reception facilities at tourist destinations, strengthening their attractiveness of tourist destinations. However, the reduction in the housing wealth effect can be attributed to the diminishing marginal propensity for consumption, which occurs despite the continuous increase in housing price. In instances where housing price experience an upsurge in the high regime, the estimated coefficient of housing price exhibits a negative trend, thereby leading to the dominance of the crowding-out effect. The reason for this is that exorbitant housing costs surpass the threshold of acceptable household earnings, which hinders the expenditure and tourism consumption of residents. Consequently, this situation is detrimental to the growth of the tourism sector. In this study, cities with high housing price are mainly located in megacities in the eastern coastal area of China, where overly high housing costs increase living costs, reduce residents’ savings income and thus decrease tourist consumption capacity.

Conclusions

The investigation of the correlation between the advancement of tourism and the fluctuation of housing price has been of significant significance in promoting sustainable tourism and the real estate sector, as well as in addressing quality-of-life concerns amid China’s swift economic growth. The objective of this investigation was to examine the reciprocal impact of tourism expansion and residential property values in 35 prominent urban centres in mainland China over the period spanning from 2006 to 2020, utilising PSTR models. Transition variables such as lnTA (levels of tourism development) and lnHP (housing price) were employed in the analysis.

The findings derived from empirical analysis provide evidence for the presence of a nonlinear correlation between tourism and housing price within the framework of the PSTR model. The development of tourism has had a favourable impact on the prices of housing. Consequently, in areas with low levels of tourism development, an increase in tourism activity will result in a corresponding rise in housing price. Conversely, in regions with high levels of tourism development, the impact of tourism on housing price will be even more pronounced. This study has discovered an inverted U-shaped correlation between housing price and tourism development. In particular, the study has found that the impact of housing price is notably positive in low housing price regimes, but the negative influence of housing price prevails in high housing price regimes.

The conclusions of this investigation hold both theoretical and practical significance. The existing body of literature primarily concentrates on the unidirectional impact of tourism development on housing price. Nevertheless, there is a dearth of research on the reciprocal influence, particularly in the Chinese setting. This research endeavour represents a ground-breaking initiative aimed at investigating the dynamic relationship between the advancement of tourism and the fluctuation of residential property values in China, thereby broadening the scope of inquiry into the interdependent tourism–housing price nexus. Second, the conventional statistical approach has yielded evidence of a linear relationship between tourism and housing price in most prior studies. This study proposes a frontier method utilising the PSTR model to comprehensively understand the bidirectional tourism–housing price nexus. The study’s novel discoveries demonstrate a nonlinear mechanism of influence that significantly contributes to the existing literature on the correlation between tourism development and housing price. The aforementioned findings offer significant perspectives for governmental bodies to strike a balance between the correlation of tourism and housing costs.

The study’s findings enable us to propose policy recommendations to policymakers aimed at fostering the sound growth of the tourism and real estate sectors. On the one hand, it is imperative for the government to integrate the advancement of the tourism sector into their decision-making process when regulating housing price, given the substantial impact that the former has on the latter’s development. For example, the government can develop related tourism real estate and tourism resorts real estate to stimulate tourism consumption. However, in recent years, certain cities that are oriented towards tourism have witnessed a significant surge in the tourism industry. This growth may result in an excessive dependence on tourism for economic expansion, real estate investments and the income levels of residents. Such a scenario can prove to be highly detrimental to the sustainable development of the regional economy. Therefore, it is imperative for policymakers to prioritise the significance of real economies in ensuring the stable and healthy functioning of the national economy, while concurrently promoting the growth of the tourism industry. On the other hand, the government should acknowledge that exorbitant and meagre housing price are detrimental to tourism development. Therefore, it is recommended that both central and local governments implement a range of policies and measures aimed at promoting stability in the housing market and ensuring that housing price remain within a range that is affordable for the local economy. In urban areas where housing price are elevated, it is advisable for the government to implement measures that curtail a sudden escalation in housing price. This is to avert a scenario where exorbitant housing costs impede tourism consumption (Liu et al., 2016; Zuo and Lai, 2020).

Limitations and recommendations

The present investigation has several limitations. First, it explores the direct relationship between tourism and housing price, yet the mechanism of transmission between the two remains insufficiently investigated. Future research should examine this indirect influence mechanism. Second, the analysis is constrained by the absence of specific data, thereby limiting our scope to the aggregate count of residences and tourists. Consequently, the study omits any discourse on the various categories of abodes (e.g. apartments and houses), the composition of tourist arrivals (i.e. domestic, inbound and outbound) and the nature of the tourist destination. Subsequent research endeavours ought to concentrate on the interrelation between tourism and housing price, with a focus on the subtype perspective. Finally, this study conducted an investigation on 35 prominent urban centres in China, utilising them as case studies, and identified a nonlinear relationship between the prices of housing and tourism. Nonetheless, the outcomes could exhibit noteworthy variations if the investigation is carried out in diverse nations. Henceforth, it is recommended that forthcoming research endeavours undertake a comprehensive comparison of the correlation between tourism and housing price on a global scale.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Jun Zhang

Note

Author biography

Jun Zhang is now a lecturer at school of tourism and urban-rural planning of Zhejiang Gongshang University in P.R. China. Zhang focuses on the studies of tourism disaster management and tourism economic field. E-mail: zhangjun025@mail.zjgsu.edu.cn

References

Balli

Flint-Hartle

, et al. (2019) Towards understanding the volatility of housing prices and exploring the tourism demand impact. Tourism Analysis 24(4): 453–465.

Biagi

Brandano

Caudill

(2016) Tourism and house prices in Italy: a latent class approach. Tourism Economics 22(5): 964–978.

Biagi

Brandano

Lambiri

(2015) Does tourism affect house prices? Evidence from Italy. Growth and Change 46(3): 501–528.

Biagi

Faggian

(2004) The effect of tourism on the house market: the case of Sardinia. In: The 44th ERSA Congress. Porto: Portugal.

Biagi

Lambiri

Faggian

(2012) The effect of tourism on the housing market. In: Perdue JSirgy

Uysal

(eds), Handbook on Tourism and Quality of Life Research. Netherlands: Springer, pp. 635–652.

Campbell

Cocco

(2007) How do house prices affect consumption? Evidence from micro data. Journal of Monetary Economics 54(3): 591–621.

Cebula

(2009) The hedonic pricing model applied to the housing market of the city of Savannah and its Savannah historic landmark district. Review of Regional Studies 39(1): 9–22.

Chen

Hwang

Chang

(2021) The effect of wealth effect and population aging on tourism expenditure. Current Issue in Tourism 25: 1852–1865.

Cheng

ACS

Fung

(2008) Financial market and housing wealth effects on consumption: a permanent income approach. Applied Economics 40(23): 3029–3038.

10.

Cheng

Zhang

(2020) Is tourism development a catalyst of economic recovery following natural disaster? An analysis of economic resilience and spatial variability. Current Issue in Tourism 23(20): 2602–2623.

11.

Churchill

Inekwe

Ivanovski

(2021) Has tourism driven house prices in Germany? Time-varying evidence since 1870. Tourism Economics 28(7): 1705–1723

12.

Dong

Hui

Jia

(2017) How does housing price affect consumption in China: wealth effect or substitution effect? Cities 64: 1–8.

13.

Duarte

Pinilla

Serrano

(2013) Is there an environmental kuznets curve for water use? a panel smooth transition regression approach. Economic Modelling 31: 518–527.

14.

Fereidouni

Tajaddini

(2017) Housing wealth, financial wealth and consumption expenditure: the role of consumer confidence. Journal of Real Estate Finance and Economics 54(2): 1–21.

15.

Friedman

(1957) A Theory of Consumption Function. Princeton, NJ: Princeton University Press.

16.

Garza

Ovalle

(2019) Tourism and housing prices in Santa Marta, Colombia: spatial determinants and interactions. Habitat International 87: 36–43.

17.

Gonzalez

Terasvirta

van Dijk

(2005) Panel smooth transition regression models SEE/EFI Working Paper Series in Economics and Finance. Stockholm, Sweden: Stockholm School of Economics, p. 604.

18.

Huang

(2020) A study on the influence of tourism development on the price of urban real estate: a case of Guiyang city in China. Price: Theory and Practice 2: 161–163.

19.

Kavarnou

Nanda

(2018) How does tourism penetration affect house prices? Evidence from Crete, Greece. Tourism Analysis 23(3): 289–301.

20.

Kim

Park

Lee

(2012) Do expectations of future wealth increase outbound tourism? Evidence from Korea. Tourism Management 33(5): 1141–1147.

21.

Kim

Lee

(2021) How do house prices affect consumption patterns across categories? Korean Economic Review 37(2): 367–398.

22.

Chen

(2014) The crowding out effect of housing price rise on residents’ consumption in China. Statistical Research 12: 32–40. (in Chinese).

23.

Shu

Tan

, et al. (2020) Does the demographic structure affect outbound tourism demand? A panel smooth transition regression approach. Journal of Travel Research 59: 893–908.

24.

Lin

Hsu

Lin

(2019) The effect of housing prices on consumption and economic growth – the case of Taiwan, Journal of the Asia Pacific Economy 24(2): 292–312.

25.

Lin

VSS

Mao

Song

(2015) Tourism expenditure patterns in China. Annals of Tourism Research 54: 100–117.

26.

Liu

(2013) Can tourism development increase house prices? An empirical evidence based on China provincial data. Tourism Science 27(2): 24–35. (in Chinese).

27.

Liu

Huang

Lin

(2016) Impact of house prices on the urban residents’ tourism consumption: an analysis based on dynamic panel data. Tourism Tribune 31(5): 27–35. (in Chinese).

28.

Liu

Xiao

Shao

(2019) Dynamic transformation mechanism and spatial heterogeneity analysis to water and land resources constraints in the Yangtze River Economic Zone. China Population, Resources and Environment 29(3): 89–98. (in Chinese).

29.

Ludwig

Slok

(2002) The impact of changes in stock prices and house prices on consumption in OECD countries. IMF Working Paper 1: 43.

30.

Fang

(2019) Will the development of tourism improve housing prices? An empirical analysis of international tourism island, Hainan. Journal of Chongqing Technology and Business University (Social Sciences Edition) 36(5): 47–53.

31.

Mikulić

Vizek

Stojčić

, et al. (2021) The effect of tourism activity on housing affordability. Annals of Tourism Research 90: 103264.

32.

Pang

(2014) The empirical research of tourism consumption function of China: discussion with Zhou Wenli and Li Shiping. Tourism Tribune 29(3): 31–39.

33.

Rosalina

Garces

LPDM

Purisima

ANC

(2015) Tourism and crime: evidence from the philippines. South Asian Studies 4(3): 565–580.

34.

Schäfer

Hirsch

(2017) Do urban tourism hotspots affect Berlin housing rents? International Journal of Housing Markets and Analysis 10(2): 231–255.

35.

Sheppard

Udell

(2016) Do Airbnb properties affect house prices. Williams College Department of Economics Working Papers 3(1): 43.

36.

Simo-Kengne

Gupta

Bittencourt

(2013) The impact of house prices on consumption in South Africa: evidence from provincial-level panel VARs. Housing Studies 28(8): 1133–1154.

37.

Yin

Tao

(2018) How do housing prices affect consumption in China? New evidence from a continuous wavelet analysis. Plos One 13(9).

38.

Thackway

MKM

Lee

, et al. (2022) Spatial variability of the 'Airbnb effect': a spatially explicit analysis of airbnb's impact on housing price in sydney. ISPRS International Journal of Geo-Information 11(1): 65.

39.

Tsai

Huang

(2016) The impact of tourism resources on tourism real estate value. Asia Pacific Journal of Tourism Research 21(10): 1114–1125.

40.

Tsui

KWH

Tan

Chow

CKW

(2016) Will air transport or tourism affect urban property prices? The case of regional cities in New Zealand. In: The 57th Annual Conference of New Zealand Association of Economists. Auckland, New Zealand, 29 Jun 2016, pp. 1–21.

41.

Yıldırım

Karul

(2021) Does tourism affect house prices in Turkey? Evidence from smooth structural break tests. International Journal of Housing Markets and Analysis 15(5): 1033–1052.

42.

Zhang

(2023) Spatial effects of tourism development on economic resilience: an empirical study of wenchuan earthquake based on dynamic spatial durbin model. Natural Hazards 115: 309–329. published online.

43.

Zhang

(2022) Exploring the effects of tourism development on air pollution: evidence from the panel smooth transition regression model. International Journal of Environmental research and Public Health 19: 8442.

44.

Zhang

Xiang

(2022) Revisiting the relationship between tourism and crime based on a dynamic spatial durbin model. Crime and Delinquency 68(11): 2146–2171.

45.

Zhang

Yang

(2021) Does tourism contribute to real estate boom? A DSGE modeling in small open economy. Journal of Hospitality and Tourism Research 45(1): 257–279.

46.

Zuo

Huang

(2018) Revisiting the tourism-led economic growth hypothesis: the case of China. Journal of Travel Research 57(2): 151–163.

47.

Zuo

Lai

(2020) The effect of housing wealth on tourism consumption in China: age and generation cohort comparisons. Tourism Economics 26(2): 211–232.