Energy Economics 133 (2024) 107516

Contents lists available at ScienceDirect

Energy Economics
journal homepage: www.elsevier.com/locate/eneeco

ESG investment and bank efficiency: Evidence from China

Qiang Cao , Tingting Zhu , Wenmei Yu *
School of Finance, Anhui University of Finance and Economics, Bengbu 233030, China

A R T I C L E I N F O A B S T R A C T

JCL code: Although environmental, social, and governance (ESG) investment is essential to achieving sustainable devel­
D20 opment, previous studies usually focused on the relation between ESG investment and non-financial firms but
G21 neglected its impact on bank efficiency. To fill this research gap, we use a stochastic frontier analysis (SFA) model
P34
to investigate the influence of ESG investment on banks' profit efficiency and to explore whether financial
Keywords: technology (fintech) can strengthen the relationship between them. Results show that increasing ESG investment
ESG
is beneficial to bank efficiency. We also find structural differences in the influence of ESG investment on bank
Bank efficiency
Fintech
efficiency. Specifically, environmental (E) and governance (G) investments enhance bank efficiency, while social
(S) variables reduce bank efficiency. In addition, fintech has a moderating effect. When the level of fintech is
high, ESG investment rises, and the increase in bank efficiency is more significant. Finally, by conducting a time-
varying analysis of bank efficiency, we find that banks with higher ESG investments are more efficient. Under the
goal of sustainable development, strengthening ESG investment and focusing on the empowering role of fintech
are essential ways for commercial banks to increase efficiency.

1. Introduction Tokyo Stock Exchange also issued the Japan Corporate Governance Code
incorporating sustainability issues and ESG element considerations into
As climate change, geopolitics, and policy uncertainty have become board responsibilities. In 2019, the National Association of Securities
important disturbances to global financial markets, investors and Dealers Automated Quotation released the ESG Reporting Guide 2.0,
financial institutions have been increasingly concerned about sustain­ which provided disclosure requirements on ESG matters.
ability (Morea et al., 2022; Iazzolino et al., 2023). Environmental, social, Driven by the notion of sustainable development, ESG investment
and governance (ESG) investment is an important way for companies to has also received great attention in China (Ji et al., 2023). The Guidelines
achieve sustainable development (Veltri et al., 2023). on Green Credit, published in 2012 by the China Banking Regulatory
The term ESG was first introduced by the United Nations (UN) in Commission (CBRC), stipulated that commercial banks should engage in
2004, which set the harmonious and sustainable development of the green credit. In 2020, the CBRC issued the Guidance on Promoting the
economy, society, and the environment in an integrated manner as its High-Quality Development of the Banking and Insurance Industries, which
core concept. In 2006, the UN-supported Principles for Responsible In­ explicitly required banks to strengthen ESG information disclosure.
vestment proposed a disclosure conceptual system for ESG investing, Good ESG investment performance helps improve a bank's reputation,
which has later become a critical catalyst for the development of the build its brand image, gain more stakeholder support, and achieve long-
field. In 2015, the UN proposed 17 Sustainable Development Goals in term bank growth.
the UN 2030 Agenda for Sustainable Development. The Paris Agreement in However, since China's ESG policy guidelines have not yet formed a
the same year also recognized the need to protect the environment. unified and effective information disclosure standard, rating agencies
In recent years, countries across the world have responded positively and data service providers have difficulties in obtaining valuable
to the UN's call. In 2010, the U.S. required listed companies to disclose corporate information, which increases the risk of commercial banks in
environmental information, such as climate change (Commission, making ESG investments. Specifically, as the policy on ESG disclosure is
2010). In 2014, the European Union issued the Non-financial Reporting only opinion-oriented and not legally binding, enterprises can choose to
Directive, which incorporated ESG into policies and regulations for the disclose relevant information according to their own development
first time. One year later, Japan's Financial Services Agency and the needs, which makes the disclosed information lack consistency.

* Corresponding author.
E-mail addresses: [email protected] (Q. Cao), [email protected] (T. Zhu), [email protected] (W. Yu).

https://doi.org/10.1016/j.eneco.2024.107516
Received 7 June 2023; Received in revised form 26 August 2023; Accepted 25 March 2024
Available online 3 April 2024
0140-9883/© 2024 Elsevier B.V. All rights reserved.
Q. Cao et al. Energy Economics 133 (2024) 107516

Furthermore, the disclosure guidelines do not specify the ESG indicators Chinese commercial banking system, we investigate the impact of ESG
or parameters to be revealed, so companies often disclose information implementation on the efficiency of different types of banks. Finally,
on criteria in their favor while seldom revealing negative indicators. Due considering the increased adoption of fintech in the banking sector, we
to the lack of micro-level policies on ESG indicators and systems, some further include fintech as a moderating variable and explore its effects
enterprises also focus their disclosures on ESG-related management on the relationship between ESG investment and bank efficiency.
policies but leave little room for implementation methods, specific ini­ The remaining sections of the paper are organized as follows. Section
tiatives, and implementation effects. 2 provides a literature review and research hypotheses. Section 3 deals
Despite these challenges, China has been on a path to strong ESG with the econometric model setting and variable descriptions. Section 4
development. Credits that are lent out in the form of green loans are contains the empirical results and analysis. Section 5 provides further
generally considered to be of higher quality and more beneficial to the research, and Section 6 contains findings and policy recommendations.
growth of bank efficiency (Cui et al., 2018b). In addition, as banks are
accelerators of the financial cycle (Altunbaş et al., 2016), research has 2. Literature review and research hypothesis
pointed out that ESG investments could reduce the climate risk facing
the bank industry (Galletta and Mazzu, 2023) and is thus conducive to 2.1. The relationship between ESG investment and bank efficiency
bank stability as well as keep systemic risks inherent in Chinese financial
industry under control. In this context, it makes sense to explore what Early research on ESG concentrated on the impact of ESG on tradi­
impact ESG investment has on bank efficiency in the Chinese banking tional enterprise performances. For instance, Cui et al. (2018a) found
sector. that actively fulfilling corporate social responsibility (CSR) could help
Besides, we note that the unique regulatory environment in China enterprises access core resources. Xie et al. (2019) indicated that en­
has led to significant differences across banks (Lee et al., 2015; Lee and terprises improving the level of ESG information disclosure could help
Lee, 2019). In the same vein, we assume that it may result in great enhance enterprise efficiency. As financial institutions such as com­
differences in bank efficiency when banks of various kinds make ESG mercial banks possess more and more liquidity and the call for their
investments. After all, in China, local banks tend to operate at the pro­ social responsibility grows in public, the research perspective has shifted
vincial level and their operations are thus limited in scale and influence. to ESG investment by banks.
In addition, there are also banks with varied ownerships. For example, Research has suggested that banks' ESG investments enhance effi­
state-owned banks, joint-stock banks, and city commercial banks are ciency through reputation spillovers, i.e., when banks make ESG in­
distinctly different in terms of the scale of operation and scope of vestments, they improve their social image and gain reputation, which
business. Therefore, in estimating the effect of ESG investments on bank in turn enhances their efficiency (Wu and Shen, 2013). Theoretically,
efficiency, the heterogeneity of the Chinese commercial banking system this impact can be achieved through two channels: external and internal.
must be taken into account. Externally, banks with a high reputation can improve their relationships
Since 2013, the Chinese banking sector has gradually increased the with external players, including regulators and investors. That is, banks
use of financial technology (fintech), which is reckoned as the applica­ with enhanced reputations are more likely to win the trust of regulators,
tion that reduces the information asymmetry between banks and firms thus reducing external regulatory pressure and improving bank effi­
and strengthens the positive impact of ESG on bank efficiency. Studies ciency. For instance, Branco and Rodrigues (2006) found that firms'
have shown that the widespread use of technologies has changed the engagement in social responsibility could help them gain access to some
traditional lending model of commercial banks (Lee et al., 2021), valuable and rare resources, such as reputation. A good reputation for
bringing new opportunities for commercial banks to make ESG in­ social responsibility can strengthen the ties with firms' external players
vestments. For example, banks could utilize big data to advance finan­ and help them negotiate with the government in a more favorable place.
cial anti-fraud and predict customer defaults. Banks can also use On the other hand, a bank with a high reputation can gain investors'
machine learning to make intelligent recommendations on financial trust and reduce financing costs. Wu and Shen (2013) showed that
products and achieve precision marketing. reputation was closely related to brand awareness. Banks that fulfill CSR
Considering that China's financial system is dominated by indirect could create a brand name and sense of identity among their clients,
financing, commercial banks have always occupied a pivotal position in which help to attract more deposits and loans, as well as generate higher
the Chinese economy. Improving bank efficiency not only enhances the non-interest income. Andrieș and Sprincean (2023) indicated that
level of competition and profitability of banks (Mateev et al., 2022), but companies with a long-term commitment to sustainability demonstrated
also contributes to the sustainability of banks. Therefore, it is worth greater transparency, thereby reducing information asymmetry on the
exploring whether ESG investments can contribute to bank efficiency part of investors. Kim et al. (2005) also noted that banks with high
and what role fintech plays between the two. quality and reputation provided their customers with a sense of high
So far, there has been a paucity of studies on the effect of ESG in­ creditworthiness. Customers are more willing to borrow from reputable
vestments on bank efficiency. Furthermore, when conducting research banks, even at higher interest rates. Agnese and Giacomini (2023)
on bank efficiency, the few existing studies often employed a two-step conducted a study of 63 banks in Europe and found that the higher the
approach (Schmidt, 2011; Afza and Asghar, 2017). However, the ESG score, the lower the cost of issuing bonds.
method is subject to model misspecification, which leads to a bias in its Internally, a bank with a higher reputation is able to attract better
estimation of both model parameters and technical efficiency, impairing prospective employees or increase the motivation and loyalty of current
the credibility of the findings of the relevant studies. Finally, few studies employees, which contributes to the efficiency of the bank. Branco and
have considered the specification of Chinese banks by property rights Rodrigues (2006) indicated that firms with a strong commitment to
and regionality. social responsibility tended to be more attractive to job applicants and
This study is to address the gap by investigating how ESG investment retain them once hired. This resulted in lower expenditure for turnover,
and its three subcomponents, namely, environmental (E), social (S), and hiring, and training. In addition, ESG investments require banks to
governance (G) investments, affect bank efficiency. In addition, we improve their corporate governance in terms of board diversity,
employ the heterogeneous stochastic frontier analysis (SFA) model of compensation management, committee independence, and sharehold­
(Battese and Coelli, 1995) in our estimation. This method adopts a one- ing structure. Strong corporate governance enhances a bank's financial
step estimation strategy to incorporate the influencing factors of effi­ performance and makes it function more efficiently. Esteban-Sanchez
ciency into the model setup, which can mitigate the bias of results that et al. (2017) examined how four aspects of corporate social performance
may be brought about by the traditional two-step method (Wang and affected financial performance and found that the corporate governance
Schmidt, 2002). Simultaneously, considering the heterogeneity of the dimension positively affected financial performance. They concluded

2
Q. Cao et al. Energy Economics 133 (2024) 107516

that effective corporate governance contributed to a trusting relation­ with the government, obtain the corresponding resource, strengthen
ship with stakeholders, reduced agency problems, and thus improved a their competitive advantages, and enhance bank efficiency. Given the
firm's operational and financial performance. Harjoto and Jo (2011) fact that governments around the world have introduced many policies
suggested that firms engaging in corporate governance reduced conflicts related to green financial development and actively incentivized com­
of interest between managers and non-investment stakeholders, thereby mercial banks to take on green responsibilities, the bank that
improving financial performance and firm value. By studying 89 com­ strengthens E investment significantly sends a positive signal to gov­
panies in the global electrical industry, Miras-Rodriguez et al. (2015) ernments and gains their support. Xun (2013) also pointed out that CSR
found that the corporate governance dimension of CSR positively could be used as a legitimate instrument to establish an excellent
affected ROA by exerting a good social impact. interactive and communicative relationship with governments as a way
However, another stream of the study points out that ESG investment to gain comparative advantages and improve company performance. In
by banks may ultimately lead to a decline in bank efficiency through the addition, Zhou et al. (2022) found that increasing the proportion of
resource substitution effect and the reputation insurance effect. green loans contributed to a decrease in the climate risk related to brown
Resource substitution theory suggests that ESG investments and other loans, thereby bringing down overall bank risk. At the same time, Zhou
business opportunities are competing for the same resources available, et al. (2022) also indicated that conducting financial businesses with a
leading to opportunity costs. Azmi et al. (2021)also pointed out that in large number of firms with high environmental standards could reduce
the case of excessive ESG investments, the bank's other businesses would the banks' financial risks and the likelihood of financial misconduct.
be substituted, which led to a reduction in the bank's operational per­ When banks carry out S investment, their market value decreases, as
formance and ultimately impaired the value of the bank. Di Tommaso S investment may increase the bank's non-interest expenses, resulting in
and Thornton (2020) found that ESG investments reduced the risk of the resource depletion, which hurts bank efficiency. For instance, by
bank but were insufficient to make up for its adverse impact on the value studying 51 banks across the globe, Mar Miralles-Quiros et al. (2019)
of the bank via resources appropriation. In addition, managers might be found that S investment negatively affected share price and suggested
engaging in socially responsible investments based on the initial inten­ that banks make S investment only to fulfill stakeholders' expectations,
tion of satisfying their interests. This led to decreased bank efficiency by which ultimately increased the firm's non-interest expenses. Fukuyama
engaging in risky behaviors through the acquired reputational insurance and Tan (2021) also found that actions such as endowments led to an
effect. Barnea and Rubin (2010) suggested that managers might over­ increase in bank non-interest expenses and impaired the bank's
invest in CSR for personal gain as this enhances their reputation as good efficiency.
global citizens. The above relationship can be detailed in Fig. 1. Based on When banks are engaged in G investment, they are able to optimize
this, we propose hypothesis 1. their governance structure, enhance transparency, reduce agency costs,
and improve bank efficiency. For instance, Soana (2011) found that
Hypothesis 1. ESG investments enhance bank efficiency through
banks' financial performance was positively impacted by effective
reputational spillovers and dampen bank efficiency through resource
governance. Esteban-Sanchez et al. (2017) indicated that good corpo­
substitution and reputational insurance effects.
rate governance reduced agency problems and maintained a better trust
relationship with stakeholders, which in turn had a beneficial effect on
2.2. The relationship between the subcomponents of ESG investment and the financial performance of the firm. Jamali et al. (2008) suggested that
bank efficiency compliance and transparency were the essential components of corpo­
rate governance that should be focused on and that disciplined corporate
When banks vigorously strengthen the E investment, for instance, via governance could improve financial performance. Hill and Jones (1992)
green loans, they seize policy opportunities, establish good relationships also found that good corporate governance could well balance the

Fig. 1. Mechanisms for the impact of ESG investment and its sub-components on bank efficiency.
(Source: Authors' elaboration.)

3
Q. Cao et al. Energy Economics 133 (2024) 107516

relationship between managers and stakeholders, reduce conflicts of investigates the efficiency by setting the inefficiency term.
interest, and lower agency costs. Based on this, we develop hypothesis 2. DEA is often used to calculate the efficiency of multiple inputs and
outputs, but it ignores the effect of random errors and thus leads to
Hypothesis 2. E and G investments are conducive to increased bank
biased results. This is especially the case when the units on the pro­
efficiency, while S investment decreases bank efficiency.
duction frontier are highly influenced by random factors, resulting in a
large efficiency error. In contrast, the SFA model considers random er­
2.3. Fintech, ESG investment and banking efficiency rors and improves the accuracy of the results. In addition, the SFA model
can effectively identify the efficiency influencing factors and analyze the
With the maturity of Internet finance, commercial banks can increase relationship between efficiency and influencing factors, which further
the impact of their ESG investments on profit efficiency with the help of enhances the practical implication of the study. Therefore, we choose
technology empowerment. Du et al. (2022) found that fintech could the SFA model.
break the resource dilemma and improve resource use efficiency by The SFA model widely used in previous studies is a two-step
empowering traditional financial institutions with technology. Specif­ approach, where efficiency is first calculated and then regressed on
ically, the introduction of fintech technologies to commercial banks the remaining variables using efficiency (Zamore et al., 2021). However,
effectively alleviates information discrepancies and information asym­ the two-step SFA model leads to biased estimates. This is because, in the
metries with enterprises. Zhan and Jing (2022) pointed out that fintech first step, the inefficiency term is usually assumed to be independently
helped external stakeholders collect and quantify soft information about and identically distributed, which is the only way to estimate the effi­
firms with its powerful data search and analysis capabilities. Huang ciency value using the approach (Jondrow et al., 1982). However, in the
(2022) also found that fintech reduced the degree of information second step, the efficiency is set as a function of variables such as ESG,
asymmetry between banks and firms and improved banks' credit allo­ which indicates that the efficiency is not independently and identically
cation efficiency. In addition, traditional ESG reports only update once a distributed, contradicting the assumption made in the first step (Greene,
year, but AI-driven analyses instantly reflect a firm's ESG performance. 2005b). Therefore, the two-step approach is fundamentally flawed. In
Nost and Colven (2022) confirmed that instantaneous ESG analyses contrast, the one-step approach can effectively circumvent this inherent
solved the problem of time lag. Zhang et al. (2022) indicated that fin­ contradiction by incorporating the influencing factors of efficiency
tech, through cloud computing, artificial intelligence and other tech­ directly into the modeling setup. The Monte Carlo simulation analyses of
nologies, effectively extracted valuable immediate information and (Wang and Schmidt, 2002) also show that the one-step approach is more
provided decision makers with multi-dimensional data references before reliable. Among the specific one-step estimation methods, one of the
signing loan contracts. Finally, fintech is able to identify greenwash more widely used is the heterogeneous SFA model developed by (Battese
behavior. Seele and Schultz (2022) pointed out that if a company is and Coelli, 1995), which is able to overcome the shortcomings of the
unable to query, analyze and disclose its real-time ESG data, then the two-stage estimation method previously used in studying bank effi­
company's efforts in ESG could easily turn to greenwash. Vergara and ciency (Agostino et al., 2023). Therefore, a one-step heterogeneous SFA
Agudo (2021) suggested that the full application of big data and AI model is used in this paper.
technologies to ESG investments increased banks' willingness to invest The SFA model consists of frontier and inefficiency analysis models.
in ESG. Based on this, we propose hypothesis 3. It is essential to set the functional form of the frontier model. Common
Hypothesis 3. Fintech further enhances the relation between ESG functional forms are the Cobb-Douglas production function and the
investments and commercial banks' profit efficiency. transcendental logarithmic production function. Although the Cobb-
Douglas production function has a linear form after logarithmic trans­
3. Econometric model setting and description of variables formation, it connotates very strict assumptions. For example, given
factor prices, the Cobb-Douglas production function requires that the
3.1. Model setting elasticities of demand and factor shares are both constants and that the
elasticities of substitution of the factors are all − 1. The transcendental
3.1.1. SFA model setup logarithmic production function, however, is a variable elasticity of the
We use the SFA model to estimate bank efficiency. The existing substitution production function that can incorporate interaction and
studies on the estimation of bank efficiency include mainly non- quadratic terms in the model to capture the interactive effects of the
parametric and parametric estimation methods. Non-parametric esti­ bank's individual factors and the potential for nonlinearities (Kumbha­
mation methods are represented by data envelopment analysis (DEA), kar, 1989). It is more flexible and inclusive and fits the characteristics of
which does not perform model setting but develops efficiency studies commercial banks. Therefore, referring to (Berger et al., 2009; Chen and
through linear programming. Xie et al. (2022) used DEA to estimate Lu, 2021), we construct a profit efficiency model for commercial banks
bank efficiency in seven Asian emerging economies. Zhong et al. (2021) as follows:
fused DEA with neural networks to build the SBM-DEA-BPNN model and

∑2 ( ) ∑ ( ) ( ) ( )
pbtit ym,it 3 wn,it 1∑2 ∑2 ym,it yq,it
ln = α0 + β m ln + γ n ln + θ 1t + β mq ln ln +
w2,it tait m=1 tait n=2 w1,it 2 m=1 q=1 tait tait
( ) ( ) ( ) ( ) ( )
1∑3 ∑3 wn,it wr,it 1 1∑2 ∑2 y w e
γnr ln ln + θ2 t2 + ηmn ln m,it ln n,it + λ1 ln it + (1)
2 n=2 r=2 w1,it w1,it 2 2 m=1 n=1 tait w1,it tait
( ) ( ) ∑ ( ) ( ) ∑ ( ) ( )
eit eit 2 ym,it eit 3 wn,it eit
λ2 ln ln + ϕ m ln ln + φ n ln ln + νit − μit
tait tait m=1 tait tait n=2 w1,it tait

found that the model was more applicable in calculating bank efficiency. α, β, γ, λ, η, ϕ, φ, θ is the parameter to be estimated. The price of
The parameter estimation method is represented by SFA, which de­ money w1 is used to standardize profits and other factor prices so that
composes the stochastic disturbance term in the model setting and the linearity of input factor prices can be set. Standardize profits, net

4
Q. Cao et al. Energy Economics 133 (2024) 107516

Table 1
Description of the main variables.
Indicator type Variable name Variables Variable definitions

Frontier analysis model

Actual profit profit before tax pbt total profit before tax (RMB million)
Outputs total loans loans total loans (RMB million)
other profitable assets oea total profitable assets other than loans (RMB million)
Inputs the price of funds fup interest expense/(total short-term financing + total long-term financing)
the price of labor empp staff costs/number of employees
the price of assets asp (operating and administrative expenses - staff costs)/Net fixed assets (RMB million/person)
Net input net equity input equ total owners' equity (RMB million)
total assets tas asset size (RMB million)

Inefficiency model
ESG investments esg SSI ESG Rating Indicators
Investment in environment env SSI E Rating Indicators
Investment in social soc SSI S Rating Indicators
Investment in governance gov SSI G Rating Indicators
Fintech fintech fintech Index

Source: Authors' elaboration.

equity inputs, and outputs with the size of total assets ta to control for 3.1.3. One-step bank efficiency model setting that considers the moderating
the effect of heteroskedasticity. As commercial banks may have negative role of fintech
pre-tax profits, we refer to (Tabak et al., 2013) to replace negative pre- To further consider the moderating role played by fintech between
tax profits with one to avoid logarithmic invalidation. ESG investment and bank profit efficiency, we add the interaction term
Parameters are subject to model flushness and symmetry constraints, between the level of fintech and bank ESG investment to the inefficiency
i.e., flushness of factor input prices, self-multiplying outputs and inputs, model, and the inefficiency model is as follows:
∑ ∑ ∑
and symmetry of cross terms: γn = 1, βmq = 0, γnr = 0,βmq = βqm ,
n μit = δ0 + δ1 esgit + δ2 fintechit + δ3 esgit * fintechit + εit (4)
γ nr = γrn .
δ0 , δ1 , δ2 , δ3 is the parameter to be estimated and esgit is the bank's
ESG investment. fintechit is a proxy variable for fintech. This variable is a
3.1.2. Setting of the heterogeneity model
dummy variable and is considered to be at a high level when the value of
The inefficiency model is as follows:
fintech is greater than the mean, fintechit is taken as 1, and at a low level
μit = δ0 + δ1 esgit + εit (2) when the value of fintech is less than the mean, fintechit is taken as 0.
The profit efficiency model for one-step commercial banks consid­
δ0 , δ1 is the parameter to be estimated and esgit is the bank's ESG
ering the moderating effect of fintech is constructed as follows:
investment.

∑2 ( ) ∑ ( ) ( ) ( )
pbtit ym,it 3 wn,it 1∑2 ∑2 ym,it yq,it
ln = α0 + β m ln + γ n ln + θ 1t + β mq ln ln +
w2,it tait m=1 tait n=2 w1,it 2 m=1 q=1 tait tait
( ) ( ) ( ) ( )
1∑3 ∑3 wn,it wr,it 1 1∑2 ∑2 y w
γnr ln ln + θ2 t2 + η ln m,it ln n,it +
n=1 mn
2 n=2 r=2 w1,it w1,it 2 2 m=1 tait w1,it (5)
( ) ( ) ( ) ∑ ( ) ( ) ∑ ( ) ( )
eit eit eit 2 ym,it eit 3 wn,it eit
λ1 ln + λ2 ln ln + ϕ m ln ln + φ n ln ln +
tait tait tait m=1 tait tait n=2 w1,it tait
νit − (δ0 + δ1 esgit + δ2 fintechit + δ3 esgit *fintechit + εit )

Therefore, the one-step commercial bank profit efficiency model is

constructed as follows:

∑2 ( ) ∑ ( ) ( ) ( )
pbtit ym,it 3 wn,it 1∑2 ∑2 ym,it yq,it
ln = α0 + β
m=1 m
ln + n=2 n
γ ln + θ 1 t + β
q=1 mq
ln ln +
w2,it tait tait w1,it 2 m=1 tait tait
( ) ( ) ( ) ( ) ( )
1∑3 ∑3 wn,it wr,it 1 2 1∑2 ∑2 ym,it wn,it eit
γ nr ln ln + θ 2t + η mn ln ln + λ1 ln + (3)
2 n=2 r=2 w1,it w1,it 2 2 m=1 n=1 tait w1,it tait
( ) ( ) ∑ ( ) ( ) ∑ ( ) ( )
eit eit 2 ym,it eit 3 wn,it eit
λ2 ln ln + ϕm ln ln + φn ln ln + νit − (δ0 + δ1 esgit + εit )
tait tait m=1 tait tait n=2 w1,it tait

5
Q. Cao et al. Energy Economics 133 (2024) 107516

3.2. Description of key variables from AAA to C. The explanatory variables in this paper are assigned a
value of 9–1 according to the grade, i.e., AAA grade is assigned 9, and C
3.2.1. Selection of bank efficiency input-output indicators grade is assigned 1.
In this paper, we refer to (Sturm and Williams, 2004) for the selection
of input-output indicators for commercial banks using the intermedia­ 3.2.3. Fintech
tion approach. The bank's efficiency is measured by its capability to We use a text mining method to construct the fintech index. First, we
transform resources into income-producing financial assets, and this construct the initial lexicon. In this paper, 16 keywords are selected from
approach focuses on banking activities that play an intermediary role, in four dimensions as the basic indicators for constructing the fintech
the form of loans that distribute depositors' deposits to borrowers. index. Second, we calculate the keyword word frequency. With the help
Therefore, the three input indicators selected for this paper are the price of the Baidu index search engine, we count the daily word frequency
of funds, the price of labor, and the price of assets. The two output in­ data of keywords. Thirdly, we synthesize the fintech index. The key­
dicators include total loans and other profitable assets, the net input words were analyzed using principal component analysis to synthesize
indicators are net equity input and total assets, and the explanatory the fintech index.1
variable is commercial banks' profit before tax. All the variables are
defined in Table 1. 3.3. Sample selection

3.2.2. ESG We use the quarterly panel data of 37 listed banks from Q4 2014 to
In this paper, we refer to (Lin et al., 2021; Wu et al., 2022) to select Q4 2021 as the research sample. The data required to calculate the profit
the ESG index developed by Sino-Securities Index Information Service efficiency of listed banks are mainly sourced from Bureau van Dijk
(Shanghai) Co. Ltd. (SSI) to measure banks' ESG. There are many (BVD) bankscope database, with some of the missing data on the number
different evaluation systems and rating standards on ESG rating data of employees being completed based on the China Research Data Service
domestically and internationally, such as Thomson Reuters, Bloomberg, Platform (CNRDS) and banks' financial statements, and some of the
and SynTao Green Finance. The SSI ESG rating is selected because it missing quarterly data being supplemented by linear interpolation and
incorporates the actual situation of the Chinese banking market, with moving average methods. Data on operating and management fees and
comprehensive data coverage, long data duration, and strong data staff costs for listed banks are sourced from the China Stock Market and
availability. The SSI ESG rating system is based on data publicly dis­ Accounting Research (CSMAR). The ESG ratings of listed banks and their
closed by commercial banks, social responsibility and sustainability three sub-data were obtained from the Flush iFind database. A total of 6
reports, government and regulatory website data, and media coverage large state-owned commercial banks, 9 joint-stock commercial banks,
data. In terms of data construction, it is divided into a three-tier rating 15 urban commercial banks, and 7 rural commercial banks are included
system, including 3 dimensions, 14 themes, 26 key indicators, and over in this paper. The sampling is representative of the Chinese banking
130 underlying indicators. The ratings are divided into nine grades, industry, as these banks cover the major commercial bank types and
account for over 80% of China's banking assets.

Table 2 4. Empirical results and analysis

Baseline model: ESG and bank efficiency.
Model 1 Model 2 4.1. Analysis of the main regression results
Coefficient t value Coefficient t value
4.1.1. Baseline regression
Frontier analysis model
The estimates on the effect of ESG investments on banks' profit ef­
yy1 26.9037 (0.95) 23.5298** (2.37)
yy2 − 5.6499 (− 0.29) − 4.2195 (− 0.49) ficiency are shown in Table 2. From Model 2, we find the coefficient of
yy11 − 4.1407 (− 1.24) − 3.6665*** (− 3.54) ESG is significantly negative, which indicates that banks improve profit
yy12 − 2.1015 (− 0.99) − 1.9172** (− 2.35) efficiency by increasing ESG investment. The finding is similar to
yy22 1.3684 (0.74) 1.1230 (1.41) (Belasri et al., 2020). This is because banks build a better image of the
yye1 4.2512*** (2.95) 4.2504*** (7.60)
yye2 2.3530** (1.98) 2.2951*** (5.18)
bank and earn a higher social reputation when they increase their ESG
ww2 20.3495*** (3.69) 19.4896*** (8.20) investments as this financial move improves relations with external
ww3 − 6.7565 (− 1.46) − 6.2178*** (− 3.02) players and gains the trust of regulators and the trust of investors. Ac­
ww22 0.0797 (1.35) 0.0676 (1.51) cording to (Wu and Shen, 2013), the investment also improves the in­
ww23 − 0.0280 (− 0.28) − 0.0268 (− 0.66)
ternal environment, boosts employee motivation, and improves
ww33 − 0.1439 (− 1.32) − 0.1401*** (− 3.39)
eta − 23.7138* (− 1.72) − 23.7031*** (− 4.72) corporate governance, generating reputation spillovers. This validates
eta2 − 5.5050*** (− 4.75) − 5.2772*** (− 8.67) hypothesis 1.
t 0.0013 (0.03) − 0.0067 (− 0.14)
yw12 − 1.4809*** (− 2.60) − 1.4227*** (− 6.00) 4.1.2. ESG investment sub-components and bank efficiency
yw13 0.5579 (1.40) 0.5255*** (3.02)
yw22 − 1.8165*** (− 3.88) − 1.7478*** (− 8.49)
To investigate the effect of the ESG investment sub-component on
yw23 0.7445* (1.87) 0.6968*** (3.88) bank efficiency, we replace the core explanatory variable ESG invest­
we2 0.1601 (0.38) 0.2101 (1.26) ment with E, S, and G investment, respectively. Table 3 shows that E and
we3 − 0.4295* (− 1.79) − 0.4544*** (− 4.36) G investments increase profit efficiency while S investment decreases it.
t1 − 0.0099 (− 0.67) − 0.0077 (− 0.40)
Further, when all three ESG sub-indicators are included, the direction
_cons − 16.8270 (− 0.15) − 16.0428 (− 0.33)
and significance of the coefficients of each sub-indicator remain un­
changed. This result corresponds to (Mar Miralles-Quiros et al., 2019). It
Inefficient model
means that E investment helps banks to seize on policy dividends while
esg − 0.0820** (− 2.21)
_cons − 1.0140** (− 2.34) 0.4967*** (2.86) reducing the risks associated with polluting loans, which is beneficial to
Usigma − 5.9439*** (− 7.98) − 5.8303*** (− 3.13) bank efficiency. This aligns with the finding of (Zhou et al., 2022). On
Vsigma − 2.3833*** (− 16.13) − 2.4184*** (− 42.46)
N 1073 1073
1
Source: Authors' elaboration. See Appendix A for details of the the selected fintech index keyword phrase
Note: ***, **, * indicates 1%, 5%, and 10% level of significance. database.

6
Q. Cao et al. Energy Economics 133 (2024) 107516

Table 3
ESG sub-items and bank efficiency.
E S G All included

Coefficient t value Coefficient t value Coefficient t value Coefficient t value

Frontier analysis model

Inefficient model

env − 52.300*** (− 5.65e+30) − 0.2123*** (− 4.54)

soc 0.1507*** (3.31) 0.2656*** (3.11)
gov − 0.2137*** (− 4.55) − 0.2102*** (− 3.81)
_cons − 11.687*** (− 1.15e+30) − 0.1099 (− 0.39) 5.3747 (.) 0.3009* (1.73)
Usigma − 97.239 (.) − 7.6768 (− 0.58) − 4.2071 (− 0.08)
Vsigma 12.8332 (.) − 2.3997*** (− 30.13) − 2.5821 (− 0.25)
N 1073 1073 1073 1073

Source: Authors' elaboration.

Note: ***, **, * indicates 1%, 5%, and 10% level of significance.

Table 4
Robustness tests.
1. Adjusting the sample 2. Re-measurement of profit efficiency 3. Greene(2005) 4. Cluster robust standard errors

Coefficient t value Coefficient t value Coefficient t value Coefficient t value

Frontier analysis model

Inefficient model

esg − 0.1080*** (− 4.95) − 0.0813*** (− 2.93) − 0.2985*** (− 3.42) − 0.1080*** (− 2.76)

_cons 0.7286*** (5.50) 0.5618*** (3.93) 1.1753*** (4.10) 0.7286*** (3.30)
Usigma − 5.5228*** (− 4.06) − 5.4730*** (− 3.55) − 3.3915 (.) − 5.5228*** (− 7.76)
Vsigma − 2.5698*** (− 33.13) − 2.3524***- (− 33.13) − 3.2382*** (− 47.68) 2.5698*** (− 14.45)
N 928 1073 1073 928

Source: Authors' elaboration.

Note: ***, **, * indicates 1%, 5%, and 10% level of significance.

Table 5
Estimation of policy effects.
Estimation of policy effects

Coefficient t value

Frontier analysis model

Inefficient model

esg − 1.7e+02*** (− 2.02e+71)

Ingroup_inyear − 1.6e+02*** (− 2.87e+71)
_cons 42.9202*** (7.86e+69)
Usigma − 2.8e+02*** (− 2.47e+69)
Vsigma − 2.4e+02*** (− 5.48e+72)
N 1073

Source: Authors' elaboration.

Note: ***, **, * indicates 1%, 5%, and 10% level of significance.

Fig. 2. Parallel trend test. Table 6

Note: The solid line (circle) represents the control group and the solid line (plus Subgroup studies.
sign) describes the treatment group. Local banks Ownership analysis
(Source: Authors' elaboration.)
Coefficient t value Coefficient t value

Frontier analysis model

the other hand, via G investment, commercial banks optimize their in­
ternal corporate governance to improve transparency and reduce agency Inefficient model
costs, thereby increasing profit efficiency (Mar Miralles-Quiros et al., esg − 0.1581*** (− 3.28) − 0.0964*** (− 3.12)
2019). Excessive S investment, however, leads to the appropriation of interactive 0.0616*** (2.63) − 0.0616*** (− 2.63)
banks' limited resources and increased expenditures, thus reducing _cons 0.7130*** (4.69) 0.7129*** (4.69)
Usigma − 4.6326*** (− 6.06) − 4.6329*** (− 6.06)
profit efficiency. This is consistent with (Fukuyama and Tan, 2021). Our
Vsigma − 2.4995*** (− 40.64) − 2.4995*** (− 40.64)
findings validate Hypothesis 2. N 1073 1073

Source: Authors' elaboration.

Note: ***, **, * indicates 1%, 5%, and 10% level of significance.
4.2. Robustness tests

First, we assume that banks' ESG investments may not always be

7
Q. Cao et al. Energy Economics 133 (2024) 107516

unsatisfactory, making banks selective in disclosing ESG information control groups satisfy the parallel trend test. In Table 5, the coefficient
and leading to a serious lack of data on ESG disclosure. In order to on Ingroup_inyear is significantly negative, indicating that the imple­
address outliers due to bias in sample selection, we delete five banks mentation of the policy leads to decreased inefficiency and increased
with serious missing ESG data volume, namely Xiamen Bank, Postal bank profit efficiency. Our finding is similar to (Luo et al., 2021). This
Savings Bank of China, Zhejiang Commercial Bank, Chongqing Rural result is in line with the baseline model results.
Commercial Bank, and Qingdao Rural Commercial Bank. We use the
remaining 32 commercial banks for the regression. As is shown in Model 5. Further research
1, Table 4, the coefficient of ESG is − 0.108 and significant. Regarding
impact direction, the result is robust. 5.1. Subgroup study
Next, we measure profit efficiency using outputs, inputs and net in­
puts. However, considering that measurement errors in the explanatory 5.1.1. Local banks
variables may also affect the final results, we reconfigure the model by The impact of ESG investment on profit efficiency may vary across
removing the net equity input and the cross-product term of the net banks. In China, local commercial banks are limited in size and re­
equity input for regression analysis. As is shown in Model 2, Table 4, the sources. Therefore, we add the local bank dummy variable and its
coefficient of ESG is − 0.0813 and significant. The effect of ESG remains interaction term with ESG investment to the SFA model. The local bank
robust. dummy is 1 if the bank is an urban or rural commercial bank, and 0 if
So far, we have estimated the results using the heterogeneous SFA otherwise. Table 6 shows the results. The coefficient on the local bank
model. However, distinguishing between individual heterogeneity and dummy variable and its interaction with ESG is significantly positive,
inefficiency components is always a difficult task when performing SFA. indicating that when local banks increase their ESG investment, they
If the heterogeneity of the production functions of different banks become less efficient. The finding is consistent with (Zhou et al., 2022).
cannot be set correctly, setting bias will be reflected in the disturbance ESG investments deteriorate profitability and reduce bank efficiency by
terms of the model, resulting in biased results. Therefore, we use the displacing banks' resources from other investments (Barnea and Rubin,
method proposed by (Greene, 2005a) to deal with individual hetero­ 2010). Local banks tend to have more limited resources compared to
geneity in panel data. We add individual effects to the frontier function national banks, and this resource substitution effect may lead to greater
to study the effect of ESG on banks' inefficiency term (Mutarindwa et al., adverse effects. In addition, compared to national banks, local banks'
2021). As is shown in Model 3, Table 4, the coefficient of ESG is − 0.2985 influence is largely confined in cities or provinces where banks are
and significant, which remains robust regarding the sign and direction of located, and their limited scope of operations makes it difficult to
the impact. develop a larger scale. This adds to difficulties in developing economies
Finally, given variances in ESG investments that may lead to unre­ of scale when banks attempt to make ESG investments. Hughes et al.
liable results in estimating the coefficients, we use cluster robust stan­ (2019) also pointed out that small community banks might face higher
dard errors for estimation (Kallel and Triki, 2022). As is shown in Model costs in regulatory compliance and technology-related aspects than
4, Table 4, the coefficient of ESG is − 0.108 and significant, which is in large banks, making it difficult to achieve economies of scale. And small
line with the benchmark model. community banks are also less efficient in utilizing investment
opportunities.

4.3. Research on policy effects 5.1.2. Ownership analysis

The influence of ESG investment on profit efficiency may also vary
To mitigate endogeneity arising from reverse causality (Dwumfour by bank ownership. In China, large state-owned banks often have high
et al., 2022), we adopt the DID model to investigate the impact of banks' efficiency by absolute advantages regarding the scale of operation and
implementation of ESG policies on profit efficiency. In this paper, we use scope of business. Therefore, we add the state-owned bank dummy
the issuance of Guidelines on Green Bond Assessment and Certification variable and its interaction term with ESG investment to the model for
Practices (Provisional) as a quasi-natural experiment. The guidelines were further analysis. The state-owned bank dummy variable is defined as 1 if
jointly issued by the central bank and the SEC in December 2017. The the bank is large state-owned or joint-stock, and 0 if otherwise. Also, in
treatment group includes high ESG investment level banks, i.e., banks Table 6, we observed a significant negative effect of the state-owned
with current ESG ratings higher than the average, and the control group bank dummy variable and its interaction term on ESG, indicating that
includes low ESG investment level banks, i.e., banks with current ESG the contribution of ESG investment to bank efficiency is more effective
ratings lower than the average. We then incorporate the policy effect in state-owned banks. This result is similar to (Huang et al., 2017; Ji
estimates into the SFA model. As shown in Fig. 2., the experimental and

Table 7
Moderating effects of fintech.
ESG E S G

Coefficient t value Coefficient t value Coefficient t value Coefficient t value

Frontier analysis model

Inefficient model

fintech 0.6843* (1.87) 2.2673 (0.70) − 0.2064** (− 2.25) 0.7538 (1.25)

esg*fintech − 0.1268* (− 1.93)
env*fintech − 2.5213 (− 0.62)
soc*fintech 0.1418** (2.45)
gov*fintech − 0.3506* (− 1.66)
_cons 0.0345 (0.29) − 1.9895 (− 0.60) 0.0885 (0.54) − 0.1037 (− 0.19)
Usigma − 4.9993*** (− 3.08) − 2.8752** (− 2.56) − 8.9034 (− 0.69) − 4.2442** (− 2.06)
Vsigma − 2.4348*** (− 38.51) − 2.4357*** (− 51.11) − 2.3914*** (− 50.57) − 2.4332*** (− 45.5)
N 1073 1073 1073 1073

Source: Authors' elaboration.

Note: ***, **, * indicates 1%, 5%, and 10% level of significance.

8
Q. Cao et al. Energy Economics 133 (2024) 107516

et al., 2023). As the Chinese government encourages the strong growth

of ESG to achieve the goal of sustainable development, state-owned
banks are more prominent in reducing risk and improving bank effi­
ciency. This is because state-owned banks are sensitive to government
policies and show more momentum in enforcing relevant laws and
regulations. Yuan et al. (2022) pointed out that given the protruding
places large state-owned banks and joint-stock commercial banks have
in China's banking industry, regulators were usually stricter and pushed
them to carry out reforms in a more thorough manner. In addition, the
objective of long-term stable operation of state-owned banks prompts
these banks to focus on sustainable development, which helps to avoid
risky investments. According to (Trinh et al., 2023), the higher the CSR
of a bank, the lower the frequency of extremely dangerous events and
the lower the systemic tail risk. Hu et al. (2023) further pointed out that
state-owned banks had higher credibility of information on ESG in­
vestments and were more likely to win investors' trust, suggestive of
strong reputation spillover effects.
Fig. 3. Comparison of ESG investment levels to profit efficiency.
5.2. Moderating effects of fintech Note: The solid line represents the low ESG investment bank efficiency and the
dashed line represents the high ESG investment bank.
Fintech has empowered commercial banks and changed the tradi­ (Source: Authors' elaboration.)
tional business model. With the profound combination of fintech and
banks, the scope of fintech applications has become more extensive. It 5.3. Comparison of ESG investment levels to profit efficiency
not only includes front-end operation management and marketing, but
also gradually combines with back-end risk management and resource We divide banks into a high ESG group and a low ESG group and
allocation, thereby increasing the profit efficiency of commercial banks. examine how the efficiency of these two groups differs over the sample
For example, commercial banks use artificial intelligence to make period. Specifically, we average the ESG, with those above the mean
intelligent recommendations on financial products and use big data for being the high ESG group and those otherwise being the low ESG group.
risk control (identifying customer defaults and anti-fraud). However, As shown in Fig. 3., over the entire sample period, banks in the high ESG
whether fintech will further facilitate or inhibit the relationship between group are generally more efficient than those in the low ESG group. This
ESG investment and bank efficiency needs further discussion. finding is similar to (Belasri et al., 2020). When banks have a high ESG
In this paper, we add the fintech dummy variable (fintech) and its investment, they build a good image, win the trust of stakeholders and
interaction term with ESG investment to the inefficiency term of the optimize their internal environment, which is conducive to improving
baseline model, as detailed in eq. (5). We construct a fintech index and bank efficiency.
calculate the mean value. If the fintech index is greater than the mean,
fintech is considered at a higher level, and the variable fintech is taken as 6. Conclusions, implications, and policy recommendations
1; otherwise, 0. As is shown in Table 7, fintech and ESG investment
interaction term coefficient is significantly negative, suggesting that the We empirically analyze the influence of ESG investment on banks'
effect of ESG investment on the profit efficiency of banks is more sig­ profit efficiency by conducting a heterogeneous SFA on the panel data of
nificant when the fintech level is higher. This is similar to the view of 37 listed banks in China from 2014 Q4 to 2021 Q4. We further test the
(Liu et al., 2021). As mentioned earlier, fintech is able to further alle­ correlation between different subcomponents of ESG investment and
viate information asymmetry between banks and external stakeholders, banks' profit efficiency as well as the moderating role of fintech in that
increase transparency, and thus enhance bank efficiency. Therefore relation. The findings are as follows. First, banks improve their profit
commercial banks should further strengthen the integration of fintech efficiency by increasing their ESG investments, which is mainly attrib­
and ESG investment. This validates Hypothesis 3. uted to the reputational spillover effect. Second, the impact of E, S, and
Further, we replace ESG investment with E, S, and G investment. G investments varies in that E and G investments increase banks' profit
Table 7 shows the regression results. It shows that fintech enhances the efficiency while S investment decreases it. Third, fintech further en­
enhancing effect of E and G investment on bank efficiency, as well as the hances the contribution of ESG investments to bank profitability.
inhibiting effect of S investment on bank efficiency. These findings are in Finally, heterogeneity is found in profit efficiency across banks in that
line with (Kimani et al., 2020; Zheng and Siddik, 2022). From the E banks with higher ESG investment have higher profit efficiency.
investment perspective, as fintech levels increase, information asym­ Our study has important implications in both theory and practice.
metry decreases and banks are more likely to identify high-quality green From the theoretical perspective, we shift to the banking industry and
companies. At this point, the risks associated with making investments extend the existing literature by exploring the impact of ESG investment
in the environment are reduced, and bank efficiency gets further on bank efficiency. In addition, we introduce fintech indicators, offering
enhanced. From an S investment perspective, the increased level of a fresh perspective for future research on the effects of big data and
fintech strengthens competition among banks, when efficient use of artificial intelligence on the connection between ESG investment and
resources is a crucial concern for banks. Devoting many resources to bank efficiency. From a practical perspective, our findings suggest that
socially responsible investments lead to resource tying up and lower ESG investment is not a good choice for local banks. Local banks should
market value, further enhancing S investment as a greater disincentive not blindly pursue expanding the scale of ESG investment. Instead, they
to bank efficiency. From the perspective of G investment, as the level of should combine their own development characteristics and explore a
fintech continues to rise, banks better identify high-quality composite sustainable development path that suits them. For example, local banks
talents, and implement different incentive mechanisms for different can consider inclusive green projects and small and micro green pro­
talents to stimulate staff motivations, thus amplifying the influence of G jects. They can also go for green credit business from the upstream and
investment on efficiency. downstream of some big projects. State-owned banks, on the other hand,
should actively make ESG investments. They should proactively take on

9
Q. Cao et al. Energy Economics 133 (2024) 107516

the social responsibility of supporting sustainable development, References

continue to implement the concept of ESG investment, and play their
guiding role in sustainable development. For example, state-owned Afza, T., Asghar, M., 2017. Efficiency of commercial banks in Pakistan: application of sfa
and value added approach. Argum. Oecon. 38 (1), 195–220. https://doi.org/
banks can further rationalize their ESG governance and management 10.15611/aoe.2017.1.07.
frameworks and improve ESG disclosure quality. At the same time, they Agnese, P., Giacomini, E., 2023. Bank’s funding costs: do esg factors really matter?
have to caution against high-risk investments conducted by managers Financ. Res. Lett. 51 (1), 1–13. https://doi.org/10.1016/j.frl.2022.103437.
Agostino, M., Ruberto, S., Trivieri, F., 2023. The role of local institutions in cooperative
for reputational insurance effects. banks? Efficiency. The case of Italy. Int. Rev. Econ. Financ. 84, 84–103. https://doi.
We make the following policy recommendations. Firstly, commercial org/10.1016/j.iref.2022.11.001.
banks should incorporate ESG investment management into the bank's Altunbaş, Y., Tommaso, C.D., Thornton, J., 2016. Is there a financial accelerator in
european banking? Financ. Res. Lett. 17 (1), 218–221. https://doi.org/10.1016/j.
daily operations and management. On the one hand, commercial banks frl.2016.03.020.
should set up a special ESG investment strategy committee in the board Andrieș, A.M., Sprincean, N., 2023. Esg performance and banks’ funding costs. Financ.
to take charge of implementing the strategic plans. On the other hand, Res. Lett. 54 (1), 1–10. https://doi.org/10.1016/j.frl.2023.103811.
Azmi, W., Hassan, M.K., Houston, R., Karim, M.S., 2021. Esg activities and banking
banks should build their own ESG evaluation system, decompose and
performance: international evidence from emerging economies. J. Int. Financ. Mark.
incorporate investment indicators into the management of ESG invest­ Inst. Money. 70 (1), 1–18. https://doi.org/10.1016/j.intfin.2020.101277.
ment. In daily operations, banks should launch staff training on ESG Barnea, A., Rubin, A., 2010. Corporate social responsibility as a conflict between
knowledge, issue practical operational guidelines, and strengthen the shareholders. J. Bus. Ethics 97 (1), 71–86. https://doi.org/10.1007/s10551-010-
0496-z.
sharing of ESG investment information. Battese, G.E., Coelli, T.J., 1995. A model for technical inefficiency effects in a stochastic
Second, banks should attach much importance to the structural dif­ frontier production function for panel data. Empir. Econ. 20 (2), 325–332. https://
ferences in ESG sub-components and vigorously strengthen E and G doi.org/10.1007/BF01205442.
Belasri, S., Gomes, M., Pijourlet, G., 2020. Corporate social responsibility and bank
investments. Considering the carbon peaking and carbon neutrality efficiency. J. Multinatl. Financ. Manag. 54 (1), 1–10. https://doi.org/10.1016/j.
goals, banks should continue to vigorously develop green credit, issue mulfin.2020.100612.
green products and channel funds to high-quality green enterprises. Berger, A.N., Hasan, I., Zhou, M., 2009. Bank ownership and efficiency in China: what
will happen in the world’s largest nation? J. Bank. Financ. 33 (1), 113–130. https://
Besides, banks should also devise effective remuneration incentive doi.org/10.1016/j.jbankfin.2007.05.016.
schemes for employees. At the same time, banks should improve their Branco, M.C., Rodrigues, L.L., 2006. Corporate social responsibility and resource-based
corporate governance and enhance the transparency of their governance perspectives. J. Bus. Ethics 69 (2), 111–132. https://doi.org/10.1007/s10551-006-
9071-z.
policies to win the goodwill of investors. Chen, X., Lu, C.-C., 2021. The impact of the macroeconomic factors in the bank
Third, commercial banks should focus on fintech empowerment and efficiency: evidence from the chinese city banks. N. Am. J. Econ. Financ. 55 (1),
strengthen the bank's ESG investment infrastructure. For one thing, 1–14. https://doi.org/10.1016/j.najef.2020.101294.
Commission, S.E, 2010. Commission Guidance Regarding Disclosure Related to Climate
commercial banks have to apply fintech in all aspects of ESG investment,
Change. Date, Available at. extension://bfdogplmndidlpjfhoijckpakkdjkkil/pdf/
including evaluation systems, risk identification, database support and viewer.html?file=https%3A%2F%2F2.gy-118.workers.dev/%3A443%2Fhttps%2Fwww.sec.gov%2Ffiles%2Frules%2Finterp%
information disclosure. For another, the banking industry should fully 2F2010%2F33-9106.pdf(accessed February 8, 2010).
leverage on emerging technologies such as machine learning and arti­ Cui, J., Jo, H., Na, H., 2018a. Does corporate social responsibility affect information
asymmetry? J. Bus. Ethics 148 (3), 549–572. https://doi.org/10.1007/s10551-015-
ficial intelligence to data-screening, establish an ESG investment infor­ 3003-8.
mation database and integrate it with the existing financial information Cui, Y., Geobey, S., Weber, O., Lin, H., 2018b. The impact of green lending on credit risk
database so as to include ESG investment project evaluation, risk iden­ in China. Sustain 10 (6), 1–16. https://doi.org/10.3390/su10062008.
Di Tommaso, C., Thornton, J., 2020. Do esg scores effect bank risk taking and value?
tification and corporate loan flow monitoring into daily business Evidence from european banks. Corp. Soc. Responsib. Environ. Manag. 27 (5),
activities. 2286–2298. https://doi.org/10.1002/csr.1964.
Du, P., Huang, S., Hong, Y., Wu, W., 2022. Can fintech improve corporate environmental,
social, and governance performance?-a study based on the dual path of internal
CRediT authorship contribution statement financing constraints and external fiscal incentives. Front. Environ. Sci. 10 (1), 1–17.
https://doi.org/10.3389/fenvs.2022.1061454.
Qiang Cao: Conceptualization, Project administration, Formal Dwumfour, R.A., Oteng-Abayie, E.F., Mensah, E.K., 2022. Bank efficiency and the bank
lending channel: new evidence. Empir. Econ. 63 (3), 1489–1542. https://doi.org/
analysis, Writing – review & editing, Data curation, Writing – original 10.1007/s00181-021-02166-5.
draft. Tingting Zhu: Software, Writing – review & editing, Formal Esteban-Sanchez, P., de la Cuesta-Gonzalez, M., Paredes-Gazquez, J.D., 2017. Corporate
analysis, Data curation. Wenmei Yu: Writing – review & editing, Formal social performance and its relation with corporate financial performance:
international evidence in the banking industry. J. Clean. Prod. 162 (1), 1102–1110.
analysis, Conceptualization, Methodology, Funding acquisition,
https://doi.org/10.1016/j.jclepro.2017.06.127.
Supervision. Fukuyama, H., Tan, Y., 2021. Corporate social behaviour: is it good for efficiency in the
chinese banking industry? Ann. Oper. Res. 306 (1–2), 383–413. https://doi.org/
Declaration of competing interest 10.1007/s10479-021-03995-4.
Galletta, S., Mazzu, S., 2023. Esg controversies and bank risk taking. Bus. Strateg.
Environ. 32 (1), 274–288. https://doi.org/10.1002/bse.3129.
None. Greene, W., 2005a. Fixed and random effects in stochastic frontier models. J. Prod. Anal.
23 (1), 7–32. https://doi.org/10.1007/s11123-004-8545-1.
Greene, W., 2005b. Reconsidering heterogeneity in panel data estimators of the
Acknowledgments stochastic frontier model. J. Econ. 126 (2), 269–303. https://doi.org/10.1016/j.
jeconom.2004.05.003.
2022 Anhui Provincial Research Preparation Program Project, Harjoto, M.A., Jo, H., 2011. Corporate governance and csr nexus. J. Bus. Ethics 100 (1),
45–67. https://doi.org/10.1007/s10551-011-0772-6.
Research on Green Finance and Green Technology Innovation Synergy Hill, C.W.L., Jones, T.M., 1992. Stakeholder-agency theory. J. Manag. Stud. 29 (2),
to Improve China's Carbon Neutral Performance, No. 2022AH050558. 131–154. https://doi.org/10.1111/j.1467-6486.1992.tb00657.x.
Hu, J., Zou, Q., Yin, Q., 2023. Research on the effect of esg performance on stock price
synchronicity: empirical evidence from china’s capital markets. Financ. Res. Lett. 55
Appendix A. Supplementary data (1), 1–7. https://doi.org/10.1016/j.frl.2023.103847.
Huang, S., 2022. Does fintech improve the investment efficiency of enterprises? Evidence
Supplementary data to this article can be found online at https://doi. from china’s small and medium-sized enterprises. Econ. Anal. Policy. 74 (1),
571–586. https://doi.org/10.1016/j.eap.2022.03.014.
org/10.1016/j.eneco.2024.107516.
Huang, J., Duan, Z., Zhu, G., 2017. Does corporate social responsibility affect the cost of
bank loans? Evidence from China. Emerg. Mark. Financ. Trade 53 (7), 1589–1602.
https://doi.org/10.1080/1540496x.2016.1179184.
Hughes, J.P., Jagtiani, J., Mester, L.J., Moon, C.-G., 2019. Does scale matter in
community bank performance? Evidence obtained by applying several new
measures of performance. J. Bank. Financ. 106 (1), 471–499. https://doi.org/
10.1016/j.jbankfin.2019.07.005.

10
Q. Cao et al. Energy Economics 133 (2024) 107516

Iazzolino, G., Bruni, M.E., Veltri, S., Morea, D., Baldissarro, G., 2023. The impact of esg Soana, M.-G., 2011. The relationship between corporate social performance and
factors on financial efficiency: an empirical analysis for the selection of sustainable corporate financial performance in the banking sector. J. Bus. Ethics 104 (1),
firm portfolios. Corp. Soc. Responsib. Environ. Manag. 30 (4), 1917–1927. https:// 133–148. https://doi.org/10.1007/s10551-011-0894-x.
doi.org/10.1002/csr.2463. Sturm, J.E., Williams, B., 2004. Foreign bank entry, deregulation and bank efficiency:
Jamali, D., Safieddine, A.M., Rabbath, M., 2008. Corporate governance and corporate lessons from the australian experience. J. Bank. Financ. 28 (7), 1775–1799. https://
social responsibility synergies and interrelationships. Corp. Gov.-Int. Rev. 16 (5), doi.org/10.1016/j.jbankfin.2003.06.005.
443–459. https://doi.org/10.1111/j.1467-8683.2008.00702.x. Tabak, B.M., Fazio, D.M., Cajueiro, D.O., 2013. Systemically important banks and
Ji, L., Sun, Y., Liu, J., Chiu, Y.-H., 2023. Environmental, social, and governance (esg) and financial stability: the case of latin america. J. Bank. Financ. 37 (10), 3855–3866.
market efficiency of china’s commercial banks under market competition. Environ. https://doi.org/10.1016/j.jbankfin.2013.06.003.
Sci. Pollut. Res. 30 (9), 24533–24552. https://doi.org/10.1007/s11356-022-23742- Trinh, V.Q., Cao, N.D., Li, T., Elnahass, M., 2023. Social capital, trust, and bank tail risk:
x. the value of esg rating and the effects of crisis shocks. J. Int. Financ. Mark. Inst.
Jondrow, J., Lovell, C.A.K., Materov, I.S., Schmidt, P., 1982. On the estimation of Money 83 (1), 1–23. https://doi.org/10.1016/j.intfin.2023.101740.
technical inefficiency in the stochastic frontier production function model. J. Econ. Veltri, S., Bruni, M.E., Iazzolino, G., Morea, D., Baldissarro, G., 2023. Do esg factors
19 (2–3), 233–238. https://doi.org/10.1016/0304-4076(82)90004-5. improve utilities corporate efficiency and reduce the risk perceived by credit lending
Kallel, H., Triki, M., 2022. Foreign Ownership, Bank Efficiency and Stability: Whether institutions? An empirical analysis. Util. Policy 81 (1), 1–11. https://doi.org/
the Institutional Quality of Countries Is Important? Early Access, pp. 1–22. https:// 10.1016/j.jup.2023.101520.
doi.org/10.1002/ijfe.2701. Vergara, C.C., Agudo, L.F., 2021. Fintech and sustainability: do they affect each other?
Kim, M., Kristiansen, E.G., Vale, B., 2005. Endogenous product differentiation in credit Sustain 13 (13). https://doi.org/10.3390/su13137012.
markets: what do borrowers pay for? J. Bank. Financ. 29 (3), 681–699. https://doi. Wang, H.J., Schmidt, P., 2002. One-step and two-step estimation of the effects of
org/10.1016/j.jbankfin.2004.05.024. exogenous variables on technical efficiency levels. J. Prod. Anal. 18 (2), 129–144.
Kimani, D., Adams, K., Attah-Boakye, R., Ullah, S., Frecknall-Hughes, J., Kim, J., 2020. https://doi.org/10.1023/a:1016565719882.
Blockchain, business and the fourth industrial revolution: whence, whither, Wu, M.-W., Shen, C.-H., 2013. Corporate social responsibility in the banking industry:
wherefore and how? Technol. Forecast. Soc. Chang. 161 (1), 1–16. https://doi.org/ motives and financial performance. J. Bank. Financ. 37 (9), 3529–3547. https://doi.
10.1016/j.techfore.2020.120254. org/10.1016/j.jbankfin.2013.04.023.
Kumbhakar, S.C., 1989. Estimation of technical efficiency using flexible functional form Wu, S., Li, X., Du, X., Li, Z., 2022. The impact of esg performance on firm value: the
and panel data. J. Bus. Econ. Stat. 7 (2), 253–258. https://doi.org/10.2307/ moderating role of ownership structure. Sustain 14 (21), 1–22. https://doi.org/
1391443. 10.3390/su142114507.
Lee, C.-C., Lee, C.-C., 2019. Oil price shocks and chinese banking performance: do Xie, J., Nozawa, W., Yagi, M., Fujii, H., Managi, S., 2019. Do environmental, social, and
country risks matter? Energy Econ. 77 (1), 46–53. https://doi.org/10.1016/j. governance activities improve corporate financial performance? Bus. Strateg.
eneco.2018.01.010. Environ. 28 (2), 286–300. https://doi.org/10.1002/bse.2224.
Lee, C.C., Ning, S.L., Lee, C.C., 2015. How does bank capital affect bank profitability and Xie, Z.K., Liu, X.L., Najam, H., Fu, Q.H., Abbas, J., Comite, U., Cismas, L.M.,
risk? Evidence from china's wto accession. China World Econ. 23 (4), 19–39. https:// Miculescu, A., 2022. Achieving financial sustainability through revenue
doi.org/10.1111/cwe.12119. diversification: a green pathway for financial institutions in asia. Sustain 14 (6),
Lee, C.-C., Li, X., Yu, C.-H., Zhao, J., 2021. Does fintech innovation improve bank 1–16. https://doi.org/10.3390/su14063512.
efficiency? Evidence from china’s banking industry. Int. Rev. Econ. Financ. 74 (1), Xun, J., 2013. Corporate social responsibility in China: a preferential stakeholder model
468–483. https://doi.org/10.1016/j.iref.2021.03.009. and effects. Bus. Strateg. Environ. 22 (7), 471–483. https://doi.org/10.1002/
Lin, Y., Fu, X., Fu, X., 2021. Varieties in state capitalism and corporate innovation: bse.1757.
evidence from an emerging economy. J. Corp. Finan. 67 (1), 1–41. https://doi.org/ Yuan, L., Zhong, Y., Lu, Z., 2022. Foreign strategic investors and bank credit risk in
10.1016/j.jcorpfin.2021.101919. China: disclosure, finance or management effects? Pac.-Basin Financ. J. 73 (1), 1–12.
Liu, Y., Saleem, S., Shabbir, R., Shabbir, M.S., Irshad, A., Khan, S., 2021. The relationship https://doi.org/10.1016/j.pacfin.2022.101762.
between corporate social responsibility and financial performance: a moderate role Zamore, S., Beisland, L.A., Mersland, R., 2021. Excessive focus on risk? Non-performing
of fintech technology. Environ. Sci. Pollut. Res. 28 (16), 20174–20187. https://doi. loans and efficiency of microfinance institutions. Int. J. Financ. Econ. 1 (1), 1–18.
org/10.1007/s11356-020-11822-9. https://doi.org/10.1002/ijfe.2477.
Luo, S., Yu, S., Zhou, G., 2021. Does green credit improve the core competence of Zhan, W., Jing, H., 2022. Does fintech development reduce corporate earnings
commercial banks? Based on quasi-natural experiments in China. Energy Econ. 100 management? Evidence from China. Sustain 14 (24). https://doi.org/10.3390/
(1), 1–11. https://doi.org/10.1016/j.eneco.2021.105335. su142416647.
Mar Miralles-Quiros, M., Luis Miralles-Quiros, J., Redondo-Hernandez, J., 2019. The Zhang, A., Wang, S., Liu, B., Liu, P., 2022. How fintech impacts pre- and post-loan risk in
impact of environmental, social, and governance performance on stock prices: chinese commercial banks. Int. J. Financ. Econ. 27 (2), 2514–2529. https://doi.org/
evidence from the banking industry. Corp. Soc. Responsib. Environ. Manag. 26 (6), 10.1002/ijfe.2284.
1446–1456. https://doi.org/10.1002/csr.1759. Zheng, G.-W., Siddik, A.B., 2022. The effect of fintech adoption on green finance and
Mateev, M., Sahyouni, A., Al Masaeid, T., 2022. Bank performance before and during the environmental performance of banking institutions during the covid-19 pandemic:
covid-19 crisis: does efficiency play a role? Rev. Manag. Sci. Early Access 1-54. the role of green innovation. Environ. Sci. Pollut. Res. 30 (1), 25959–25971. https://
https://doi.org/10.1007/s11846-022-00611-y. doi.org/10.1007/s11356-022-23956-z.
Miras-Rodriguez, M.D., Carrasco-Gallego, A., Escobar-Perez, B., 2015. Has the csr Zhong, K.Y., Wang, Y.F., Pei, J.M., Tang, S.M., Han, Z.L., 2021. Super efficiency sbm-dea
engagement of electrical companies had an effect on their performance? A closer and neural network for performance evaluation. Inf. Process. Manag. 58 (6), 1–13.
look at the environment. Bus. Strateg. Environ. 24 (8), 819–835. https://doi.org/ https://doi.org/10.1016/j.ipm.2021.102728.
10.1002/bse.1848. Zhou, X.Y., Caldecott, B., Hoepner, A.G.F., Wang, Y., 2022. Bank green lending and
Morea, D., Mango, F., Cardi, M., Paccione, C., Bittucci, L., 2022. Circular economy credit risk: an empirical analysis of china’s green credit policy. Bus. Strateg. Environ.
impact analysis on stock performances: An empirical comparison with the euro stoxx 31 (4), 1623–1640. https://doi.org/10.1002/bse.2973.
50® esg index, 14 (2), 1–20. https://doi.org/10.3390/su14020843.
Mutarindwa, S., Siraj, I., Stephan, A., 2021. Ownership and bank efficiency in africa: true
Qiang Cao is an Associate Professor with a PhD in Economics, School of Finance, Anhui
fixed effects stochastic frontier analysis. J. Financ. Stab. 54 (1), 1–12. https://doi.
University of Finance and Economics, Bengbu, Anhui, China. His research interests include
org/10.1016/j.jfs.2021.100886.
machine learning and applied econometrics. E-mail:[email protected]
Nost, E., Colven, E., 2022. Earth for ai: a political ecology of data-driven climate
initiatives. Geoforum 130 (1), 23–34. https://doi.org/10.1016/j.
geoforum.2022.01.016. Tingting Zhu is a master's student, School of Finance, Anhui University of Finance and
Schmidt, P., 2011. One-step and two-step estimation in sfa models. J. Prod. Anal. 36 (2), Economics, Bengbu, Anhui, China. Her research interests include machine learning and
201–203. https://doi.org/10.1007/s11123-011-0228-0. prediction research. E-mail: [email protected]
Seele, P., Schultz, M.D., 2022. From greenwashing to machinewashing: a model and
future directions derived from reasoning by analogy. J. Bus. Ethics 178 (4),
Wenmei Yu is an Associate Professor with a PhD in Economics, School of Finance, Anhui
1063–1089. https://doi.org/10.1007/s10551-022-05054-9.
University of Finance and Economics, Bengbu, Anhui, China. Her research interests
include international finance and prediction. E-mail: [email protected]

11