An environmentally related policy impact analysis considering wind effect: evidence from suspending old coal-fired generators in South Korea

ABSTRACT

This study measures the causal effect of an individual, aged coal-fired generator on PM2.5 concentration. For this, the policy of shutting down old coal-fired power plants in Korea as a quasi-natural experiment was examined. This study adopts the difference-in-differences technique and controls both for wind effects and overseas pollutant effects. The results show that the shut-down policy reduces daily PM2.5 concentration by 2.04–3.65 μg/m3 across the number of generators suspended, and the share of wind effects in the total policy effects is estimated as about 50%. In addition, overseas effects account for about 20% (5.64 μg/m3) of regional PM2.5 concentration on average. These results imply that although the closure of old coal-fired power plants can reduce average PM2.5 concentration, the policy alone is not enough to expect a remarkable improvement in air quality when PM2.5 concentration is high. Thus, these findings suggest that the Korean government must consider domestic policies and cross-border cooperative policies to resolve the high PM2.5 concentration issue.

KEYWORDS:

• Policy experiment
• PM2.5 concentration
• coal-fired power plant
• difference-in-differences

JEL CLASSIFICATION:

• Q53
• Q58

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

¹ Before COVID-19 pandemic, The frequency of fine dust advisory issues increased from 39 in 2016 to 590 in 2019. A fine dust advisory is issued when average hourly PM2.5 concentration exceeds 75 μg/m3 for more than two hours.

² Appendix includes institutional background, the description about data, summary statistics, characteristics of the treatment and the control group, the suspension schedule of two generators, the detailed explanation regarding identification strategies, and the daily PM2.5 concentration trend of two groups.

³ To identify wind direction effects, this study generated a dummy representing wind blowing from the power plant based on the difference between the azimuth of each station relative to Yeongdong power plants and that of wind blowing to each station. Thus the dummy equals one if the difference is less than 45, and it equals zero if the difference is larger than 45.

⁴ Table A4 in Appendix includes results of robust check based on models including various variables related to wind.