Measurement model setting

Since the time of implementing the "river chief system + procurator" in various regions is not consistent and even some administrative regions have not yet implemented the policy, this paper used a time-varying DID model. DID (difference-in-difference) is based on a counterfactual framework to assess the change in the observed Factor y in both cases of policy occurrence and nonoccurrence, and it is widely used in public policy evaluation because it can more accurately assess the net effect of policy implementation on individuals. Based on the weekly water quality statistics and real-time water quality monitoring system of the China General Environmental Monitoring Station from 2015 to 2022, this study used the time-varying DID model to test the impact mechanism of the "river chief system + procurator" policy on the water ecological environment in river and lake basins. By referring to the method of Thorsten and Alexey [18], a time-varying DID model was constructed:

Specifically, WQ refers to the water quality of the water quality monitoring point of country i in year t, and it is measured by the water quality class index and four water pollutant indicators, including the hydrogen ion concentration index, dissolved oxygen, potassium permanganate, and ammonia nitrogen. γ_i is the individual fixed effect, μ_t is the time fixed effect, ε_it is the disturbance term, and control_it includes population, GDP, GDP per capita, public revenue, public expenditure, the share of the primary sector, the share of the secondary sector, and sewage treatment rate as the control variables. The core explanatory variable R_it in this paper is the interaction term between the individual dummy variable and the time dummy variable and will be used to identify national water quality monitoring sites where the "river chief system + procurator" policy has been implemented in year t. In addition, the model estimation and parallelism trend tests all used robust standard errors.