The Short-Run Effect of a Local Fiscal Squeeze on Pollution Abatement Expenditures: Evidence from China’s VAT Pilot Program

Abstract

Introduced in 2012, China’s value-added tax (VAT) pilot program gradually replaced business tax (BT) with VAT. It has created a large fiscal squeeze for the local government since 75% of VAT revenue goes to the central government. Employing a difference-in-differences estimator with continuous treatment intensity, we find that this fiscal squeeze has a negative effect on pollution abatement expenditures. Moreover, private firms in eastern regions are less responsive to this shock than those in the rest of China due to having better regulated local governments. We also find that this effect is smaller in magnitude if the firm owner is younger, more educated or has industrial and political connections compared to her respective counterparts. This fiscal squeeze reduces pollution abatement expenditures more in regions with higher fiscal stress, looser environmental regulation, and lower pollution abatement costs. Further exploration shows that, in response to this fiscal squeeze, local governments have adopted several tools to compensate for revenue loss, including increasing tax enforcement and loosening environmental regulation.

Appendices

Appendices

1.1 Appendix A

The current study adopts two distinct measures of environmental regulation (ER). They are separately evaluated in columns (2) and (3) in Table 16. The first measure concerns three representative indicators of regulation: wastewater, sulfur dioxide, and dust (Zhao and Sun 2016; Ren et al. 2018). We standardize the original data on all three indicators since the raw data are not comparable. Then, the value of each single indicator is mapped into the interval [0, 1]. In this way, we remove the restriction of different units. The min–max normalization method is applied. We evaluate the ratio of each city’s level to the average level. If the ratio is larger than 1, then this city’s weighted average emissions of these three representative indicators are higher than the average of all cities, and the regulation in that city is also looser than average. Finally, this measure is adopted as its inverse value. Therefore, if fiscal squeeze indeed loosens environmental regulation as expected, we should see a negative coefficient for this measure. It is calculated as follows:

$$ER_{1i} = \frac{1}{{\frac{1}{3}\mathop \sum \nolimits_{j = 1}^{3} W_{j} UE_{ij}^{n} }}$$

(3)

where $UE_{ij}^{n} = \frac{{UE_{ij} - \min \left( {UE_{j} } \right)}}{{\max \left( {UE_{j} } \right) - \min \left( {UE_{j} } \right)}}$, and $W_{j} = UE_{ij} /UE_{j}^{avg}$. $UE_{ij}^{n}$ is the rescaled (min–max normalized) emission of pollutant j in city i during our sample period. $\min \left( {UE_{j} } \right)$ and ${\text{max}}\left( {UE_{j} } \right)$, respectively, denote the minimum and maximum emissions of pollutant j of all cities in our sample. $UE_{j}^{avg}$ denotes the average emissions of pollutant j in all cities. $W_{j}$ represents the weight for each pollutant j.

The second measure of environmental regulation in our study focuses on the rates of abatement of sulfur dioxide and dust particles (Maisseu and Voss 1995). A higher rate of abatement indicates more stringent regulation in the city. We add the weight of each city, which is represented by the amount of emissions for each pollutant. This weighting is necessary because the amount of emissions varies greatly across cities, and the amount of emissions within each city also varies across different pollutants. If this measure of a city is higher, then we can conclude that firms in this city have, on average, a higher rate of abatement for the two representative indicators compared to other cities and are subject to a more stringent level of environmental regulation. If the fiscal squeeze indeed loosens environmental regulation as expected, we should also see a negative coefficient of this measure. It is calculated as follows:

$$ER_{2i} = \frac{1}{2}\mathop \sum \limits_{j = 1}^{2} P_{j} R_{ij}^{n}$$

(4)

where $R_{ij}^{n} = \frac{{R_{ij} - {\text{min}}\left( {R_{j} } \right)}}{{\max \left( {R_{j} } \right) - {\text{min}}\left( {R_{j} } \right)}}$. $R_{ij}^{n}$ is the rescaled (min–max normalized) rate of abatement of pollutant j (sulfur dioxide or dust particle) in city i during our sample period. ${\text{min}}\left( {R_{j} } \right)$ and ${\text{max}}\left( {R_{j} } \right)$, respectively, denote the minimum and maximum rates of abatement of pollutant j of all cities in our sample. $P_{j}$ represents the weight for each pollutant j. Specifically, $P_{j} = \frac{{D_{ij} }}{{\mathop \sum \nolimits_{i} D_{ij} }}/\frac{{GDP_{ij} }}{{\mathop \sum \nolimits_{i} GDP_{ij} }}$, where $D_{ij}$ denotes the amount of emissions of pollutant j for city i. This weighting is necessary because the amount of emissions varies greatly across cities, and the amount of emissions within each city also varies across different pollutants.

1.2 Appendix B

In this section, we provide graphical evidence showing that environmental regulation was loosened during our sample period. The following figure depicts that the ratio of environmental expenditures to total GDP has gradually decreased since 2012, indicating that environmental regulation has loosened in response to the fiscal squeeze (Qi and Zhang 2014; Bai et al. 2019; Hao et al. 2018) (Fig. 5).

Fig. 5

The ratio of environmental expenditure to GDP

1.3 Appendix C

In this section, we provide present descriptive statistics on several variables that are analyzed later in our paper (Table 17).

Table 17 Statistical summary