Can Mega Sporting Events Promote Urban Green Transformation? Evidence from China

Abstract

With the adoption of the concept of sustainable development, mega sports events (MSEs) are gradually aiming to be greener and to lower their carbon footprints. As such, does the hosting of MSEs provide an opportunity for urban green transformation? Taking MSEs in China as a quasi-natural experiment and using prefecture-level panel data from 2000 to 2020, we empirically explored the effects and mechanisms of MSE hosting on the green economic development in urban areas by applying a time-varying difference-in-differences (DID) approach. The findings revealed that the hosting of MSEs can significantly increase the green total factor productivity (GTFP) of the host city, indicating that MSEs promote the green transformation of urban areas. The results of the mechanism analysis showed that hosting MSEs promotes urban green transformation through three channels: encouraging residents to choose public transport, promoting the development of the digital economy, and upgrading of the industrial structure. In addition, the effect of hosting MSEs on green transformation is stronger in cities where the public is more concerned about the environment. We argue that host city governments should implement the concept of green sports when organizing MSEs, increase public concern about the environment, focus on the development of urban public services, and vigorously develop tertiary industries to drive urban green transformations.

1. Introduction

Mega sports events (MSEs) are characterized by a large scale, high levels of international participation, and extensive global media coverage. MSEs include all types of comprehensive sporting events such as the Olympic Games, Asian Games, and Youth Olympic Games, as well as specific competitions with international influence such as the FIFA World Cup and the World Athletics Championships [1]. In the context of rapid globalization and the transformation of urban governance, the impact of MSEs on host cities not only exceeds that of the sports competition, but also provides all-encompassing contributions to city visibility, economic growth, and cultural diffusion [2]. Historically, during the process of organizing MSEs, host cities often excessively consumed resources to amplify the event scale and economic impact, resulting in negative externalities such as ecological degradation, increased waste, and traffic congestion [3]. For example, the 1964 Tokyo Olympic Games, which necessitated USD 3 billion in urban renewal, extensively damaged the city’s natural ecological landscape. Similarly, the hosting of the 1992 France Winter Olympics destroyed over 30 hectares of native forests, markedly diminishing local biodiversity [4]. The worsening of pollution and increases in carbon emissions present critical challenges to global climate health and human habitats. Against this background, advanced concepts such as the green economy, sustainable development, and carbon neutrality have been developed and implemented [5,6,7] to minimize ecological and environmental costs while achieving economic growth and committing to the green transformation of cities.

The evolution towards green sports is increasingly becoming a strategic objective for cities hosting MSEs in response to the need for sustainable economic development. This approach integrates local ecological protection with sports and cultural initiatives into a cohesive strategic framework. A pivotal document in this context is the Olympic Agenda 2020, which was unveiled by the International Olympic Committee (IOC) in 2014. This agenda identifies sustainability as one of the three core pillars for the future development of MSEs [8]. Consequently, the implementation of green sports practices has emerged internationally as a prominent research topic, encompassing areas such as Olympic Games legacy management [9], the low-carbon construction of venues [10], renewable energy use [11], carbon footprint assessment, and sustainable operational strategies [12]. China has vigorously pursued the development of a green economy given China’s commitment to reduce its carbon intensity at the 15th Conference of the Parties to the United Nations Framework Convention on Climate Change in Copenhagen in 2009. China has thus recently been actively bidding for and preparing to host MSEs with international influence to demonstrate the Chinese government’s capacity for social governance on the international stage. Furthermore, the operation of these MSEs has been strategically aligned with sustainable development principles. The Chinese government issued a policy document entitled “Carbon Neutral Implementation Guidelines for MSEs”, which explicitly requires cities hosting MSEs to anticipate the impact of the event on the environment and strengthen local ecological protection during the event, to advance the green sport development concept. Specific green sports measures have been implemented by China in organizing MSEs. The 2008 Beijing Olympics Games were seminal as China’s first Olympics. The theme of this event was “Green Olympics, Scientific Olympics, and Humanistic Olympics”, emphasizing the integrated and co-developed relationship among sports, culture, and ecology [13]. The Chinese government departments strengthened the local transportation network during the 2014 Nanjing Youth Olympic Games, providing comprehensive coverage with new-energy buses at the competition venues [14]. Subsequently, for the hosting of the 2022 Beijing Winter Olympic Games, various positive measures, including low-carbon venue construction, low-carbon transportation systems, and ecological preservation, were implemented, fully meeting energy demands using green and clean electricity for the event [15]. Empirical research has focused on analyzing the impact of various MSEs in China on environmental pollution indicators, such as air pollution [16], environmental performance [17], and carbon emissions [4].

Our purpose in this study was to explore the effects of MSEs on green economic growth in host cities and the mechanisms of these effects from the prefecture–city-level perspective in China. We also aimed to explain whether the long-term impact of hosting MSEs on host cities promotes or inhibits green transformation. This study contributes to the existing literature in several important aspects: First, although the path and feasibility of hosting MSEs in terms of sustainable development from a specific technological perspective have been extensively studied by scholars and government departments, empirical evidence remains scarce regarding the broader impact of MSEs on the green transformation of host cities at regional and national levels. Second, numerous studies have theoretically explored the green and low-carbon measures implemented during the MSEs held in China and their outcomes, and some scholars have empirically demonstrated how the undesirable outputs of MSEs increase environmental pollution. However, a gap remains in the literature regarding integrating the desirable outputs of MSEs, reflected in economic indicators, into the empirical framework used to fully assess the influence of MSEs on the green economy. No study has yet been devoted to exploring whether the MSEs held in specific regions or countries within a certain time interval actually provide opportunities for the green transformation of the host cities. Unlike previous studies, we incorporated environmental pollution and economic indicators into a unified framework to measure the green total factor productivity (GTFP). We considered the MSEs held in multiple host cities in China at different times as a quasi-natural experiment, and we applied a time-varying difference-in-differences (DID) model to explore the impact of MSEs on the host city’s GTFP. This methodology provided empirical evidence showing whether the hosting of MSEs contributes to urban green transformation. Finally, we further discussed the heterogeneity of the impacts and explored the mechanisms through which MSEs contribute to urban green transformation, with the aim of exploring the reasons and channels for organizing MSEs to realize urban green transformation. We aimed to fill the existing research gaps with this study, share insights from Chinese experiences with other countries or cities organizing MSEs, and advocate for the global realization of the dual goals of green sports and sustainable economic development.

The rest of this paper is organized as follows: Section 2 provides a literature review; Section 3 describes the methodology and data used in this study; Section 4 presents the results of the empirical analysis; and Section 5 outlines the conclusions and policy recommendations.

2. Literature Review

2.1. The Environmental Effects of MSE

During the critical period of economic restructuring and environmental governance, scholars have not reached a consensus on the economic and ecological impacts of MSEs. MSEs enhance national image, stimulate economic growth, and boost tourism revenues [18,19]; however, MSEs are also linked to adverse outcomes such as increased residential prices, air pollution, and ecological damage [20,21]. Empirical studies are increasingly focusing on the environmental impacts of MSEs. Mirella et al. [16] examined the effects of pollution control measures and meteorological factors before and after the hosting of the Nanjing Youth Olympic Games in China on the concentrations and compositions of six air pollutants. They found that the strict environmental regulations during the event effectively reduced pollution and increased local air quality, which is consistent with the results obtained by Zhao et al. [22]. Liu and Ogunc [23] demonstrated a notable increase in air quality during both the hosting and post-hosting periods of the 2008 Beijing Olympics Games, and Triantafyllidis and Davakos [24] empirically investigated the relationship between the number of spectators at rugby matches and air pollution levels, finding that events with larger audiences tended to increase pollutant levels. From the environmental performance perspective, Long, Chen, and Park [17] developed a frontier distance function model to assess the contribution of the 2008 Beijing Olympics Games to urban environmental efficiency. Their findings revealed that the phases before, during, and after the event had different effects in different regions. Chen, Long, and Salman [25] employed a quasi-natural experiment to empirically assess the impact of the successful 2010 bid and the 2014 hosting of the Nanjing Youth Olympic Games on urban energy efficiency. The results demonstrated that the event substantially increased the energy efficiency of the host city, Nanjing, and had notable spatial spillover effects. Wu et al. [26] estimated how government interventions such as energy substitution, traffic management, and temporary air pollution control contributed to carbon reductions during the preparations for the Beijing Olympics Games. Ito, Higham, and Cheer [27] explored the carbon emission reductions that occurred during the 2020 Tokyo Olympics Games and estimated that the COVID-19 pandemic reduced the carbon emissions of the Olympics by nearly 130,000 tons due to the decrease in international visitors. Zhang et al. [4] applied the synthetic control method to show that the Nanjing Youth Olympic Games notably increased Jiangsu Province’s carbon emissions. They further combined the synthetic control method with the LMDI method to identify different channels through which the event produced carbon increases and reductions.

2.2. The Drivers of Urban Green Transformation

The simultaneous development of resources, environment, and economy has an integral role in driving the green economy [28,29,30]. The GTFP index, which integrates energy–environment factors into traditional economic models, serves as a prevalent metric for measuring green economic growth in empirical studies. This indicator is crucial for assessing whether a city is on the green transformation path [31,32].

The factors driving urban green transformation encompass economic, social, and environmental dimensions [33]. These include the inherent demand for inclusive urban green development and the influences of the external environment, policy, and technology [34]. Traditional urban development models often sacrifice the environment for economic growth at the external environmental level. Faced with challenges such as air pollution, water resource contamination, and ecosystem degradation [35], a model must be developed that promotes balanced development across economic, social, and environmental levels [36]. At the policy level, environmental regulations serve as critical macro-control tools for governments to achieve urban green transformation. With the introduction of the Porter hypothesis [37], scholars have recognized the profound impact of environmental regulations on green economic growth. Governments are constraining the development of inefficient and highly polluting industries within cities by enacting laws, levying taxes, and implementing other command-and-control policies, which encourage businesses to increase operational efficiency and reduce emissions [38]. Additionally, preventive policies such as traffic management, the promotion of new energy sources, and carbon trading schemes are guiding urban economic activities and fostering environmental awareness among businesses and citizens to realize urban green transformation [39]. Bian et al. [40] examined the direct and indirect impact effects of establishing a carbon emissions trading system on fostering the green transformation of Chinese cities. Green technology innovation has strongly contributed to the formation of clusters of science and technology, which play an important role in ameliorating environmental pollution [41,42]. Both government and enterprises promote urban eco-efficiency through increased R&D expenditure in green technology innovations [37]. Yang et al. [43] discussed the link between industrial structural upgrades, GTFP increases, and carbon emission reductions, finding that industrial rationalization curbs carbon emissions. Furthermore, upgrading the industrial structures of cities can lead to the elimination of obsolete production processes, promote the transition of the energy structure from high energy consumption to energy savings [44], and advance the industrial structure [45]. Additionally, government policies focused on optimization and business environment improvements can attract foreign direct investment (FDI) [46]. In response to the principles of green and low-carbon development, governments are more inclined to more stringently scrutinize foreign investments according to the pollution halo hypothesis [47]. High-quality FDI enhances the technological and managerial capabilities of host cities through green knowledge spillover, thereby promoting urban green transformations [48].

In summary, scholars have mainly assessed the impact of MSEs on the economic or ecological indicators of host cities from a singular perspective. However, studies of the effects of hosting MSEs on both the economy and ecological environment are lacking. Many studies have been conducted on the factors driving urban green transformation, studies quantifying the impact of hosting MSEs on this transformation have been mainly theoretical. Therefore, we employed the GTFP, which reflects both economic and environmental factors, as an explained variable and applied a time-varying DID model to analyze whether the MSEs held in China have effectively promoted green economic growth and facilitated the green transformation of the host cities. We discussed the heterogeneity and mechanisms of these effects to find the reasons why MSEs affect the green transformation of the host cities. Our findings provide evidence supporting the green development of sports and the development of policies for global climate change governance.

3. Methodology and Data

3.1. Model Settings

The hosting of MSEs provides a quasi-natural experimental setting, which potentially exogenously impacts urban green transformation. In this study, several MSEs held in China during 2000–2020 were taken as study objects, and a counterfactual control group was constructed using cities that had not hosted any MSEs, A time-varying DID model was employed to assess the impact of hosting MSEs on the urban GTFP due to the differences in the timing of MSEs in China. We chose a fixed-effects model instead of a random-effects model because the p-value of the Hausman test was 0.000. We constructed the following two-way fixed-effects regression model following the approach proposed by Beck and Levine [49]:

$${GTFP}_{it}=a_0+{\alpha }_1{DID}_{it}+{\alpha }_2{Control}_{it}+{\mu }_i+{\eta }_t+t+{\epsilon }_{it}$$

(1)

$${DID}_{it}={City}_i\times {Time}_t$$

(2)

where the city dummy variable ${City}_i$ took values of 1 and 0 for the treatment group of cities hosting MSEs and the control group, respectively. The time dummy variable ${Time}_t$ took a value of 0 before the event and 1 after; ${DID}_{it}$ represents the explanatory variable indicating whether city $i$ hosted an MSE in period $t$. The regression coefficient ${\alpha }_1$ indicates the actual effect of hosting an MSE on the city’s green economic growth. A positive ${\alpha }_1$ indicates that the green transformation of the city hosting the MSE was promoted compared with that of other cities. The explained variable ${GTFP}_{it}$ denotes the GTFP of city $i$ in period $t$, ${Control}_{it}$ denotes the set of control variables, $a_0$ is the constant term, ${\mu }_i$ represents individual fixed effects, ${\eta }_t$ represents time fixed effects, $t$ captures the global macroeconomic trends, and ${\epsilon }_{it}$ is the error term.

3.2. Variable Selection and Descriptive Statistics

3.2.1. Explained Variable

In this study, GTFP served as a proxy variable for urban green economic growth that integrated economic growth and environmental factors into a unified framework for measurement. The objective of optimizing the GTFP is to maximize economic output while minimizing undesired outputs such as energy consumption and environmental pollution. Unlike traditional DEA models, the non-radial SBM-DEA model effectively incorporates undesirable outputs [50]. The GML index measures productivity changes based on the DEA model using time-series data [51]. Therefore, we synthesized the strengths of both models to develop the SBM-GML index model, drawing on the empirical strategy reported by Lin et al. [52]. This model was used to evaluate the city-level GTFP in China; the input and output indices used in the analysis are presented in

Table 1. Input and output indicators for calculating GTFP.

Variable Type	Variable Name	Data Source
Input indicators	Capital	The perpetual inventory method is used to measure the capital stock
	Labor	Year-end employment
	Energy	Annual electricity consumption
Desirable output indicators	GDP	Deflating GDP for each year with 2000 as the base period
Undesirable output indicators	Waste Water	Industrial wastewater discharge
	Smoke	Industrial soot emissions
	Sulfur Dioxide	Industrial sulfur dioxide emissions

.

3.2.2. Explanatory Variable

In this study, the explanatory variable comprised six MSEs held during the period of 2000 to 2020 that were recognized by the IOC: the 2007 Changchun Asian Winter Games, the 2008 Beijing Olympics Games, the 2009 Harbin World University Winter Games, the 2010 Guangzhou Asian Games, the 2011 Shenzhen World University Summer Games, and the 2014 Nanjing Youth Olympic Games. Consequently, the treatment group in this study included the host cities of Changchun, Beijing, Harbin, Guangzhou, Shenzhen, and Nanjing. The ${City}_i$ values of these cities were set to 1, and their ${Time}_t$ values for the year after the event were set to 1. The control group comprised other prefecture-level cities in China that did not host any MSEs. We excluded Tianjin, Shenyang, Qinhuangdao, Shanghai and Qingdao as co-host cities of the 2008 Beijing Olympics and Foshan, Dongguan and Shanwei as co-host cities of the 2010 Guangzhou Asian Games in the control group to avoid the interference effect from co-host cities that also participated in these events. We set the ${City}_i$ value of the control group to zero.

3.2.3. Control Variables

Guided by the studies of Chen et al. [25], Long, Chen and Park [17], and Zhang et al. [4], we identified urban population density (POP), gross domestic product per capita (PERGDP), foreign direct investment (FDI), pollution control investment (PCI), R&D expenditure (R&D), human capital (HC), and public policy (POLICY) as the control variables in the empirical process. Specifically, we used the number of college students enrolled in the city as a measure of HC, and we used a dummy variable for whether the city had implemented a low-carbon pilot policy as a proxy for POLICY.

We only used the annual panel data of 288 prefecture-level cities in China from 2000 to 2020 for empirical analysis owing to the substantial amount of missing data in some cities. The data sources included the official websites of the General Administration of Sport of China and the Chinese Olympic Committee, China City Statistical Yearbook, China Regional Economic Statistical Yearbook, provincial and municipal statistical yearbooks, and the statistical bulletins of national economic and social development for each city. Individual missing values were filled using linear interpolation or mean imputation. Additionally, we algorithmized all the variables except the explanatory variable, R&D, and POLICY to mitigate the potential heteroscedasticity issues in the variables.

Table 2. Statistical description.

Type	Variable	Symbol	Obs	Mean	Std. Dev.	Min	Max
Explained variable	Green total factor productivity	GTFP	6048	0.6593	0.0375	0.5431	0.7730
Explanatory variable	Six MSEs held in China	DID	6048	0.0094522	0.096771	0	1
Control variables	Urban population density	POP	6048	6.4631	0.9777	2.6390	9.5506
	Gross domestic product per capita	PERGDP	6048	9.9240	0.9025	4.6051	13.0556
	Foreign direct investment	FDI	6048	9.1084	2.4068	0	14.9412
	Pollution control investment	PCI	6048	10.1275	1.8735	0	15.3734
	R&D expenditure	R&D	6048	0.1636	0.0498	0.0069	0.3647
	Human capital	HC	6048	4.5855	1.0897	1.0612	9.7486
	Public policy	POLICY	6048	0.52662	0.4993	0	1

provides the descriptive statistics for all variables.

Table 3. Correlation matrix.

	GTFP	DID	POP	PERGDP	FDI	PCI	R&D	HC	POLICY
GTFP	1.0000
DID	0.1503 *	1.0000
POP	0.3698 *	0.0454 *	1.0000
PERGDP	0.7365 *	0.1403 *	0.1473 *	1.0000
FDI	0.6969 *	0.1461 *	0.4336 *	0.6289 *	1.0000
PCI	0.4411 *	0.1068 *	0.1919 *	0.4152 *	0.3947 *	1.0000
R&D	0.0355 *	−0.0359 *	−0.1694 *	−0.0395 *	−0.0832 *	−0.0077	1.0000
HC	0.4075 *	0.1402 *	0.3301 *	0.4721 *	0.4568 *	0.2224 *	−0.2015 *	1.0000
POLICY	0.3738 *	0.0766 *	0.1802 *	0.3372 *	0.3248 *	0.8648 *	0.0144	0.1976 *	1.0000

* significant at 10%.

presents the correlation matrix between the main variables.

4. Results of Empirical Analysis

4.1. Baseline Regression

Table 4. Baseline regression results.

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
Variables	GTFP	GTFP	GTFP	GTFP	GTFP	GTFP	GTFP	GTFP
DID	0.0433 ***	0.0368 ***	0.0264 ***	0.0222 ***	0.0221 ***	0.0224 ***	0.0201 ***	0.0201 ***
	(0.0037)	(0.0035)	(0.0031)	(0.0030)	(0.0030)	(0.0030)	(0.0030)	(0.0030)
POP		0.0117 ***	0.0091 ***	0.0071 ***	0.0069 ***	0.0070 ***	0.0063 ***	0.0063 ***
		(0.0004)	(0.0004)	(0.0004)	(0.0004)	(0.0004)	(0.0004)	(0.0004)
PERGDP			0.0201 ***	0.0164 ***	0.0158 ***	0.0166 ***	0.0143 ***	0.0143 ***
			(0.0006)	(0.0006)	(0.0006)	(0.0006)	(0.0007)	(0.0007)
FDI				0.0038 ***	0.0037 ***	0.0037 ***	0.0034 ***	0.0034 ***
				(0.0002)	(0.0002)	(0.0002)	(0.0002)	(0.0002)
PCI					0.0012 ***	0.0012 ***	0.0012 ***	0.0009 ***
					(0.0002)	(0.0002)	(0.0002)	(0.0003)
R&D						0.0433 ***	0.0495 ***	0.0492 ***
						(0.0067)	(0.0067)	(0.0067)
HC							0.0026 ***	0.0026 ***
							(0.0003)	(0.0003)
POLICY								0.0016 *
								(0.0009)
Constant	0.6246 ***	0.5477 ***	0.3845 ***	0.4019 ***	0.3990 ***	0.3844 ***	0.3984 ***	0.4008 ***
	(0.0014)	(0.0031)	(0.0054)	(0.0052)	(0.0052)	(0.0057)	(0.0059)	(0.0061)
N	6048	6048	6048	6048	6048	6048	6048	6048
R-squared	0.598	0.649	0.718	0.740	0.742	0.744	0.747	0.747
Time trend	YES	YES	YES	YES	YES	YES	YES	YES
Province FE	YES	YES	YES	YES	YES	YES	YES	YES
Time FE	YES	YES	YES	YES	YES	YES	YES	YES

Standard errors in parentheses, * significant at 10%; *** significant at 1%.

presents the results of the baseline regression, with GTFP as the explained variable and the six MSEs held in China during the period of 2000 to 2020 as the explanatory variables. In Table 4, column (1) provides the individual impacts of hosting MSEs on urban green transformation. For the results in columns (2) to (7), we gradually added control variables compared with those in column (1). The coefficient of the explanatory variable is significantly positive at the 1% level in column (7), demonstrating that hosting MSEs substantially enhanced the green transformation of the host cities. Specifically, the regression coefficient of 0.0201 suggests that, all else being equal, hosting an MSE increased the green economic growth of the host city by an average of 2.01%. Regarding the results of the control variables, the coefficients of population density, GDP per capita, FDI, pollution control investment, R&D expenditure, and human capital were all significantly positive, affirming their supportive role in fostering urban green transformation.

4.2. Robustness Test

4.2.1. Parallel Trend Test

The DID model was predicated on the parallel trend assumption, which posited that in the absence of policy shocks from hosting MSEs, the trends in GTFP for both the treatment and control groups should remain consistent. Host cities typically commence preparatory activities four years prior to the event given the extensive scale, high profile, and significant impact of MSEs. Therefore, the parallel trend assumption of the DID model needed to be verified. To address this, we constructed a dynamic effects test model to detect the endogeneity issues that may arise from divergence in the green economic growth trends between the host and other cities during the pre-event preparation phase. This model was formulated as follows:

$${\mathrm{G}\mathrm{T}\mathrm{F}\mathrm{P}}_{it}={\beta }_0+{\sum }_{k=-5}^{k=5}{\beta }_j{DID}_{it}^k+\gamma {Control}_{it}+u_i+{\eta }_t+t+{\epsilon }_{it}$$

(3)

where $k$ represents the number of years between time $t$ and the year in which the MSE was hosted in this city. The dummy variable ${DID}_{it}^k$ reflects the policy shock of hosting the MSE, with ${DID}_{it}^k$ being set to 1 for the cities in the treatment group from four years before to four years after the MSE, and 0 otherwise. Additionally, we adopted the conventional method of setting the dummy variable ${\mathrm{D}\mathrm{I}\mathrm{D}}_{it}^{-1}$ for the year immediately preceding the MSE as the base period. This approach was used to circumvent potential multicollinearity issues.

Although preparatory activities for MSEs theoretically enhance the GTFP of host cities, the results in Figure 1 of the parallel trend test indicate no significant pre-event differences in the GTFP between the treatment and control groups. This result is similar to that of a previous study that empirically demonstrated the impact of MSEs on air pollution [53]. However, the impact of this policy was significantly positive in the post-event period, affirming the parallel trend hypothesis and that hosting MSEs can facilitate the green transformation of cities.

4.2.2. Sample Change

The co-host cities theoretically possess the potential to undergo green transformation because the cities co-hosting the MSE also organize a portion of the event. We included the co-host cities in the treatment group for empirical discussion to evaluate whether baseline regression results were influenced by sample selection bias. A new explanatory variable named DID_xb was generated to assess the robustness of the findings. The regression results in columns (1) and (2) in

Table 5. Regression results with sample change and considering lagged effects.

	Sample Change		Considering Lag Effects
Variables	GTFP	GTFP	GTFP	GTFP
DID_xb	0.0233 ***	0.0116 ***
	(0.0027)	(0.0022)
L.DID			0.0435 ***	0.0195 ***
			(0.0040)	(0.0032)
POP		0.0062 ***		0.0063 ***
		(0.0004)		(0.0004)
PERGDP		0.0142 ***		0.0144 ***
		(0.0007)		(0.0007)
FDI		0.0035 ***		0.0035 ***
		(0.0002)		(0.0002)
PCI		0.0009 ***		0.0009 ***
		(0.0003)		(0.0003)
R&D		0.0503 ***		0.0477 ***
		(0.0069)		(0.0069)
HC		0.0027 ***		0.0027 ***
		(0.0004)		(0.0004)
POLICY		0.0019 **		0.0013
		(0.0009)		(0.0009)
Constant	0.6246 ***	0.4014 ***	0.6266 ***	0.4003 ***
	(0.0014)	(0.0062)	(0.0014)	(0.0063)
N	6048	6048	6048	6048
R-squared	0.590	0.743	0.584	0.739
Time trend	YES	YES	YES	YES
Province FE	YES	YES	YES	YES
Time FE	YES	YES	YES	YES

Standard errors in parentheses, ** significant at 5%; *** significant at 1%.

demonstrate that the impact of hosting MSEs on urban GTFP remained significantly positive even after including co-host cities. Notably, the regression coefficient decreased from 0.0201 in the baseline to 0.0116 with the inclusion of co-host cities, indicating that the hosting of MSEs plays a stronger role in promoting the green transformation of the host cities. Additionally, the effect remained statistically significant across all host cities, including co-host cities.

4.2.3. Considering Lag Effects

We evaluated whether the one-period-lagged effect of hosting MSEs on urban green transformation remained significant, thereby testing the robustness of the empirical evidence. The regression results in columns (3) and (4) in Table 5 show that the lagged regression coefficients of the explanatory variables, named L.DID, remained significantly positive, with only a slight deviation from the baseline coefficient of 0.0201. The signs and magnitudes of the control variable coefficients minimally changed, reinforcing the robustness of the baseline regression conclusions.

4.2.4. Adjustment Time Interval

The sample time period in the baseline regression of 2000 to 2020 was selected, to explore whether the impact of hosting MSEs on GTFP varied with the changes in the sample time span. To assess this, the sample interval was adjusted by sequentially excluding the first year, the first two years, the last year, and the last three years, setting the intervals as 2001–2020, 2002–2020, 2000–2019, and 2000–2018, respectively. The regression results are shown in

Table 6. Regression results with adjustment time interval.

	(1) 2001–2020	(2) 2002–2020	(3) 2000–2019	(4) 2000–2018
Variables	GTFP	GTFP	GTFP	GTFP
DID	0.0198 ***	0.0196 ***	0.0206 ***	0.0213 ***
	(0.0030)	(0.0030)	(0.0032)	(0.0034)
POP	0.0063 ***	0.0064 ***	0.0062 ***	0.0060 ***
	(0.0004)	(0.0004)	(0.0004)	(0.0004)
PERGDP	0.0144 ***	0.0144 ***	0.0143 ***	0.0142 ***
	(0.0007)	(0.0007)	(0.0007)	(0.0007)
FDI	0.0035 ***	0.0035 ***	0.0034 ***	0.0035 ***
	(0.0002)	(0.0002)	(0.0002)	(0.0002)
PCI	0.0009 ***	0.0011 ***	0.0009 ***	0.0009 ***
	(0.0003)	(0.0003)	(0.0003)	(0.0003)
R&D	0.0477 ***	0.0444 ***	0.0509 ***	0.0527 ***
	(0.0069)	(0.0071)	(0.0069)	(0.0071)
HC	0.0027 ***	0.0026 ***	0.0026 ***	0.0027 ***
	(0.0004)	(0.0004)	(0.0004)	(0.0004)
POLICY	0.0013	0.0007	0.0017 *	0.0019 **
	(0.0009)	(0.0010)	(0.0009)	(0.0009)
Constant	0.4006 ***	0.4018 ***	0.4016 ***	0.4024 ***
	(0.0063)	(0.0065)	(0.0062)	(0.0064)
N	5760	5472	5760	5472
R-squared	0.740	0.728	0.744	0.739
Time trend	YES	YES	YES	YES
Province FE	YES	YES	YES	YES
Time FE	YES	YES	YES	YES

Standard errors in parentheses, * significant at 10%; ** significant at 5%; *** significant at 1%.

: columns (1)–(4) show the estimation results for 2001–2020, 2002–2020, 2000–2019, and 2000–2018, respectively. The results showed that after adjusting the time interval of the sample, the sign direction and significance of the regression coefficients for the organization of an MSE remained consistent, and the significantly positive correlation with GTFP was maintained.

4.2.5. Endogenous Treatment

Based on the assumption that the current value of the explanatory variable does not affect the current error term, using its lagged-one-period value for regression analysis can help mitigate potential endogeneity issues. Nonetheless, the omitted variables related to GTFP may still create endogeneity issues. In this analysis, the lagged-one-period value of the explanatory variable was used as the instrumental variable. Three approaches, two-stage least squares (2SLS), generalized method of moments (GMM), and limited information maximum likelihood (LIML), were applied to examine the effects. The regression outcomes are presented in

Table 7. Regression results with endogenous treatment.

	(1) 2SLS	(2) GMM	(3) LIML
Variables	GTFP	GTFP	GTFP
DID	0.0200 ***	0.0200 ***	0.0200 ***
	(0.0032)	(0.0032)	(0.0032)
POP	0.0063 ***	0.0063 ***	0.0063 ***
	(0.0004)	(0.0004)	(0.0004)
PERGDP	0.0144 ***	0.0144 ***	0.0144 ***
	(0.0007)	(0.0007)	(0.0007)
FDI	0.0035 ***	0.0035 ***	0.0035 ***
	(0.0002)	(0.0002)	(0.0002)
PCI	0.0009 ***	0.0009 ***	0.0009 ***
	(0.0003)	(0.0003)	(0.0003)
R&D	0.0477 ***	0.0477 ***	0.0477 ***
	(0.0068)	(0.0068)	(0.0068)
HC	0.0027 ***	0.0027 ***	0.0027 ***
	(0.0003)	(0.0003)	(0.0003)
POLICY	0.0013	0.0013	0.0013
	(0.0009)	(0.0009)	(0.0009)
Constant	0.4279 ***	0.4279 ***	0.4279 ***
	(0.0082)	(0.0082)	(0.0082)
N	5760	5760	5760
R-squared	0.740	0.740	0.740
Time trend	YES	YES	YES
Province FE	YES	YES	YES
Time FE	YES	YES	YES

Standard errors in parentheses, *** significant at 1%.

; columns (1) to (3) present the 2SLS results, GMM results, and LIML results, respectively. The findings confirm that after considering endogeneity issues, hosting an MSE still positively affected urban GTFP at the 1% statistical level, proving that the conclusion that the hosting of MSEs significantly and positively contributed to urban green transformation is robust.

4.3. Heterogeneity Discussion

With the spread of information and increasing public awareness, urban residents have grown more concerned about environmental factors such as air and water quality, green space coverage, and waste management. This shift has compelled city governments to adopt measures for improving the urban environment. Public environmental concern extensively affects reductions in urban carbon emissions and enhancements in urban green economic growth [42,52]. We further distinguished the heterogeneity of the impact of hosting MSEs on urban green transformation in cities with different levels of public environmental concern. The Baidu search index for “haze” was selected as a proxy variable for public environmental concern following the methodology of Li, Yang, and Li [54]. As the calculation of GTFP includes the unexpected output indicators of smoke and sulfur dioxide, the environmental perception of the term “haze” is comparatively high in comparison with that of other terms such as “environmental pollution” that relate to the environment. Urban residents can directly perceive haze severity in their daily lives, so the search frequency of the term “haze” more accurately reflected public concern for environmental issues. In this study, public environmental concern was categorized according to the mean value, generating samples with high (Public_H) and low (Public_L) public environmental concern for subsample regression analysis. The results are shown in

Table 8. Regression results of heterogeneity discussion in public environmental concern.

	Public_H		Public_L
Variables	GTFP	GTFP	GTFP	GTFP
DID	0.0526 ***	0.0205 ***	0.0447 ***	0.0194 ***
	(0.0051)	(0.0042)	(0.0078)	(0.0062)
POP		0.0068 ***		0.0055 ***
		(0.0006)		(0.0005)
PERGDP		0.0169 ***		0.0120 ***
		(0.0011)		(0.0009)
FDI		0.0034 ***		0.0037 ***
		(0.0003)		(0.0002)
PCI		0.0008 *		0.0012 ***
		(0.0005)		(0.0003)
R&D		0.0339 ***		0.0550 ***
		(0.0111)		(0.0088)
HC		0.0025 ***		0.0023 ***
		(0.0005)		(0.0005)
POLICY		0.0002		0.0033 ***
		(0.0015)		(0.0012)
Constant	0.6220 ***	0.3837 ***	0.6239 ***	0.4214 ***
	(0.0065)	(0.0114)	(0.0014)	(0.0079)
N	2562	2562	3486	3486
R-squared	0.534	0.707	0.526	0.701
Time trend	YES	YES	YES	YES
Province FE	YES	YES	YES	YES
Time FE	YES	YES	YES	YES

Standard errors in parentheses, * significant at 10%; *** significant at 1%.

, where column (1) reports the regression results for the Public_H sample, column (2) adds control variables to those of column (1), column (3) displays the regression results for the Public_L sample, and column (4) introduces control variables to those used for column (3). The analysis revealed that green economic growth is more strongly encouraged in cities with higher public environmental concern that host MSEs, with relatively higher marginal returns.

We further divided Chinese cities into coastal groups (City_C) and inland groups (City_L) for subsample regression according to the regional division on the map, exploring the regional heterogeneity in the impact of hosting MSEs. The regression results are presented in

Table 9. Regression results of heterogeneity discussion on coastal cities.

	City_C		City_L
Variables	GTFP	GTFP	GTFP	GTFP
DID	0.0575 ***	0.0294 ***	0.0333 ***	0.0148 ***
	(0.0062)	(0.0055)	(0.0046)	(0.0037)
POP		0.0077 ***		0.0060 ***
		(0.0013)		(0.0004)
PERGDP		0.0147 ***		0.0141 ***
		(0.0022)		(0.0007)
FDI		0.0045 ***		0.0035 ***
		(0.0007)		(0.0002)
PCI		0.0023 ***		0.0007 ***
		(0.0006)		(0.0003)
R&D		0.0335 **		0.0460 ***
		(0.0166)		(0.0074)
HC		0.0019 *		0.0025 ***
		(0.0010)		(0.0004)
POLICY		−0.0079 ***		0.0032 ***
		(0.0024)		(0.0010)
Constant	0.6471 ***	0.3739 ***	0.6202 ***	0.4070 ***
	(0.0034)	(0.0211)	(0.0015)	(0.0065)
N	987	987	5061	5061
R-squared	0.633	0.752	0.578	0.732
Time trend	YES	YES	YES	YES
Province FE	YES	YES	YES	YES
Time FE	YES	YES	YES	YES

Standard errors in parentheses, * significant at 10%; ** significant at 5%; *** significant at 1%.

. Column (1) reports the regression results for the City_C sample, and column (2) introduces control variables to the result in column (1). Similarly, column (3) displays the regression results for the City_L sample, with column (4) incorporating control variables. The analysis revealed that hosting an MSE significantly enhances the GTFP of coastal cities, demonstrating higher efficiency in promoting urban green transformation than in inland cities.

4.4. Mechanism Exploration

Ceder and Perera [55] used two major cities as examples of how the hosting of MSEs increases the connectivity of public transportation. Concurrently, governmental departments use these events to actively encourage social participation through various initiatives, thereby enhancing public awareness of environmental protection to increase the preference for green public transportation [56]. Ding and Liu [57] illustrated the positive impact of enhanced transportation infrastructure on China’s green economic development from 2008 to 2020. Therefore, we argue that MSE can increase the urban GTFP by encouraging residents to choose public transport. Additionally, Holland [58] argued that hosting MSEs creates opportunities for developing the Internet and social media strategies. Wang et al. [59] confirmed the significant impact of the energy used by the Internet and the digital economy on green economic growth across 30 provinces in China. Based on these findings, we suggest that enhancing the digital economy is an effective channel through which MSEs can boost a city’s GTFP. Furthermore, Solberg and Preuss [60] and Ristova [61] illustrated the importance of MSEs in promoting tertiary industries such as tourism and increasing revenue through, for example, accommodation and catering services. Xu and Zhou [62] posited that upgrading the industrial structure enhances the development of China’s green economy. From these insights, we suggest that industrial upgrading serves as a catalyst through which MSEs boost the GTFP of cities. Overall, encouraging residents to choose public transport (PT), digital economy (DE), and industrial upgrading (IU) are identified as the three pivotal mechanisms through which MSEs realize urban green transformation. We referred to Squicciarini [63] for further explorations of these mechanisms, then the steps for testing these mechanisms were set up as follows:

First, we verified the effect of the organization of MSE on the mechanism variable with the following equation:

$${mechanism}_{it}={\alpha }_0+{\alpha }_1{DID}_{it}+{\alpha }_2{Control}_{it}+{\mu }_t+{\eta }_t+t+{\epsilon }_{it}$$

(4)

${mechanism}_{it}$ contains three mechanism variables: PT, DE, and IU. Second, we examined the impact of the mechanism variables on urban GTFP, supplementing the theoretical discussion of how these variables influence the explained variables with the following formula:

$${GTFP}_{it}={\alpha }_0+{\gamma }_1{mechanism}_{it}+{\alpha }_2{Control}_{it}+{\mu }_t+{\eta }_t+t+{\epsilon }_{it}$$

(5)

Third, we employed the year-end actual number of publicly operated electric buses as a proxy variable for PT, the quantity of Internet access as a proxy variable for DE, and the proportion of the tertiary industry as a proxy variable for IU.

Table 10. Regression results of mechanism exploration.

	PT		DE		IU
Variables	GTFP	GTFP	GTFP	GTFP	GTFP	GTFP
DID	1.0224 ***		0.5205 ***		0.1812 ***
	(0.1078)		(0.0871)		(0.0270)
PT		0.0080 ***
		(0.0004)
DE				0.0155 ***
				(0.0004)
IU						0.0096 ***
						(0.0015)
POP	0.3357 ***	0.0036 ***	0.1814 ***	0.0035 ***	0.0186 ***	0.0061 ***
	(0.0142)	(0.0004)	(0.0115)	(0.0004)	(0.0036)	(0.0004)
PERGDP	0.5011 ***	0.0104 ***	0.1925 ***	0.0114 ***	−0.1569 ***	0.0159 ***
	(0.0237)	(0.0007)	(0.0192)	(0.0006)	(0.0059)	(0.0007)
FDI	0.0907 ***	0.0028 ***	0.1137 ***	0.0017 ***	0.0229 ***	0.0033 ***
	(0.0067)	(0.0002)	(0.0055)	(0.0002)	(0.0017)	(0.0002)
PCI	0.0427 ***	0.0005 **	0.0337 ***	0.0003	−0.0019	0.0009 ***
	(0.0092)	(0.0002)	(0.0075)	(0.0002)	(0.0023)	(0.0003)
R&D	0.2489	0.0471 ***	1.1923 ***	0.0307 ***	−0.2735 ***	0.0517 ***
	(0.2393)	(0.0064)	(0.1934)	(0.0060)	(0.0600)	(0.0067)
HC	0.3450 ***	−0.0000	0.2401 ***	−0.0010 ***	0.0914 ***	0.0019 ***
	(0.0123)	(0.0004)	(0.0099)	(0.0003)	(0.0031)	(0.0004)
POLICY	−0.0341	0.0019**	0.0349	0.0011	−0.0081	0.0017 *
	(0.0318)	(0.0009)	(0.0257)	(0.0008)	(0.0080)	(0.0009)
Constant	−3.6837 ***	0.4291 ***	−3.6291 ***	0.4558 ***	4.3485 ***	0.3572 ***
	(0.2179)	(0.0060)	(0.1762)	(0.0057)	(0.0547)	(0.0090)
N	6048	6048	6048	6048	6048	6048
R-squared	0.706	0.766	0.818	0.796	0.471	0.747
Time trend	YES	YES	YES	YES	YES	YES
Province FE	YES	YES	YES	YES	YES	YES
Time FE	YES	YES	YES	YES	YES	YES

Standard errors in parentheses, * significant at 10%; ** significant at 5%; *** significant at 1%.

presents the results of the study of the mechanisms through which MSEs affect urban green transformation. Columns (1) and (2) report the mechanisms of influence of PT. Column (1) displays the estimation results of Equation (4), where PT is the explained variable and MSE hosting is the explanatory variable, suggesting that the hosting of MSEs significantly promotes PT. Column (2) reports the estimation results of Equation (5), where GTFP is the dependent variable and PT is the independent variable. The significantly positive coefficient of PT suggests that its increase promotes urban green transformation, supporting our theoretical hypothesis and confirming that encouraging residents to choose public transportation is an effective mechanism through which MSEs facilitate urban green transformation. Columns (3) and (4) illustrate the regression results for DE, demonstrating that MSE hosting not only promotes the DE development but also enhances the green economy of host cities. Similarly, columns (5) and (6) suggest that IU serves as a mechanism through which MSEs promote urban green transformation.

5. Discussion

5.1. Comparison with Existing Results

We constructed an SBM-GML model to measure the GTFP of Chinese prefecture-level cities from 2000 to 2020 by integrating economic growth and environmental factors into a unified framework. We empirically examined the impacts of hosting MSEs on GTFP and the mechanisms through which they occur using a time-varying DID model with two-way fixed effects. Our study aimed to fill the gap in our knowledge of the long-term impact of hosting MSEs on the GTFP of Chinese host cities. We drew a series of conclusions that provide new empirical evidence for the practice of the concept of green sports and the impact of hosting MSE on the green transformation of Chinese cities. In addition, the results of the heterogeneity analysis and mechanism exploration in this study provide references for governmental departments to formulate the appropriate public policies to promote urban green transformation.

Some findings in the literature on the economic or environmental impacts of MSEs echo those of our study, with a portion of the literature highlighting the economic benefits, such as stimulating economic growth, boosting tourism, and increasing urban identity of hosting MSEs [2,64]. Other studies note the potential to reduce air pollution concentrations during the hosting of MSE [16,53] and to enhance environmental and energy performance [17,25]. In contrast to these studies that individually measured economic benefits or environmental indicators, this study provides a new perspective by comprehensively examining the simultaneous economic and environmental benefits of MSE under the framework of sustainable development.

Another interesting finding from our heterogeneity discussion is that choosing to host MSEs in a coastal Chinese city where the public is highly concerned about the environment can more strongly promote local green transformation. We think that the possible reason for this finding is that these cities not only have geographic advantages but also tend to have stronger economic foundations, public awareness, and experience in urban governance. Consequently, these cities are more likely than the lesser-developed inland regions to be able to achieve sustainability in terms of event concepts, venue construction, legacy management, and environmental regulation during the hosting of MSEs. In addition, our results showed that encouraging residents to choose public transportation, enhancing the level of the digital economy, and industrial upgrading are the three mechanisms through which hosting MSEs promotes urban green transformation. These results prove the important roles of public transportation, the digital economy, and industrial upgrades in the process of promoting the growth of a city’s green economy [57,59,62].

5.2. Limitations of the Study and Future Research

We acknowledge some limitations arising from a number of constraints in this study. First, the scope of this study was limited to the six MSEs held in China and recognized by the IOC from 2000 to 2020, but could not include the most recent and influential MSEs held in China in the past two years, such as the Beijing Winter Olympics and the Hangzhou Asian Games, due to data unavailability. In further studies, researchers should aim to obtain more up-to-date data to comprehensively capture the long-term impact of MSEs on urban green transformation. Second, our findings may not be fully applicable to other countries with different geographic and socio-economic environments. For example, the conclusion that hosting MSEs in coastal cities where public environmental concern is high has a stronger urban green transition effect may only apply to countries similar to China, with highly developed coastal city economies. In future research, we will collect panel data from more countries for comparative analysis, focusing on exploring the heterogeneous effects of MSEs on GTFP in different countries and regions to improve the generalizability of the findings. Third, the current methodology used to address the endogeneity of the impact of MSEs on urban green transformation is still immature, so an exploration of more effective instrumental variables is required to obtain more precise estimates of the impact. Finally, further research can incorporate qualitative research methods to gain a deeper insight into the mechanisms through which MSEs influence urban green transformation, and empirically explore other potential impact mechanisms and moderating factors, such as urban green technological innovation and environmental regulations. We will endeavor to study the contribution of major sports events to urban green transformation and the channels of impact from a more refined perspective.

6. Conclusions and Policy Recommendations

MSEs have a sustained impact on the economic, social, and environmental aspects of host cities, so discussions of their effects on urban green transformation are theoretically and practically important. We conducted an empirical analysis using the panel data of Chinese cities from 2000 to 2020, using a time-varying DID model to assess the impacts of six MSEs held in China that were recognized by the IOC on host city GTFP. The empirical conclusions passed a series of robustness tests. Additionally, we explored the heterogeneity and mechanisms underlying the influence of MSEs on urban green transformation. The empirical results indicated the following: (1) MSE hosting significantly increases the level of green economic growth in the host city, enhancing the GTFP of the host city by 2.01% on average. (2) The heterogeneity in the results indicated that organizing MSEs strongly promotes the green transformation of cities where the public is more concerned about the environment and in coastal locations. (3) The results of exploring the mechanisms through which these effects occur demonstrated that the hosting of MSEs can drive urban green transformation through three channels: encouraging residents to choose urban public transportation, enhancing the level of the digital economy, and promoting industrial upgrading. Based on the findings of this study, the following policy recommendations are proposed:

• With the international emphasis on the Olympic spirit, an increasing number of cities will have the opportunity to host various types of MSEs. The host countries and cities should embrace the long-term strategic goal of green and sustainable development. They should seize the opportunity to host MSEs with international influence to implement the IOC’s principles of sustainable development, strategies, roadmaps, and environmental protection standards. Simultaneously, they should innovate their institutional mechanisms, strengthen the urban governance capacity of government departments, and create a sustainable management model in line with their national conditions. The overall aim would be to minimize the negative impacts of MSE on the ecological environment while stimulating economic growth, successfully achieving both green sports and sustainable economic development, to promote the green transformation of the host city.
• Our empirical evidence showed that the Chinese government should prioritize coastal cities with high levels of public environmental concern when selecting host cities for MSEs. Choosing to hold MSEs in these cities can maximize the driving force of MSEs on urban green transformation while benefiting from international influence.
• The governments of host cities in China should actively promote the concept of green sports, thereby enhancing public environmental concern during the events. Additionally, they should implement effective policy measures to encourage residents to use public transportation, advance the digital economy, and expand tertiary industries. By increasing infrastructure quality, expanding the coverage of public services, adjusting the industrial structure, and upgrading the functions of cities, full advantage can be taken of the economic and social benefits of MSE. This accelerates urban renewal towards increased green efficiency, further enhancing the level of green economic development of the city and facilitating urban green transformation.