The coordination pattern of tourism efficiency and development level in Guangdong Province under high-quality development

China’s economy has entered a stage of high-quality development, making the transformation and upgrading of tourism, as well as improvements in quality and efficiency, essential. Guangdong Province is a leading tourism economy in China, with rapid growth in cultural and tourism industries and sustained top national rankings in added value and tourism revenues. However, the COVID-19 pandemic caused sharp declines in tourists in 2020 with only partial recovery in 2021, and there exist significant inter-city gaps in tourism efficiency and development level. The study focuses on Guangdong’s 21 cities to understand and promote coordinated development of tourism efficiency and development level, thereby enhancing regional tourism competitiveness. The research aims to measure tourism development levels (2000–2020), evaluate and decompose tourism efficiency, analyze spatiotemporal dynamics of regional differences and spatial structures, and construct and assess a coupling coordination model between tourism efficiency and development level.

Prior research on tourism efficiency has evolved across four dimensions: (1) Content has moved from single-efficiency evaluations (e.g., management, operational, transportation efficiency) to comprehensive assessments such as ecological efficiency, poverty alleviation efficiency, and regional tourism efficiency. (2) Methods have shifted from qualitative evaluations to quantitative multi-model approaches, including DEA, SBM-Malmquist, and DEA-SNA. (3) Spatial scales have transitioned from national/provincial/mega-regional levels to more meso-scale analyses (e.g., national scenic spots, A-level attractions). (4) Analytical depth has progressed from describing spatiotemporal variation to exploring driving mechanisms involving natural environment, economic development, resource endowment, transportation, and institutional factors. Despite these advances, limitations remain: many studies give insufficient joint consideration to both efficiency and development level under the high-quality development paradigm; classic exploratory spatial analyses often emphasize cross-sectional spatial association, missing temporal dynamics; and static approaches without the time dimension reveal limited regional differences. The literature underscores the need for integrated spatiotemporal analyses and coupled evaluations of tourism efficiency and development level.

Study area and period: 21 cities of Guangdong Province (Pearl River Delta, Eastern, Western, Northern Guangdong) from 2000 to 2020. Methods: (1) Measurement of tourism development level using total tourist headcount and total tourism revenue (deflated by CPI with 2000 as base), with indicator weights derived via the entropy method. (2) Tourism efficiency measurement and decomposition using the DEA-BCC model and the Malmquist productivity index to obtain changes in comprehensive efficiency (EFFCH), technological progress (TECHCH), pure technical efficiency (PECH), scale efficiency (SECH), and total factor productivity (TFPCH). (3) Exploratory spatiotemporal data analysis of local spatial structure using LISA time paths (length and curvature capturing the dynamics and volatility of local spatial dependence) and LISA spatiotemporal leap analysis based on a Markov transition framework to classify transitions into four types (I–IV) indicating combinations of self and neighborhood stability/leaps. (4) Coupling coordination degree model between tourism efficiency and development level: coupling degree C computed from subsystem indices f(x) and g(y) with k=2, and coupling coordination D = C × T (T = α f + β g; α=β=0.5). The coupling coordination was classified into severe disorder, moderate disorder, basic coordination, moderate coordination, and high coordination with thresholds 0.05, 0.10, 0.15, and 0.20. Indicators: Inputs—labor approximated by the number of employees in the tertiary industry; capital proxied by numbers of 3A (or three-star) and above tourist attractions, star-rated hotels, and travel agencies (tourism land data unavailable). Outputs—total tourist arrivals and total tourism revenue (deflated). Data sources: Guangdong Provincial Statistical Yearbook, Tourism Yearbook, and Statistical Bulletins of National Economic and Social Development (circa 2000–2002); administrative boundary vectors from national mapping authorities. Missing data were supplemented using indicator smoothing. Software: ArcGIS 10.2 for spatial classifications and mapping.

- Tourism development level: The logarithmic mean value increased from 0.012 (2000) to 0.067 (2020), showing a steady upward trend. Natural break thresholds for development scale were 0.0175, 0.03, 0.0425, and 0.0622; the overall mean was 0.047 (medium scale). City counts across scales were 6, 7, 3, 3, and 2, with 90.47% falling in small to medium scales. High-value scale areas were concentrated in cities such as Zhanjiang and Foshan. - Malmquist results (2000–2020 averages): TFPCH = 1.2003 (overall intensification increasing), EFFCH = 1.0253, TECHCH = 1.1682, PECH = 1.0026, SECH = 1.0000. Periods of TFP decline included 2004, 2006–2008, and 2011–2014, associated with shocks such as SARS, the financial crisis, and haze events. Comprehensive efficiency improvement was not the main driver of productivity growth; rather, technological progress led the gains. The mean pure technical efficiency level was 0.816, with the change index >1 in 13 of 20 intervals (mean 1.025), indicating fluctuating but improving technical efficiency; scale efficiency level averaged 0.856, with SECH = 1.000, suggesting unchanged scale efficiency. Under largely unchanged scale efficiency, comprehensive efficiency showed fluctuating declines at times, indicating suboptimal factor configuration and resource use. - LISA time paths: For tourism efficiency, mean path curvature = 10.424; the maximum LISA time path length occurred in Guangzhou (27.925) and the minimum in Meizhou (3.246). For development level, mean path curvature = 6.813; Shenzhen (19.309) and Zhuhai (11.076) were high, Meizhou was lowest (1.250). Development level’s local spatial structure was slightly more volatile in dependence direction than efficiency, with higher curvature in PRD and lower in peripheral regions. - LISA spatiotemporal leap analysis: Tourism efficiency experienced spatiotemporal leaps in 18 cities (85.71%), indicating active transitions and instability in local spatial structure; three cities (14.29%) showed no leap (type IV), evidencing transfer inertia. Five cities (23.81%) showed synergistic leaps (type III). The number of HH-type cities rose from three to five, indicating strengthened high-efficiency agglomeration. Tourism development level showed leaps in 15 cities (71.43%), with six cities (19.35%) showing no leap (type IV) and two cities (9.53%) showing collaborative leaps (type III). HH-type cities increased from two to four; LL-type decreased from four to two, suggesting increased agglomeration at higher development levels and dispersion among lower levels. - Coupling and coordination: Mean coupling coordination increased from 0.115 (2000) to 0.136 (2010) and 0.159 (2020), indicating a shift from severe disorder toward basic coordination. High coordination areas became more concentrated in the PRD. Some high-initial-level regions (e.g., Guangzhou, Shenzhen) showed gradual or continuous regression in coupling coordination. Spatial divergence patterns of coupling degree and coupling coordination were similar, but the diffusion of high coupling coordination areas was slower and more substantial than that of high coupling degree areas. Overall, most cities improved from dysfunction to coordination over time.

Findings indicate pronounced spatial heterogeneity in tourism efficiency across Guangdong, with the PRD exhibiting higher values and peripheral regions lagging. Technological progress, rather than improvements in comprehensive efficiency, primarily sustained productivity growth, while largely unchanged scale efficiency constrained gains. Local spatial structures of efficiency were unstable and prone to positional changes, whereas development levels exhibited stronger path dependence and locking. Coupling and coordination between tourism efficiency and development level improved overall, yet core cities sometimes regressed, suggesting congestion or over-concentration effects that may inhibit efficiency gains and spillovers. The slower diffusion of coupling coordination, compared to coupling degree, points to limited effectiveness of regional cooperation mechanisms. Policy implications include accelerating core city cluster construction, enhancing inter-city collaboration, optimizing allocation and flow of key factors (population, capital, technology), and leveraging PRD spillovers to reduce regional disparities. Strengthening coordinated regional development, protecting ecological assets, and deepening “tourism+” integration with other industries can further enhance quality and efficiency. The positive linear relationship between coordination and the levels of tourism efficiency and development underscores that higher coordination aligns with more developed, reasonable, and healthy tourism systems.

(1) Tourism comprehensive efficiency from 2000 to 2020 shows evident spatial differences, with high values concentrated in the Pearl River Delta and lower values in eastern, western, and northern Guangdong. Technological progress is the main driver of productivity growth; when scale efficiency remains unchanged, comprehensive efficiency may fluctuate downward, reflecting suboptimal factor configurations and the need for transformation and upgrading. (2) At the local scale, tourism efficiency is relatively unstable and susceptible to positional changes among municipal units, whereas the tourism development level is more stable due to path dependence. Both internal factors and neighboring influences shape local spatial structures, with internal drivers playing a stronger role. (3) Coordination between tourism efficiency and development level exhibits a positive linear relationship with both outcomes; higher coordination corresponds to more advanced and healthier tourism development. Spatially, high coordination aligns with PRD high-value distributions. (4) Despite progress, spatial polarization persists and unit characteristics remain scattered, driven by uneven development foundations and intra-regional disharmony. Strengthening regional integration, building collaborative development circles, and promoting factor flows from core to peripheral areas are needed to realize trickle-down effects and coordinated development. Future research should refine evaluation systems by incorporating broader indicators of sustainability, ecological constraints, transportation convenience, human and policy factors, and innovation capacity to provide a more comprehensive measurement of high-quality tourism development.

- Indicator constraints: City-level data on tourism-specific labor and capital were limited; labor was proxied by total tertiary industry employment, and capital by counts of 3A+ attractions, star-rated hotels, and travel agencies. Tourism land and fixed investment statistics were unavailable. - Data processing: Some missing values were supplemented via indicator smoothing, and revenue was deflated using CPI (base year 2000), which may introduce measurement uncertainty. - Scope of quality dimensions: Due to data availability, measurement focused on economic outputs (arrivals, revenues) and did not fully incorporate environmental, social, or governance dimensions of high-quality development. - Spatial-temporal modeling simplifications: LISA-based transitions and Markov classifications capture dynamics but may not account for all underlying causal mechanisms or policy shocks, and equal weighting (α=β=0.5) in the coupling model assumes equal importance of subsystems.