Will World Cultural Heritage Sites Boost Economic Growth? Evidence from Chinese Cities

Abstract

Cultural heritage is closely related to the economy. However, most studies focus on the relationship between the cultural heritage and tourism economy, instead of on the overall economy. This paper estimated the effect of the World Cultural Heritage(s) (WCH) acquisition on economic growth in 242 Chinese cities from 2004 to 2017, based on multiple variations of the difference-in-differences method. Our results show that the WCH acquisition can boost economic growth in local cities. In addition, research and development investments, appearance patents, and exports are three plausible channels for the WCH acquisition to spur cities’ economic growth. Connecting to high-speed rail is not necessary for a city to gather economic benefits from the WCH acquisition. Being a smart city can increase the economic enhancement capacity of the WCH acquisition.

1. Introduction

Cultural heritage is closely linked to various areas of sustainable development goals, such as environmental, social, and economic issues [1]. Specifically, World Cultural Heritage(s) (WCH) sites play significant roles in economic activities, such as the tourism economy, culture economy, circular economy, and low-carbon economy [2,3,4]. However, the impact of the WCH acquisition on economic growth has not been thoroughly attempted.

Previous literature takes cultural heritage as a source for the economic development of tourism [5,6]. However, the impact of cultural heritage on the overall economy has not been measured. This paper examined the causal impact of the WCH acquisition on economic growth by using yearly data from 242 Chinese cities from 2004 to 2017. We applied multiple variations of the difference-in-differences (DID) method to identify the causal effect of the WHC acquisition with falsification tests confirming the validity.

This study confirmed that the WCH acquisition can accelerate a city’s economic development. On average, a WCH acquisition in a city leads to a 3.8% increase in GDP per capita for a city. The results are robust after changing measurements of the WCH acquisition, adding tourism-related control variables, using sub-samples, and conducting falsification tests. In addition, we found that cultural heritage boosts economic growth through R&D investments, appearance patents, and exports.

We next investigated whether the effect of the WCH acquisition on economic performance persists across all types of cities. First, in terms of the stimulating effect of WCH on overall economic performance, non-capital cities seem to benefit more than provincial capitals. In addition, the WCH acquisition can still significantly and positively affect economic growth in cities without high-speed rail. Finally, we found that the economic boosting impact of the WCH acquisition in smart cities is higher than in other cities.

There are three primary contributions of this study. Firstly, our paper contributes to the general literature on economic growth [7,8,9,10,11,12,13,14]. Most papers link cultural heritages to urbanization [15,16,17,18,19], economic structure [20,21,22], household consumption, and fixed investments [23,24]. Our paper shows that cultural heritages also play a significant role in economic growth. Similarly, cultural heritage has been found to have significant positive effects on international tourism in EU countries [5,25]. Particularly in China, according to a recent empirical study, intangible cultural heritage contributes to international tourism [26]. We complement these studies by showing the impact of cultural heritage sites on the growth of the overall economy.

Secondly, our paper also contributes to the literature regarding the relationship between cultural heritage and economic issues. The examples are studies in Sicily Island (Italy) [27], in Southern Moravia (Czech Republic) [28], and in Iran [29]. A study in Lijiang (a Chinese ancient town) showed that cultural heritages can affect urban location policy [30]. Another study in Shanghai (a Chinese metropolis) showed that cultural heritages can also affect regeneration projects [31]. We complement these studies by covering most Chinese cities and investigating the causal effect of cultural heritages on economic growth. On the one hand, this study used panel data from 242 sample cities from 2004 to 2017. On the other hand, to conduct causal inferences, we made use of the quasi-natural experiments provided by the WCH acquisition.

Thirdly, viewed broadly, this paper contributes to the literature linking cultural heritages to several non-economic indicators. For example, the non-economic indicators are creativity [27,32,33], digitization [34], political control [30], urban space [35,36], and place branding [37]. Our paper studied three channels through which cultural heritage may affect economic growth—R&D input, appearance patents, and exports. An empirical study analyzed cultural heritage and creativity in Italian provinces [32]; our paper complements this study. We found that the WCH acquisition significantly increases Chinese cities’ R&D input and the number of appearance patents. A recent study suggested that cultural heritage, considered as a tool for place branding, can make a destination appealing to visitors and build regional branding opportunities [37]. Our finding that the WCH acquisition can increase export opportunities echoes this study.

The rest of the paper proceeds as follows. A brief literature review is provided in the next section. The data and methodology are presented in Section 3. Section 4 presents the main results. Section 5, Section 6 and Section 7 present the results of robustness checks, channel exploration, and heterogeneity analysis, respectively. Section 8 delivers the discussion and Section 9 concludes the paper.

2. Literature Review

This paper is related to the literature on cultural heritage and the role of cultural heritage in promoting economic development. First, the benefits of cultural heritage to cities’ economy are supported by numerous studies [38,39]. UNESCO’s experience provides compelling evidence suggesting that cultural heritage contributes a great deal to economic growth [40]. Cultural heritage affects economic activities through the adaptive reuse of urban industrial buildings [41], the spatial planning of cultural heritage and urban landscapes [35,36], and political control [30]. From the view of Chinese cities, China’s 4000-year-old cultural and philosophical heritage has had a profound influence on the development of a sustainable economy [39]. Cultural heritage has been seen to bolster a new economy in Shanghai [31], Lijiang [30], Suzhou [36], and cities along the Silk Road [37]. The idea that cultural heritage can affect the overall economy has gained support from numerous scholars. However, few quantitative studies have examined the causal effect of cultural heritage on the overall economy.

Second, the tourism economy is the focus of most studies relating to the economic influence of cultural heritage. Cultural heritage is the principal motivation of tourism, and tourists seek to experience the cuisines, handicrafts, and religions of different cultures [42]. For example, evidence from South Africa [43] and the Dominican Republic [44] shows that the potential contribution of urban tourism to economic regeneration and inclusion cannot be ignored. A case study of Southern Moravia (Czech Republic) [28] suggests that cultural heritage fosters local and regional economic development, especially within rural areas. In Lijiang, China, cultural heritage is helpful in the process of tourism commodity production [45]. For cities and regions along the Silk Road, heritage tourism can be a driving force for economic growth through sustainable tourism development and place branding [37]. Particularly, as for quantitative research, using dynamic panel data from 27 EU member states over the 2008–2018 period, a study investigated the role of UNESCO cultural heritage on tourism development [25]. Based on the DID methodology, using city-level data from 2000 to 2018, a study found that intangible cultural heritage listings create a statistically significant increase in income from international tourism [26]. However, quantitative studies have been limited to the impact of cultural heritage on the tourism economy, and few empirical quantitative studies have focused on the impact of cultural heritage on the overall economy.

Third, according to the existing literature, besides tourism, the impact of cultural heritage on the economic development of cities may relate to different aspects of economic activities, such as creativity [27,32] and place brand [31,37]. Two recent Italian cases of cultural regeneration—an immersive and interactive museum and a farm-related cultural park—present the hybridization of urban cultural heritage with creativity [27]. Data from the Cultural and Creative Cities Monitor show that the cultural and creative factors of cities, with spillover effect, can boost the entire economic system in the local area [33]. In addition, taking the Silk Road as an example, local cultural heritages are decisive in the process and implementation of regional branding and economic growth [37]. Digitization is also essential for economic activities inspired by cultural heritage, such as the “Shadow Story” industry (collaborative digital storytelling) in China [46] and the path renewal dynamics in the Kyoto kimono cluster [34]. Various aspects affect the impact of cultural heritage on overall economic growth, but few have been systematically and quantitatively studied.

However, while the impact of cultural heritage on economic activity has been widely discussed, researchers have devoted few resources to assessing the actual impact of cultural heritage on overall economic growth. The channels through which cultural heritage affects regional economic growth are not well understood. In this paper, we evaluated the causal impact of WCH acquisitions on city-level economic growth and explored the channels through which WCH acquisitions can have an impact on city-level economic growth.

3. Data and Methodology

This section describes the main variables and methods used in our analysis. Key variables include WCH acquisition, economic growth, and other city-level characteristics. We explain how they are defined for each of the main classes of variables and where the corresponding data come from. In this paper, we used DID specification with fixed-effect models.

3.1. WCH Acquisition Data

In this paper, we measured WCH acquisition in three ways. Firstly, to examine the impact after first WCH acquisition, we defined a dummy variable ($WCHdummy$) with a value of one for cities in the years after the first WCH acquisition and with 0 for others. Secondly, the number of WCH acquisitions accumulated ($No.ofWCH$) in the city helps to reflect the overall quantitative situation of WCH acquisition at city level. Thirdly, considering the relative exogeneity of the data at the province level, we also used the variable ($No.ofWCHinProv.$), which means the number of WCH acquisitions accumulated in the province where the sample city is located, to replace the variable ($No.ofWCH$). The year and city of each WCH acquisition in China are available from the UNESCO World Heritage convention website at http://whc.unesco.org/ (accessed on 31 December 2021). There are 16 sample cities with at least one WCH acquisition at the start of the sample period. There are 27 sample cities with at least one WCH acquisition and 207 sample cities without any WCH acquisitions at the end of the sample period; see Figure 1. We made a related video showing how the cities in the treatment group changed over the sample period. Please see the video in the Supplementary Material.

3.2. Economic Growth Data

We controlled real GDP ($GDP$) and real GDP per capita ($GDPperCapita$) to capture the effect of a city’s economic growth status. The main data source was the China City Statistical Yearbook, covering the years 2005 to 2018. Both were deflated by the CPI with 2005 as the base year and were transformed into the logarithmic form. Real GDP per Capita is the city-level nominal GDP divided by the city’s average population at the end of the year.

3.3. City-Specific Control Variables Data

We used eight city-specific control variables. The main data source was the China City Statistical Yearbook covering the years 2005 to 2018. Missing data were supplemented from the National Bureau of Statistics of China (www.stats.gov.cn (accessed on 31 December 2021)).

The first control variable was the proportion of government expenditure to GDP (Government expenditure), since the existing empirical studies [7,8,9] indicate that government expenditure is one of the important drivers of long-term economic growth. Moreover, considering the contribution of foreign direct investment to economic growth [10,11,12,13,14], we used the proportion of foreign direct investment to GDP ($Foreigndirectinvestment$) as a control variable. Since many scholars have long documented the effect of urbanization on economic growth [15,16,17,18,19], the proportion of the non-agricultural employed population to the total employed population in the city ($Urbanization$) entered the groups of city-specific control variables. Moreover, we controlled the proportion of the output value of the secondary industry to GDP ($Secondaryindustryratio$) to capture the positive effect of industrial structure on economic growth. Industrial structure rationalization is shown to be conducive to the promotion of regional economic growth [20,21,22]. Higher education is key to balanced and sustained economic development [47,48,49]; this study also controlled the proportion of the number of students in colleges and universities to the registered population at the end of the year ($Education$). Based on the empirical influences of household consumption and fixed investments on economic growth in China [23,24], we used the proportion of total retail sales of consumer goods to total wages of employees ($Consumption$) and the growth rate of fixed assets ($Growthrateoffixedassets$) to control the effect of consumption and fixed investment. In addition, to control the effect of regional population on economic growth, we added the variable $Population$ in the logarithm form. All variables discussed are summarized in

Table 1. Description of variables.

Variables	Definition	Sources
WCH acquisition
$WCHdummy$	A dummy variable with a value of 1 cities in years after the first the World Cultural Heritage (WCH) acquisition and 0.	The UNESCO World Heritage convention website, http://whc.unesco.org/ (accessed on 31 December 2021)
$No.ofWCH$	The number of WCH acquisition accumulated in the city.
$No.ofWCHinProv.$	The number of WCH acquisition accumulated in the province where the sample city is located.
Economic growth
$ln\left(GDP\right)$	GDP ($100millionYuan$); deflated by the CPI with 2005 as the base year and in the logarithm form.	Statistical Yearbook covering the years 2005 to 2018. Missing data were supplemented from the National Bureau of Statistics of China (www.stats.gov.cn (accessed on 31 December 2021)).
$ln\left(GDPperCapita\right)$	GDP per Capita ($Yuan$); deflated by the CPI with 2005 as the base year and in the logarithm form.
Control Variables
$Governmentexpenditure$	The proportion of government expenditure to GDP (%)	Statistical Yearbook covering the years 2005 to 2018. Missing data were supplemented from the National Bureau of Statistics of China (www.stats.gov.cn (accessed on 31 December 2021)).
$Foreigndirectinvestment$	The proportion of Foreign direct investment to GDP (%)
$Urbanization$	The proportion of non-agricultural employed population to total employed population in the city (%)
$Secondary industry ratio$	The proportion of output value of the secondary industry to GDP (%)
$Education$	The proportion of the number of students in colleges and universities to the registered population at the end of the year ($0.01\%$)
$Consumption$	The proportion of total retail sales of consumer goods to total wages of employee (%)
$Growthrateoffixedassets$	The growth rate of fixed assets; deflated by the CPI with 2005 as the base year (%)
$ln\left(Population\right)$	The registered population at the end of the year($TenThousandPersons$); in the logarithm form.

for reference.

3.4. Descriptive Statistics

Our data construction results in an imbalanced panel of data linking WCH acquisition data, economic growth data, and additional time-varying variables at the city level for the period 2004–2017. The data consist of 3322 city-year observations of 242 cities in 26 Chinese provinces.

Table 2. Descriptive statistics.

Variables	Mean	Stand. Dev.	Min.	Max.	Obs.
A. WCH acquisition
$WCHdummy$	0.106	0.307	0	1	3322
$No.ofWCH$	0.111	0.419	0	6	3322
$No.ofWCHinProv.$	1.491	1.341	0	7	3322
B. Economic growth
$ln\left(GDP\right)$	15.972	0.909	13.104	18.914	3322
$ln\left(GDPperCapita\right)$	10.059	0.683	7.865	12.807	3322
C. Control variables
$Governmentexpenditure$	15.163	6.964	2.624	68.761	3322
$Foreigndirectinvestment$	2.090	2.108	0.001	18.189	3322
$Urbanization$	96.946	6.770	26.030	99.990	3322
$Secondaryindustryratio$	48.525	10.070	14.950	85.920	3322
$Education$	1.729	2.317	0.008	13.112	3322
$Consumption$	3.735	1.327	0.223	10.759	3322
$Growthrateoffixedassets$	18.597	19.218	−74.215	244.027	3322
$ln\left(Population\right)$	5.929	0.604	3.743	7.269	3322

provides the summary statistics of the key variables.

3.5. Methodology

We used the DID specification to assess the relationship between WCH and economic growth, based on the following regression setup:

$$Y_{c,t}=\alpha +\beta \times WC{H}_{c,t}+\gamma \times X_{c,t-1}+A_c+B_t+{\epsilon }_{c,t},$$

(1)

where $Y_{c,t}$ is a measure of economic growth for city c in year t, $A_c$ and $B_t$ are vectors of city and year dummy variables that account for city and year fixed effects, $X_{c,t}$ is a set of time-varying control variables, and ${\epsilon }_{c,t}$ is the error term. $WC{H}_{c,t}$ represents a measure of WCH acquisition for city c in year t. The method is classical DID if $WC{H}_{c,t}$ is a dummy variable, i.e., $WCHdumm{y}_{c,t}$, and the method changes to DID with a discrete treatment if the $WC{H}_{c,t}$ is the other discrete variable, such as $No.ofWC{H}_{c,t}$ and $No.ofWCHinProv{.}_{c,t}$. Clearly, $\beta$, indicating the impact of WCH on economic growth, is the key parameter of interest. A positive and significant $\beta$ suggests that WCH helps to promote economic growth, while an insignificant or significantly negative $\beta$ indicates that WCH is unrelated to or negatively related to economic growth.

The DID estimation technique is known to be highly suitable while estimating the effect of sharp changes in economic environment [26]. Several recent related papers using DID models are presented in appendix

Table A1. Examples of related papers using difference-in-differences(DID) model.

Author(s), Year	Main Content
Tan et al., 2022 [26]	Tourism-enhancing effect of cultural heritage listing.
Gao and Su, 2019 [53]	The role of city honor designations in influencing tourism performance.
Yao et al., 2023 [86]	Causal link between World Heritage inscription and local tourism outcomes.
Gao and Su, 2021 [87]	The impact of quality disclosure on the city tourism economy.
Di Matteo, 2022 [88]	Effects of high-speed rail infrastructure on tourism density in Italy.
Zhang and Zhang, 2023 [89]	The causal effect of tourist attractions on economic growth.
Nawaz et al., 2021 [90]	The mean causal effects of a treatment on an outcome of the determinants of scaling up green financing and climate change mitigation.

. WCH acquisition is not meant to develop local areas, but such acquisition does affect economic activities in the local city [50]. In other words, WCH acquisition can be viewed as a sharp and exogenous change in a city and, hence, a DID model is highly appropriate. City c in year t with at least one WCH acquisition can thus be categorized into the treatment group. City c in year t without any WCH acquisition can be regarded as a sample in the control group. Specifically, if a city applied its first WCH successfully during the sample period, then this city can be categorized into the control group before this acquisition year and into the treatment group in and after this acquisition year.

4. Main Results

Our empirical analysis rests on the assumption that the cross-city timing of the first WCH acquisition is not affected by economic growth conditions. Figure 2 shows that the economic growth patterns are similar for cities with at least one WCH acquisition and cities without any WCH acquisition.

Moreover, Figure 3 shows that neither the level of economic growth before the first WCH acquisition nor its rate of change prior to first WCH acquisition explain the timing of the first WCH acquisition. Figure 3A shows a scatter plot of the average economic growth level prior to the first WCH acquisition and the year of the first WCH acquisition. Figure 3B shows a scatter plot of the average change in economic growth level prior to the first WCH acquisition and the year of the first acquisition. The t-statistics for the correlations in Figure 3A,B are $-1.25$ and $-0.12$, respectively. Thus, the first WCH acquisition of cities turns out to be exogenous to the state of economic development.

Table 3. The effect of WCH acquisition on economic growth.

Dep. Var.:	(1)	(2)	(3)	(4)
	ln(GDP)	ln(GDP per Capita)
$No.ofWCH$	0.030 ***	0.038 ***
	(0.010)	(0.010)
$WCHdummy$			0.192 ***
			(0.067)
$No.ofWCHinProv.$				0.009 *
				(0.005)
$Consumption$	−0.016 ***	−0.022 ***	−0.033	−0.022 ***
	(0.006)	(0.007)	(0.023)	(0.007)
$Growthrateoffixedassets$	0.001 ***	0.001 ***	−0.002 ***	0.001 ***
	(0.000)	(0.000)	(0.000)	(0.000)
$Governmentexpenditure$	−0.006 ***	−0.002	0.047 ***	−0.002
	(0.002)	(0.001)	(0.005)	(0.001)
$Foreigndirectinvestment$	0.002	0.007 **	-0.012	0.007 **
	(0.003)	(0.003)	(0.012)	(0.003)
$Urbanization$	0.004 **	0.005 **	0.017 ***	0.005 **
	(0.002)	(0.002)	(0.005)	(0.002)
$Education$	0.032 ***	0.004	0.140 ***	0.005
	(0.009)	(0.010)	(0.028)	(0.010)
$Secondaryindustryratio$	0.015 ***	0.017 ***	0.023 ***	0.017 ***
	(0.001)	(0.001)	(0.004)	(0.001)
$ln\left(Population\right)$	0.497 ***	−0.194 **	1.295 **	−0.199 **
	(0.084)	(0.096)	(0.495)	(0.095)
N	3045	3045	3045	3045
R${}^2$	0.947	0.928	0.727	0.928
No. of cities	242	242	242	242

Notes: *** denotes 1%, ** 5%, and * 10% significance, respectively. All regressions include city and year fixed effects. Robust standard errors are in parentheses.

reports the main results of Equation (1), showing the effect of WCH acquisition on economic growth. The key explanatory variable of interest is the number of WCH acquisitions accumulated in the city ($No.ofWCH$). The dependent variables in Columns 1 and 2 are $ln\left(GDP\right)$ and $ln\left(GDPperCapita\right)$, respectively. Particularly, the results in Column 2 show that, on average, a WCH acquisition in a city leads to a 3.8% increase in GDP per capita for a the city. In this study, the results in Columns 1 and 2 are used as baseline results.

5. Robustness Checks

5.1. Measurements of WCH Acquisition

Firstly, in Table 3, based on the baseline model of Column 2, Column 3 uses a dummy variable, $WCHdummy$, to replace the key independent variable,$No.ofWCH$. The result in Column 3 suggests that GDP per capita in cities with at least one World Heritage site is 19%. Our baseline findings are insensitive to alternative measures of WCH acquisition status. The results are consistent with the baseline estimates in terms of the sign and magnitude of the coefficients of the WCH acquisition state indicators. Indeed, the model in Column 3 of Table 3 is the classical DID method, representing that a city’s first WCH acquisition has boosted economic growth; the model in Column 2 of Table 3 is a DID method with a discrete treatment, representing that the WCH acquisition’s role in boosting the economy increases with the number of WCH acquisitions.

Secondly, we also considered the possibility that the application of WCH is an endogenous decision outcome of the local municipal government. Local governments focused on economic growth will be given more financial support and human resources to successfully apply the WCH by hiring additional experts, detecting cultural implications, and connecting with international organizations. To address this issue, we created a new variable, $No.ofWCHinProv.$, which counts the number of WCH acquisitions in the province that a city belongs to. While a WCH acquisition in a local city may be endogenous, a WCH acquisition in a province can be more exogenous to the economic growth of the local city. The results in Column 4 of Table 3 still support the positive impact of WCH acquisition on a city’s economic growth, although the magnitude of the effect of WCH acquisition at a provincial level is considerably smaller. With an increase in the number of WCH acquisitions in the province that a city belongs to, the economic growth would increase by 0.9%.

5.2. Additional Control Variables

The WCH acquisition may have a positive impact on both domestic and inbound tourism and hence drive economic growth. However, many scholars suggest that WCH acquisition does not promote tourism growth [51]. For example, the entry of Italy into the World Cultural Heritage List does not affect the growth of tourism demand of destination cities, which have a mature tourism market [52]. Empirical evidence from Chinese domestic data also shows that WCH acquisition would not significantly affect a city’s domestic tourism revenue and inbound tourism revenue [53]. Therefore, we in generally believe that the development of tourism would not affect the role of WCH acquisition in promoting regional economic growth.

Empirically, based on the baseline model of Column 2 in Table 3, Columns 1–4 in

Table 4. Robustness check: Additional control variables and subsamples.

Dep. Var.:	Additional Control Variable				Subsamples
	(1)	(2)	(3)	(4)	(5)	(6)	(7)
	ln(GDP per Capita)
$No.ofWCH$	0.047 ***	0.048 ***	0.048 ***	0.048 ***	0.039 ***	0.042 ***	0.021 **
	(0.015)	(0.015)	(0.014)	(0.014)	(0.010)	(0.010)	(0.011)
$ln\left(Domestictourists\right)$	0.043 **
	(0.018)
$ln\left(Domestictourismrevenue\right)$		0.043 ***
		(0.012)
$ln\left(Inboundtourists\right)$			−0.001
			(0.005)
$ln\left(Inboundtourismrevenue\right)$				−0.001
				(0.005)
N	2276	2274	2313	2298	444	1061	2107
R${}^2$	0.930	0.929	0.929	0.930	0.948	0.935	0.829
No. of cities	242	242	242	242	35	84	242
Controls	YES	YES	YES	YES	YES	YES	YES

Notes: *** denotes 1% and ** 5% significance, respectively. All regressions include city and year fixed effects. Robust standard errors are in parentheses. Other control variables in models are $Consumption$, $Growthrateoffixedassets$, $Governmentexpenditure$, $Foreigndirectinvestment$, $Urbanization$, $Education$, $Secondaryindustryratio$, $ln\left(Population\right)$ and Constant. Column 5 limits cities with at least one WCH in sample period. Column 6 limits cities with at least one WCH or on the Tentative List Submission in sample period. Column 7 excludes samples before 2009.

are augmented with different additional variables related to tourism. The additional control variables are the number of domestic tourists ($ln\left(Domestictourists\right)$), domestic tourism revenue ($ln\left(Domestictourismrevenue\right)$), the number of international inbound tourists ($ln\left(Inboundtourists\right)$), as well as international inbound tourism revenue ($ln\left(Inboundtourismrevenue\right)$).

5.3. Sub-Samples

The results in Table 4 indicate that our baseline finding that WCH acquisition improves city economic growth is robust. On the one hand, there could be a reverse causality problem between WCH and economic growth. Cities with WCH are more likely to develop culture and attract investment, where economic development rates are higher than in other cities. To address this issue, we estimated our baseline model by limiting cities with at least one WCH in the sample period. We further limited cities with at least one WCH (http://whc.unesco.org/ (accessed on 31 December 2021)) or that were on the Tentative List Submission (https://whc.unesco.org/en/tentativelists/state=cn (accessed on 31 December 2021)) in the sample period, and the results are presented in Columns 5 and 6 of Table 4. The coefficients of $No.ofWCH$ measures are significantly positive as our baseline results, as is the magnitude.

On the other hand, the 2008 Beijing Olympic Games, as a mega-event, may have had a significant cultural and economic impact on Chinese cities [54]. We also re-estimated our baseline model by excluding samples prior to 2009. The results are presented in Column 7 of Table 4. The results are again in agreement with our baseline findings, with minor differences in magnitude.

5.4. Falsification

Our analysis assumed that the economic growth patterns of cities with WCH acquisitions were similar prior to the first WCH acquisition. However, if the local government anticipates a successful WCH application, a city may revise its economic growth pattern and related policies. In such a scenario, the validity of the DID approach is subject to whether economic growth is systematically different between cities with at least one WCH acquisition in the years prior to the first WCH acquisition. We used a falsification test to check whether the pattern of economic growth in cities with/without WCH acquisition follows a common trend before the first WCH acquisition in a city. For a city that successfully applied the first WCH in year t, we assumed it had successfully applied the first WCH in year $t-k$, where k is a positive integer.

We only included observations from 2004 to year $t-1$ for any city with at least one WCH acquisition in this falsification exercise. Meanwhile, we included observations from 2004 to 2017 for all cities without any WCH acquisitions. The following model investigated whether there were systematic differences in the trend of economic growth between the two groups of cities from year $t-k$ to year $t-1$.

$$Y_{c,t}={\alpha }_c+{\beta }_t+\rho Pseudo\_WC{H}_{i,t-k}+\sigma X_{c,t-1}+{\epsilon }_{c,t},$$

(2)

where $Pseudo\_HSR\_dumm{y}_{i,t-k}$ is a binary variable, taking a value 1 if city i had its first WCH in year $t-k$. We set the value of k to be 2 and 6 and perform five sets of falsification tests.

The results are reported in

Table 5. Pseudo-WCH acquisition in pre-treatment periods.

Dep. Var.:	(1)	(2)	(3)	(4)	(5)
	ln(GDP per Capita)
$Pseudo\_WCH\_dummy(t-2)$	0.016
	(0.039)
$Pseudo\_WCH\_dummy(t-3)$		0.026
		(0.040)
$Pseudo\_WCH\_dummy(t-4)$			0.034
			(0.048)
$Pseudo\_WCH\_dummy(t-5)$				0.052
				(0.060)
$Pseudo\_WCH\_dummy(t-6)$					0.052
					(0.069)
N	2767	2767	2767	2767	2767
R${}^2$	0.928	0.928	0.928	0.928	0.928
No. of cities	225	225	225	225	225
Controls	YES	YES	YES	YES	YES

Notes: All regressions include city and year fixed effects. Robust standard errors are in parentheses. Other control variables in models are $Consumption$, $Growthrateoffixedassets$, $Governmentexpenditure$, $Foreigndirectinvestment$, $Urbanization$, $Education$, $Secondaryindustryratio$, $ln\left(Population\right)$, and Constant.

. Columns 1–5 show the results of Equation (2) assuming the first WCH acquisition to be 2, 3, 4, 5, and 6 years earlier than the actual acquisition year. None of the coefficients for the imaginary WCH acquisition indicator, $Pseudo\_HSR\_dummy$, is statistically significant at the 10% level. This insignificance indicates that the economic growth rates of the cities before and after the first WCH acquisition were not significantly different from each other. Thus, the application of the DID approach to examine the effect of WCH on the economic growth of cities was validated.

6. Channels

We have found that WCH acquisitions have a robust and positive effect on the economic growth of a city. This section investigates three channels through which the WCH acquisition may affect economic growth.

The first channel is R&D investments, which means that WCH motivates investments in R&D in the economic production process, and this then leads to higher economic development. Empirical results show that the utilization of cultural heritage, as well as the design and production of related cultural products, require extra R&D investments [27,53,55]. Both theoretical and empirical literature support that R&D investments are crucial for economic growth [56,57]. On the theoretical front, many classical theoretical models illustrate that the function of R&D investments is as a growth engine [58,59,60,61]. On the empirical front, the importance of R&D investments for economic growth has been supported by a long history [57,62]. Recently, an empirical study from Chinese provinces during 2000–2010 suggests that the effects of total R&D investments can promote technological upgrading, capital deepening, and economic growth [63]. Additionally, a new perspective points out that a positive shock to R&D investments was positively connected with economic growth in the boom period [64].

The second channel is appearance patents. In the view of production output, the elements of cultural heritage can be absorbed in the improvement of product appearances [46,65]. There is reasonably persuasive evidence that patents have important effects on economic growth in the industries and time periods that have been studied. Evidence from 54 manufacturing industries in up to 72 countries for the period of 1981–2000 show that patent rights, leading to factor accumulation and raising productivity, are associated with faster economic growth [66]. Even environment-related patents play significant roles in sustainability and economic growth [67]. What is more, evidence from the EU suggests that R&D investment can be transformed into innovation and, eventually, innovation into economic growth; innovation is always measured by patent-related indicators [68].

The third channel is exports. Considering the cultural trade network, cultural heritage can also increase economic growth by improving cultural diversity and competitiveness in the global market [33,37,69]. Study of the lower-efficiency non-export sector and the higher productivity export sector can explain the effect of exports on economic growth [70,71,72]. Empirical studies show that increasing exports of goods and services is one of the effective solutions to maintaining high economic growth rates in Vietnam [73], the United Arab Emirates [74], and China [75].

We estimated the following regression model to investigate the three channels.

$$Channe{l}_{c,t}={\alpha }_c+{\beta }_t+\gamma WCHc,t+\sigma X_{c,t-1}+{\epsilon }_{c,t},$$

(3)

where $Channel$ refers to three channel variables ($ln\left(RD\right)$, $apatent$, and $ln\left(export\right)$). The estimation results are presented in

Table 6. Regression results of channels.

Dep. Var.:	(1)	(2)	(3)	(4)	(5)	(6)	(7)
	ln(RD)	apatent	ln(export)	ln(GDP per Capita)
$No.ofWCH$	0.154 ***	1.355 *	0.147 **	0.027 ***	0.037 ***	0.035 ***	0.038 ***
	(0.059)	(0.700)	(0.062)	(0.010)	(0.010)	(0.010)	(0.010)
$ln\left(RD\right)$				0.068 ***
				(0.008)
$apatent$					0.001 ***
					(0.000)
$ln\left(export\right)$						0.024 ***
						(0.006)
N	3044	3045	3045	3044	3045	3045	3045
R${}^2$	0.881	0.032	0.301	0.934	0.929	0.930	0.928
No. of cities	242	242	242	242	242	242	242
Controls	YES	YES	YES	YES	YES	YES	YES

Notes: *** denotes 1%, ** 5%, and * 10% significance, respectively. $ln\left(RD\right)$ is the R&D expenditure of cities in RMB 10,000, deflated by the CPI with 2005 as the base year and in the logarithm form; the data source is the China City Statistical Yearbook covering the years 2005–2018. $apatent$ is a score of the number of appearance patents published (ranging from 0 to 100); the data source is the China Innovation and Entrepreneurship Index (CIEI) from the Peking University Open Research Data Platform. $ln\left(export\right)$ is the export of cities in RMB 10,000, deflated by the CPI with 2005 as the base year and in the logarithm form; the data source is the China City Statistical Yearbook covering the years 2005–2018. All regressions include city and year fixed effects. Robust standard errors are in parentheses. Other control variables in models are $Consumption$, $Growthrateoffixedassets$, $Governmentexpenditure$, $Foreigndirectinvestment$, $Urbanization$, $Education$, $Secondaryindustryratio$, $ln\left(Population\right)$ and Constant.

.

As shown in Columns 1–3 of Table 6, the dependent variables are $ln\left(RD\right)$, $apatent$, and $ln\left(export\right)$, respectively, and the coefficients of $No.ofWCH$ are all significantly positive. According to Columns 4–5 in Table 6, the coefficients of $No.ofWCH$ are smaller than those in the baseline results (Column 7 in Table 6). As is evident, WCH acquisition has a significantly positive impact on cities’ R&D expenditure, appearance patents innovation, and exporting, which may lead to more economic growth as expected.

7. Heterogeneity Analyses

In this section, we investigated the impact of WCH acquisition across different types of cities. We firstly compared provincial capital cities to non-provincial capital cities, then cities with high-speed rail to those without, and finally smart and non-smart cities.

First, we examined how the impact of WCH acquisitions on economic growth varies between provincial capital cities and non-provincial cities. The estimation results in Columns 1 and 2 of

Table 7. Heterogeneity analyses: political, traffic, and intelligent factors.

Dep. Var.:	(1)	(2)	(3)	(4)	(5)	(6)
	ln(GDP per Capita)
	Capital City		With High-Speed Rail		Smart City
	Yes	No	Yes	No	Yes	No
$No.ofWCH$	0.025 **	0.062 **	0.020 ***	0.150 **	0.076 ***	0.034 ***
	(0.009)	(0.030)	(0.006)	(0.062)	(0.027)	(0.008)
N	330	2715	998	2047	425	2620
R${}^2$	0.944	0.929	0.795	0.939	0.464	0.935
No. of cities	26	216	175	236	89	242
Controls	Yes	Yes	Yes	Yes	Yes	Yes

Notes: *** denotes 1%, and ** 5% significance, respectively. All regressions include city and year fixed effects. Robust standard errors are in parentheses. Other control variables in models are $Consumption$, $Growthrateoffixedassets$, $Governmentexpenditure$, $Foreigndirectinvestment$, $Urbanization$, $Education$, $Secondaryindustryratio$, $ln\left(Population\right)$, and Constant.

show that the impact of WCH acquisition on economic growth is higher in non-provincial capital city cities than in capital cities. The results are quite reasonable. Because the provincial capital city initially had a considerably more developed economy, the acquisition of WCH accelerated the economic development of the capital city. However, the economic acceleration of WCH acquisitions for provincial capitals is lower than for non-provincial capitals.

Second, we examined whether high-speed rail makes a difference to the impact of the WCH acquisition on economic growth. In Columns 3 and 4 of Table 7, we present the estimation results in cities with and without high-speed rail. Cities with high-speed rail have a lower coefficient of economic growth than cities without high-speed rail, but the economic boost is still statistically significant. Cities without high-speed rail can still achieve economic growth through WCH acquisitions.

Lastly, we split cites by smart cities and others; the estimation results are contained in Columns 5 and 6 of Table 7. The coefficients of the variable WCH acquisition are higher in smart cities than in non-smart cities. The results suggest that the WCH acquisition can boost economic growth faster in smart cities than in non-smart cities.

8. Discussions

Firstly, in this study, we identified a causal effect of cultural heritage on overall economic growth. There exists an extensive sociological work exploring the association between World Cultural Heritage sites and tourism economy—other than the overall economic status—such as conservation and tourism promotion [4], tourism demand [52], sustainable tourism management [51], and resource curse in tourism economies [2]. As for realizing the causal inference of cultural heritages on the tourism economy, much evidence supports the tourist-enhancing effect of World Cultural Heritage [76,77], while some empirical results are conflicting [53,78,79]. This paper moved beyond this literature by providing a causal estimate of the impact of cultural heritage on overall economic growth, rather than simply tourism economic development.

Secondly, our study adds additional quantitative evidence of the causal effect on the economy of cultural heritage by showing that the R&D channel, appearance innovation channel, and exports channel exist. This study was motivated by and complements several previous case studies relating cultural heritage to R&D, appearance innovation, and exports. For example, case studies indicate the importance of R&D in the sustainability of intangible cultural heritage products (i.e., such as bamboo basketry manufacturing) [55], as well as the inspiration originated by cultural heritage in production and service processes (i.e., creative and collaborative digital storytelling) [46]. Recently, a new theory of Embodied Cognition has been proposed to standardize and guide the design practice of cultural heritage, suggesting that the design activity of cultural heritage is affected by three dimensions of heritage environment, body constitution, and cognitive activities [80]. Additionally, the appearance innovation channel is consistent with a study suggesting that innovative design includes analyzing the appearance modeling characteristics, decorative element characteristics, and composition form characteristics of cultural heritage [65]. A recent study supports that intangible cultural heritage projects can increase the international tourism economy by way of a cultural brand development channel, and the probability of gaining China’s Excellent Tourism City designation is used to measure the cultural brand [26]. In our study, the export channel shows that WCH acquisition can enhance economic growth by acquiring export opportunities. This finding is in line with a study from 191 questionnaire responses supporting the importance of the territorial brand among export and tourist consumers [81].

Lastly, the current study also shows evidence of heterogeneity in the causal effect of world cultural heritage on economic growth, according to political, traffic, and intelligent factors. Different aspects of the heterogeneity in the causal effect of cultural heritage on tourism economic issues in terms of city characteristics, development level, heritage site measures, and other issues have also been reported in the literature [82]. Similarly, our results show that non-provincial capitals with WCH acquisitions have higher economic growth than provincial capitals. Amenities such as transportation and hospitality are related to the protection of World Heritage sites and the development of the urban economy [53,83]. Indeed, according to our results, smart cities can reap greater economic growth through World Heritage site inclusion. Previous literature has shown that transportation services, such as high-speed rail, can reasonably be considered allies of economic growth [84,85]. However, our findings confirm that the boost to economic growth from WCH acquisitions remains in cities without high-speed rail links and is higher than in cities with any high-speed rail links. This can be explained by the fact that, in order to meet the goal of sharing the outstanding universal value of a World Heritage site, when applying for WCH, the cities have developed their transportation facilities [53].

9. Conclusions

This paper used city-level panel data to investigate the impact of WCH acquisitions on economic growth in cities. The data cover the majority of Chinese cities and span the period 2004–2017. The timing of the first WCH acquisition and pre-existing economic status figures and falsification tests suggest that the economic growth patterns of the cities before the first WCH acquisition were similar and there were no systematic differences. Thus, we adopted a DID approach to capture the causal effect of WCH acquisition.

Our results suggest that the first WCH acquisition in a city can boost economic growth, and its role in boosting the economy increases with the number of WCH acquisitions. To better understand this finding, we explored three channels through which WCH acquisitions affect economic growth. The results show that WCH acquisition has motivated more R&D, brought more innovation, and stimulated cities’ exports. More importantly, the heterogeneity analysis shows that cities without high-speed rail can still achieve economic growth through WCH acquisitions, and that WCH acquisitions in smart cities have a higher positive effect than in non-smart cities.

This study confirmed that WCH acquisition can boost economic growth at the city level through relevant case studies and qualitative research in Chinese cities, and has implications for infrastructure construction and policy making at the city level. First, it is necessary to invest in R&D activities and support the application of patents related to the cultural heritage of a city. Second, we can use cultural heritage to raise the profile of cities and create a business card for the city. We should actively support city-specific physical product production enterprises and promote product export capacity with the help of cultural heritage brands. Third, to improve the convenience of cultural tourism and promote commercial tourism in the city, we should strengthen the construction of smart cities.

Our study has several limitations and thus leaves some topics open for future exploration. First, due to the limited availability of city-level data, the time period covered in this study is from 2004 to 2017. We were unable to capture the economic impact of WCH during and after the COVID-19 pandemic. After patiently waiting for quarantines, lock-downs, and social distancing measures to ease, shoppers have been overindulging in travel, spending, and consumption. However, as research and development work, production activities, and business travel services return to normal, the pattern of irregular consumption fluctuations may return to normal, so our study is also important for today’s Chinese cities and still has modern significance. Second, this paper suggests that the ability of WCH acquisition to enhance the city-level economy is not only due to directly increasing domestic and international tourism revenue, but also through three channels, such as increasing investment in R&D, increasing numbers of appearance patents, and increasing exports. However, there is no direct evidence to show which companies are better at using WCH to improve innovation and export capacity. As such, it would be an interesting topic for future research.