China, a major manufacturing hub and global trade player, faces a critical water scarcity challenge coupled with severe water pollution. Industrial production is a significant contributor to this pollution. Existing literature primarily examines the aggregate effects of trade on pollution, overlooking firm-level heterogeneity. This paper addresses this gap by investigating how firms respond to trade openness concerning water pollution, focusing on the role of intra-industry agglomeration and firm productivity. The study's central question is how firm productivity, influenced by agglomeration and export activity, affects water pollution. The hypothesis is that high-productivity firms, benefiting from agglomeration and trade, can ultimately reduce water pollution through resource reallocation, a mechanism consistent with the PRR hypothesis. This firm-level analysis contrasts with previous research that often relies on aggregated national or regional data, thereby offering a more nuanced understanding of the trade-pollution nexus in China.

The paper reviews existing literature on trade and environmental economics, focusing on studies examining the relationship between trade and various pollutants. Earlier research often considered aggregated effects at national or regional levels, using hypotheses like the environmental Kuznets curve, competitiveness hypothesis, and pollution haven hypothesis. More recent studies emphasize firm-level heterogeneity, proposing the PRR hypothesis, the distressed and dirty industry hypothesis, and the pollution offshoring hypothesis. This paper specifically tests the PRR hypothesis for water pollution, differentiating it from studies focusing on other pollutants or using aggregated data. The literature indicates varied effects of trade on pollution, depending on factors like economic development stage, industry type, and the specific pollutant considered. While some studies suggest benefits to the environment from free trade at an aggregate level, others highlight country-specific or pollutant-specific relationships. This paper aims to add to this body of knowledge by providing a firm-level analysis of water pollution in China, focusing on the channels through which trade affects pollution.

The study employs a two-stage empirical strategy using a firm-level panel dataset from 2000 to 2012, combining data from the Annual Survey of Industrial Firms (ASIF), Annual Environmental Survey of Polluting Firms (AESPF), and China Statistical Yearbook (CSY). The first stage investigates the impact of intra-industry agglomeration on firm productivity, using a model that includes agglomeration index, firm characteristics (export intensity, capital-labor ratio, wages), and city-level characteristics (graduates, FDI). The second stage examines the relationship between firm productivity, export activity, and water pollution. This model includes firm productivity (measured as labor productivity), its squared term, export intensity, capital-labor ratio, wages, their squared terms, revenue, water usage, water purification capacity, and relative values of capital-labor ratio and water usage. A three-stage least squares (3SLS) estimation technique is used to address endogeneity issues arising from the two-stage structure. Further analysis involves interaction tests (using fixed effects) to explore how productivity interacts with export effects, composition effects, and technique effects. Finally, marginal effects analysis using figures are employed to further understand the conditional relationships between variables. The agglomeration index is calculated using firm employment data to represent intra-industry agglomeration. Export intensity is measured as the ratio of export value to total revenue.

The empirical analysis reveals an inverted U-shaped relationship between firm productivity and water pollution. Intra-industry agglomeration positively affects firm productivity, but its effect diminishes after a certain level. The key finding is that higher-productivity firms, particularly exporters, eventually generate less water pollution. Initially, higher productivity correlates with higher pollution, but beyond a certain productivity threshold, the relationship reverses, supporting the PRR hypothesis. The study finds that the export effect on water pollution is conditional on firm productivity and number of employees. At low productivity levels, the export effect is positive; but as productivity increases, it turns negative. The analysis of the effect of number of employees shows a negative interaction with export intensity, meaning increases in employees lessen the positive effect of exports on water pollution. This suggests that resource reallocation (specifically, labor) between firms plays a crucial role in the pollution reduction process. Regional disparities are also evident, with the eastern region exhibiting higher water pollution levels compared to other regions. The interaction analysis between exporting agglomeration and capital-labor ratio reveals a non-linear relationship. This means that while initially exporting agglomeration may lead to increased pollution due to resource reallocation towards exporting firms, beyond a certain agglomeration level, the effect reverses, indicating increased productivity and pollution reduction.

The findings strongly support the PRR hypothesis for water pollution in China. The inverted U-shaped relationship between productivity and pollution, along with the interaction effects and marginal effect analysis, highlight that trade, when coupled with high productivity, can induce pollution reduction through efficient resource reallocation. The observed regional disparities indicate the need for region-specific policies. The non-linear relationship found in the interaction analysis emphasizes the importance of considering different stages of agglomeration and their impact on pollution. The study extends the understanding of the trade-environment relationship to the firm-level, providing insights into the mechanisms through which trade affects pollution.

This study empirically validates the PRR hypothesis in the context of water pollution in China, demonstrating that trade liberalization can lead to pollution reduction through productivity-driven resource reallocation. Policy implications include prioritizing productivity enhancement, encouraging exports with stringent pollution controls, and designing region-specific policies that address the water pollution disparities. Future research can delve deeper into intra-industry trade at the firm level and incorporate green technology innovation data to investigate its role in pollution reduction.

The study's limitations include the use of labor productivity as a proxy for firm productivity instead of total factor productivity (TFP), which would require data from a more limited period. The unavailability of complete import data limited the analysis of intra-industry trade. The reliance on firm-level data might not capture all aspects of the national water pollution picture. However, the results, despite these limitations, still provide valuable insights.