Market integration is usually expected to help regions grow. But in China’s Pearl River Delta city cluster, this study finds a strikingly different pattern: integration is linked to slower growth locally, with little detectable effect on neighboring cities. The Pearl River Delta city cluster contains nine cities: Guangzhou, Shenzhen, Foshan, Huizhou, Dongguan, Zhuhai, Zhaoqing, Jiangmen, and Zhongshan.

It also has a highly developed economy and strong transport and port advantages. In 2021, the PRD urban agglomeration created 80.876 percent of Guangdong Province’s total GDP using 30 percent of its land area, with a total GDP value of 1,518.022 billion United States dollars.

City clusters are an important force in regional economic development, and the PRD city cluster is an important engine for Guangdong and China. At the same time, the PRD still has unbalanced and insufficient development among cities, an inadequate market economy system, and bottlenecks in the spatial flow of goods and factors.

Existing research on market integration and economic growth is relatively rich, but spatial econometric models are used less often to analyze spatial effects. Most studies examine integration between countries, while less attention has been paid to integration within a country, especially within a city cluster.

The study examines nine PRD cities and constructs a relationship model using GDP per capita as the explanatory variable and market integration as an explanatory variable. The analysis also includes economic openness, human capital, consumption level, government spending scale, and social capital stock as control variables.

Finally, spatial analysis models are used for fixed-effects and decomposition-effects analyses. The core explanatory variable is the Market Integration Total Index. Market integration is divided into goods, labor, and capital market integration.

The relative price index method measures commodity market integration and capital market integration for nine PRD cities from 2010 to 2019. The absolute deviation method of average wages calculates labor market integration, while the coefficient of variation method determines the weights used to obtain the total market integration index.

The data mainly come from the Guangdong Provincial Statistical Yearbook for 2010 through 2019. The study covers nine cities and includes GDP per capita, the Market Integration Total Index, and the control variables. The variables include imports and exports per capita, college students per ten thousand people, retail sales relative to GDP, government fiscal expenditure relative to GDP, and social fixed capital investment relative to GDP.

Macroeconomic variables among PRD cities may be spatially linked, so the study determines whether spatial autocorrelation exists before conducting spatial econometric modeling. The study measures Moran index values for GDP per capita in the PRD from 2010 through 2019 to examine overall regional correlation.

The text says GDP per capita is significantly and spatially positively correlated globally, while Table 3 reports Moran index values below zero that pass one-percent or five-percent significance tests. Figure two uses Moran scatterplots to show local spatial autocorrelation across the PRD city cluster.

Panels A through D report GDP per capita, with negative Moran indices in every year: negative zero point four eight seven in 2010, negative zero point four six three in 2013, negative zero point four nine zero in 2016, and negative zero point four eight six in 2019. Panels E through H show market integration, with indices of zero point three five nine, zero point three three four, negative zero point one four three, and zero point zero three zero, revealing how spatial clustering changes over time.

The study uses scatterplots for 2010, 2013, 2016, and 2019, and a LISA plot with a five-percent significance level to examine local concentration in specific cities. In general, the local centers of economic growth in the PRD urban agglomeration show a spreading trend.

Figure three maps local LISA agglomerations across the PRD city cluster, separating GDP per capita in panels A through D from market integration in panels E through H for 2010, 2013, 2016, and 2019. The GDP maps show no significant local spatial clustering in these years, while market integration displays localized patterns: Zhuhai is a significant high-high cluster in 2010; Huizhou, Dongguan, and Shenzhen form significant high-high clusters in 2013; and Foshan and Zhongshan show significant low-low correlations.

For market integration, the PRD city cluster shows localized agglomeration in 2010, 2013, 2016, and 2019. Zhuhai entered a significant high-high agglomeration type in 2010. In 2013, Huizhou, Dongguan, and Shenzhen showed significant high-high correlation, while Foshan and Zhongshan showed significant low-low correlations.

The Spatial Durbin Model is constructed to study spatial effects from neighboring cities’ economic growth and neighboring market-integration levels on the observed cities’ economic growth. Table 5 shows that the LM statistic passes the five-percent significance test, indicating that both the SEM and SLM models are applicable and that the more general SDM model can be selected.

Wald and LR tests show that the SDM models do not degenerate into SEM or SLM models. Table five reports four specification tests for the PRD urban agglomeration’s spatial model. The LM statistic is one hundred seventy-two point forty-five, while Hausman, LR, and Wald report forty-one point twenty-eight, one hundred four point eleven, and twenty-four point thirteen; each has a p-value of zero point zero zero zero, marked with three asterisks.

Together, these results support using a general spatial Durbin model and fixed effects for the analysis. The study compares space-fixed, time-fixed, and space-time double-fixed SDM models. The space-time double-fixed model has the largest R-squared value, 0.0571, indicating the best fit among these models.

It also has the smallest sigma-squared value and the largest log-likelihood value, indicating stability and the best explanatory ability according to the text. Table six compares spatial Durbin model estimates under space-fixed, time-fixed, and space-time, or dual, fixed effects.

It reports both main coefficients and spatially lagged, or W-x, coefficients, with standard errors in parentheses; for example, l n Open is positive and marked with three significance stars in all three specifications. The authors use fit statistics to compare these models, including R-squared values of zero point zero one nine seven, zero point zero zero one one, and zero point zero five seven one, alongside log-likelihoods of one hundred fifty-eight point three one five five, twelve point zero zero nine one, and one hundred eighty-one point eight four eight five.

In the space-time double-fixed model, the estimated coefficient of market integration is minus 48.2086 and is significant at the ten-percent level, indicating that market integration is not conducive to regional economic growth. The lagged market-integration term is 0.0172 and insignificant, indicating no obvious effect on economic growth in surrounding areas.

The text gives possible reasons: resources concentrate in Guangzhou and Shenzhen, creating a siphon effect, while administrative, factor, and public-service barriers hinder circulation and spillovers. Table seven decomposes the spatial Durbin model’s effects into direct effects within a region, indirect effects from neighboring regions, and total effects.

For market integration, the direct effect is minus fifty-one point seven eight nine three, the indirect effect is twenty-four point two nine one six, and the total effect is minus twenty-seven point four nine seven seven, reported as insignificant in the surrounding text. The table also reports significant total effects for openness, human capital, and government, highlighting how regional and neighboring influences contribute differently to economic growth.

The market-integration direct effect on economic growth is negative, the indirect effect is positive, and the total effect is negative and insignificant. The text also states that market integration inhibits regional economic growth, while higher market integration in neighboring areas has an insignificant promoting effect.

The study finds that market integration negatively impacts regional economic development and has a negligible effect on surrounding areas. Its recommendations are to eliminate market trade barriers, reduce local protectionism, enhance infrastructure, and create an efficient collaboration mechanism.

The central finding is that market integration restrained regional economic growth and produced negligible benefits for surrounding areas. The paper therefore emphasizes removing trade barriers, reducing local protectionism, improving infrastructure, and building stronger collaboration across cities.