Received: 5 October 2019 Revised: 2 April 2020 Accepted: 24 May 2020
DOI: 10.1002/hpm.3012
R E S E A R CH A R T I C L E
Endangering China's environmental health security
goals through negative environmental investor
behaviours
Lulin Zhou1,2 | Maxwell O. Antwi1 | Henry A. Antwi2,3 |
Ama Boafo-Arthur2,4 | Tehzeeb Mustafa2,3
1School of Management, Jiangsu University,
Summary
Zhenjiang, China
2Center for Health and Public Policy China sees the need to maximise its environmental health
Research, Jiangsu University, Zhenjiang, China security as a major priority in its sustainable development
3Institute for Systems Engineering, Jiangsu agenda. This is at the heart of China's “ecological civilisation”
University, Zhenjiang, China
and “beautiful China” dream. One of the objectives of this
4School of Continuing and Distance
dream is to sensitize investors to invest in health and envi-
Education, University of Ghana, Accra, Ghana
ronmental stocks to support environmental health goals.
Correspondence
However, both the Shanghai and the Shenzhen stock mar-
Henry A. Antwi, School of Management,
Jiangsu University, 301 Xuefu Road, kets continue to witness contemporaneous movement
Zhenjiang, 212013, China. (herding behaviour) by investors from environmental stock to
Email: asanteantwi2@yahoo.com
perceived safer stocks and this is stifling the growth of the
Funding information environmental health sector due to capital deprivation. Our
Government of Jiangsu Province, Grant/
paper evaluates the significance and potential effect of this
Award Number: 1102129C; Post-doctoral
Foundation of China, Grant/Award Numbers: herding trend among environmental stocks using a collection
2018T60508, 2018M511705; National
of sophisticated econometric models namely, the state-space
Natural Science Foundation of China, Grant/
Award Numbers: 71904066, 71974079 model, enhanced state-space model, the cross-sectional SD
(CSSD) and the cross-sectional absolute deviation (CSAD).
The models are used to evaluate firm-level data collected
from the 80 environmental stocks indexed by the KGRM
MSCI China IMI Environment 10/40 Index. Three of the
models confirm the presence of endemic negative (herding)
investor behaviour among environmental stocks in China and
this threatens the sustainability of environmental stock capi-
tal to promote China's environmental health goals. We have
proposed measures to ameliorate the risks posed by such
negative contemporaneous investor behaviours.
1398 © 2020 John Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/hpm Int J Health Plann Mgmt. 2020;35:1398–1411.
ZHOU ET AL. 1399
K E YWORD S
environmental, health, herding behaviour, investor, stock market
1 | INTRODUCTION
The effect of environmental degradation on human health and well-being has gained a self-perpetuating momentum
across the globe since the early 1960s. Many countries have responded to the global call enacting various environ-
mental health regulations, policies programs and thresholds.1 In most of these countries, both public and private
organisations are required by domestic and international laws to design enduring sustainable mechanisms to mini-
mize environmental health externalities associated with their operations.2 Since the United Nations Conference on
Environment and Development (Rio de Janeiro Earth Summit, 1992), new frontiers in ecological health security have
emerged beyond the previously known areas of environmental concern.3 Before this Earth Summit in 1992, environ-
mental health security programs among countries focused mainly on natural and few built environmental factors that
affect human health, for example, sanitation.4 Over time, and in response to the UN Decade of Education for Sus-
tainable Development (2005-2014), a new sphere in environmental health security emerged to shift the paradigm to
radically promote development that seeks to meet human development goals without compromising on the natural
systems, resources and ecosystem upon which the global economy and society depends.4
Central to the new sphere in sustainable development is environmental health security. This interest is also due
to the increase in the records of industry-induced health casualties over the world. The focus of new environmental
concerns on environmentally responsible behaviours at all level of development keeps inspiring advanced research in
pro-environment programs in China and many other emerging countries.5 China's ecological civilization program
requires all manufacturing enterprises to adopt environmentally responsible manufacturing (ERM) practices to pre-
vent, detect and respond to environmental-led calamities.6 China's industries are consciously but systematically
readjusting their operational systems, skills, strategies, structures and shared values to positively alter the changing
dynamics of the ecosystem and its resulting environmental health concerns.7
Most of the environmentally responsible business practices being pursued in China are economically driven
system-wide and integrated strategies that seek to eliminate or reduce waste streams associated with supply chain,
product and service design, manufacturing processes, product use and disposal of materials.8 Fundamental in this
endeavour is the recognition that waste and pollution may have far-reaching environmental health security implica-
tions. Despite the benefits of the previous generation of sustainable techniques that emerged in the early 1980s to
support industries (just in time, total quality management, process reengineering, time-based competition, etc.), these
alone have not been sufficient to safeguard concerns for the protection of the environment9 hence need for alterna-
tive approaches. Further, as more and more industrial clusters emerge across China, another move to imbed sustain-
able environmental health practices is driving collaborative networks so that group of firms can optimise
environmentally responsible behaviours through shared resources. For example, Biotech, Medtech and Pharmaceuti-
cal firms such as Jiangsu Hengrui Medicine, Fosun Pharmaceutical, Sinopharm Group, Harbin Pharmaceuticals,
Guangzhou Pharmaceutical Holdings Limited in Jiangsu, Guangdong, Heilongjiang, Shanghai and Jiangxi Provinces
have been successful in promoting greater environmental health security through self-regulating collaborative net-
work of environmentally responsible policies, programs and standards.10 These industry-led initiatives have engen-
dered more exceptional public support, renewed political commitment from top leaders and elicited massive
government subsidies for these industries among others.10
Despite this seemingly successful effort in China's quest to secure its environmental health security goals, the
strategies and policies are not without pitfalls. China believes that a flourishing environmentally stock market is
essential to get the needed capital to consolidate environmental health goals.11 President's Xi's “ecological
 10991751, 2020, 6, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/hpm.3012 by University of Ghana - Accra, Wiley Online Library on [06/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
1400 ZHOU ET AL.
civilisation” and “beautiful China” dream are founded on a strong public–private partnership through capitalisation of
the environmental stock market.
However, Ji and Zhang12 observed that the environmental stock market in China is not attractive to Chinese
and foreign investors alike. In the last decade, the market has not been stable as investors move away in herds to
perceived safer stocks intermittently in response to both internal and external shocks. For example, high environ-
mental tariff, tight ecological inspections, closed-down of non-compliant producers have affected investor confi-
dence in the sector.
Further, the regulations for new energy vehicles announced by the government in September 207 helped the
dominance and surge in the stock price of BYD (Chinese electronic car manufacturer).12 Yet this sudden decision
failed to give non-Chinese makers enough time to prepare, such as building new plants and designing new models,
effectively ensuring fines or forcing them to buy high-priced credits from Chinese manufacturers once the regula-
tions go into effect.12 When the stock market growth is created by such an artificial process as in the case of BYD,
external and institutional investors become apprehensive as they do not know the next move of the state, which can
potentially affect the profitability of their investments in the environmental stock market.13
The logical thing is for the investors to flee in herds (en-mass) to safer investments making the sector vulnerable
to investor herding behaviour.14 A recent study by KraneShares (a company that sponsors an exchange-traded fund
attached to the China environmental health theme) reported on the challenges of the environmental stock market
due to frequent environmental stock market fluctuations (returns and prices). Similarly, Broadstock & Cheng15 also
noted that over 50% of ecological stocks in China are overweighted and vulnerable to investor shifts when policies
become less favourable. This tendency has the potential to increase market uncertainty and further compound the
rate of mass and unexpected herding of investors from the environmental health market in China. This pattern is
suspected to be responsible for the nearly 6.4% decline in the KraneShares MSCI China Environment Index from
June 2018 to April 2019.
Moreover, as the rate of economic growth begins to slow down in China, environmental investors are more con-
cerned with the need to preserve their capital and have begun deserting long-term investment strategies in the envi-
ronmental stock market. They are instead opting for more stable and short-term markets such as the healthcare
industry.16 China's culture is highly collectivist as such investors tend to influence many others when they decide to
shun a particular market. If this happens in the environmental stock market, the capital needed to stimulate environ-
mental health security goals will reduce.17 Several studies have already been conducted regarding herding behaviour
in other stock markets in many countries including China, but not much has been done in the case of the environ-
mental stock market in China and its repercussions on environmental health security. These studies have also dif-
fered in the methodologies used as well as the outcomes they identified.
For example, Christie and Huang17; Chang et al18 and Demirer and Kutan19 studied herding behaviour based on
rational asset pricing conducted on securities in the United States, Europe and China, respectively. These studies
observed that herding behaviour contributes significantly to the return on securities during market recessions.
Demirer et al20 also analysed the existence of investor herding behaviour among 60 highly capitalised stocks based
on a cross-sectional SD model. The study did not find herding behaviour among investors during period of economic
stress. A similar study was conducted in the crude oil and heating oil capital markets by Uchida and Nakaga.21 The
authors used a local-stochastic volatility (LSV) model known as the cross-sectional variability of factor sensitivities to
analyse data on daily returns of the stock market. Chen et al14 developed it for the Japanese securities market and
their results confirmed that investor do herd on the stock market. In the healthcare sector, Lulin et al22 explored the
effect of investor herding behaviour on the Chinese stock markets, whereas Tripathi et al23 used herding techniques
to study quarantine behaviour during the epidemic to support control. Similar industry-level studies by (Hwang and
Salmon,24 Chang,25 and Flammer26) are documented in the existing literature with conflicting outcomes based on
the differences in the analytical procedure used.
This study is seeking to find out whether the environmental stock market in China also experiences investor's
herding behaviour. This is done by assembling a collection of four sophisticated econometric models, namely the
 10991751, 2020, 6, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/hpm.3012 by University of Ghana - Accra, Wiley Online Library on [06/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
ZHOU ET AL. 1401
state-space model, enhanced state-space model, the cross-sectional SD (CSSD) and the cross-sectional absolute
deviation (CSAD). We then compare the results of the four methods to determine whether they prove that indeed,
there is investor's herding behaviour in the environmental stock market in China. If this is the case, we can then
argue that environmental health security in China is endangered by harmful investor behaviour on the stock market.
To the best of our knowledge, this study is the first to apply the econometric model to determine investor's
herding behaviour in the environmental stock market. Moreover, this study is innovative because we apply the
models to firm-level data instead of industry-level data, which have predominantly been employed in the earlier
studies. We also use the daily equally weighted index returns data from the selected environmental stocks on the
Shenzhen, and Shanghai stock markets. The Hong Kong stock market is included for standardisation purposes. The
three stock markets are essential for this study because they have a high presence of both domestic and foreign
investors. Previous researchers have posited that foreign investors with higher information asymmetry tend to herd
more than local investors. We briefly explain our methodology, the source of data and analytical procedure and pre-
sent the empirical. We conclude the study by explaining the environmental health policy implications and future
research direction.
2 | MATERIALS AND METHODS
2.1 | Data selection
We collected the data on the 80 environmental stocks indexed on the MSCI China IMI Environment 10/40 from two
main sources. Firstly, performance data were collected from May 2015 to August 2017 from the Shenzhen and
Shanghai stock exchanges since the KGRM MSCI China IMI Environment 10/40 Index was not established within
the period. Data from September 2017 to April 2019 were, however, collected from the KGRM MSCI China IMI
Environment 10/40 Index. Since its launch, the KGRM MSCI China IMI Environment 10/40 Index has become a
benchmark for investors that identify with China's sustainability vision and wants to encourage environmentally
responsible business practices.27
We applied a strict data inclusion and exclusion criteria to eliminate outliers, clerical errors, empty and partially
recorded information as well as other glaring mistakes in the data. The final results yielded a total of 116 800 aver-
age daily returns of 80 environmental stocks for the period under consideration. The 80 stocks were further grouped
into the four sectors, namely power generation, materials, automobile and environmental solution per MSCI classifi-
cations. We studied the stocks in groups in order to quickly identify herding syndromes in a homogeneous trade
where the members face the same decision problem.
The daily return of the securities was computed based on the equally weighted portfolio for the 80 stocks, and
the summary statistics is presented in Table 1.
Table 1 also shows the proportion of both domestic and foreign investors in the four sectors. This helps to verify
the claim that investor groups (where trade is sufficiently homogeneous) are more susceptible to herd than individual
investors. Generally, Table 1 shows persistent growth in the number of foreign investors in the four sectors from
2015 to 2019.
Table 2 shows the summary statistics for the log of the average daily return, the average number of companies
and return dispersion used in computing the individual statistics for the sectors. Panel A shows that the average daily
returns are positive for the four sectors. Panel B, on the other hand, shows the summary statistics for daily cross-
sectional SDs within each sector. Both Panel A and Panel B show that the volatility cross-section in the automobiles
sector is relatively high.
 10991751, 2020, 6, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/hpm.3012 by University of Ghana - Accra, Wiley Online Library on [06/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
1402 ZHOU ET AL.
TABLE 1 Percentage of share holders in the market by sector (%)
Power generation (%) Materials (%) Automobile (%) Environmental solution (%)
Domestic Domestic Foreign Domestic Domestic Foreign Domestic Domestic Foreign Domestic Domestic Foreign
Year individual institution investor individual institution investor individual institution investor individual institution investor
2015 92.8452 3.1066 3.0482 93.7272 1.7056 4.1872 92.4242 0.392 6.9638 91.4242 0.392 7.2638
2016 92.7472 0.784 6.3986 90.7676 1.539 6.7232 83.1522 0.5096 14.308 80.2522 0.5096 18.308
2017 88.6312 0.8526 9.5162 88.3372 1.6272 9.9572 80.9484 0.4606 18.231 79.9484 0.4606 18.631
2018 87.4552 0.8918 10.653 89.0134 1.3234 9.2232 85.9366 0.735 12.3284 84.9366 0.735 13.3284
2019 83.7116 0.8918 14.3966 73.6902 1.882 24.4278 81.8892 0.5292 16.5914 80.8892 0.5292 17.5914
 10991751, 2020, 6, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/hpm.3012 by University of Ghana - Accra, Wiley Online Library on [06/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
ZHOU ET AL. 1403
TABLE 2 Summary of the statistics
Sector # Firms # Obs. Mean SD
for the average daily returns and the
cross-sectional SDs Panel A: Average daily return
Power generation 16 23 360 0.02% 2.41%
Materials 32 46 720 0.02548 2.3559
Automobiles 17 24 820 0.02744 2.7901
Environmental solutions 15 21 900 0.01372 2.6185
Panel B: Cross-sectional standard deviation
Power generation 1.53% 0.89%
Materials 1.94922 0.6909
Automobiles 1.957 0.7105
Environmental solutions 2.07662 0.5968
2.2 | Analytical procedure
Four different but related econometric models were employed to test the presence of investor's herding behaviour
among the 80 environmental stocks on the Shenzhen, Shanghai and the Hong Kong Stock Exchanges. These were
two return dispersion models (CSSD and CSAD) and two state spaced models. The first return dispersion models
namely cross-sectional SD (CSSD) measures the presence of herding behaviour as follows:
uvuffiPffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiu Nt ðr −r Þ2i,t p,t
CSSDt= i=1 ð1Þ
N−1
where n denotes the number of companies in the aggregate market portfolio, whereas ri,t denotes the stock return
observed on the company i for day t and rp,t is the cross-sectional average of the n returns in the market portfolio for
day t. This measure also denotes the individual security returns dispersion close to the market average. The key idea
in this model is that herd behaviour prevents the return of securities from deviating away from the overall market
return since it is assumed that individuals invest largely based on the collective actions in the market. Another advan-
tage of this approach is that it assumes that herding behaviour may largely occur during extreme market movements
as investors will usually opt for market consensus during such period. The return dispersion measure in Equation (1)
during market stress is estimated using a regression equation
CSSDt = α+ βDD
L
t + β D
U
U t + εt ð2Þ
where DLt = 1, if the aggregate market portfolio return on day t lies in the lower tail of the return distribution; 0 other-
wise, and DUt = 1, if the aggregate market portfolio return on day t lies in the upper tail of the return distribution;
0 otherwise. Since the formation of herd behaviour means imitating the market consensus, obtaining negative and
statistically significant βD (for down markets) and βU (for up markets) coefficients implies the presence of herding
behaviour among the market participants. This method has been used successfully in earlier studies by Huang,
HuiYing et al,28 Ji and Zhang14 and returned good results. The second return dispersion model used for this study is
proposed by Chang et al.24 This model uses the cross-sectional absolute deviation of returns (CSAD) as a proxy to
measure return dispersion (CSAD). This idea is expressed mathematically as
1XN
CSADt = j rN i,t−rm,t j ð3Þ
i=1
 10991751, 2020, 6, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/hpm.3012 by University of Ghana - Accra, Wiley Online Library on [06/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
1404 ZHOU ET AL.
This model disagrees with the assumption of the capital asset pricing model (CAPM). Instead, it
believes that return dispersion and market returns have a linear relationship and that is expressed mathe-
matically as
CSADt = α+ γ1jrm:tj+ εt ð4Þ
The argument of Chang et al24 who is the originator of this model is that market recession leads to the formation
of a non-linear relationship between market return and return dispersion instead of the linearity assumption pro-
posed in the CAPM. To date, this model has been applied to different studies with a robust outcome. For example,
Akerlof and Shiller29 successfully applied this model to detect herding behaviour robustness of reference in different
industries and markets. This study adopts but modifies the quadratic form of Chang et al' s26 general quadratic rela-
tionship between CSADt and rm,t as follows:
CSADt = α+ γ1jrm:tj+ γ2r2m:t + εt ð5Þ
This model measures herd behaviour by looking at the proportional increases in the cross-sectional absolute
deviation (CSAD) during extreme market movements. Herd formation will show in this case with a negative and sta-
tistically significant non-linear coefficient γ2 and vice versa.
The second model used to measure herding behaviour among environmental stocks on the Chinese stock mar-
kets are the state-space models. Consistent with earlier works of Hwang and Salmon,22 the measurement of herd
behaviour in state-space models has based on cross-sectional variability of factor sensitivities instead of returns. In
the first case as proposed by Hwang and Salmon 2004, the relative dispersion of the betas of the securities in the
market is estimated as follows
ErðritÞ= βimtEtðrmtÞ ð6Þ
where rit and rmt are the excess returns on asset i and the market at the time t, respectively, βimt is the systematic risk
measure, and Et() is the conditional expectation at the time t. In equilibrium, only β is needed to price an asset i. The
fundamental idea in this third model is that investors do not seek equilibrium relationship of Equation (5) in a herd
behaviour condition but match their assets with the market return and this gives a bias β term and a commensurable
expected rate of return. This is measured as follows
Ebt ðritÞ b
ð Þ = βimt = βimt−hmtðβE imt−1Þ ð7Þt rmt
where Ebt ðritÞ and βbimt are the market's biased short-run conditional expectation on the excess returns of asset i and
its beta at the time t, and hmt is a latent herding parameter that changes over time, hmt≤1, and conditional on market
fundamentals. As noted by Hou & Li,30 herding behaviour is present when (0 < hmt<1) then some degree of herding
exist on the market, but the hmt defines its intensity. For the total asset, a measure of herding behaviour is estimating
using the SD of βbimt which is formulated as follows
  rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Std bc βimt = Ec ðβimt−h 2mtðβimt−1Þ−1Þ = Ec ðβimt−1Þ2 ð1−hmtÞ= StdcðβimtÞð1−hmtÞ ð8Þ
  where EC( ) represents the cross-sectional expectation, Std βbc imt is the time-varying interval in response to herding
behaviour in the market. In its log form the equation is expressed as follows
 10991751, 2020, 6, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/hpm.3012 by University of Ghana - Accra, Wiley Online Library on [06/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
ZHOU ET AL. 1405
  
log Std βbc imt = μm + υmt ð9Þ
 
where μm = E[log[Stdc(βimt)]] and υ ii~mt d 0,σ2mυ the first state-space model is expressed mathematically as
  
log Std βbc imt = μm +Hmt + υmt, ð10Þ
 
Hmt =ϕmHmt−1 + ηmt, where Hmt = logð1−hmtÞ and ηmt  iid 0,σ2mη ð11Þ
Regarding the third model, the focus is basically on the dynamism in the movement patterns in the latent state
variable, Hmt, the state-space model equation. Herding behaviour is deemed to exist in the market when σ2mη =0 and
significant, which implies that Hmt = 0 for all t. A unique aspect of this model for measuring herding behaviour is that
herding behaviour Hmt is viewed as a dynamic process that evolves. When the ϕ is also significant, it means the
structure of herding behaviour is autoregressive.
The fourth model to estimate the presence of herding behaviour in the environmental stock market in China and
which is an enhanced or augmented state-space model seeks to increase the first state-space model to ensure the
robustness of inference. This is done by simply adding market volatilitylogσmt to Equation (8) and return, rmt, as an
independent variable to create an augmented state-space model expressed mathematically as;
  
log Std bc βimt = μm +Hmt + cm1logσmt + cm2rmt + υmt: ð12Þ
The interpretation of this model follows consistently with the interpretation of the first state model.
3 | RESULTS
In Equation (2), the Newey-West heteroskedasticity and autocorrelation consistent standard errors were used to
estimate the CSSD (herding behaviour) among environmental stocks in China due to the significant variation in dis-
persion and strong correlation. The MSCI China Environment Index represents the market while the lower and upper
one and five percentile of the market returns represents market stress. The results of Equation (2) presented in
Table 3 show that the (βD) values are positive and statistically significant and this validates the capital asset pricing
model that equity return dispersion is high when prices are rising.
TABLE 3 Coefficient of regression for CSSD = α+ β DL Ut D t + βUDt + εt
Market return in the extreme upper/ Market return in the extreme upper/lower
Return Dispersions lower 1% of the return distribution 5% of the return distribution
Sectors α βD βU α βD βU
Power generation 1.45% 0.14% 0.194%** 1.33% 0.454%*** 0.169%***
R2 (0.442) (0.405) (0.357) (0.456)
Materials 1.819 0.699*** 0.442*** 1.6895 0.541*** 0.451***
R2 (0.471) (0.490) (0.577) (0.327)
Automobile 1.758 0.167 0.026 1.7218 0.497*** 0.160***
R2 (0.519) (0.514) (0.457) (0.404)
Environmental Solutions 2.073 0.146 0.149* 2.048 0.449*** 0.142***
R2 (0.493) (0.414) (0.457) (0.470)
 10991751, 2020, 6, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/hpm.3012 by University of Ghana - Accra, Wiley Online Library on [06/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
1406 ZHOU ET AL.
Table 4 presents the estimate of herding behaviour based on the alternative CSAD non-linear Equation (4). Con-
sistent with the earlier suggestions by Chang et al (2000), three different regression models were run for each sector.
The first regression model used the entire sample, whereas the remaining two regressions were applied only to the
restricted data on the fluctuations in the market index (up and down markets streams). Using different models help
to detect an asymmetric effect in investor's herding behaviour. The higher R2 values of the non-linear CSAD model
show that it fits better than the CSSD model; hence, its estimates are more credible. Thus, the outcome in Table 4 is
more reasonable than those in Table 3 since the proportion of variation in the herding attributed to the independent
variables is higher. Since the first non-linear term (γ2) in the CSAD model is statistically significant, that estimate is
used to infer herding behaviour, and it prevails in the four sectors analysed.
The γ2 in Table 4 is negative and statistically significant, and that suggests a non-linear and declining relationship
between market return and equity return dispersion. The regression model applied to the restricted data (up and
down markets streams) shows that investors herd during market losses for all the sectors. Table 5, on the other hand,
presents the result of herding behaviour using the first state-space model. Similar to the CSAD, the first state-space
model shows a high persistent, Hmt which is strong and statistically significant, and that confirms the presence of
investor herding behaviour. The ϕ̂m estimates are highly persistent improve, and the σmη values are large and statisti-
cally significant. This further confirms the presence of herding behaviour among environmental stocks across the
sectors. The results from the second state-space model in Table 6 also affirm the robustness of inference of herd
behaviour noted among the environmental stocks using the CSAD and the first state-space model. In the second
model, two additional market variables (market return and market volatility) are added to better estimate the degree
of herding, given the state of the market. The results show that the inclusion of market returns and market volatility
variables does not affect the statistical significance of Hmt. Secondly, changes in the volatility of factor sensitivities or
dispersion in the betas for the 80 environmental stocks denoted by Std βbc imt could be attributed to investors'
herding behaviour, and not just changes in market fundamentals.
Table 6, on another hand, shows that the coefficients for the market vo latility term (logσmt) are positive and sta-
tistically significant. This means that the volatility of factor sensitivitiesStd βbc imt , in all the four sectors, grows in the
same direction with market volatility. That observation affirms previous studies such as References 9,12,31,33 and
34 that investors tend to herd during highly volatile market conditions. This notwithstanding, the results in Table 6
show that the dispersion of the betas Stdc β
b
imt , for all stocks in the materials and environmental solution sectors, is
significantly higher than in the other sectors. This is not surprising because in Table 1, the materials and environmen-
tal solutions sectors have the highest numbers of foreign investors and this may support previous findings34,35 that
foreign investors tend to herd more than domestic investors. Another observation is the relatively high value for ϕ̂m
the materials and environmental solutions sectors. Besides, having more foreign investors than the others, the com-
bined percentage of stocks within these sectors is approximately 59%. This high volume and value of trade make the
two sectors highly susceptible to investor's pro-cyclical behaviour.
4 | DISCUSSION OF FINDINGS
The objective of this research is to examine whether the environmental stock market investors in China exhibit this
kind of behaviour and to what extent. We are inspired by the view that the stability of the environmental stock mar-
ket in China is critical to the attainment of the overall sustainability objectives of the world's largest producer of car-
bon emissions and for that matter its environmental health security. The behaviour of investors in the environmental
stock market can potentially affect China's environmental health security goals. Herding behaviour, which is a con-
temporaneous loss of investor confidence and movement from environmental stocks to perceived safer stocks by
investors, is a major challenge in the realisation of China's environmental health goals. This study applied an
 10991751, 2020, 6, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/hpm.3012 by University of Ghana - Accra, Wiley Online Library on [06/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
ZHOU ET AL. 1407
TABLE 4 Coefficient of regression for CSADt = α+ γ1jr 2m:tj+ γ2rm:t + εt
Absolute deviation Whole sample Down market (Rm < 0) Up market (Rm > 0)
Sector α γ1 γ2 α γ1 γ2 α γ1 γ2
Power generation 1.016 0.162*** −0.017*** 0.981 0.117*** −0.029*** 1.02508 0.095** −0.003
R2 (0.593) (0.732) (0.605) (0.467) (0.673) (0.645)
Materials 1.196 0.193*** −0.013** 1.150 0.165*** −0.024*** 1.09858 0.117*** 0.0004
R2 (0.791) (0.582) (0.701) (0.504) (0.714) (0.544)
Automobile 1.111 0.404*** −0.044*** 1.153 0.371*** −0.056*** 1.12308 0.121*** −0.029***
R2 (0.671) (0.638) (0.604) (0.613) (0.573) (0.727)
Environmental Solutions 1.214 0.129*** −0.040*** 1.327 0.171*** −0.049*** 1.39258 0.169*** −0.017***
R2 (0562) (0.561) (0.619) (0.654) (0.589) (0.789)
 10991751, 2020, 6, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/hpm.3012 by University of Ghana - Accra, Wiley Online Library on [06/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
1408 ZHOU ET AL.
TABLE5 State space model 1 forIndustry μ ϕm σmυ σmη
log Stdc β
b
imt = μm +Hmt + υmt and
Power −0.675*** 0.940*** 0.075*** 0.099***
Hmt = ϕmHmt−1 + ηmt
generation (0.667) (0.620) (0.657) (0.654)
Materials −0.645*** 0.940*** 0.019*** 0.091***
(0.675) (0.604) (0.574) (0.552)
Automobile −0.570*** 0.945*** 0.047*** 0.091***
(0.496) 90.561) (0.752) (0.520)
Environmental −0.647*** 0.957*** 0.040*** 0.067***
Solutions (0.496) (0.561) (0.752) (0.520)
  
TABLE 6 State-space model 2 log Std βbc imt = μm +Hmt + cm1logσmt + cm2rmt + υmt and Hmt = ϕmHmt−1 + ηmt
Industry μ ϕm σmυ σmη logσm rm
Power generation 0.451*** 0.941*** 0.075*** 0.097*** 0.659*** −0.005
R2 (0.704) (0.070) (0.904) (0.727) (0.754) (0.640)
Materials 0.879*** 0.971*** 0.019*** 0.091*** 0.672*** 0.005
R2 (0.607) (0.575) (0.676) (0.775) (0.560) (0.706)
Automobile −0.747*** 0.944*** 0.047*** 0.091*** −0.151*** 0.046
R2 (0.572) (0.741) (0.891) (0.795) (0.667) (0.510)
Environmental Solutions 0.897*** 0.977*** 0.040*** 0.067*** 0.776*** 0.027
R2 (0.665) (0.721) (0.584) (0.650) (0.614) (0.704)
ensemble of more sophisticated econometric models to firm-level data of 80 environmental stocks classified into
four main sectors by the KGRM MSCI China IMI Environment 10/40 Index.
Apart from the cross-sectional SD model, the remaining three robust models namely the cross-section absolute
deviation, state-space model and augmented state-space model yield evidence of endemic herding tendencies
among the environmental stocks on the Chinese stock market. The literature presented earlier shows an enormous
amount of studies that warn of the dire consequence of persistence herding behaviour on any stock market. For
example, the destabilising effect of investor herding behaviour on stock price synchronicity as validated in Reference
18 and that potentially stimulate sharp fluctuation in environmental stock prices en masse. The impact of such
abrupt changes is loss of investor confidence, disinterest and diversion of capital to safer securities.29 The resulting
investor disinterest deprives the sector of the needed capital and this can stagnate the sector and that may result in
substantial negative ramification for other “sustainability dependent” sectors, including food security and public
health safety. While China has achieved high economic growth in the last two decades, and has successfully lifted
many of its citizens out of poverty, a quarter of its population has access to substandard tap water, 122 million Chi-
nese still lives below the international poverty line and over 100 million have low-level social security and this can be
compounded by weak environmental market. A decade ago, Akerlof and Shiller29 warned about ramifying effect of
this impending danger as a result of economic behaviour of human beings similar to animalistic instincts. Secondly, a
weak private sector support for China's ecological civilization and environmental health goals due to investor disin-
terest may force China to divert resources intended for other critical social sectors to meet the sustainability objec-
tives in response to excessive global pressure on China. There is also the danger of developing an environmentally
responsible society that is overly controlled by the state as it may result in higher environmental taxes and charges
with paradoxical consequences.30 In the BYD case recounted in earlier sections of the paper, government must
 10991751, 2020, 6, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/hpm.3012 by University of Ghana - Accra, Wiley Online Library on [06/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
ZHOU ET AL. 1409
desist from sudden and seemingly bais announcement of strict regulations that tends to create doubt in the minds of
foreign investors. The state must find alternative ways to provide support for local industries to compete at par with
international ones instead of creating artificial conditions for them to excel. This may avert a potential and an
unprecedented crash in the environmental stock market if herding behaviour persists among investors. Moreover,
China's 13th 5-year Plan lacks precision on public–private consensus building and investor sensitisation to avert any
panic response to changes in market conditions.20 As domestic investors are less likely to herd, equipping them to
reduce better exposure to fluctuations is needed to support the sustainable goals. Further, China must use its experi-
ence and leadership in sustainable development policies to highlight its readiness to engage the global investor com-
munity and reduce the unintended opaqueness and uncertainty that usually hover around media portrayals of
China's investment terrain.31 This may help foreign investors to confidently explore all the available opportunities
the environmental stock market offers them to the hilt.
5 | CONCLUSIONS
The destructive health effect of global warming and other environmental damages is well documented at the beginning
of this paper. For example, drywalls that are cheaply produced in China have the potential to emit high levels of sulphur
gas that corrodes electrical wiring and also causes breathing problems, bloody noses and headaches for building occu-
pants. A recent study conducted by Scientific American found out that household contaminants such as phthalates
which is a hazardous chemical emitted by vinyl floors, cause autism in kids. Similarly, women exposed to high levels of
polybrominated diphenyl ether (PBDE) flame retardants common in cushions, carpet padding and mattresses have a
97% chance of acquiring poisonous chemicals in their bloodstreams and can avert or abort the pregnancy or suffer
from other fertility issues, as a result. These health impacts of poor environmental control cannot be disregarded.
Significantly, China's effort at meeting its environmental health goals through sustainable development occupied
several pages of the inaugural address of President Xi Jinping at the 18th National Congress of the Communist Party
of China, July 2018. This “beautiful China” dream that outlines China's ecological civilisation has become a global
blueprint document not only for China but also its nearest neighbours and friends. However, the availability of capi-
tal and the confidence of investors are critical to the attainment of this objective. The extent to which herding or
procyclical behaviour exists on the Chinese environmental stock market was examined in this paper. The results con-
firm the presence of endemic herding behaviour on the environmental stock market, and this can threaten the stabil-
ity of the environmental health security once it depends on funds from environmental stock investments. We note
that investors mostly herd during a period of market stress, and this is more prevalent among foreign and group of
investors. We have proposed measures to improve the risk posed by such behaviours. Further research is needed in
other countries, especially among the BRICS using the same models to validate the generalizability of this observed
herding behaviour among environmental stock investors.
ACKNOWLEDGEMENTS
The authors are indebted to the National Natural Science Foundation of China (Grants nos. 71974079, 71904066)
for supporting this project. The Post-doctoral Foundation of China (2018T60508, 2018M511705 and 2018T60508)
and the Post-doctoral Foundation of Government of Jiangsu Province number 1102129C also supported this
project.
CONFLICT OF INTEREST
The authors declare no potential conflict of interest.
ETHICS STATEMENT
This article does not require any human/animal subjects to acquire such approval.
 10991751, 2020, 6, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/hpm.3012 by University of Ghana - Accra, Wiley Online Library on [06/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
1410 ZHOU ET AL.
ORCID
Henry A. Antwi https://orcid.org/0000-0002-6595-2964
REFERENCES
1. Shen K, Li L, Liu J, Chen C, Liu J. Stocks, flows and emissions of DBDPE in China and its international distribution
through products and waste. Environ Pollut. 2019;250(3):79-86.
2. Zhu M, Feng Q, Zhang M, et al. Effects of topography on soil organic carbon stocks in grasslands of a semiarid alpine
region, northwestern China. J Soils Sediments. 2019;19(4):1640-1650.
3. Wang S, Guo Z. A study on the co-movement and influencing factors of stock markets between China and the other
G20 members. Int J Finance Econ. 2019;6(8):178-191.
4. Zhao Z, Zhang X, Dong S, et al. Soil organic carbon and total nitrogen stocks in alpine ecosystems of Altun Mountain
National Nature Reserve in dry China. Environ Monit Assess. 2019;191(1):40.
5. Song L, Zhang C, Han J, Chen WQ. In-use product and steel stocks sustaining the urbanization of Xiamen, China.
Ecosyst Health Sustainability. 2019;5(1):110-123.
6. Li C, Hu Z, Tang L. Re-examining the Chinese A-share herding behaviour with a Fama-French augmented seven-factor
model. Appl Econ. 2019;51(5):488-508.
7. Hong S, Gan P, Chen A. Environmental controls on soil pH in the planted forest and its response to nitrogen deposition.
Environ Res. 2019;172:159-165.
8. Yao J, Ma C, He WP. Investor herding behaviour of the Chinese stock market. Int Rev Econ Finance. 2014;29:12-29.
9. Yu L, Fung HG, Leung WK. Momentum or contrarian trading strategy: which one works better in the Chinese stock
market. Int Rev Econ Finance. 2019;62(9):87-105.
10. Thompson WC. When politics, the environment, and advocacy compete-environmental security in the South China
Sea. The Environment-Conflict Nexus. Cham: Springer; 2019:103-118.
11. Syriopoulos T, Bakos G. Investor herding behaviour in globally listed shipping stocks. Marit Policy Manage. 2019;46(5):
545-564.
12. Banga J. The green bond market: a potential source of climate finance for developing countries. J Sustainable Finance
Invest. 2019;9(1):17-32.
13. Baulkaran V. Stock market reaction to green bond issuance. J Asset Manage. 2019;1(1):95-128.
14. Ji Q, Zhang D. How much does financial development contribute to renewable energy growth and upgrading of energy
structure in China. Energy Policy. 2019;128:114-124.
15. Broadstock DC, Cheng LT. Time-varying relation between black and green bond price benchmarks: macroeconomic
determinants for the first decade. Finance Res Lett. 2019;29:17-22.
16. Cai W, Lee E, Xu AL, Zeng CC. Does corporate social responsibility disclosure reduce the information disadvantage of
foreign investors? J Int Accounting, Auditing, Taxation. 2019;34:12-29.
17. Christie WG, Huang RD. Following the pied piper: do individual returns herd around the market? Financ Anal J. 1995;
51:31-37.
18. Chang EC, Cheng JW, Khorana A. An examination of herd behaviour in equity markets: an international perspective.
J Bank Finance. 2000;24:1651-1679.
19. Demirer R, Kutan AM. Does herding behaviour exist in Chinese stock markets? J Int Financ Mark Inst Money. 2006;16:
123-142.
20. Demirer R, Kutan AM, Zhang H. Do ADR investors herd?: evidence from advanced and emerging markets. Int Rev Econ
Finance. 2014;30:138-148.
21. Uchida H, Nakagawa R. Herd behaviour in the Japanese loan market: evidence from bank panel data. J Financ Intermed.
2007;16:555-583.
22. Lulin Z, Antwi HA, Wang W, Yiranbon E, Marfo EO, Acheampong P. The effect of herd formation among healthcare
investors on health sector growth in China. Int J Equity Health. 2016;15(1):113.
23. Hwang S, Salmon M. Market stress and herding. J Empir Finance. 2004;11:616.
24. Chang Y. Green finance in singapore: barriers and solutions. Handbook of Green Finance: Energy Security and Sustainable
Development. Singapore: Springer; 2019.
25. Flammer C. Green bonds: effectiveness and implications for public policy. Environmental and Energy Policy and the Econ-
omy. 2020;1(1):95-128.
26. Fonseka M, Rajapakse T, Richardson G. The effect of environmental information disclosure and energy product type on
the cost of debt: evidence from energy firms in China. Pacific-Basin Finance J. 2019;54:159-182.
27. Hou X, Li S. The price of official-business cooperation: Evidence from the stock market reaction to “Hunting the Tigers”
in China. China Finance Rev Int. 2019;10:52-74.
 10991751, 2020, 6, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/hpm.3012 by University of Ghana - Accra, Wiley Online Library on [06/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
ZHOU ET AL. 1411
28. Hui-Ying L, Qiang Y, Sharma G. Herding behaviour in C2C e-commerce: empirical investigation in China. Paper pres-
ented at: Management Science and Engineering (ICMSE), 2010 International Conference on. 2010. pp. 33–9.
29. Akerlof, G. A, Shiller. R. Animal Spirits: how Human Psychology Drives the Economy, and why it Matters for Global Capital-
ism. Princeton, New Jersey, USA: Princeton University Press; 2009.
30. Hou X, Li S. The price of official-business collusion. China Finance Review International. 2019;10(1):52-74. http://dx.doi.
org/10.1108/cfri-07-2018-0114.
31. Xing K, Yang X. How to detect crashes before they burst: evidence from the Chinese stock market. Physica A. 2019;
121:392.
32. Alhaj-Yaseen YS, Rao X. Does Asymmetric Information Drive Herding? An Empirical Analysis. Journal of Behavioral
Finance. 2019;20(4):451-470.
33. Indars ER, Savin A, Lublóy A . Herding behaviour in an emerging market: Evidence from the Moscow Exchange. Emerg-
ing Markets Review. 2019;38:468-487.
34. Zhu L, Li H, Zheng D. Institutional industry herding in China. The Chinese Economy. 2020;53(3):246-264.
35. Qasim M, Hussain E, Mehboob I, Arshad M. Impact of herding behavior and overconfidence bias on investors’ decision-
making in Pakistan. Accounting. 2019;5(2):81-90.
How to cite this article: Zhou L, Antwi MO, Antwi HA, Boafo-Arthur A, Mustafa T. Endangering China's
environmental health security goals through negative environmental investor behaviours. Int J Health Plann
Mgmt. 2020;35:1398–1411. https://doi.org/10.1002/hpm.3012
 10991751, 2020, 6, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/hpm.3012 by University of Ghana - Accra, Wiley Online Library on [06/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License