Does ESG Certification Improve Price Efficiency in the Chinese Stock Market?

Retrieved from the Forbes website. https://​www.​forbes.​com/​.

In 2019, 99.76% of investors in the Chinese stock market are individual investors, and they hold about 31.6% of the A-share market capitalization. This data is from “the 2019 National Stock Market Investor Status Survey Report”, released by the Securities Association of China.

Hou and Moskowitz (2005) propose three ways to measure the pricing efficiency performances. Considering D2 is weighted with lags in construction, and D3 is adjusted for accuracy by adding standard errors to the base of D2, we do not report the results based on D3 in this paper for brevity. These results can be found in the online Appendix or request from the authors.

For our choice of a 1-week price delay, Hou and Moskowitz (2005) argue that the price delay is best measured on a weekly frequency, as there is little variation at the monthly level and too much noise at the daily level.

The ESG 300 index refers to excluding the 20% of listed companies with the lowest ESG scores in the CSI 300 index, use remaining stocks as index samples to provide performance benchmarks for ESG investment. The ESG 100 refers to selecting 100 stocks with higher ESG ratings from the CSI 300 index, and the ESG 40 refers to the selection of 40 stocks with higher ESG ratings from the SSE 180 Corporate Governance Index. The ESG300 index sample is adjusted every six months, and the sample adjustment time is the next trading day on the second Friday of June and December each year. The ESG100 index sample is adjusted every six months, and the sample adjustment time is the next trading day on the second Friday of January and July each year. The proportion of the sample adjusted each time generally is less than 20% unless the original recalled ratio exceeds 20%. The ESG40 index sample is adjusted every six months, and the sample adjustment time is the next trading day on the second Friday of January and July each year. The proportion of samples adjusted each time generally is less than exceed 10% unless the original recalled ratio exceeds 10%. For old samples with the same rating before and after the adjustment, stocks ranked in the top 60% of the average daily total market value are retained first.

As reported in Table 4, all the F-stats are big and significant at the 1% level, which indicates the explanation power of our regression is high and has a good measure for our study. Only the r-squared based on the pricing efficiency proxy |ρ| is relatively low, which might be influenced by the way to calculate this variable (using only one lagged market return). Other r-squares based on D1 and D2 are near to 0.08, which is relatively high in the monthly panel data regression for the asset pricing studies (see in He and Fang, 2019; Gao et al., 2020; Cao et al., 2021 who also find the r-squared is less than 0.1). Combing the high value of F-stats in these tables, we can draw a conclusion that our regressions are reasonably set and can be used to study the influence of ESG certification on pricing efficiency.

The probabilities of informed trading reflect various types of transactions using private information. The lower the value of \(VPIN\) indicates a higher degree of information symmetry.