Equity Market Reaction to Green Bond Issuance 
Announcement 
Global Empirical Evidence of Corporate Green Bonds in 2013–2021 
 
 
 
 
 
Master's thesis  
in Accounting and Finance 
 
 
Author: 
Arttu Vuokko 
 
Supervisor: 
Professor Mika Vaihekoski 
 
15.2.2023 
Turku 
  
  
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
The originality of this thesis has been checked in accordance with the University of Turku 
quality assurance system using the Turnitin Originality Check service.  
 Master's thesis 
 
Subject: Accounting and Finance 
Author: Arttu Vuokko 
Title: Equity Market Reaction to Green Bond Issuance Announcement: Global Empirical 
Evidence of Corporate Green Bonds in 2013–2021 
Supervisor: professor Mika Vaihekoski 
Number of pages: 71 pages + appendices 2 pages 
Date: 15.2.2023 
Sustainable development (SD) is becoming increasingly prominent in business and finance, and 
especially environmental challenges are receiving growing concern. What often falls outside the 
discussion is the substantial amount of capital required to achieve the desired sustainability goals. 
Assistance to the situation is providing the capital market, which enables the allocation of capital 
to the desired targets to mitigate and combat sustainability challenges. In particular, the trend is 
reflected in the debt capital markets, where multiple sustainable debt instruments have emerged. 
Out of all of the current ones available in the markets, green bonds have become the most popular 
in recent years. They are bonds whose proceeds must be used to finance or re-finance so-called 
green projects that aim to promote environmental issues. 
This thesis examines the equity market reaction to green bond issuance announcements of listed 
corporates. The second objective is to evaluate possible linkages between abnormal returns and 
green bond and issuer characteristics. The scope of the study is global, and the final sample 
consists of 564 green bond observations from 31 countries and ten economic sectors between 
2013–2021. The research methods are an event study and regression analysis. The results are 
compared between different subsamples and regions to make the analysis comprehensive. In 
addition to the empirical study, the thesis provides an extensive and timely literature review on 
the topic.   
The results of the event study are not entirely unambiguous but tend to lean more toward a positive 
equity market reaction. In particular, the positive reaction is visible in the pre-event window 
suggesting possible information leakage prior to the event, although the results are statistically 
insignificant. The result is more positive and significant for non-financial corporates, first-time 
issuances, and issuers from emerging & developing markets. Regionally, the strongest and most 
positive reaction is for non-financial corporates in Asia-Pacific, with the cumulative average 
abnormal return (CAAR) of 1.025% over a 10-day event window. The regression analysis 
suggests no statistically significant links between the abnormal returns and the green bond and 
issuer characteristics. The same regression results apply to financial and non-financial corporates 
and different regions. 
 
Key words: green bond, sustainable finance, sustainable debt instrument, event study, 
regression analysis 
  
Pro gradu -tutkielma 
 
Oppiaine: Laskentatoimi ja rahoitus 
Tekijä: Arttu Vuokko 
Otsikko: Osakemarkkinoiden reaktio vihreiden joukkovelkakirjojen liikkeellelaskusta 
ilmoittamiseen: Globaali empiirinen tutkimus yritysten vihreistä joukkovelkakirjoista vuosina 
2013–2021 
Ohjaaja: professori Mika Vaihekoski 
Sivumäärä: 71 sivua + liitteet 2 sivua 
Päivämäärä: 15.2.2023 
Kestävästä kehityksestä on yhä näkyvämpää liiketoiminnassa ja rahoituksessa. Erityisesti 
ympäristöhaasteet ovat kasvava huolenaihe. Keskustelun ulkopuolelle jää usein merkittävä 
pääoman määrä, joka tarvitaan haluttujen kestävyystavoitteiden saavuttamiseksi. Apua 
tilanteeseen tuovat pääomamarkkinat, jotka mahdollistaa pääoman kohdentamisen haluttuihin 
kohteisiin kestävyyshaasteiden lieventämiseksi ja torjumiseksi. Suuntaus näkyy erityisesti 
velkapääomamarkkinoilla vastuullisten velkainstrumenttien kehittymisenä. Kaikista 
markkinoiden nykyisistä vastuullisista velkainstrumenteista, vihreät joukkovelkakirjat ovat 
nousseet erityiseen suosioon viime vuosina. Ne ovat joukkovelkakirjoja, joiden tuotot on 
käytettävä ympäristökysymyksiä edistävien, niin sanottujen vihreiden hankkeiden, 
rahoittamiseen tai uudelleenrahoitukseen. 
Tässä opinnäytetyössä tarkastellaan osakemarkkinoiden reaktiota pörssiyhtiöiden vihreiden 
joukkolainojen liikkeeseenlaskuilmoituksiin. Toisena tavoitteena on arvioida mahdollisia 
yhteyksiä epänormaalien tuottojen sekä vihreiden joukkovelkakirjojen ja liikkeeseenlaskijan 
ominaisuuksien välillä. Tutkimus on maailmanlaajuinen, ja lopullinen otos koostuu 564 vihreästä 
joukkovelkakirjalainahavainnosta, 31 maasta ja kymmeneltä talouden sektorilta, vuosina 2013–
2021. Tutkimusmenetelmät ovat tapahtumatutkimus ja regressioanalyysi. Jotta analyysi olisi 
mahdollisimman kattavat, tuloksia verrataan eri osanäytteiden ja maantieteellisten alueiden 
välillä. Empiirisen tutkimuksen lisäksi tutkielma tarjoaa aiheesta laajan ja ajankohtaisen 
kirjallisuuskatsauksen.   
Tapahtumatutkimuksen tulokset eivät ole täysin yksiselitteisiä, mutta ne kallistuvat enemmän 
kohti positiivista osakemarkkinoiden reaktiota. Positiivinen reaktio näkyy erityisesti tapahtumaa 
edeltävässä ikkunassa, mikä viittaa mahdolliseen tietovuotoon ennen tapahtumaa, vaikkakin 
tulokset ovat tilastollisesti merkityksettömiä. Tulos on myönteisempi ja merkittävämpi 
rahoitusalan ulkopuolisille yrityksille, ensikertaisille liikkeellelaskuille sekä kehittyvien 
markkinoiden liikkeeseenlaskijoille. Aluekohtaisesti reaktio on voimakkain ja myönteisin Aasian 
ja Tyynenmeren alueen rahoitusalan ulkopuolisille yrityksille, joiden keskimääräinen 
kumulatiivinen epänormaali tuotto on 1,025 prosenttia kymmenen päivän tapahtumaikkunan 
aikana. Regressioanalyysi ei viittaa tilastollisesti merkittäviin yhteyksiin epänormaalien tuottojen 
sekä vihreän joukkovelkakirjan ja liikkeeseenlaskijan ominaisuuksien välillä. Samat regressio 
tulokset koskevat rahoitusalan ja rahoitusalan ulkopuolisia yrityksiä sekä eri maantieteellisiä 
alueita. 
 
Avainsanat: vihreä joukkovelkakirja, vastuullinen rahoitus, vastuullinen velkainstrumentti, 
tapahtumatutkimus, regressionanalyysi 
  
TABLE OF CONTENTS 
1 Introduction 8 
1.1 Motivation 8 
1.2 Research questions and goals 10 
1.3 Structure 11 
2 Background 13 
2.1 Sustainable finance 13 
2.2 Debt capital markets and sustainable finance 15 
3 Green bonds 18 
3.1 Definition 18 
3.2 Global green bond market development 21 
3.3 Current principles, standards and regulations 23 
3.4 Challenges 26 
3.5 Previous studies 28 
4 Data and research methods 32 
4.1 Data 32 
4.2 Research methods 37 
4.2.1 Event study 37 
4.2.2 Regression analysis 42 
5 Results and discussion 47 
5.1 Event study 47 
5.2 Regression analysis 55 
6 Conclusions 61 
References 63 
Appendices 72 
Appendix 1 Used market indices 72 
Appendix 2 Results from event study using USD as common currency 73 
 
  
LIST OF FIGURES 
Figure 1 Annual green bond issuances in November 2022 21 
Figure 2 Geographical heat map of green bond issuances in November 2022  23 
Figure 3 Event study structure 38 
Figure 4 Abnormal returns using local currency vs. USD 48 
 
LIST OF TABLES 
Table 1 Sustainable debt instruments 17 
Table 2 Current green bond types 20 
Table 3 Screening criteria for green bonds 33 
Table 4 Green bond sample by country and economic sector 34 
Table 5 Summary statistics of green bonds and issuers 36 
Table 6 Explanatory variables 43 
Table 7 Descriptive statistics for regression variables 44 
Table 8 Correlation matrices 46 
Table 9 Daily abnormal returns 49 
Table 10 Cumulative average abnormal returns 50 
Table 11 Comparison between subsamples 51 
Table 12 Comparison between regions 54 
Table 13 Regression results for non-financial group 56 
Table 14 Regression results for financial group 58 
Table 15 Regional regression results 60 
 
  
LIST OF ABBREVIATIONS 
Abbreviation Definition 
AAR Average Abnormal Return 
ABS Asset-Backed Security 
AR Abnormal Return 
CAAR Cumulative Average Abnormal Return 
CAR Cumulative Abnormal Return 
CBI Climate Bonds Initiative 
DCM Debt Capital Market 
EIB European Investment Bank 
EMH Efficient Market Hypothesis 
ESG Environmental, Social, and Governance 
GDB Green Bond Principle 
ICMA International Capital Market Association 
IMF The International Monetary Fund 
IPCC International Panel on Climate Change 
IPSF International Platform on Sustainable Finance 
MPT Modern Portfolio Theory 
MSCI Morgan Stanley Capital International 
OLS Ordinary Least Squares 
RI Total Return Index 
ROA Return on Assets 
SD Sustainable Development 
SDG Sustainable Development Goal 
SIFMA Securities Industry and Financial Markets Association 
SPV Special Purpose Vehicle 
SRI Socially Responsible Investing 
TEG Technical Expert Group 
TRBC The Refinitiv Business Classification 
UN United Nations 
VIF Variance Inflation Factor 
 
 
8 
1 Introduction 
1.1 Motivation 
Sustainable development (SD) has gained a significant role in modern society. According 
to Brundtland Commission’s (1987) report, it is “development that meets the needs of the 
present without compromising the ability of future generations to meet their own needs.”. 
In general, although not entirely unanimously, the term binds together environmental, 
social, and economic challenges that the world is facing (Schoenmaker & Schramade 
2018, 3). To address these challenges, in 2015, 193 countries in the United Nations (UN) 
agreed on 17 Sustainable Development Goals (SDGs). The main objective was to 
encourage governments, corporations, and civil society to tackle the defined broad issues 
of sustainability with a deadline of 2030. (UN General Assembly 2015; Thompson 2021, 
20–23.) 
Out of all the areas of SD, the environmental challenges have received considerable 
attention recently. The concern has been fueled by the UN Climate Report, published by 
the Intergovernmental Panel on Climate Change (IPCC) on 28 February 2022, stating the 
devastating consequences of rapid global warming. The trend can also be detected from 
the increasing amount of pledges to reach net zero carbon emission targets. According to 
Taking Stock’s (2021) global assessment of net zero targets, 61% of nations, 9% of states 
and regions, 13% of cities, and 21% of corporates have made some form of commitment 
to net zero. The annual sales of the corporates with net zero commitments included in the 
assessment account for nearly USD 14 trillion, equivalent to around one-fifth of the sales 
of the world’s 2,000 largest listed corporates. Furthermore, in August 2022, MSCI (2022) 
reports the cumulative number of listed corporates, with some type of net zero targets, 
being 3,152. While there may be many disagreements about the motives of pursuing net 
zero, and many tend to fall short from the claims1, the increase in its popularity indicates 
a growing interest in mitigating climate change. 
What often falls outside the discussion and news coverage is how much capital is required 
to achieve the SDGs. According to The New Climate Economy’s (2014) estimation, 
around USD 90 trillion of infrastructure investments are needed between 2015 and 2030 
to achieve global sustainable development and climate objectives. (Green Finance 
 
1 See e.g. New Climate Institute (2022) and UN High-Level Expert Group (2022).  
9 
 
Initiative 2016; Thompson 2021, 13.) In addition, McKinsey & Company’s (2022) 
analysis suggests that a total of around USD 275 trillion in cumulative spending on 
physical assets is required between 2021 and 2050 to reach the net zero scenario by 2050. 
It is equivalent to over USD 9 trillion each year. Also, according to Kharas and 
McArthur’s (2019) estimate, in 2015, the global public sector SDG spending was already 
around USD 20 trillion per year and is expected to reach USD 33 trillion or more by 2030. 
Although these are just a few examples, and the exact amount is challenging to estimate, 
the overall consensus is clear: the amount of required capital is substantial. 
Assistance to the situation is providing the capital market, which serves as a powerful 
mechanism to enable the financing of SDGs by providing ease to capital mobilization and 
deployment. Capital markets can be divided into equity and debt markets, from which 
debt markets are significantly larger. (World Economic Forum 2019.) What gives debt 
capital markets a particularly vital role in financing ongoing and new sustainable projects 
is their financing structure. A typical debt-to-equity ratio of an infrastructure project is 
70/30. The proportion of debt is usually even higher for sustainable projects, such as 
renewable energy, energy efficiency and low-emission vehicle projects, with a ratio of 
75/25. (McKinsey & Company 2013; OECD 2015.) In the past, these projects have been 
funded by general bank loans and balance sheet financing. However, as the transition to 
a more sustainable world requires a growing amount of capital, the capabilities of post-
financial crisis companies are exceeding, and the need for more debt capital is increasing. 
(Thompson 2021, 239–240.) 
The vital role of debt capital markets as an SDG financing enabler is shown well in the 
development of different sustainable debt instruments. They are used to raise debt capital 
for SD-aligned operations. Especially one of these instruments has gained broad attention 
in the past few years. It is called a green bond, and it was first introduced to the public in 
2007 when European Investment Bank (EIB) issued the world’s first green bond (EIB 
2022a). Green bonds are virtually equal to conventional bonds, but the proceeds must be 
used to finance so-called ‘green projects’. These projects aim to contribute to the 
environmental challenges of SD. For the project to be classified as green, it must meet 
certain criteria. (CBI 2019; ICMA 2022.) Although the use of proceeds is immensely 
limited for the issuer, the green bond market has only continued to grow exponentially. 
According to Climate Bonds Initiative (CBI) (2022a), in 2021, the global green bond 
market reached an annual value of over USD half trillion, maintaining its trend of 10 
10 
consecutive years of market expansion. Moreover, the market is expected to reach USD 
1 trillion by 2022 and USD 5 trillion by 2025. However, to get to this target, collective 
actions are needed. For example, policies need to be accelerated, the green project 
pipeline needs to be expanded, and the capital flow from advanced economies to 
emerging markets needs to be augmented. (CBI 2022d.) 
Green bonds are not only intriguing from the perspective of the corporates and 
governments issuing them. The growing interest in sustainability is also evident among 
investors, as they are increasingly more interested in the overall impact of their 
investments. The mere financial return is no longer enough, as for many, the influence of 
investment in the environment and climate has reached equal importance. Thus, for 
investors interested in sustainability, green bonds offer an opportunity to diversify and 
manage risk while fulfilling the desired sustainable outcomes. (Thompson 2021, 240.) 
While green bonds may raise concerns about greenwashing2, for issuers, they also serve 
as a tool to signal their commitment to sustainability to the investors (Flammer 2021). 
Consequently, there is some evidence that issuing green bonds would lead to a positive 
stock price reaction for listed corporates (see e.g. Baulkaran 2019; Wang et al. 2020; Tang 
& Zhang 2020; Flammer 2021). 
The booming market and the growing popularity of green bonds make them a compelling 
and valuable research topic in the field of finance. Because the green bond is a relatively 
new financial instrument, market developments and changes are happening rapidly. A 
continuous examination is needed to keep up with the changes and their effects. By 
making good use of the constantly growing amount of new data, it is possible to increase 
the amount of knowledge on which the decisions can be based. This thesis aims to 
contribute to the finance literature on green bonds, particularly from the point of view of 
listed corporates and investors.   
1.2 Research questions and goals 
This thesis examines the impact of green bonds on the stock returns of the listed corporate 
issuers. In other words, the aim is to study how the equity market reacts to the green bond 
issuance announcements. In addition, different characteristics of the green bonds and the 
listed issuers, which could explain the possible abnormal returns, are examined. The focus 
 
2 See Financial Times (2022). 
11 
 
is on the global green bond market. Although there are some previous studies on the topic, 
the conclusions are not consistent. By using the new available data, it is possible to find 
new meaningful observations. The goal is also to make the analysis as comprehensive as 
possible. Thus, two main research questions for this thesis are: 
1. Do green bonds have an impact on the stock returns of the listed corporate issuers? 
2. Are there links between green bond and issuer characteristics and abnormal 
returns?  
A two-step empirical analysis is conducted to answer these questions. The first part 
utilises an event study method. The event study examines whether green bonds create 
abnormal returns to the issuers and aims to answer the first research question. The date 
of the first issuance announcement serves as the investigated event. The announcement 
day, rather than the actual issuance day, represents better the time when the information 
entered the market. The results are analysed comprehensively between various 
subsamples, such as regions and sectors. In particular, the focus is on Europe, Asia-Pacific 
and North America, and the financial and non-financial sectors. Also, a comparison 
between the effects of the first-time and subsequent issuance is conducted. The 
comparisons allow a deeper analysis of the results by highlighting key differences 
between the scrutinised subsamples.  
The second part of the empirical analysis is a cross-sectional regression analysis. It is 
performed to answer the second research question as it allows for highlighting specific 
characteristics that are linked to the possible abnormal returns. To be precise, the 
characteristics to be considered are those of green bonds and the listed corporate issuers. 
In this part also, the results are analysed between regions and sectors. On top of answering 
the main research questions, the goal is to provide an extensive and up-to-date literature 
review of green bonds and related concepts. 
1.3 Structure 
To provide a comprehensive understanding of green bonds and their origins, Chapter 2 
contains background to the topic. First, the concept of sustainable finance is outlined. 
Second, the relationship between the debt capital market and the sustainable finance 
movement is discussed. Other sustainable debt instruments, excluding green bonds, are 
also presented in detail. Chapter 3 focuses on green bonds. The green bonds are defined 
12 
in depth, and the current green bond types are presented, along with the global green bond 
market development and current green bond standards. Also, the challenges related to 
green bonds and their future market development are discussed. Before moving to the 
chapters covering the empirical study, similar previous studies are summarized. Chapter 
4 introduces the data and research methods used in the empirical analysis, and the final 
results are displayed and discussed in Chapter 5. Chapter 6 summarizes the thesis and 
provides ideas for future research. 
13 
 
2 Background 
2.1 Sustainable finance 
The acknowledgement that, through investing, it is possible to impact the world and 
surroundings beyond the financial benefits is not a newly discovered idea. It has been 
around for centuries. Even though the modern form of finance can be considered to have 
started in the 1950s3, the concept of aligning monetary investments with personal values 
dates, at least, to the Quakers who aimed to align their purchases and investments with 
their religious values in 17th century England. (Bugg-Levine & Emerson 2011, 5–6; 
Boatright 2014, 150.) In modern finance, this ideology is strongly related to the 
sustainable finance movement, which extends to the different dimensions of 
sustainability. Thus, one major objective of modern sustainable finance is to contribute 
to sustainable development (SD) by considering the effects of finance, including investing 
and lending, on economic, social, and environmental issues (Schoenmaker & Schramade 
2018, 4). 
Sustainable finance has many related terms, which are often treated synonymously, 
including, for example, green finance; climate finance; responsible banking; and 
environmental, social, and governance (ESG). However, arguably there are differences 
between the scopes of the terms as each can be considered to cover diverging dimensions 
of SD. For example, green finance is usually referred to when only the environmental 
aspect is in consideration, and thus, green bonds can be classified as green financial 
products under green finance. ESG, a widely used term in finance, covers almost all 
dimensions of SD. It also often refers to quantified scores measuring the level of 
sustainability. Generally, the term sustainable finance is considered to cover all of the 
dimensions, and it is also treated so in this thesis. (Thompson 2021, 2–3.) ESG is mainly 
used to refer to measurable scores. Furthermore, this thesis classifies green bonds as 
sustainable debt instruments to ensure consistency and clarity. 
Sustainable finance can be seen to be partly incompatible with traditional finance theory. 
One big trend among investors is socially responsible investing (SRI), also known as 
impact investing. It refers to an investing strategy which accounts for the consequences 
 
3 Miller (2000) refers to the year 1952 as a “big bang” of modern finance, as it is when Harry Markowitz 
published an article called “Portfolio Selection” in The Journal of Finance. 
14 
of the investments to SD, and it essentially leads to the pruning of some investees and 
favouring others. (Bugg-Levine & Emerson 2011, 5; Boatright 2014, 150–151.) From the 
perspective of Markowitz’s (1952) modern portfolio theory (MPT), a rational investor 
maximizes expected returns at a given level of risk by using diversification, done by 
selecting various securities for the portfolio. Markowitz argues that for diversification to 
be optimal, one should utilize all of the available securities in the market. Therefore, 
excluding non-sustainable securities from the portfolio selection is contrary to the MPT. 
Although sustainable finance and traditional theories may have some incompatibilities, a 
valid claim is that the mere financial benefit is not the full power of capital. For some, 
being able to influence matters far beyond the traditional views is what activates the full 
potential. (Bugg-Levine & Emerson 2011, 5.) Thus, sustainable finance acknowledges 
that financial systems have a leading role in allocating capital to sustainable corporates 
and projects and accelerating the transition to more a sustainable future (Schoenmaker & 
Schramade 2018, 4).  
Many can be perceived to be intrigued by the full power of capital, as sustainable finance 
has continued to grow in popularity among researchers and financial market participants. 
According to Luo et al. (2022), the Web of Science (WoS) database contained globally 
3,786 research articles on some fields of sustainable finance between 2000 and 2021. 
They also discovered that the annual number of research papers has continued to grow 
consecutively almost every year during this period, with the amount being the highest in 
2021. Furthermore, according to PwC’s (2021) global investor survey, in 2021, 79% of 
the respondents consider environmental, social, and governance (ESG) risks and 
opportunities as important factors in making investment decisions. Also, 49% would sell 
their investments if the target corporation is not showing enough action in addressing 
ESG issues. The growing demands of investors are also reflected in corporate actions, 
and KPMG’s (2022) global sustainability reporting survey suggests that 96% of G250 
corporates and 79% of N100 corporates report on sustainability or ESG matters.4  
Policymakers and various global organizations are also issuing regulations, standards, 
and measures to promote the development of sustainable finance. According to European 
Policy Center (EPC) (2022), global, regional, and local initiatives play a crucial role in 
 
4 G250 refers to the world’s 250 largest corporates by revenue based on the Fortune 500 ranking, and N100 
refers to a worldwide sample of the top 100 corporates by revenue in 58 countries, territories, and 
jurisdictions providing a more broad-based snapshot of sustainability reporting (KPMG 2022). 
15 
 
achieving the full potential of sustainable finance by providing transparency and 
efficiency to the markets. A major goal of the initiatives is to ensure that investments 
classified as 'sustainable' truly contribute to SDG objectives. Another important aim is to 
prevent greenwashing, where entities make false or misleading claims about the positive 
environmental impact of a product, service, or activity (Thompson 2021, 30, 249). 
According to International Shareholder Services (ISS) (2022), globally, the EU is a 
pioneer in regulatory initiatives, but North America and Australia have also significantly 
increased their regulatory efforts. Also, Asia has accelerated the growth in new initiatives. 
Although plenty has already been achieved to promote and reinforce sustainable finance, 
much remains to be done, especially as geopolitical tensions and other macroeconomic 
issues have led to increased challenges in 2022. Significant objectives include, for 
example, reducing complexity and creating harmonization. (ISS 2022; EPC 2022.) 
2.2 Debt capital markets and sustainable finance 
The sustainable finance movement is reflected particularly in the debt capital market 
(DCM). It is a place where different debt instruments, mainly bonds, are traded. Debt 
instruments, also known as fixed-income assets, can be divided into two main categories. 
These are organization-guaranteed bonds and asset-backed securities (ABS). 
Organization-guaranteed bonds, typically issued by governments and corporates, are 
backed (collateralized) entirely by the issuing organization and raise capital for general 
purposes. The interest payments of ABSs are backed by a specified pool of assets, which 
are often placed in a corporate structure called a special purpose vehicle (SPV). For 
example, the issuers can bundle a pool of loans, securitize their revenues, and use that 
income as collateral for an ABS. On top of these two main categories, a broad range of 
hybrid structures are available in the DCM. While, for issuers, the debt instruments put 
up a possibility to borrow debt capital to finance operations, they also provide an 
opportunity to offset liabilities, generate returns,  and diversify portfolios for investors. 
(Thompson 2021, 238, 273.) 
In addition to the DCM, the overall capital markets include the equity market (stock 
market), where stocks of listed corporates are traded. However, the DCM is substantially 
larger. According to SIFMA’s (2022) capital market fact book, in 2021, the global equity 
16 
market capitalization was USD 124.4 trillion5, whereas the global debt outstanding was 
USD 126.9 trillion. Although with this comparison, the markets seem to have a similar 
size, the issuance volume in the DCM is significantly higher. The SIFMA report augments 
this statement by disclosing the global debt issuance being USD 26.8 trillion in 2021, 
whereas the equity issuance was only USD 1,042 billion. One reason for this can be 
considered to be the pecking order theory, according to which corporates prefer debt 
financing over equity financing if external capital is needed (Myers & Majluf 1984; 
Myers 1984).  
As a result of the merger of sustainable finance ideology and the DCM, new sustainable 
debt instruments, including green bonds, have gained conspicuous popularity. Table 1 
presents the current sustainable debt instruments and their market sizes and shares as at 
the end of 2021. For now, green bonds are excluded, as they are presented in depth in the 
next chapter. After green bonds, the second most popular sustainable debt instrument is 
a social bond, used to finance projects with positive social outcomes such as 
socioeconomic advancements, food security, or employment generation and thus 
contribute to the social dimension of SD (ICMA 2021b.). The third most popular is a 
sustainability bond, which has originated from the perception that green projects 
(typically financed by green bonds) can have social co-benefits and vice versa. The fourth 
most popular sustainability-linked bond is, by contrast with the other instruments, a 
behaviour-based product.6 It means that the financial characteristics of the bond are tied 
to pre-set sustainability targets and can therefore vary according to how the issuer 
achieves the targets. (ICMA 2020; Bloomberg 2021.) Transition bonds are the latest 
newcomers and are mainly found in highly polluting sectors, such as mining and aviation, 
which do not fall into the existing definitions of green but are crucial in a transition to net 
zero (CBI 2022b). 
 
 
 
 
 
 
5 Equivalent to the total market capitalization of the listed corporates globally (SIFMA 2022). 
6 All the other sustainable debt instruments can be classified as activity-based products (Bloomberg 2021). 
17 
 
Table 1 Sustainable debt instruments (modified from ICMA 2020; ICMA 2021b; 
ICMA 2021c; CBI 2022b) 
Instrument Definition 
Market size 2021 
(USD billion) 
% of 
total  
Social Bond Proceeds exclusively applied to 
finance or re-finance so-called 
‘social projects’, which aim to 
achieve greater social benefits.  
 
223.2 20.87% 
Sustainability Bond Proceeds exclusively applied to 
finance or re-finance projects 
that are a combination of both 
green and social projects with 
co-benefits in either direction.  
 
200.2 18.72% 
Sustainability-Linked Bond A forward-looking 
performance-based instrument 
for which the financial and/or 
structural characteristics can 
vary depending whether the 
issuer achieves predefined 
sustainability objectives. 
 
118.8 11.11% 
Transition Bond Financing operations that aim to 
decarbonize an activity or 
support the issuer in its 
transition to Paris Agreement 
alignment.  
 
4.4 0.41% 
 
The market in which the sustainable debt instruments operate can be referred to as the 
sustainable debt market. In recent years, it has grown exponentially, led by green bonds. 
According to a global report by Climate Bonds Initiative (CBI) (2022b), the total 
sustainable debt market reached a value of almost USD 1.1 trillion in 2021. Compared to 
2020, the market saw an increase of 46%. At the end of 2021, the cumulative number of 
all sustainable debt instruments placed on the market was 16,697. From these, 5,999 were 
issued in 2021, accounting for 35% of the total. Although the share of the sustainable debt 
in the overall DCM is still relatively small, it has the potential to increase its significance 
in the future if the same growth trajectory continues. However, challenges to the 
sustainable debt market, and DCM as a whole, are posed by several factors, such as the 
continued Ukraine invasion, high inflation, soaring interest rates, and the growing 
concern of global recession (CBI 2022b; Bloomberg 2022a).  
18 
3 Green bonds 
3.1 Definition 
Of the sustainable debt instruments discussed earlier, green bonds are by far the most 
popular and subject to particular attention and, thus, the main topic of this thesis. In 2021, 
they accounted for 49% of the total sustainable debt market, with a market size of USD 
522.7 billion (CBI 2022b). Regardless of the recent traction, thus far, there is no globally 
accepted universal definition for green bonds. Probably the most widely accepted and 
adopted definition is from International Capital Market Association’s (ICMA) (2022) 
Green Bond Principles (GBPs). They define green bonds as 
any type of bond instrument where the proceeds or an equivalent amount will 
be exclusively applied to finance or re-finance, in part or in full, new and/or 
existing eligible green projects. 
The eligible green projects can be related to, but are not limited to, pollution prevention, 
clean transportation, renewable energy, green buildings, climate change adaptation, 
biodiversity, or sustainable water management. Although there are several possible 
projects to which the proceeds can be applied, the main requirement is that they are related 
to the promotion of environmental issues. (ICMA 2022.) According to the CBI, in 2021, 
the three largest categories for the use of proceeds were energy, buildings, and transport, 
covering 81% of the whole year's total. Given that in 2022, energy, transport, and 
production & construction were the world's top 3 most polluting industries7, capital can 
be seen to be allocated as desired and expected (The Eco Experts 2022). 
Although green bonds have similar features as plain vanilla bonds8, four generally agreed 
distinguishing aspects have been specified. Currently, these are (ICMA 2022; Thompson 
2021, 241): 
1. Use of proceeds: The proceeds should be used to finance projects with green 
outcomes. 
2. Process for project evaluation and selection: The issuers should communicate the 
environmental suitability objectives, the process of determining the fit of the 
 
7 The level of pollution is measured by the annual greenhouse gas (GHG) emissions (The Eco Experts 
2022). 
8 Plain vanilla bond refers to the most basic version of bonds. 
19 
 
projects, and complementary information on processes used to identify and 
manage risk associated with the selected projects. 
3. Disclosure and management of proceeds: The allocation of funds raised from 
green bonds should be independently audited and made easily accessible to 
stakeholders. 
4. Reporting: The issuers should evaluate and report the environmental impacts of 
green bonds.  
It is essential to point out that these are only recommendations and not requirements for 
the issuers. However, they intend to, for example, increase the investors' confidence. In 
particular, the issuers' transparent communication and reporting will most likely increase 
the trust of investors favouring sustainable investment strategies. (Thompson 2021, 241.) 
To date, ICMA (2022) has identified four distinct green bond types.9 The types are 
Standard Green Use of Proceeds Bonds, Green Revenue Bonds, Green Project Bonds, 
and Secured Green Bonds. They all have specific characteristics and differ by the 
intended use and risk profiles. Table 2 assembles the current types with their descriptions. 
Standard Green Use of Proceeds Bond is the standard form, and so far, the majority of 
green bonds issued are this type. In most cases, these types of debt instruments reduce 
credit risk and the risk of default, as the bond is backed by the issuer’s entire balance 
sheet, giving rating agencies and investors more confidence. The collateral of Green 
Revenue Bonds is the project-linked cash flows (revenue streams), whereas the Green 
Project Bonds are backed by the green project(s) assets and balance sheet(s). They are 
similar to the extent that the credit exposure of the investors is linked directly with the 
success of the project(s) that the bonds are financing. Therefore, they may have a higher 
risk profile, which is usually acknowledged, with higher return requirements. (Thompson 
2021, 241–242; ICMA 2022.) 
Secured Green Bonds can be categorised into Secured Green Collateral Bonds and 
Secured Green Standard Bonds. It depends on whether the net proceeds are exclusively 
applied to project(s) securing the specific bond only (secured collateral bond) or to 
project(s) of the issuer, originator or sponsor, where such projects may or may not be 
 
9 The latest update and specification to the green bond types was done in June 2022. 
20 
securing the specific bond in whole or in part (secured standard bond). This category type 
may include but is not limited to covered bonds, secured notes, securitisations, asset-
backed commercial paper, and other secured structures. (ICMA 2022.) There are also 
similar, but slightly different, green bond type classifications available.10 However, they 
will not be covered in this thesis since the classification provided by ICMA is considered 
widely respected (Refinitiv 2022a).  
Table 2 Current green bond types (modified from ICMA 2022; CBI 2022c) 
Type Definition 
Standard Green Use of Proceeds Bond An unsecured full recourse-to-the-issuer debt 
obligation where the bond is backed by the 
issuer’s entire balance sheet.  
 
Green Revenue Bond A non-recourse-to-the-issuer debt obligation 
where the project-linked pledged cash flows of 
the revenue streams, taxes, fees etc. are 
collateral for the debt. 
 
Green Project Bond Can be issued for a single or multiple green 
project(s), and the recourse is to the green 
project(s) assets and balance sheet(s) with or 
without potential recourse to the issuer. 
 
Secured Green Bond A secured debt obligation collateralized by a 
pool of eligible green projects (assets). 
 
 
A few sub-categories have also emerged under green bonds. Practically, they are 
classified as green bonds but focus more explicitly on the promotion of specific 
environmental issues. The most notable ones include climate bonds and blue bonds. The 
proceeds of climate bonds are used to mitigate or adapt to the effects of climate change, 
and the proceeds of blue bonds are used to support sustainable marine and fisheries 
projects. Technically, most of the green bonds could be classified as climate bonds since 
the majority of proceeds have been used to finance projects related to climate change, 
while blue bonds are a smaller sub-category. (Thompson 2021, 244–245; ICMA 2022.) 
Overall, as the green bond market, and the sustainable debt market as a whole, continues 
to develop rapidly, the definitions can see changes shortly.   
 
10 See e.g. CBI 2022c. 
21 
 
3.2 Global green bond market development 
The first-ever green bond was issued in 2007 by a multilateral development bank, 
European Investment Bank (EIB), and since then, the market has grown immensely. The 
EIB’s green bond portfolio alone had grown to USD 24.5 billion by 2017, covering 160 
green projects in 46 countries. (Thompson 2021, 243–244; EIB 2022a.) Figure 1 
illustrates the exponential growth of the global green bond market. From 2007 to 2012, 
the annual issuance was still relatively small, and the growth was limited. During this 
period, the issuers were mostly multilateral development banks, such as EIB, which are 
supranational institutions set up by sovereign states (OECD 2015; EIB 2022b). In 2013, 
the first corporate issuers joined the market, and since then, the growth has been 
explosive, with 2021 being the record-breaking year with over USD half trillion annual 
issuance volume. Today, the green bond market issuers include, for example, corporates 
(financial and non-financial), countries (sovereign issuers), local governments, and 
development banks. Out of the current issuer types, corporates are the most vigorous, and 
at the end of 2021, they accounted for 44% of the cumulative green bond volumes. (CBI 
2022b.) 
 
Figure 1 Annual green bond issuances in November 2022 (data retrieved from Refinitiv Eikon)  
 
 
22 
In 2021, 73% of the total green bond volume originated from developed markets, while 
the rest came from emerging markets and supranational issuers. Figure 2 displays the 
geographical distribution of the green bond issuances measured by the USD amount 
issued, highlighting the current top 3 green bond issuing regions. These are Europe, Asia-
Pacific, and North America, out of which Asia-Pacific has experienced the strongest 
recent growth in green bond issuances, mostly led by China. In 2021, Asia-Pacific was 
the second biggest region measured by cumulative green bond issuances, surpassing 
North America for the first time in green bond market history. For now, Europe has been 
the most dominant region, contributing 50% of the total annual volume in 2021. However, 
the leading role might be diminishing in 2022 as Asia-Pacific green bond supply has 
increased. As shown in Figure 2, green bonds have not so far gained popularity in Africa, 
even though it is one of the territories most affected by climate change. Some of the 
reasons include, for example, less developed capital markets and smaller-scale loan 
needs. However, it will be interesting to follow in which direction the green bond market 
in Africa, as well as in other emerging markets, will develop in the future. (S&P Global 
2022a; CBI 2022b.) 
Country-wise, at the end of 2021, the US had the largest cumulative green bond issuance 
volume of USD 304 billion, followed by China with USD 199 billion. However, China 
will likely surpass the US as the number one green bond issuer shortly, as the supply 
keeps increasing rapidly in China (Bloomberg 2022b).11 The US green bond market is 
typically characterized, by a large number of issuers, with relatively small-sized deals. 
The market in China has recently seen a rise in non-financial corporate issuers, although 
one-third of the issuers are still financial institutions which lend money to other 
corporates' relevant projects. From European countries, France has traditionally been the 
pioneer, but as also shown in Figure 2, Germany has taken the number one position 
(Thompson 2021, 242). Germany was also the world's leading green bond issuer in Q3 
2022, with a volume of USD 15.89 billion. Recently, Italy has also raised its head with 
growing green bond volumes on the European market. (S&P Global 2022b; CBI2022b.) 
 
11 The heat map in Figure 2 suggests that China would have already exceeded the US in the amount issued. 
The data is retrieved from Refinitiv Eikon at the end of November 2022. However, while writing this thesis, 
no official records were found. 
23 
 
 
Figure 2 Geographical heat map of green bond issuances in November 2022 (data retrieved 
from Refinitiv Eikon) 
 
Forecasts and expectations for the green bonds market are high. According to CBI 
(2022a) forecast, the amount of annual green bond volume is expected to reach a value 
between USD 900 billion and USD 1 trillion by the end of 2022. After the first annual 
trillion, the next significant milestone would be to hit an annual volume of USD 5 trillion 
by 2025. However, this goal is very optimistic, and to reach it, governments, 
policymakers, investors, and issuers have to contribute to the market development. (CBI 
2022a.) Also, the overall market situation is unfavourable to the forecasts and 
expectations. Macroeconomic factors such as the continued Ukraine invasion, high 
inflation, hiking interest rates, the European energy crisis, and growing concern about the 
recession are threatening the entire world economy, including green bonds (Bloomberg 
2022a; IMF 2022a; CBI 2022e). Thus, CBI (2022e) has reported a 22% year-on-year 
(YoY) decline in green bond issuances at the end of Q3 2022. 
3.3 Current principles, standards and regulations 
To date, there is no uniform global standard for green bonds. However, as the global green 
bond market keeps growing rapidly, the need for new robust standards, principles and 
regulations increases immensely. Also, the lack of universal agreement on common 
approaches can be considered to be holding back the development of the green bond 
market. One major goal of the standards and principles is to ensure transparency and 
24 
create trust and credibility in the market. For the participants, they also enable a better 
overall understanding and awareness of the market. (Thompson 2021, 245, 249.) Despite 
the lack of globally harmonized standards and regulations, some regions and nations have 
supported green bond market development through regional and local approaches. 
Principles and frameworks are also offered by some associations, namely the 
International Capital Market Association (ICMA) and Climate Bonds Initiative (CBI). 
Probably the most well-known and widely accepted principles for green bonds are the 
ICMA’s (2022) Green Bond Principles (GBPs), which have also been referred to several 
times in this thesis. First published in 2014, the principles offer green bond issuers a 
collection of voluntary frameworks and outline the best practices. GBPs also serve 
investors by providing transparent and necessary information to assess the environmental 
impact of green bond investments. The third group, which they also support, are the 
underwriters for whom the principles offer vital steps that will facilitate transactions that 
preserve market integrity. The newest edition of GBPs was published in June 2021 and 
further developed in June 2022. However, since the GBPs are only a voluntary framework 
and not a formal regulatory standard, it is evident that different applications of the 
principles may occur. (Thompson 2021, 250–252, 256; ICMA 2022.)  
To supplement the GBPs, particularly the certification of green bonds, the CBI (2019) 
has developed Climate Bonds Standards. The purpose is not to offer competing or 
alternative practices to GBPs but to build on its principles, providing even more detailed 
descriptions. Thus, it is fully aligned with the GBPs. One of the reasons for the creation 
of these complementary standards has been the criticism for the lack of detail of the GBPs.  
The standard is divided into pre-issuance requirements and post-issuance requirements. 
The pre-issuance requirements are those that the issuer must meet when applying for a 
certification ahead of issuance, and the post-issuance requirements are those that the 
issuer must meet when seeking continued certification following the issuance. If the issuer 
meets the requirements, they have the possibility for an independent assurance provided 
by the Approved Verifier. The assurance is beneficial for the relationship between the 
issuer and the investors. On top of the certification specification, the Climate Bonds 
Standard also offers detailed criteria for the use of proceeds. Overall, the standard is a 
more rigorous addition to the GBPs. (CBI 2019; Thompson 2021, 256–257.) 
25 
 
The latest significant news on the development of green bond standards and principles is 
from China. The China Green Bond Standard Committee published new voluntary China 
Green Bond Principles in July 2022. Like the CBI’s Climate Bonds Standard, they refer 
to and have similar attributes to the GBPs by ICMA but are more directed to domestic 
green bond issuances. China’s green bond principles make four core components for the 
issuance of green bonds, which are the use of proceeds, management of the proceeds, 
project evaluation and selection, and duration of information disclosure. Although the 
principles are aimed at domestic issuances, the intention is partly to make more of the 
green bonds in China globally recognized. A significant part of this new update is that 
there have been plans for rule changes that would make these policies mandatory for 
exchange-traded bonds. There are speculations that if this rule change comes to force, 
China could become a leader in global green bond regulation. (Sustainable Fitch 2022; 
Reuters 2022.) 
Over the past few years, green bond taxonomies have become an increasingly important 
topic of discussion. According to CBI (2022b, 2022g), a taxonomy is “a classification 
system that identifies activities, assets or revenue segments that deliver on key 
environmental objectives”. They aim to provide guidance and, more importantly, clarity 
to the market participants on which activities or assets are eligible for sustainable 
investments and thus support the development of the green bond market. For decades, the 
eligibility of assets for inclusion in ESG and other sustainable products has been 
determined by principles, definitions, and classification systems, such as ESG scores, 
offered by agents in the private sector. The new taxonomies, however, have been put 
forward by public actors to form a more top-down approach to determining green 
activities. The taxonomies are generally publicly available, granular, and science-based. 
In time, the taxonomies have become more detailed and mandatory for market 
participants.  (CBI 2022b, 2022g.) 
Currently, the main actors in the field of green taxonomy are China and the EU. Out of 
these, China was the first to put forward a mandatory domestic taxonomy for the issuance 
of green bonds in 2015. It goes by the name of the Green Bond Endorsed Project 
Catalogue or China Taxonomy.12 In the EU, the European Commission established a 
 
12 The latest version of the China Taxonomy was published in April 2021 and was updated to align more 
closely with global definitions (PBOC 2021; CBI 2022b). 
26 
Technical Expert Group (TEG) on sustainable finance in 2018, intending to develop the 
EU-wide taxonomy and other related action plans.13 The TEG published the final report 
on EU Taxonomy in March 2020, containing recommendations related to the overarching 
design of the taxonomy. However, the taxonomy is still not fully complete and is being 
developed in several stages. The latest development step is the Climate Delegated Act, 
which provides taxonomy criteria on climate change mitigation measures, and was put to 
force in January 2022. (TEG 2020; European Commission 2022a; European Commission 
2022b.) Although the EU and China are the two current main taxonomy agents, 
taxonomies are also either in discussion, development, or draft elsewhere. For example, 
at the beginning of 2022, taxonomies were in discussion in Australia and Mexico, in 
development in Brazil, Canada, India, Japan and Kazakhstan, and in the draft in South 
Africa and the UK. (Thompson 2021, 257–258; CBI 2022b, 2022g.) 
3.4 Challenges 
At first glance, the proliferation of green bonds appears to be an adequate and good thing. 
However, there are still various challenges to overcome concerning green bonds. For 
example, the maintenance of market integrity and transparency becomes an essential 
factor as the market is growing and developing at a rapid pace. Thus, it is crucial to 
evaluate the green bond phenomenon from a critical aspect and highlight the current 
challenges that the market is facing. The previously discussed principles, standards, and 
regulations play a significant role in ensuring that the development goes in the desired 
direction. Despite the recent developments, a lot remains to be done. (Thompson 2021, 
241; Financial Times 2022.)  
One broadly highlighted significant challenge is the increasing risk of greenwashing. It 
means that the issuers may seek to take advantage of the investors, following the impact 
investing guidelines, by labelling the issued bonds as ‘green’ but not committing to the 
promotion of environmental issues (Thompson 2021, 30, 249). This can also be described 
as a ‘moral hazard’ problem caused by asymmetric information between the issuers and 
the investors. The rapid expansion of the green bond market has only accelerated the 
concerns of both regulators and investors. (Financial Times 2022.) Although principles, 
 
13 The TEG also published a recommendation for a voluntary EU Green Bond Standard (EU GBS) in June 
2019 and updated them in March 2020 (European Commission 2022c). However, as of writing this thesis, 
it is not in application.  
27 
 
standards, and regulations are already widely used, they are not yet sufficient to eliminate 
the threat of greenwashing altogether. According to CBI (2022f), in 2022, 3 in every 4 
dollars from green bond issuances met the best practice climate standards. However, as 
there still are so-called ‘self-labelled’ green bonds that are not aligned with the current 
standards and are not externally verified, the risk of greenwashing persists. Of course, the 
mitigation of greenwashing also requires greater criticalness on the part of investors, but 
responsibility cannot be entirely theirs to carry. 
One of the problems of the current standards is the voluntary nature of their use, which 
can lead to governance and legitimacy deficits. Consequently, there have been 
discussions about whether mandatory standards, instead of voluntary ones, should be 
applied to prevent greenwashing better.14 However, there are also fears that mandatory 
standards would increase the costs of issuing green bonds, and affect the willingness to 
use them as a source of green financing. (Financial Times 2022.) China’s recent 
reformation of their national green bond principles to make them more globally aligned 
and the possible update to make them mandatory may serve as a possible benchmark for 
the best practices in the future. Furthermore, if the new approach is put to force and found 
to work well, it could also be applied in other nations. Therefore, China has the 
opportunity to act as a kind of pioneer in the development of the green bond market. 
Speaking of China, it has particularly high stakes regarding green bonds. China's 
president Xi Jinping's government has the twin goals of peaking emissions by 2030 and 
achieving carbon neutrality by 2060. (Bloomberg 2022b; Reuters 2022.) If these goals 
are achieved, according to Climate Action Tracker (2020), it would be the biggest single 
reduction in global warming projections in history. However, the current reality with 
green bonds, especially in China, is that it is almost impossible to know how the capital 
is spent and whether it is having the desired impact. This issue is evident, for example, in 
the case of financial institution issuers that use the proceeds from green bonds to lend 
capital to relevant projects, meaning that they do not have their own projects to finance 
directly. In China, financial institutions cover one-third of the issued green bonds. In 
 
14 See e.g. European Parliament (2021). 
28 
addition, the Chinese green bond market also suffers from the risk of greenwashing and 
general uncertainty.15 (Bloomberg 2022b.) 
Globally, there is a need for better harmonization of the green bond market. For investors, 
the existence of several diverging standards can lower confidence and increase transaction 
costs as there is a need to assess multiple different standards. However, global 
harmonization is easier said than done and requires vigorous dialogue among all the 
market participants. (German Development Institute 2017.) It would be particularly 
desirable to standardize the taxonomy globally to have a universal and unambiguous 
understanding of what green bonds are and what they can be used for. Multiple separate 
taxonomies can lead to market fragmentation, limit the functioning of the global green 
bond market and make international mobility of green capital challenging. So far, this has 
not been achieved, but there are some collective efforts, such as the International Platform 
on Sustainable Finance (IPSF), developed by the EU and China to serve as a foundation 
for a common ground taxonomy.16 (CBI 2022g.) However, as the green bond market 
grows, the puzzle becomes increasingly complex. Therefore, the action should be as swift 
as possible. 
3.5 Previous studies 
As the green bond market evolves at an accelerating pace and in the presence of several 
challenges, it is interesting to assess how investors view green bond issuances. Besides, 
the investors who buy the bonds, are the ones lending the money to the issuers. One way 
to assess this is to examine the impact of green bond issuance announcements on the stock 
returns of the listed corporate issuers. It essentially describes how the equity market reacts 
to green bonds and whether investors see the issuance as a positive or negative sign under 
current market conditions. This research approach also makes it possible to assess 
whether the listed issuers, on top of gaining debt financing for operations, get extra 
benefits from issuing green bonds. For example, issuing green bonds might lead to 
increased investor engagement, as green bonds might serve as a signal for the 
environmental commitment of the issuer (Flammer 2021). There are some previous 
studies on this topic, and most of them, although not all, have found green bonds to have 
 
15 According to Bloomberg (2022b), in a rare interview, even the Chinese PBOC Governor Yi Gang has 
warned against greenwashing, low-cost arbitrage, and green project fraud in China's green bond market. 
16 IPFS is a forum for dialogue between policymakers to increase the amount of private capital being 
invested in environmentally sustainable investments (European Commission 2022d). 
29 
 
a positive impact on the stock returns of the issuer. Some studies have also examined 
whether there are any green bond or issuer characteristics that would be linked to these 
abnormal returns.  
Baulkaran (2019) studied the stock market reaction to green bond issuance 
announcements with observations mainly from Europe but also from Canada, the US, 
China, and Australia. The final sample consisted of 54 green bonds issued by public 
corporates, and the research method was an event study. The issuance announcement 
dates were had-collected for each observation using news articles and company press 
releases. The results suggest that the stock market reaction is positive and statistically 
significant around the announcement date. A regression analysis was also conducted, to 
further examine the link between abnormal returns and bond and firm characteristics. The 
only bond characteristic linked with the returns was the coupon, and the result suggested 
that a higher coupon resulted in lower abnormal returns. Out of the examined firm 
characteristics, operating cash flows, firm size, and growth opportunities had statistically 
significant linkage. (Baulkaran 2019.) 
Lebelle et al. (2020) examined corporate green bond issuances globally. The initial 
sample was 2,079 green bond issuances of 190 individual corporate issuers from 2009 to 
2018. The final sample for the empirical study contained 475 green bonds issued by 145 
individual public corporates. The research method was an event study, and the observed 
event was the green bond issuance announcement date. Contrary to the majority of the 
results from similar studies, the results suggest that the market reacts negatively to the 
green bond issuance announcement. Thus, the results imply that green bond issuances 
convey unfavourable information about the issuing corporation to the investors. In 
addition, the negative response was stronger for first-time issuances compared to 
subsequent issuances, and for issuers in developing markets compared to emerging 
markets. They also found that financial corporates experienced a stronger negative impact 
than non-financial corporates, although the magnitude of the differences was small. They 
also examined the dependency of abnormal returns on issuer characteristics and found 
leverage, book-to-market ratio, and growth to have a statistically significant linkage. 
(Lebelle et al 2020.) 
Tang and Zhang (2020) examined green bonds globally from the shareholder's benefit 
perspective. Their initial sample was 1,510 observations from 28 countries from 2007 to 
30 
2017. Of these, 241 green bonds were issued by 132 unique public corporates. The chosen 
research method was an event study, and the observed event was the issuance 
announcement date. They found green bond issuance announcements to have a positive 
and significant impact on the stock prices of public issuers. The impact was stronger for 
non-financial corporate issuers compared to financial institution issuers. Also, the 
response was stronger for first-time issuances. In addition, they found green bonds to 
increase institutional ownership and improve stock liquidity after the issuance. They 
concluded that by issuing green bonds, it is possible to attract more positive media 
exposure, and through this, investors that follow the guidelines of impact investing are 
more likely to invest in the issuing company. (Tang & Zhang 2020.) 
Glavas (2020) examined stock price reactions to green bond issuance announcements 
globally. The data consisted of 302 corporate green bonds and 478 conventional bonds, 
which were added to the analysis to allow comparison. The research method was an event 
study. The results implied a significant and positive reaction to all bond issuance 
announcements. However, the reaction was higher for green bond issuances compared to 
conventional bond issuances. This result suggests that the announcement of the green 
bond issuance contains value-generation information and is perceived as a value-creating 
event. In addition, the positive stock reaction was stronger after the Paris Agreement17 in 
2015, suggesting that the agreement has been key to the interest of investors in green 
bonds. (Glavas 2020.) 
Wang et al. (2020) examined the market reaction to green bond issuances in China, the 
largest emerging debt market. They also studied whether green bonds have a pricing 
premium over conventional bonds. They constructed a comprehensive sample of Chinese 
corporate green bonds issued between 2016 and 2019. The final sample consisted of 159 
green bonds and 297 conventional bonds. Of these, 48 green bonds and 75 conventional 
bonds, were issued by publicly listed issuers. The research method for observing the 
market reaction was an event study. After conducting the event analysis for both green 
and conventional bonds, they found that the difference in abnormal returns on the 
announcement date was insignificant. However, for longer time windows around the 
event, the cumulative abnormal returns were significant, positive and higher for green 
 
17 The Paris Agreement is a legally binding global treaty on climate change adopted by 196 countries in 
Paris on 12 December 2015 (UNFCCC 2022). 
31 
 
bonds. Thus, the results suggest that the equity market reacts positively to the green bond 
issuance announcement in China. In addition, the pricing effect for green bonds was found 
to be positive and significant. (Wang et al. 2020.) 
Flammer (2021) examined corporate green bonds globally between 2013 and 2018. The 
initial sample consisted of 1,189 corporate green bonds. Of these, 565 were issued by 
public corporates. An event study method was used to study the equity market reaction to 
the green bond issuance announcements. The final sample for the event study consisted 
of 384 issuer-day green bond observations. The results suggest that the market response 
was positive and significant to the issuance announcement. Also, it was stronger for first-
time issuers and green bonds, which were certified by independent third parties. In 
addition, an interesting finding was that companies that issued green bonds improved 
their environmental performance, meaning higher environmental ratings and lower CO2 
emissions, following the issuance, suggesting that there is no sign of greenwashing. The 
green bond issuers also experienced an increase in ownership by green and long-term 
investors. (Flammer 2021.) 
As presented above, previous studies suggest largely consistent results from the equity 
market response to the green bond issuance announcement. Overall, the results imply a 
positive reaction. However, since not all results are unanimous, further research is 
necessary. Also, the green bond market is constantly generating new data and the market 
experiences regulatory changes rapidly. Therefore, examining the equity market response 
to the green bond issuance announcements remains meaningful, and studying this topic 
is important for both investors and issuers. 
32 
4 Data and research methods 
4.1 Data  
The data is retrieved from Refinitiv Eikon. The green bonds data in Refinitiv Eikon is 
originally from Climate Bonds Initiative (CBI) (Refinitiv 2021). For now, there is no 
centralized database for green bonds. ICMA (2017) provides a summary of the available 
green bond database providers, the major ones being Bloomberg, Environmental Finance, 
Dealogic, and CBI. Lebelle et al. (2020) add Trucost, owned by S&P Global, to this list. 
Out of these, the CBI's database is the oldest, established in 2013, and it is also the most 
comprehensive provider of green bond data. A green bond has to be aligned with the 
GBPs to be included in the database, adding credibility to the data. Thus, CBI’s database 
has been utilized in several previous studies on green bonds (See e.g. Tang & Zhang 2020; 
Lebelle et al. 2020; Tolliver et al. 2020). 
The data is global, and the focus is on green bonds issued by listed corporates between 
2013 and 2021. The green bonds data is retrieved using Refinitiv’s Government Corporate 
Bonds – Advanced Search app. Issuer Type is set to Corporate, and Bond Type to Bond.18 
Green bond criterion is set to ‘Yes’, and Prospectus is needed to be available.19 With these 
selections, the initial sample retrieved from the app, consisted of a total of 2,761 
individual corporate green bonds between 2013 and 2021. Table 3 displays further 
screening criteria of the green bond sample. Since the study examines the green bond 
issuance announcements’ effect on stock returns, criterion No. 1 is set to exclude private 
issuers and asset types such as unit trusts and preferred shares from the sample. Criteria 
No. 2 and 3 are set to ensure that the sample is reliable and allows for a comprehensive 
analysis. Criterion No. 4 certifies that the data contains only pure green bonds and not, 
for example, sustainability-linked bonds. Criterion No. 5 is essential for the chosen 
research method. Since some corporations issue multiple green bonds on the same date, 
similar to Flammer (2021), criterion No. 6 is set to include only issuer-day observations 
in the sample. 
 
 
 
18 Certificates of Deposits and Commercial papers are excluded. 
19 Prospectus is a legal document that provides, for example, details of the use of proceeds. 
33 
 
Table 3 Screening criteria for green bonds 
Criterion Description 
No. 1 Issuer listed in the stock market with the asset type being ordinary 
share.  
 
No. 2 Required green bond data: First announcement date, Issuance date, 
Amount issued, Coupon, Maturity, Currency & Use of Proceeds. 
 
No. 3 Required issuer data: Issuer Name, TRBC Sector Classification, 
Domicile, Accounting Data (Assets, Liabilities etc.) & Refinitiv ESG 
score. 
 
No. 4 Issuer ESG Label, the original label used by the issuer for a green bond 
in the prospectus or official source, is available and set to ‘Yes’. 
 
No. 5 Stock prices are available at least 282 trading days prior and 20 trading 
days after the First Announced Date. 
 
No. 6  Only unique issuer-day observations are included.  
 
The final green bond sample contains 564 green bond issuances from 31 countries by 272 
publicly listed corporate issuers. Table 4 presents the observations by country and 
economic sector. Regionally, based on the amount issued, Europe is the first (USD 123.4 
billion), North America is the second (USD 63.8 billion), and Asia-Pacific is the third 
(USD 45.9 billion). The regional distribution corresponds fairly well to the volume 
distribution of the 2021 green bond market, where Europe contributed half of the volume 
of the year, Asia-Pacific around one quarter, and North America just under one-fifth. At 
the end of 2021, the cumulative volume of green bonds issued by financial and non-
financial corporates was around USD 704 billion. Not all of these corporates are listed. 
(CBI 2021b.) Given that, in this study, only listed corporates are included, the sample can 
be considered rather comprehensive, with a total amount of USD 234.3 billion issued. 
However, it is also good to point out that the data availability has imposed some 
constraints. 
 
34 
Table 4 Green bond sample by country and economic sector 
This table presents the green bond sample used in the empirical study by country and economic sector, 
sorted by the amount issued. Others row in Panel A includes countries with an amount issued under USD 
1 billion, including UAE, Turkey, Hong Kong, South Africa, Switzerland, New Zealand, India, Brazil, and 
Chile. Panel B presents the sample distribution by The Refinitiv® Business Classification (TRBC) 
Economic Sectors. In total, there are 13 different TRBC Economic Sectors.  
  
Amount Issued 
(USD Billion) 
% of 
Total 
# of Green 
Bonds 
% of 
Total 
Panel A: By Country 
US 62.7 26.77 % 98 17.38 % 
China (Mainland) 30.6 13.07 % 43 7.62 % 
France 25.8 11.03 % 44 7.80 % 
Germany 23.1 9.84 % 45 7.98 % 
Spain 19.1 8.15 % 39 6.91 % 
Italy 14.6 6.24 % 24 4.26 % 
Japan 11.0 4.69 % 83 14.72 % 
Norway 8.4 3.58 % 24 4.26 % 
United Kingdom 7.1 3.03 % 13 2.30 % 
Sweden 6.0 2.57 % 57 10.11 % 
Denmark 4.7 2.01 % 7 1.24 % 
Austria 2.9 1.24 % 11 1.95 % 
Ireland 2.3 0.96 % 3 0.53 % 
Australia 2.2 0.95 % 4 0.71 % 
Belgium 1.8 0.75 % 3 0.53 % 
Portugal 1.8 0.75 % 2 0.35 % 
Finland 1.7 0.72 % 7 1.24 % 
Netherlands 1.6 0.67 % 2 0.35 % 
Taiwan 1.4 0.61 % 26 4.61 % 
Canada 1.1 0.47 % 3 0.53 % 
Poland 1.1 0.46 % 3 0.53 % 
Greece 1.0 0.43 % 2 0.35 % 
Other 2.4 1.01 % 21 3.72 % 
Panel B: By Economic Sector 
Financials 115.7 49.37 % 230 40.78 % 
Utilities 53.1 22.67 % 92 16.31 % 
Real Estate 34.9 14.90 % 143 25.35 % 
Technology 9.5 4.04 % 20 3.55 % 
Consumer Cyclicals 8.6 3.68 % 15 2.66 % 
Industrials 4.8 2.03 % 34 6.03 % 
Basic Materials 3.1 1.34 % 13 2.30 % 
Consumer Non-Cyclicals 2.2 0.94 % 8 1.42 % 
Energy 2.2 0.92 % 8 1.42 % 
Healthcare 0.2 0.10 % 1 0.18 % 
Total 234.3 100 % 564 100 % 
35 
 
The economic sector classification in Table 4 is based on the TRBC (The Refinitiv 
Business Classification) Sector Classification, which covers over 250 000 securities in 
130 countries to 5 levels of granularity. According to Refinitiv, it is the most 
comprehensive sector and industry classification available. A significant part of the 
sample is from the Financials sector. It consists of four different business sectors, which 
are Banking & Investment Services, Insurance, Collective Investments, and Investment 
Holding Companies. (Refinitiv 2022b.) A difference between the non-financial and 
financial sectors is that the non-financial issuers use the proceeds from green bonds to 
finance green projects directly, whereas financial issuers use the proceeds to finance 
projects of other corporates' through loans (green lending) (see e.g. Lebelle et al. 2020; 
Wang et al. 2020; Fatica et al. 2021). Especially in Asia, the green bond market is 
currently dominated by banks in the financial sector (Taghizadeh-Hesary et al. 2021; 
Bloomberg 2022b). The majority of the observations from China, Germany, Spain, Japan, 
Norway, the United Kingdom, Sweden, Denmark, Austria, Taiwan, Canada, and Poland 
are from the financial sector. Also, Ireland, Australia, Belgium, UAE, South Africa, 
Switzerland, India, and Chile have observations only from financial corporates. This 
indicates that the financial sector is also prominent outside of Asia.   
Panel A in Table 4 presents the summary statistics of the included green bonds. The mean 
of the amount issued is USD 415.4 million, but the amounts vary substantially. Compared 
to the total green bond market, CBI (2022b) reports the average size of a green bond being 
USD 250 million in 2021. The large mean of the final sample is explained by the large 
green bond issuances by Chinese financial corporates, with the biggest being USD 4.1 
billion. The sample's largest non-financial corporate green bond is issued by an 
automobile manufacturer Ford Motor Company in the US. An interesting observation is 
a green bond with 1000-year maturity issued by the Danish energy corporation Orsted in 
2019. Orsted is the world's largest offshore wind farm developer and issued a 1000-year 
maturity bond also earlier in 2017 (Insider 2017). The coupon rates of the green bonds 
are scattered moderately evenly between the minimum and maximum values. The sample 
includes 29 zero-coupon green bonds.  
Panel B in Table 4 presents summary statistics of the individual issuers. The total amount 
of issuers is 272. The size of the issuers varies substantially, measured both by market 
capitalization and total assets, and thus, the variance of the size measures are large. The 
leverage ratios (debt-to-equity) are overall moderate, with a mean of 0.73. Moreover, all 
36 
the ratios are under one, meaning that the issuers have more equity over debt. The return 
on assets (ROA) is quite low for all the issuers. For some, the ROA is negative, indicating 
that some issuers may be generating losses. Tobin’s Q is used to measure growth 
opportunities. For many issuers, it is under 1, suggesting that the market value is lower 
than the book value of assets. (Tobin 1969, 1978; Baulkaran 2019.) Overall, the size 
measures, leverage, ROA, and Tobin’s Q of the issuers in this study are largely in line 
with those reported by, for example, Baulkaran (2019), Lebelle et al. (2020), and Tang 
and Zhang (2020). To a large extent, the ESG and pillar scores are high for the issuers, 
although some are surprisingly low. One could expect the green bond issuers to have a 
higher environmental rating. However, an interesting finding is that the environmental 
pillar score does not differ substantially from the other pillars. 
Table 5 Summary statistics of green bonds and issuers  
This table presents the means, medians, minimums, maximums, and standard deviations of the green bond 
and issuer characteristics of the final sample observations. Panel A displays the green bond characteristics, 
and Panel B the issuer characteristics. All the monetary values are in USD. Maturity does not include the 
entire sample since four of the observed green bonds are perpetual bonds meaning they do not have a 
maturity date. The issuer characteristics are from the end of the fiscal year 2021. Leverage is defined as 
total liabilities divided by total assets. ROA (Return on Assets) is defined by income after taxes for the 
fiscal year divided by the average of total assets. Tobin's Q is a measure of growth opportunities and is 
defined as the market value of equity plus the book value of debt divided by the book value of assets. ESG 
score and pillar scores are Refinitiv ESG Scores. All the issuers are listed corporates with publicly traded 
shares. 
USD N Mean Median  Stdev  Min Max 
Panel A: Green Bonds  
Amount Issued (Million) 564 415.406 350.000 479.112 0.006 4143.704 
Maturity at Issue (years) 560 10.175 6.669 42.614 0.722 1000.000 
Annual Coupon % 564 1.789 1.375 1.557 0.000 8.850 
Panel B: Issuers  
Market Cap (Billion) 272 38.878 10.632 159.529 0.237 2403.239 
Total Assets (Billion) 272 290.723 39.174 705.855 0.742 5536.969 
Leverage 272 0.727 0.736 0.185 0.247 0.967 
ROA 272 0.037 0.021 0.056 –0.044 0.557 
Tobin's Q 272 0.734 0.653 0.701 0.051 7.156 
ESG score 272 66.353 70.202 17.477 17.497 95.392 
Environmental Pillar 272 69.211 72.915 21.317 0.000 99.200 
Social Pillar 272 67.482 72.845 20.767 10.347 98.315 
Governance Pillar 272 61.067 65.773 21.556 6.988 97.004 
 
37 
 
4.2 Research methods 
4.2.1 Event study 
The event study method has a long history in finance research. The first forms of event 
studies can be seen published as early as the early 1930s. However, Ball and Brown 
(1968) and Fama et al. (1969) can be considered pioneers of its current form. An event 
study is an applicable way to measure the effects of specific events on different financial 
variables. The event can be, for example, an announcement related to stock splits, 
dividends or mergers and acquisitions. (MacKinlay 1997; Brooks 2014, 634.) In this 
study the event under scrutiny is the first announcement date of the green bond issuance 
available in the Refinitiv Eikon database. The frequency of the data for event study can 
be weekly, daily, or monthly. However, daily is the most used frequency for studies in 
the literature as it has been shown to have greater power to detect abnormal performance. 
(MacKinlay 1997; Brooks 2014, 635.) The frequency of choice for this study is also daily, 
which comforts the nature of the event.   
Figure 3 illustrates the structure of an event study. 0 represents the day of the event, which 
in this case, is the first announcement date. The estimation window determines the 
expected or normal return for security if the event did not happen. The length of the 
estimation window for this study is [–281, –30], equivalent to 252 trading days (one 
trading year) before the event. The event window comprises the days to be examined to 
capture the possible abnormal returns around the event. The length of the event window 
can vary, and it is typical to examine a few different windows. In this study, the chosen 
event windows are a 41-day window [–20, +20], 21-day window [–10, +10], 11-day 
window [–5, +5], and 3-day window [–1, +1]. Out of these, [–5, +5] is considered as the 
baseline event window. In addition, a 10-day pre-event window [–10, –1] before the event 
and an 11-day post-event window [+10, +20] after the event are observed. The pre-event 
window aims to examine whether there is a possible leakage of information prior to the 
announcement. On the contrary, the post-event window aims to examine whether there is 
a possible delay in the equity market reaction. (MacKinlay 1997; Brooks 2014, 635–636.) 
38 
 
Figure 3 Event study structure (modified from MacKinlay 1997)  
 
The underlying aim of an event study is to determine if the cross-sectional distribution of 
returns is abnormal at the time of a specific event (Kothari & Warner 2007). According 
to Efficient Market Hypothesis (EMH), it is impossible to gain abnormal returns if the 
markets are informationally efficient, as all the asset prices reflect all available 
information (Pilbeam 2018, 229). Thus, event studies are testing whether the efficient 
market hypothesis holds. A simple way of calculating abnormal returns is  
 𝐴𝑅𝑖𝑡 =  𝑅𝑖𝑡– 𝐸(𝑅𝑖𝑡), (1) 
where 𝐴𝑅𝑖𝑡 is the abnormal return, 𝑅𝑖𝑡 is the actual return, and 𝐸(𝑅𝑖𝑡) is the expected or 
normal return for security 𝑖 for time period 𝑡. This approach can be referred to as the 
average return model. (Armitage 1995; MacKinlay 1997; Brooks 2014, 636.)  
Numerous more sophisticated approaches have also been created for event studies to 
increase the robustness of the results. In this thesis, the utilized method is a statistical 
market model. The market model is probably the most common approach for estimating 
abnormal returns. It is also widely used in previous green bond event studies (see e.g. 
Baulkaran 2019; Lebelle et al. 2020; Flammer 2021). In the market model, the expected 
(or normal) return is estimated using a one-factor ordinary least squares (OLS) regression. 
The market model for security 𝑖 is defined as 
 𝑅𝑖𝑡 = 𝛼𝑖 + 𝛽𝑖𝑅𝑚𝑡 + 𝜀𝑖𝑡 (2) 
 𝐸(𝜀𝑖𝑡 = 0) 
 𝑣𝑎𝑟(𝜀𝑖𝑡) = 𝜎𝜀𝑖
2 , 
where 𝛼𝑖, 𝛽𝑖 and 𝜎𝜀𝑖
2  are the parameters of the market model, 𝑅𝑖𝑡 and  𝑅𝑚𝑡 are the returns 
for security 𝑖 and market portfolio 𝑚 for time period 𝑡, and 𝜀𝑖𝑡 is the zero mean error term. 
Typically, major stock indices are used to proxy for the market portfolio. (MacKinlay 
39 
 
1997; Brooks 2014, 637.) In this study, a major stock market index of each country under 
scrutiny is utilized. The used country-specific indices are listed in the appendices.  
Returns for securities and markets are calculated using the Total Return Index (RI) 
retrieved from Refinitiv Datastream. RI accounts for dividends by assuming that they are 
re-invested, which mitigates possible large price fluctuations caused by dividend 
payments. The returns are calculated using a trade-to-trade approach, where the 
calculation is done from non-missing price days.20 Another option would be to use 
lumped returns which consist of trade-to-trade returns on non-missing price days and zero 
on missing price days. However, Campbell et al. (2010) suggest that using trade-to-trade 
returns is sufficient in the multi-county setting. If the event day, the first announcement 
date, is a non-trading day, the event is assumed to be the next trading day. The returns are 
calculated using both local currencies and USD, and the event study is conducted with 
both currencies.21  
The returns are calculated as daily logarithmic returns (log-returns) with the following 
equation: 
 𝑅𝑡 = 100% × ln (
𝑅𝐼𝑡
𝑅𝐼𝑡−1
), (3) 
where 𝑅𝐼𝑡 is the total return index for trading day 𝑡 and 𝑅𝐼𝑡−1 is the total return index for 
𝑡– 1. The benefit of log-returns is that they are analytically more trackable and can be 
interpreted as continuously compounded returns, which is also a crucial attribute for the 
event study. Also, log-returns are more likely to be normally distributed, which comforts 
the assumptions of statistical techniques. Thus, they are commonly used in the academic 
finance literature. (Strong 1992; Brooks 2014, 7–8.)  
The OLS estimators for market model parameters α̂, ?̂?and ?̂?2 are  
 ?̂?𝑖 =
∑ (𝑅𝑖𝑡– ?̂?𝑖)(𝑅𝑚𝑡−?̂?𝑚)
𝑇1
𝑡=𝑇0+1
∑ (𝑅𝑚𝑡−?̂?𝑚)2
𝑇1
𝑡=𝑇0+1
 (4) 
 
20 When using Refinitiv Datastream's RI for calculating returns, it is important to note that it ignores market 
holidays (non-trading days). One way to deal with this is to use the function 
X(RI)*IF#(X(P#S),NNA,ONE), which gives a negative value to the non-trading days. After this, the 
negative values can be highlighted and excluded from the data. 
21 Campbell et al. (2010) report that the use of the local-currency market model abnormal returns is 
sufficient in a multi-country setting. However, according to Aktas et al. (2004), the use of a common 
currency, more specifically the USD, does not have a major impact on the inferences of the results. 
40 
 ?̂?𝑖 =  ?̂?𝑖– ?̂?𝑖?̂?𝑚 (5) 
 ?̂?𝜀𝑖
2 =
1
𝐿1−2
∑ (𝑅𝑖𝑡 − ?̂?𝑖 − ?̂?𝑖𝑅𝑚𝑡)
2𝑇1
𝑡=𝑇0+1
, (6) 
where 𝐿1 is the length of the estimation window, and  ?̂?𝑖 and ?̂?𝑚 are means of the returns 
during estimation window for security 𝑖 and market portfolio 𝑚. After estimating the 
parameters, the values are used to calculate the abnormal returns. The equation for 
abnormal return (AR) is  
 𝐴𝑅𝑖𝑡 = 𝑅𝑖𝑡– (?̂?𝑖  +  ?̂?𝑖𝑅𝑚𝑡). (7) 
In the market model, the abnormal returns are disruptions from normal returns estimated 
through the market portfolio. By using the ARs it is also possible to calculate average 
abnormal returns (AAR) by  
 𝐴𝐴𝑅𝑡 =
1
𝑁
∑ 𝐴𝑅𝑖𝑡
𝑁
𝑖=1 , (8) 
where 𝑁 is the number of the observed events. (Armitage 1995; MacKinlay 1997; Brooks 
2014, 637–638.) 
To conclude the overall pattern of abnormal returns in the specified event window, a 
cumulative abnormal return (CAR) over a multi-period window can be calculated. 
Equation for CAR from time 𝑡1to 𝑡2 is 
 𝐶𝐴𝑅𝑖(𝑡1, 𝑡2) = ∑ 𝐴𝑅𝑖𝑡
𝑡2
𝑡=𝑡1
. (9) 
Thus, CAR sums daily abnormal returns together in the specified event window. In this 
study, CAR is used to summarize the results of each examined event window. By using 
CARs, we can calculate cumulative average abnormal returns (CAARs) with the 
following equation: 
 𝐶𝐴𝐴𝑅𝑖(𝑡1, 𝑡2) =
1
𝑁
∑ 𝐶𝐴𝑅𝑖(𝑡1, 𝑡2)
𝑁
𝑖=1 , (10) 
where 𝑁 is the number of individual observations. CAAR can also be calculated as the 
sum of AARs over a specific window. 
41 
 
In the market model, the null hypothesis, 𝐻0, is that the abnormal return is zero meaning 
that the event has no impact on the stock returns. That is, the abnormal returns are 
normally distributed with a zero conditional mean and conditional variance of 
 𝜎2(𝐴𝑅𝑖𝑡) =  𝜎𝜀𝑖
2 +
1
𝐿1
[1 +
(𝑅𝑚𝑡−?̂?𝑚)
2
?̂?𝑚
2 ] (11) 
where 𝜎𝜀𝑖
2  is the disturbance variance from Equation (2). (Fama et al. 1969; MacKinlay 
1997; Brooks 2014, 638–640) The zero mean null hypothesis also applies for this event 
study. 
Two different statistical tests are applied to test the significance of the event study results. 
The first is the parametric t-test. The cross-sectional t-test is given by  
 𝑡 =
𝐴𝐴𝑅/𝐶𝐴𝐴𝑅
𝜎/√𝑁
 ~N(0,1), (12) 
where the numerator is either AAR or CAAR, and the denominator is the standard error 
(S.E.) of the numerator, calculated by dividing standard deviation of the observations 𝜎 
with the square root of the number of observations 𝑁. (Brown and Warner 1985, Brooks 
2014, 639–640; Stock & Watson 2020, 113.)  
Campbell et al. (2010) have reported that in multi-country settings, the non-parametric 
tests, such as generalized sign and rank tests, appear to be more powerful than the 
commonly utilized parametric tests such as t-tests. Thus, the second statistical test is the 
non-parametric generalized sign test introduced by Cowan (1992). It examines whether 
the number of securities with positive CARs (or ARs) in the event window (or on an 
observed day) exceeds the number expected in the absence of abnormal performance. The 
expected number is based on the amount of positive abnormal returns in the 252-day 
estimation period, 
 ?̂? =
1
N
∑
1
252
∑ 𝑆𝑖𝑡
−30
𝑡=−281
𝑁
𝑖=1 ,  
where 
 𝑆𝑖𝑡 = {
1 𝑖𝑓 𝐴𝑅𝑖𝑡 > 0 
0 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
  
The generalized sign test uses the normal approximation to the binomial distribution with 
estimated parameter ?̂?. The final test statistic is 
42 
 𝑍𝐺𝑆 =
𝑤−𝑁?̂?
√𝑁?̂?(1−?̂?)
 , (13) 
where 𝑤 is the number of securities in the event window (or on the observed day) with 
positive CAR (or AR). (Cowan 1992; Campbell et al. 2010.) 
4.2.2 Regression analysis 
To further examine the cumulative abnormal returns, a regression analysis is conducted. 
The aim is to identify possible significant links between the different green bond and 
issuer characteristics and CARs. In other words, by choosing relevant characteristics, the 
goal is to explain the positive or negative equity market reaction more comprehensively. 
Regression analysis also allows us to interpret which type of corporates benefit or 
disbenefit from the green bond issuance announcements. (Brooks 2014, 640.) A similar 
approach has also been utilized in previous green bond studies, and some links have been 
identified (see e.g. Baulkaran 2019; Lebelle et al. 2020). Due to the diverging nature of 
the use of proceeds in the financial sector, the financial and non-financial corporates are 
examined as separate groups. Also, this enables highlighting possible differences between 
the financial and non-financial corporate green bond issuers. 
By using ordinary least squares (OLS) regression, the following regression is estimated: 
 𝐶𝐴𝑅𝑖(𝑡1, 𝑡2) = 𝛽0 + 𝛽1𝐴𝑀𝑇𝐼𝑆𝑆𝑈𝐸𝐷𝑖 + 𝛽2𝐶𝑂𝑈𝑃𝑂𝑁𝑖 + 𝛽3𝑀𝐴𝑇𝑈𝑅𝐼𝑇𝑌𝑖 +
                                         𝛽4𝐹𝐼𝑅𝑆𝑇_𝐷𝑖 + 𝛽5𝑆𝐼𝑍𝐸𝑖 + 𝛽6𝐿𝐸𝑉𝐸𝑅𝐴𝐺𝐸𝑖 +
                                         𝛽7𝑃𝑅𝑂𝐹𝐼𝑇𝐴𝐵𝐼𝐿𝐼𝑇𝑌𝑖 + 𝛽8𝐺𝑅𝑂𝑊𝑇𝐻𝑖 + 𝛽9𝐸𝑆𝐺𝑖 +
                                         𝐶𝑜𝑢𝑛𝑡𝑟𝑦_𝐹𝐸 + 𝑌𝑒𝑎𝑟_𝐹𝐸 + 𝜀𝑖 ,       (14) 
where the dependent variable 𝐶𝐴𝑅𝑖 is the cumulative abnormal return for an event 
window. For robustness, different event windows for CARs are tested. The explanatory 
variables are different green bond and issuer characteristics. Green bond characteristics 
include 𝐴𝑀𝑇𝐼𝑆𝑆𝑈𝐸𝐷, 𝐶𝑂𝑈𝑃𝑂𝑁, 𝑀𝐴𝑇𝑈𝑅𝐼𝑇𝑌, and 𝐹𝐼𝑅𝑆𝑇_𝐷, out of which the last is a 
dummy variable, also known as a qualitative variable. Issuer characteristics include 𝑆𝐼𝑍𝐸, 
𝐿𝐸𝑉𝐸𝑅𝐴𝐺𝐸, 𝑃𝑅𝑂𝐹𝐼𝑇𝐴𝐵𝐼𝐿𝐼𝑇𝑌, 𝐺𝑅𝑂𝑊𝑇𝐻, and 𝐸𝑆𝐺. 𝛽0, 𝛽1, … , 𝛽9 are the coefficient 
estimates which quantify the effect of each explanatory variable. 𝐶𝑜𝑢𝑛𝑡𝑟𝑦_𝐹𝐸 and 
𝑌𝑒𝑎𝑟_𝐹𝐸 are country and year fixed effect dummy variables (Brooks 2014, 529). 
The fixed effects are added to account for the leftover variation that is due year and 
country-specific factors. Similar approach was used by, for example, Lebelle et al. (2020). 
43 
 
𝜀𝑖 is an error term that sweeps up any influences on the CARs that are not captured by the 
explanatory variables. (Brooks 2014, 135, 157, 684.)  
Table 6 Explanatory variables 
Green bond characteristics 
AMTISSUED The natural logarithm of the green bond amount issued in USD 
million. 
COUPON The coupon rate of the green bond. 
MATURITY The natural logarithm of the maturity of the green bond in years. 
FIRST_D Dummy variable equal to 1 if the green bond is the first for the 
issuer and 0 if the green bond is subsequent. 
Issuer characteristics 
SIZE The natural logarithm of the issuer’s total assets the fiscal year prior 
to the issuance announcement in USD million. 
LEVERAGE The ratio between issuer’s total liabilities and total assets the fiscal 
year prior the issuance announcement. 
PROFITABILITY Measured by Return on Assets (ROA) the fiscal year prior to the 
issuance announcement. ROA is calculated as the income after 
taxes for the fiscal period divided by the average total assets at the 
beginning and at the end of the year. 
GROWTH Measured by Tobin’s Q the fiscal year prior to the issuance 
announcement. Tobin’s Q is calculated by summing up the market 
value of equity and the book value of debt and dividing it by the 
book value of assets. 
ESG The Refinitiv® ESG Score. 
 
Table 6 provides detailed descriptions of the explanatory variables. Logarithmic 
transformation is done to the AMTISSUED, MATURITY and SIZE variables, as it 
reduces heteroscedasticity and makes skewed distribution closer to a normal distribution 
(Brooks 2014, 34). The measure used as a proxy for the profitability of the issuer in the 
PROFITABILITY variable is the return on assets (ROA). Similarly to Baulkaran (2019), 
Tobin’s Q is used as a proxy for the issuer's growth opportunities in the GROWTH 
variable. The currency for the regression analysis is USD, meaning that the CARs used 
44 
as the dependent variable are from an event study conducted in USD.22 Thus, all the 
explanatory variables are calculated using USD to make the values more uniform and 
comparable. 
Table 7 Descriptive statistics for regression variables  
This table presents descriptive statistics for the regression variables. Panel A displays the descriptive 
statistics for the group with non-financial corporates. Panel B displays the descriptive statistics for the group 
with financial corporates. 
  Mean Median Stdev Kurtosis Skewness Min Max 
Panel A: Non-Financials (N = 332) 
AMTISSUED 5.249 5.365 1.236 –0.950 –0.291 1.920 7.824 
COUPON 1.941 1.875 1.444 0.113 0.614 0.000 7.637 
MATURITY 2.016 1.900 0.690 10.689 1.610 0.700 6.900 
FIRST_D 0.449 0.000 0.498 –1.969 0.207 0.000 1.000 
SIZE 9.801 9.597 1.288 0.460 0.684 6.419 13.470 
LEVERAGE 0.613 0.606 0.141 –0.510 0.014 0.258 0.948 
PROFITABILITY 0.043 0.033 0.040 2.724 0.896 –0.122 0.205 
GROWTH 1.080 0.928 0.657 14.812 3.354 0.324 5.816 
ESG 60.755 64.905 20.313 –0.318 –0.693 5.861 93.934 
Panel B: Financials (N = 228) 
AMTISSUED 5.288 6.215 1.904 6.460 –1.899 –5.139 8.329 
COUPON 1.560 0.953 1.692 3.673 1.759 0.000 8.850 
MATURITY 1.749 1.700 0.540 4.685 0.487 –0.300 3.500 
FIRST_D 0.399 0.000 0.491 –1.844 0.415 0.000 1.000 
SIZE 12.752 12.929 1.630 –0.969 –0.405 8.543 15.447 
LEVERAGE 0.906 0.932 0.076 6.471 –2.573 0.558 0.967 
PROFITABILITY 0.009 0.006 0.010 12.478 2.843 -0.007 0.074 
GROWTH 0.289 0.194 0.249 9.070 2.486 0.054 1.836 
ESG 71.194 74.982 17.506 0.085 –0.910 20.825 94.811 
 
Table 7 displays the descriptive statistics for the regression variables. The statistics are 
presented for both non-financial and financial sector groups. The green bond 
characteristics are similar to both groups, except for the MATURITY variable. The 
maturity of green bonds issued by non-financial corporates tends to be higher than those 
by financial corporates. A higher positive kurtosis and skewness of some variables 
indicates that the distributions have a long and ‘fat’ right tails (Brooks 2014, 66–67). 
From the issuer characteristics, the LEVERAGE variable has a notable difference 
between the two groups. The financial sector corporates are more leveraged than the non-
 
22 The descriptive statistics for the used CARs are displayed in appendices. 
45 
 
financial corporates. Higher leverage is common for financial corporates as they typically 
borrow capital from those with a surplus to invest and lend to those in need (Ingves 2014). 
Thus, they also typically issue green bonds to lend capital for other corporates’ green 
projects (Lebelle et al. 2020; Wang et al. 2020; Fatica et al. 2021). Other noteworthy 
differences between the groups are in GROWTH and ESG variables. The non-financial 
corporates have higher Tobin’s Q values indicating bigger growth opportunities. The ESG 
scores are lower for the non-financial sector, suggesting that ESG creditors see them as 
less sustainable, although the scores are still relatively high.   
Table 8 presents correlation matrices for both financial and non-financial sector 
regression groups. The correlation matrix serves as one way of estimating 
multicollinearity in the regression model. An assumption in the OLS estimation is that 
the explanatory variables are not correlated with one another. If a correlation is present, 
removing a correlated variable from the regression would cause the coefficients of the 
other variables to change. However, in practice, the correlation between variables will be 
non-zero, and a problem occurs only when there is a high correlation. Perfect 
multicollinearity would arise when two or more variables have an exact relationship 
meaning that one regressor is a perfect linear function of the other regressor(s). In this 
case, computing the OLS estimates would be impossible. (Brooks 2014, 217–218; Stock 
& Watson 2020, 226.) However, this is not the case in this study.   
Although perfect multicollinearity is not present, some of the variables have a higher 
correlation with each other. For the non-financial group, AMTISSUED and SIZE have a 
correlation of 0.55. It can be explained by the fact that larger corporates tend to issue 
more sizeable green bonds as they might have bigger-scale green projects that need debt 
financing. The financial group has a few strong correlations between the issuer 
characteristics. The highest correlation is between PROFITABILITY and LEVERAGE, 
which suggests that when financial corporates’ leverage increases, the profitability 
decreases and vice versa. A strong correlation is also detected between GROWTH, 
LEVERAGE and PROFITABILITY, indicating that when the leverage ratio increase, 
growth opportunities decrease, whereas an increase in profitability increases growth 
opportunities. Also, the SIZE variable is positively correlated with LEVERAGE, 
meaning that the bigger the bank the higher the leverage ratio, and negatively correlated 
with PROFITABILITY and GROWTH. To further address the multicollinearity, the 
variance inflation factors (VIFs) are discussed in the result section. 
46 
Table 8 Correlation matrices  
This table presents correlation matrices for the explanatory variables of the two regression groups. Panel A displays a correlation matrix for the variables used in regressions for 
non-financial corporates. Panel B displays a correlation matrix for the variables used in regressions for financial corporates. 
Panel A: Non-Financials 
  (1) (2) (3) (4) (5) (6) (7) (8) (9) 
(1) AMTISSUED 1         
(2) COUPON 0.077 1        
(3) MATURITY 0.423*** 0.059 1       
(4) FIRST_D 0.009 –0.105* –0.036 1      
(5) SIZE 0.547*** –0.124** 0.325*** –0.044 1     
(6) LEVERAGE 0.269*** 0.023 0.089 –0.003 0.341*** 1    
(7) PROFITABILITY –0.124** –0.078 –0.095* –0.079 –0.092* –0.314*** 1   
(8) GROWTH 0.051 –0.095* –0.001 0.057 –0.044 –0.318*** 0.391*** 1  
(9) ESG 0.365*** –0.155*** 0.205*** –0.061 0.183*** –0.129** 0.205*** 0.114** 1 
Panel B: Financials 
  (1) (2) (3) (4) (5) (6) (7) (8) (9) 
(1) AMTISSUED 1         
(2) COUPON 0.016 1        
(3) MATURITY 0.057 –0.063 1       
(4) FIRST_D 0.166** 0.039 –0.009 1      
(5) SIZE 0.169** –0.229*** –0.032 –0.250*** 1     
(6) LEVERAGE –0.016 –0.043 –0.036 –0.087 0.473*** 1    
(7) PROFITABILITY –0.022 0.191*** –0.051 0.081 –0.517*** –0.853*** 1   
(8) GROWTH 0.053 0.055 0.001 0.117*** –0.494*** –0.782*** 0.727*** 1  
(9) ESG –0.080 –0.326*** 0.296*** –0.279*** 0.239*** –0.098 –0.065 0.035 1 
Significance levels: * p < 10%  ** p < 5%  *** p < 1% 
47 
 
5 Results and discussion 
5.1 Event study 
The event study was conducted using two approaches. The first approach was calculating 
market model abnormal returns in local currencies, and the second was the same with the 
difference of converting all the values into USD. Figure 4 illustrates the differences 
between average abnormal returns (AARs) and cumulative average abnormal returns 
(CAARs) over the 41-day event window [–20, +20] using local currency and USD 
approaches. Using USD as a common currency to calculate the abnormal returns in a 
multi-country setting appears to make the results slightly inflated. Overall, the AARs 
follow the same pattern between both approaches with only slight differences. The 
inflated effect of the common currency is illustrated better by the CAARs. CAARs 
calculated with values converted to the USD are more extreme than the ones calculated 
using local currencies. However, overall, the results between the two approaches have 
only minor differences and lead to similar inferences. Therefore, this chapter presents the 
rest of the event study results from the local currency approach.23 
In addition to reflecting differences in the use of local and common currency in the event 
study, Figure 4 illustrates the initial results for the main objective of the event study, the 
equity market reaction to the green bond issuance announcement. AAR is negative on the 
announcement date 0. However, AAR spikes nine days before the event and stays positive 
for a few days after, which might indicate some information leakage before the event. 
The reason can be that investors in the equity market have received information about 
green bond issuance from other communication channels and thus anticipated the event 
before the official announcement. On the other hand, the reason may also relate to the 
accuracy of the first announcement date available in the Refinitiv Eikon database. 
Regardless, the equity market reaction appears to be positive. The positive reaction is 
visualized better with the CAARs over the event window [–20, +20], which stays positive 
from seven days prior to five days after the event. However, it is important to point out 
that the positive impact is very short-term, and the cause for the positive CAARs around 
the event is the positive spike in AARs prior to the announcement. 
 
23 The overall results from the USD approach are presented in the appendices. 
48 
 
Figure 4 Abnormal returns using local currency vs. USD 
 
Table 9 provides a more in-depth representation of the daily abnormal returns over the 
11-day baseline window [–5, +5] for the total green bond sample. Five days before the 
green bond issuance announcement, the AAR is significant and positive, as indicated in 
the graphs above, with a value of 0.121%. The parametric t-test for the result is significant 
at a 10% significance level, and the non-parametric generalized sign test is significant at 
a 5% level. The sign test practically tests whether the number of positive ARs departs 
from the expected amount of positive values estimated from the estimation period. Hence, 
the 'Pos:Neg' column shows that the number of positive values on the day –5 is greater 
than for the other days. Otherwise, the values for daily AARs are small and statistically 
insignificant, although, on the event date, the AAR is negative with a value of –0.075%. 
The standard deviations are similar for all the days and do not suggest significantly 
different reactions in the daily ARs. 
 
 
 
49 
 
Table 9 Daily abnormal returns 
This table presents the daily abnormal returns (ARs) and their means (AARs), medians, standard deviations, 
minimums, and maximums for the 11-day baseline event window [–5,+5] for the total green bond sample 
(N=564). The values are displayed in percentages. Additionally, the number of positive and negative 
abnormal returns for each day is reported in a separate column. The last two columns present the statistical 
tests for the daily abnormal returns. The t-test is the cross-sectional parametric test, and the sign test is the 
non-parametric generalized sign test.   
Day AAR Median Stdev Min  Max Pos:Neg t-test Sign test 
–5 0.121 0.059 1.559 –10.065 8.820 302:262 1.850* 2.049** 
–4 –0.091 –0.105 1.374 –5.378 6.936 265:299 –1.568 –1.067 
–3 0.053 –0.070 1.445 –5.344 8.679 270:294 0.868 –0.646 
–2 0.006 0.008 1.332 –7.059 6.764 286:278 0.107 0.702 
–1 –0.017 0.048 1.458 –8.461 5.903 287:277 –0.278 0.786 
0 –0.075 –0.054 1.438 –6.774 7.179 270:294 –1.240 –0.646 
+1 0.008 0.047 1.381 –5.641 9.400 294:270 0.136 1.375 
+2 0.040 0.047 1.501 –7.255 13.537 293:271 0.628 1.291 
+3 –0.061 –0.029 1.500 –12.260 7.695 275:289 –0.973 –0.225 
+4 0.079 0.013 1.418 –5.331 10.034 286:278 1.327 0.702 
+5 –0.083 –0.062 1.379 –7.602 6.520 268:296 –1.427 –0.815 
Significance levels: * p < 10%  ** p < 5%  *** p < 1% 
 
Table 10 aggregates the AARs in the form of CAARs for four event windows, a pre-event 
window, and a post-event window. The results are not statistically significant, which is 
expected as only one result is significant in Table 9. Despite the insignificance, they 
provide some additional indication of the equity market reaction over different intervals. 
For example, the potential information leakage before the green bond issuance 
announcement is supported by and detectable from the 10-day pre-event window [–10, –
1]. The pre-event window CAAR is positive with a value of 0.224%. The CAARs of other 
windows are considerably smaller and range from both sides to zero. However, only one 
of the actual event windows has a positive CAAR. It is a 21-day window [-10, +10], 
which also benefits from the possible information leakage. The baseline event window [–
5, +5] and the shortest event window [–1, +1] both have negative CAARs indicating a 
negative equity market reaction near the observed green bond issuance announcement 
date. 
 
 
 
50 
Table 10 Cumulative average abnormal returns 
This table presents the cumulative average abnormal returns (CAARs) for different windows for the total 
green bond sample (N=564). Medians, standard deviations, minimums, and maximums for the cumulative 
abnormal returns (CARs) are also presented. The values are displayed in percentages. [–10, –1] is the 10-
day pre-event window and [+10, +20] is the 11-day post-event window. The other windows are the actual 
event windows. Additionally, the number of positive and negative CARs is displayed. The last two columns 
present the statistical tests for the cumulative abnormal returns in each event window. The T-test is the 
cross-sectional parametric test, and the sign test is the non-parametric generalized sign test. 
Window CAAR Median Stdev Min  Max Pos:Neg t-test Sign test 
[–10, –1] 0.224 0.277 4.692 –22.370 25.704 292:272 1.136 1.207 
[–20, +20] –0.061 0.018 9.568 –52.049 33.242 282:282 –0.153 0.365 
[–10, +10] 0.021 0.153 6.973 –49.601 37.901 285:279 0.073 0.617 
[–5, +5] –0.020 0.119 4.698 –22.275 27.474 292:272 –0.101 1.207 
[–1, +1] –0.084 –0.082 2.328 –12.672 13.020 272:292 –0.860 –0.478 
[+10, +20] 0.021 –0.142 4.406 –23.085 26.021 270:294 0.112 –0.646 
Significance levels: * p < 10%  ** p < 5%  *** p < 1% 
 
Overall, the findings from the event study using the total green bond sample are not 
unambiguous. However, there is a fairly strong indication that the shares of the issuers 
experience positive abnormal returns before the green bond issuance announcement date. 
This suggests that the equity market reacts positively to the issue of the green bond, but 
there may be some kind of anticipation or information leakage regarding the 
announcement. Also, as mentioned earlier, the date available in Refinitiv Eikon may be 
inaccurate. Baulkaran (2019) also reports a negative AAR on the event date, but all the 
CAARs for observed event windows are positive. Thus, the results resemble the findings 
of this study in some ways by suggesting information leakage before the announcement.  
The results are compared between different subsamples and regions to utilize the full 
potential of the global dataset and make the event study analysis more comprehensive. 
First, Table 11 compares the event study results by splitting the total sample in three ways. 
Panel A displays the results for financial and non-financial sector corporates. Similarly to 
Tang and Zhang (2020) and Lebelle et al. (2020), the comparison is done due to the 
diverging natures in the use of proceeds. For the most part, the results are contradictory 
and suggest a more positive and significant reaction to the green bond issuance 
announcement for non-financial corporates. It is seen especially over the shortest event 
window [–1, +1], where financial corporates experience a negative abnormal return of –
0.295% at a 10% significance level measured with the t-test. However, a positive reaction, 
although much more significant for the non-financials, is detectable for both groups in 
51 
 
the pre-event window indicating information leakage. The results resemble the findings 
by Tang and Zhang (2020), suggesting a stronger positive reaction for non-financial 
corporates.   
Table 11 Comparison between subsamples 
This table presents the cumulative average abnormal returns (CAARs) for different subsamples over 
different windows. The subsamples are formed by splitting the total sample based on different 
characteristics. Panel A displays the CAARs for financial and non-financial sectors. The split is based on 
The Refinitiv® Business Classification (TRBC) Economic Sectors. Panel B displays the CAARs for first-
time and subsequent green bond issuances. Panel C displays CAARs for green bonds issued in advanced 
economies and for emerging markets & developing economies. The split is based on International Monetary 
Fund’s (IMF) country classification (IMF 2022b). The CAAR values are displayed in percentages. The T-
test is the cross-sectional parametric test, and the sign test is the non-parametric generalized sign test. 
Panel A: Financial Sector vs. Other Sectors 
 Financial Sector (N = 230) 
 Non-Financial Sectors (N = 334) 
  CAAR t-test Sign test   CAAR t-test Sign test 
[–10, –1] 0.066 0.229 –0.231  0.333 1.246 1.760* 
[–20, +20] 0.372 0.638 1.088  –0.360 –0.656 –0.429 
[–10, +10] 0.214 0.503 0.956  –0.111 –0.277 0.009 
[–5, +5] –0.082 –0.255 –0.231  0.023 0.092 1.760* 
[–1, +1] –0.295 –1.810* –1.155  0.061 0.505 0.337 
[+10, +20] –0.062 –0.218 –1.287   0.077 0.315 0.228 
Panel B: First-time vs. Subsequent 
 First-time (N = 242) 
 Subsequent (N = 322) 
  CAAR t-test Sign test   CAAR t-test Sign test 
[–10, –1] 0.206 0.663 1.224  0.238 0.932 0.537 
[–20, +20] –0.035 –0.058 –0.062  –0.081 –0.151 0.537 
[–10, +10] 0.229 0.521 0.324  –0.135 –0.342 0.537 
[–5, +5] 0.267 0.842 1.995**  –0.235 –0.937 –0.132 
[–1, +1] –0.227 –1.605 –0.962  0.023 0.169 0.202 
[+10, +20] 0.014 0.051 –0.448   0.026 0.103 –0.467 
Panel C: Advanced Economies vs. Emerging & Developing Markets 
 Advanced Economies (N = 502) 
 Emerging & Developing (N = 62) 
  CAAR t-test Sign test   CAAR t-test Sign test 
[–10, –1] 0.217 1.050 1.453  0.286 0.433 –0.496 
[–20, +20] –0.218 –0.514 –0.332  1.206 0.958 2.050** 
[–10, +10] –0.098 –0.321 0.293  0.990 1.010 1.031 
[–5, +5] –0.161 –0.776 0.293  1.119 1.763* 2.813*** 
[–1, +1] –0.103 –1.000 –0.243  0.071 0.232 –0.751 
[+10, +20] 0.038 0.196 –0.511   –0.123 –0.210 –0.496 
Significance levels: * p < 10%  ** p < 5%  *** p < 1% 
 
52 
Panel B in Table 11 compares the results between first-time and subsequent green bond 
issuances. A similar comparison was done by Tang and Zhang (2020), Lebelle et al. 
(2020) and Flammer (2021). A notable difference is seen over the baseline event window 
[–5, +5], where the CAAR is positive and significant for the first-time issuances with a 
value of 0.267% at a 5% significance level measured with the sign test. For the subsequent 
issuances, the result is insignificant but negative with a value of –0.235%. The results 
resemble the studies by Tang and Zhang (2020) and Flammer (2021), suggesting a more 
positive reaction for first-time issuances. It can be expected as the information of green 
bonds is entering the market for the first time in the first-time issuance. However, the 
result is the opposite when viewing the window [-1, +1], which may be due to the early 
reaction. Also, when looking at the pre-event window, the reaction is positive for both 
groups and even slightly higher for the subsequent issuances. It, again, supports the 
information leakage finding. 
Similarly to Lebelle et al. (2020), Panel C in Table 11 compares the results between 
advanced economies and emerging & developing markets. Again, both of the groups 
experience positive CAARs over the pre-event window. However, the event windows 
reveal a substantial difference. Advanced economies experience negative CAARs over 
all the event windows, whereas emerging & developing markets have much higher and 
positive CAARs. In the baseline window [–5, +5], the positive CAAR of 1.119% is also 
statistically significant at a 10% level measured with the t-test and significant at a 1% 
level measured with the sign test. Although the figures reported by Lebelle et al. (2020) 
are negative for both groups, the results are consistent in that the negative reaction was 
less negative for emerging & developing markets. However, it is also worth noting that 
the advanced economy group contains 502 observations, while the development & 
emerging markets group contains only 62. As the number of observations increases in the 
event study, abnormal returns tend to approach zero, while in smaller samples possible 
outliers may cause some bias in the result. 
To the best of my knowledge, no previous studies have reported regional differences in 
the equity market reaction to the green bond issuance announcement. Thus, Table 12 
presents event study results for the three leading regions involved in the green bond 
markets. These are Europe, Asia-Pacific, and North America. Since the diverging nature 
of financial corporates' use of proceeds, on top of providing the total results, the results 
are also reported excluding the financial sector to add robustness. Panel A  displays the 
53 
 
results of the equity market reaction to European green bond issuance announcements. 
The results are overall insignificant and negative. Even the earlier identified positive 
reaction over the pre-event window is practically non-existent. Thus in Europe, the equity 
market seems to react negatively to the green bond issuance announcement. One reason 
could be investor scepticism considering the validity of green bonds in Europe. Investors 
may not be confident that the issuers would use the proceeds as they should. It may be 
due, for example, to the voluntary nature of different standards. On the other hand, Europe 
can already be seen as a forerunner of the green transition24 and as a result, investors may 
not have a particular incentive to find new, sustainable investments from Europe. 
The results of North America presented in Panel C are generally in line with the results 
observed from Europe. The results are not statistically significant and are negative for all 
event windows. However, there is a great difference between pre-event and post-event 
windows, with both showing positive reactions. The CAAR over the pre-event window 
could indicate that investors in North America do react positively to the green bond 
issuance announcement but with anticipation. Although the CAARs are also positive over 
the post-event window, the negative sign test value proposes that there are more negative 
than positive CARs. Thus, the positive mean is likely caused by some exceptionally 
positive abnormal returns after the issuance. Overall in North America, the equity market 
reaction is positive over days before and negative around the event, although statistically 
insignificant. 
A significant difference is distinguished in the equity market reaction in Asia-Pacific, 
reported in Panel B. For Asia-Pacific, the CAAR is positive over all the windows except 
for the shortest event window for the total sample. The baseline window [–5, +5] indicates 
a positive CAAR of 0.484% at a 10% significance level measured with the sign test. 
Furthermore, when excluding the financial corporates, the CAAR jumps to 1.025% at a 
5% significance level measured with the t-test and a 10% level measured with the sign 
test. Also, the CAAR of 1.433% over the event window [–10, +10] for the sample 
excluding financials is significant at a 10% level measured with the t-test. The results 
from Asia-Pacific are mostly led by observations from China. It has recently done the 
most substantial updates to the national green bond standards, and it seems that the 
investors have more confidence in the Chinese green bond issuer. China is also part of 
 
24 See e.g. Forbes (2021). 
54 
the emerging markets, which partly explains the positive CAAR observations in the 
emerging & developing markets in Table 11. Overall, the results indicate a strong positive 
equity market reaction to green bond issuance announcements in Asia-Pacific. Thus, the 
result is consistent with the findings of Wang et al. (2020), suggesting a positive reaction 
in China.  
Table 12 Comparison between regions 
This table presents the cumulative average abnormal returns (CAARs) for different regions over different 
windows. The results are shown for the total samples and samples excluding the financial sector. Panel A 
displays the CAARs for Europe, Panel B for Asia-Pacific, and Panel C for North America. The CAAR 
values are displayed in percentages. The t-test is the cross-sectional parametric test, and the sign test is the 
non-parametric generalized sign test. 
Panel A: Europe 
 Total (N = 289)  Excluding Financial Sector (N = 153) 
 CAAR t-test Sign test  CAAR t-test Sign test 
[–10, –1] 0.047 0.177 0.458  –0.003 –0.008 0.796 
[–20, +20] –0.406 –0.699 –0.013  –1.074 –1.287 –0.175 
[–10, +10] –0.307 –0.722 0.340  –0.885 –1.459 0.149 
[–5, +5] –0.215 –0.775 –0.601  –0.415 –1.188 0.472 
[–1, +1] –0.049 –0.361 –0.484  0.080 0.512 0.472 
[+10, +20] –0.205 –0.766 0.105  –0.174 –0.453 0.957 
Panel B: Asia-Pacific 
 Total (N = 162)  Excluding Financial Sector (N = 96) 
 CAAR t-test Sign test  CAAR t-test Sign test 
[–10, –1] 0.406 1.002 0.167  0.893 1.502 0.865 
[–20, +20] 0.845 1.149 1.267  1.503 1.372 0.660 
[–10, +10] 0.678 1.254 0.167  1.433 1.799* 0.252 
[–5, +5] 0.484 1.427 1.739*  1.025 2.082** 1.886* 
[–1, +1] –0.115 –0.609 –0.462  0.198 0.799 0.048 
[+10, +20] 0.302 0.903 –0.777  0.367 0.728 –0.565 
Panel C: North America 
 Total (N = 101)  Excluding Financial Sector (N = 84) 
 CAAR t-test Sign test  CAAR t-test Sign test 
[–10, –1] 0.444 1.017 1.621  0.230 0.472 1.393 
[–20, +20] –0.528 –0.577 –0.569  –1.159 –1.265 –1.226 
[–10, +10] –0.139 –0.213 0.426  –0.440 –0.657 –0.353 
[–5, +5] –0.368 –0.691 1.223  –0.358 –0.727 0.738 
[–1, +1] –0.158 –0.673 0.227  –0.157 –0.599 –0.135 
[+10, +20] 0.290 0.738 –0.768  0.192 0.530 –0.353 
Significance levels: * p < 10%  ** p < 5%  *** p < 1% 
 
 
55 
 
5.2 Regression analysis 
The regression model was tested with three different event windows for the non-financial 
and financial groups to highlight possible links between the cumulative abnormal returns 
(CARs) caused by the green bond issuance announcement and several green bond and 
issuer characteristics. Additionally, the regression model was tested for the three leading 
regions in the green bond markets to highlight possible regional differences. White’s test 
was used to test the regressions for heteroscedasticity (White 1980). Heteroscedasticity 
is present if the variance of the errors is not constant throughout the sample. If this was 
the case, any inferences made could be misleading. (Brooks 2014, 181, 185.) For the most 
part, White’s test did not suggest an alarming level of heteroscedasticity. However, for a 
few of the regressions, some heteroscedasticity was detected. The White’s test results are 
reported separately in each result table. Multicollinearity was already addressed with the 
correlation matrices. However, to further analyse the multicollinearity of the regression 
models, the variance inflation factors (VIFs) of the independent variables were 
interpreted. Overall, the VIFs did not show signs of harmful multicollinearity25.  
The regression results for the non-financial group are presented in Table 13. Regression 
1 includes all the explanatory variables, whereas Regressions 2 and 3 are conducted to 
capture the effect of the green bond (2) and issuer (3) characteristics on the CARs. For 
the most part, the results are insignificant. The GROWTH variable has some significance 
when using CARs for the event window [–10, +10] as the dependent variable. However, 
the reported White’s tests indicate the presence of heteroscedasticity, and therefore, these 
results may be misleading. Also, the AMTISSUED, SIZE and LEVERAGE show some 
significance in Regression 1. However, when the bond and issuer characteristics are 
regressed separately, the significance of the results diminishes. Also, the R-squared (𝑅2), 
which measures how well the explanatory variables explain the variance of CARs, is very 
low for all the tested regressions (Stock & Watson 2020, 223). Thus, the results in Table 
13 suggest that there are no significant links between the CARs and the characteristics of 
green bond and issuers in the non-financial group.  
 
25 A general rule of thumb is that a VIF over 10 is a sign of harmful multicollinearity (CFI 2022).  
56 
Table 13 Regression results for non-financial group 
This table presents the regression results for non-financial sector observations. Regression 1 includes all the variables, Regression 2 includes only green bond characteristics, 
and Regression 3 only issuer characteristics. T-statistics are in parentheses. χ² is the chi-squared test statistic from White's heteroscedasticity test.   
  Regression 1   Regression 2   Regression 3 
  [–10, +10] [–5, +5] [–1, +1]   [–10, +10] [–5, +5] [–1, +1]  [–10, +10] [–5, +5] [–1, +1] 
Intercept –0.113 –0.165 –0.010  –0.195 –0.095 0.026  –0.150 –0.285 –0.045 
 (–0.243) (–0.573) (–0.070)  (–0.435) (–0.342) (0.194)  (–0.335) (–1.024) (–0.328) 
AMTISSUED –0.004 –0.006* –0.002*  –0.000 –0.002 –0.001  
   
 (–0.797) (–1.927) (–1.673)  (–0.043) (–0.687) (–1.261)  
   
COUPON –0.002 0.001 0.000  –0.003 0.000 –0.000  
   
 (–0.583) (0.330) (–0.355)  (–0.882) (–0.040) (–0.470)  
   
MATURITY 0.000 –0.001 0.001  0.002 0.001 0.001  
   
 (–0.007) (–0.139) (0.401)  (0.270) (0.134) (0.594)  
   
FIRST_D 0.005 0.003 –0.004  0.003 0.002 –0.004  
   
 –0.639 (0.654) (–1.435)  (0.375) (0.457) (–1.493)  
   
SIZE 0.007* 0.003 0.000      0.006 0.001 –0.000 
 (1.661) (1.331) (0.245)      (1.573) (0.356) (–0.438) 
LEVERAGE –0.008 0.038* 0.010      –0.013 0.033 0.007 
 (–0.219) (1.788) (0.978)      (–0.367) (1.547) (0.702) 
PROFITABILITY –0.019 –0.052 –0.040      –0.005 –0.029 –0.024 
 (–0.162) (–0.711) (–1.106)      (–0.043) (–0.404) (–0.679) 
GROWTH –0.015** –0.004 –0.001      –0.015** –0.005 –0.002 
 (–2.102) (–0.842) (–0.567)      (–2.214) (–1.185) (–0.997) 
ESG 0.000 0.000 0.000      0.000 0.000 0.000 
  –0.293 (1.172) (1.112)           (0.024) (0.429) (0.691) 
Observations 332 332 332  332 332 332  332 332 332 
𝑅2 0.032 0.040 0.030  0.004 0.003 0.014  0.027 0.026 0.013 
χ² (White’s test) 18.616* 11.386 13.300  6.068 5.946 6.619  14.792** 6.224 11.333 
Country FE Yes Yes Yes  Yes Yes Yes  Yes Yes Yes 
Year FE Yes Yes Yes  Yes Yes Yes  Yes Yes Yes 
Significance levels: * p < 10%  ** p < 5%  *** p < 1% 
57 
 
Table 14 presents the regression results for the financial corporate group. Similarly to the 
non-financial group, for the most part, the results are insignificant. However, using the 
event windows [–10, +10] and [–1, +1] as the dependent variables, the SIZE variable 
shows significance at a 10% level in Regressions 1 and 3. The positive coefficients for 
the firm size, measured by the issuer’s total assets, indicate that the positive equity market 
reaction is stronger for the green bond issuance announcements when the financial 
corporates are larger. Another variable with statistical significance is ESG, measured by 
ESG scores. The coefficient is very close to zero, suggesting that the ESG scores have no 
linkage to the CARs whatsoever. Again, however, the R-squared (𝑅2) is very low for all 
the tested regressions. It indicates that the chosen characteristics have a low explanatory 
power over the abnormal returns experienced from the green bond issuance 
announcements and that the effect size is very small.  
Overall, there is a notable lack of statistical significance in Tables 13 and 14. Thus, it 
seems that for the non-financial and financial groups, the CARs around the green bond 
issuance announcement have no clear dependency on the chosen characteristics. The 
finding is somewhat contrary to the previous research. For example, Lebelle et al. (2020) 
results suggest that issuers with more growth opportunities have reduced negative 
reactions to green bond issuance. Also, Baulkaran (2019) found growth opportunities to 
have a positive and statistically significant linkage between the CARs. In addition, 
Although, from the bond characteristics, a higher coupon was found to have a negative 
linkage between the abnormal returns. In this study, growth opportunities seemed to have 
some significance for the non-financial group, but with the presence of heteroscedasticity, 
the results cannot be interpreted with confidence.   
 
 
 
 
 
 
58 
Table 14 Regression results for financial group 
This table presents the regression results for financial sector observations. Regression 1 includes all the variables, Regression 2 includes only green bond characteristics, and 
Regression 3 only issuer characteristics. T-statistics are in parentheses. χ² is the chi-squared test statistic from White's heteroscedasticity test.   
  Regression 1   Regression 2   Regression 3 
  [–10, +10] [–5, +5] [–1, +1]   [–10, +10] [–5, +5] [–1, +1]  [–10, +10] [–5, +5] [–1, +1] 
Intercept –1.204** –0.504 –0.062  –0.441 –0.162 0.051  –0.986* –0.404 –0.125 
 (–2.072) (–1.155) (–0.270)  (–0.987) (–0.485) (0.291)  (1.868) (–1.015) (–0.598) 
AMTISSUED 0.000 0.002 –0.000  0.001 0.002 –0.000  
   
 (0.114) (0.883) (–0.266)  (0.275) (1.026) (–0.111)  
   
COUPON 0.002 –0.000 –0.001  0.001 –0.001 –0.001  
   
 (0.798) (–0.132) (–0.894)  (0.336) (–0.616) (–1.347)  
   
MATURITY –0.008 –0.005 –0.001  –0.009 –0.005 –0.000  
   
 (–0.933) (–0.818) (–0.147)  (–1.036) (–0.809) (–0.078)  
   
FIRST_D 0.002 0.007 0.001  0.003 0.007 –0.000  
   
 (0.173) (0.989) (0.369)  (0.297) (1.090) (–0.115)  
   
SIZE 0.007* 0.004 0.002      0.006* 0.004 0.003* 
 (1.780) (1.387) (1.642)      (1.751) (1.577) (1.873) 
LEVERAGE –0.054 –0.110 –0.037      –0.036 –0.120 –0.044 
 (–0.431) (–1.164) (0.748)      (–0.289) (–1.294) (–0.894) 
PROFITABILITY –0.414 –0.366 0.304      –0.274 –0.485 0.250 
 (–0.448) (–0.528) (0.828)      (–0.306) (–0.719) (0.703) 
GROWTH –0.002 –0.018 –0.016      –0.000 –0.013 –0.016 
 (0.061) (–0.831) (–1.394)      (–0.009) (–0.620) (–1.409) 
ESG –0.000** –0.000 –0.000      –0.001** –0.000* –0.000 
  (–2.294) (–1.402) (–0.212)      (–2.105) (–1.808) (–0.627) 
Observations 228 228 228  228 228 228  228 228 228 
𝑅2 0.04 0.041 0.036  0.008 0.020 0.009  0.033 0.029 0.032 
χ² (White’s test) 9.494 7.069 12.356  7.955 6.010 6.127  7.107 4.469 10.367 
Country FE Yes Yes Yes  Yes Yes Yes  Yes Yes Yes 
Year FE Yes Yes Yes  Yes Yes Yes  Yes Yes Yes 
Significance levels: * p < 10%  ** p < 5%  *** p < 1% 
59 
 
To analyse the linkage between the CARs and the chosen characteristics further, and to 
make the regression analysis more comprehensive, the top three green bond regions are 
examined separately. Table 15 presents the regression for the different regions, with both 
non-financial and financial groups. The results seem to follow the same trend as the 
statistical significance for the green bond and issuer characteristics is very low. The only 
significant results at a 10% significance level are for variables FIRST_D and GROWTH 
in Europe. First-time issuances have a positive coefficient, and growth opportunities have 
a negative coefficient for European financial corporates. However, since White’s test 
indicates some heteroscedasticity for the financial group regression, the results need to be 
interpreted with caution. However, for the non-financial group in Europe, the growth 
opportunities also have a negative coefficient. That is, non-financial corporates with 
higher growth opportunities experience lower equity market reactions to green bond 
issuance announcements in Europe. The reason can be, for example, that investors may 
have scepticism about the sustainability of growth-oriented corporates. In Asia-Pacific 
and North America, no significant linkages are detected. 
In total, results of 24 different regression models are presented. Overall, if anything, the 
regression analysis results suggest that there is a very minimal linkage between the CARs 
from the green bond issuance announcement and the bond and issuer characteristics. In 
other words, the results suggest that investors are not particularly interested in some 
characteristics of a green bond or their issuers over others. Thus, even if investors may 
show some interest in listed corporates that issue green bonds, their interest does not seem 
to reach other factors which would lead to significantly higher or lower abnormal returns. 
For robustness, regressions were also tested without fixed effects and with different 
amounts of explanatory variables. However, there were no great differences in the results 
that would lead to different inferences. The result applies to all the groups analysed. 
 
 
 
 
 
60 
Table 15 Regional regression results 
This table presents regional regression results for non-financial and financial sectors. (1) represents the non-
financial sector regression and (2) the financial sector regression. T-statistics are in parentheses. χ² is the 
chi-squared test statistic from White’s heteroscedasticity test. 
[–5, +5] Europe   Asia-Pacific   North America 
  (1) (2)   (1) (2) 
 (1) (2) 
Intercept –0.030 0.523**  –0.347** –0.192  0.178 –1.326 
 (–0.143) (2.134)  (–2.488) (–0.927)  (1.226) (–0.706) 
AMTISSUED –0.003 0.000  –0.009 0.006  –0.014 –0.011 
 (–0.631) (0.189)  (–1.169) (1.297)  (–0.851) (–0.507) 
COUPON –0.001 –0.004  0.009 –0.001  0.001 0.121 
 (–0.315) (–1.130)  (1.451) (–0.149)  (0.222) (1.395) 
MATURITY –0.000 0.005  –0.001 –0.008  0.000 –0.123 
 (–0.008) (0.554)  (–0.078) (–0.944)  (0.030) (–1.878) 
FIRST_D –0.000 0.018*  0.016 0.006  –0.016 0.143 
 (–0.012) (1.699)  (1.430) (0.693)  (–1.514) (1.415) 
SIZE 0.003 0.007  0.003 –0.001  –0.008 0.114 
 (0.600) (1.473)  (0.427) (–0.409)  (–1.139) (1.943) 
LEVERAGE –0.002 –0.097  0.037 0.119  0.04 –2.262 
 (–0.057) (–0.741)  (0.899) (0.661)  (0.867) (–1.153) 
PROFITABILITY –0.118 0.295  –0.190 0.608  0.370 –2.795 
 (–1.165) (0.349)  (–0.975) (0.384)  (2.326) (–0.198) 
GROWTH –0.027* –0.050*  –0.004 0.031  –0.010 –0.124 
 (–1.718) (–1.659)  (–0.473) (0.915)  (–1.038) (0.583) 
ESG 0.000 –0.000  0.000 –0.000  0.000 0.004 
  (0.983) (–1.206)   (1.243) (–1.415)   (0.072) (0.664) 
Observations 151 134  96 66  84 17 
𝑅2 0.080 0.108  0.211 0.133  0.134 0.620 
χ² (White’s test) 12.010 17.420*  21.453** 12.051  7.969 12.643 
Country FE Yes Yes  Yes Yes  Yes Yes 
Year FE Yes Yes  Yes Yes  Yes Yes 
Significance levels: * p < 10%  ** p < 5%  *** p < 1% 
 
61 
 
6 Conclusions 
This thesis has examined how the equity market reacts to green bond issuance 
announcements of listed corporates. In other words, the aim was to see how investors in 
the equity market react to the information about green bond issuances. The scope of the 
study was global. The second objective was to investigate possible links between the 
abnormal returns caused by the green bonds and different bond and issuer characteristics. 
First, the sustainable finance movement was discussed, to give the reader an explicit 
background to the topic. Second, the relationship between sustainable finance and the 
debt capital markets was considered, and after this, the current sustainable debt 
instruments, other than green bonds, were introduced.   
The green bonds were discussed extensively by first giving a detailed definition and then 
moving on to the global green bond market development. The green bond market has seen 
considerable popularity growth in recent years. It has also led to an increasing need for 
better principles, standards and regulations. Thus, to give a better understanding of the 
current state of green bonds, the most notable green bond principles, standards and 
regulations were discussed. Since the green bond is still a new and evolving sustainable 
debt instrument, there are also existing challenges and problems, such as the risk of 
greenwashing. These were also discussed, by also highlighting the differences in different 
parts of the world. Before moving to the empirical study, similar previous studies were 
introduced. 
In total, 564 corporate green bond observations from 31 countries and ten economic 
sectors were selected for the empirical study with specific criteria. The first part of the 
study was an event study using a market model approach. It was conducted in two ways 
by using local currencies and USD as a common currency. The results suggest that both 
approaches lead to similar results with no diverging inferences. The equity market 
reaction to the green bond issuance announcement is not entirely unambiguous but tends 
to lean more toward the positive reaction. A key finding is that the listed issuers 
experience positive CARs over the 10-day pre-event window [–10, –1] when analysing 
the total sample. The same result was also seen in the examined subsamples, although not 
statistically significant. A stronger and more positive reaction was detected for the non-
financial corporates, first-time issuances, and issuers in emerging & developing markets. 
Regionally, the most significant finding was a strong positive reaction for the non-
62 
financial corporates in Asia-Pacific. The worst reaction was detected in Europe. The 
regression results overall showed little to no linkage between the CARs and the green 
bond and issuer characteristics. The result was the same for non-financial and financial 
corporates and different regions. 
The results of this study had some similarities to the previous studies. However, it also 
discovered some new findings and highlighted more regional differences. The overall 
findings suggest that the equity market reaction can be either positive or negative 
depending on, for example, the issuer's sector and geographical location. In addition, the 
different characteristics of bonds and issuers do not appear to have a strong link to 
abnormal returns. However, the results are not yet entirely clear, and the need for future 
research persists as the amount of available green bond data increases. The rapid changes 
in the green bond market regulations are also constantly changing the operational 
environment. Possible improvements in the quality of available data will also likely 
enhance future research. To date, the data imposes some limitations on the quality, as 
there are multiple providers with some inconsistencies. In the future, other sustainable 
debt instruments might also serve as an intriguing research topic in finance as they grow 
and develop alongside green bonds. 
63 
 
References 
Aktas, N. – de Bodt, E. – Roll, R. (2004) Market response to European regulation of 
business combinations. Journal of Financial and Quantitative Analysis. Vol. 39 
(4), 731–757. 
Armitage, S. (1995) Event study methods and evidence on their performance. Journal of 
Economic Surveys, Vol. 9 (1), 25–52. 
Ball, R., – Brown, P. (1968) An empirical evaluation of accounting income numbers. 
Journal of Accounting Research, 159–178. 
Baulkaran, V. (2019) Stock Market Reaction to Green Bond Issuance. Journal of Asset 
Management, Vol. 20 (5), 331–340. 
Bloomberg (2021) The niche, not-so-niche sustainable debt market. 5.11.2021 
<https://www.bloomberg.com/professional/blog/the-niche-not-so-niche-
sustainable-debt-market/>, retrieved 14.12.2022. 
Bloomberg (2022a) Global Debt Market Lost at Least $75 Billion of Business in 2022. 
13.12.2022. <https://www.bloomberg.com/news/articles/2022-12-13/global-
debt-market-lost-at-least-75-billion-of-business-in-
2022?leadSource=uverify%20wall>, retrieved 13.12.2022. 
Bloomberg (2022b) A $300 Billion Bond Market Holds the Key to Solving the Climate 
Crisis. 13.11.2022. <https://www.bloomberg.com/graphics/2022-china-green-
bonds/?leadSource=uverify%20wall>, retrieved 15.12.2022. 
Boatright, J. R. (2014) Ethics in Finance, 3rd Edition. John Wiley & Sons, New York. 
Brooks, C. (2014) Introductory Econometrics for Finance, 4th Edition. Cambridge 
University Press.  
Brown, S. – J.  Warner, J. B. (1985) Using daily stock returns: The case of event studies. 
Journal of Financial Economics, Vol. 14 (1), 3–31.  
Brundtland Commission (1987) Our Common Future. Oxford University Press, Oxford. 
Bugg-Levine, A. – Emerson, J. (2011) Impact Investing: Transforming how we make 
money while making a difference. John Wiley & Sons, New York. 
Campbell, C. J. – Cowan, A. R. – Salotti, V. (2010) Multi-country event-study methods. 
Journal of Banking & Finance, Vol. 34 (12), 3078–3090. 
CBI (2019) Climate Bonds Standard Version 3.0. December 2019. 
<https://www.climatebonds.net/files/files/climate-bonds-standard-v3-
20191210.pdf>, retrieved 28.3.2022. 
64 
CBI (2022a) $500bn Green Issuance 2021: Social and Sustainable Acceleration: Annual 
Green $1tn in Sight: Market Expansion Forecasts for 2022 and 2025. 31.1.2022. 
<https://www.climatebonds.net/2022/01/500bn-green-issuance-2021-social-and-
sustainable-acceleration-annual-green-1tn-sight-market>, retrieved 15.4.2022. 
CBI (2022b) Sustainable Debt Global State of the Market 2021. April 2022. 
<https://www.climatebonds.net/resources/reports/sustainable-debt-global-state-
market-2021>, retrieved 6.12.2022. 
CBI (2022c) Explaining Green Bonds. 
<https://www.climatebonds.net/market/explaining-green-bonds>, retrieved 
15.11.2022. 
CBI (2022d) 5 steps to $5trillion by 2025. October 2022. 
<https://www.climatebonds.net/files/reports/cbi_5steps_to_5trillion_03d_0.pdf>
, retrieved 10.12.2022. 
CBI (2022e) Sustainable Debt Market: Summary Q3 2022. November 2022. 
<https://www.climatebonds.net/files/reports/cbi_susdebtsum_highlq32022_final.
pdf>, retrieved 10.12.2022. 
CBI (2022f) Market integrity: 3 in every 4 dollars from green bonds meet best practice 
climate standards. December 2022. 
<https://www.climatebonds.net/files/reports/cbi_excluded_bonds_01c.pdf>, 
retrieved 19.12.2022. 
CBI (2022g) Global Green Taxonomy Development, Alignment and Implementation. 
February 2022. 
<https://www.climatebonds.net/files/reports/cbi_taxonomy_ukpact_2022_01f.pd
f>, retrieved 19.12.2022. 
CFI (2022) Variance Inflation Factor (VIF). 5.12.2022. 
<https://corporatefinanceinstitute.com/resources/data-science/variance-inflation-
factor-vif/>, retrieved 1.1.2023. 
Climate Action Tracker (2020) China going carbon neutral before 2060 would lower 
warming projections by around 0.2 to 0.3 degrees C. 23.9.2020. 
<https://climateactiontracker.org/press/china-carbon-neutral-before-2060-would-
lower-warming-projections-by-around-2-to-3-tenths-of-a-degree/>, retrieved 
19.12.2022. 
Cowan, A. R. (1992) Nonparametric event study tests. Review of Quantitative Finance 
and Accounting, Vol. 2 (4), 343–358. 
65 
 
EIB (2022a) 15 years of EIB green bonds: leading sustainable investment from niche to 
mainstream. 5.7.2022. <https://www.eib.org/en/press/all/2022-308-15-years-of-
eib-green-bonds-leading-sustainable-investment-from-niche-to-mainstream>, 
retrieved 10.12.2022. 
EIB (2022b) Multilateral development banks. 
<https://www.eib.org/en/about/partners/development-banks/index.htm>, 
retrieved 15.12.2022. 
EPC (2022) Achieving the full potential of sustainable finance: The role of national, 
European and International initiatives. 13.7.2022. 
<https://www.epc.eu/content/PDF/2022/Green_finance_DP.pdf>, retrieved 
13.12.2022. 
European Commission (2022a) EU taxonomy for sustainable activities. 
<https://finance.ec.europa.eu/sustainable-finance/tools-and-standards/eu-
taxonomy-sustainable-activities_en>, retrieved 19.12.2022. 
European Commission (2022b) EU taxonomy compass. 
<https://ec.europa.eu/sustainable-finance-taxonomy/home>, retrieved 
19.12.2022. 
European Commission (2022c) European Green Bond Standard. 
<https://finance.ec.europa.eu/sustainable-finance/tools-and-standards/european-
green-bond-standard_en>, retrieved 19.12.2022. 
European Commission (2022d) International Platform on Sustainable Finance. 
<https://finance.ec.europa.eu/sustainable-finance/international-platform-
sustainable-finance_en>, retrieved 19.12.2022. 
European Parliament (2021) DRAFT REPORT 2021/0191(COD) on the proposal for a 
regulation of the European Parliament and of the Council on European green 
bonds. 30.11.2021. <https://www.europarl.europa.eu/doceo/document/ECON-
PR-700638_EN.pdf>, retrieved 19.12.2022. 
Fama, E. F. – Fisher, L. – Jensen, M. – Roll, R. (1969) The Adjustment of Stock Prices 
to New Information. International Economic Review, Vol. 10 (1), 1–21. 
Fatica, S. – Panzica, R. – Rancan, M. (2021) The pricing of green bonds: are financial 
institutions special? Journal of Financial Stability, Vol. 54, 100873. 
Financial Times (2022) Fears rise over ‘greenwash’ bonds. 21.3.2022. 
<https://www.ft.com/content/178449a7-8897-4359-b23a-e85524c3e227>, 
retrieved 10.12.2022. 
66 
Flammer, C. (2021) Corporate Green Bonds. Journal of Financial Economics, Vol. 142 
(2), 499–516. 
Forbes (2021) Three Critical Areas Where Europe Leads The Green Rush. 19.10.2021. 
<https://www.forbes.com/sites/cognizant/2021/10/19/three-critical-areas-where-
europe-leads-the-green-rush/>, retrieved 1.1.2023. 
German Development Institute (2017) Upscaling Green Bond Markets: The Need for 
Harmonised Green Bond Standards. December 2017. <https://www.idos-
research.de/uploads/media/BP_12.2017.pdf>, retrieved 19.12.2022. 
Glavas, D. (2020) Green regulation and stock price reaction to green bond issuance. 
Finance, Vol. 41 (1), 7–51. 
Green Finance Initiative (2016) Globalizing Green Finance: The UK as an International 
Hub. November 2016. <https://www.cbd.int/financial/gcf/uk-
hubgreenfinance.pdf>, retrieved 12.4.2022. 
ICMA (2017) Summary of Green Bond Database Providers. < 
https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&cad=r
ja&uact=8&ved=2ahUKEwjukvCW4ez7AhXBYPEDHSdzDLcQFnoECBMQA
Q&url=https%3A%2F%2Fwww.icmagroup.org%2Fassets%2Fdocuments%2FR
egulatory%2FGreen-Bonds%2FGreen-Bond-Databases-Summary-Document-
190617.pdf&usg=AOvVaw3o47CwagmEEnMaFzZTPGSc>, retrieved 
9.12.2022. 
ICMA (2020) Sustainability-Linked Bond Principles: Voluntary Process Guidelines. June 
2020. <https://www.icmagroup.org/assets/documents/Regulatory/Green-
Bonds/June-2020/Sustainability-Linked-Bond-Principles-June-2020-
171120.pdf>, retrieved 22.2.2022. 
ICMA (2021b) Social Bond Principles: Voluntary Process Guidelines for Issuing Social 
Bonds. June 2021. <https://www.icmagroup.org/assets/documents/Sustainable-
finance/2021-updates/Social-Bond-Principles-June-2021-140621.pdf>, retrieved 
22.2.2022. 
ICMA (2021c) Sustainability Bond Guidelines. June 2021. 
<https://www.icmagroup.org/assets/documents/Sustainable-finance/2021-
updates/Sustainability-Bond-Guidelines-June-2021-140621.pdf>, retrieved 
22.2.2022. 
ICMA (2022) Green Bond Principles: Voluntary Process Guidelines for Issuing Green 
Bonds. June 2021 (with June 2022 Appendix I). 
67 
 
<https://www.icmagroup.org/assets/documents/Sustainable-finance/2022-
updates/Green-Bond-Principles_June-2022-280622.pdf>, retrieved 10.12.2022. 
IMF (2022a) World economic outlook: Countering the cost-of-living crisis. October 
2022. <https://www.imf.org/en/Publications/WEO/Issues/2022/10/11/world-
economic-outlook-october-2022>, retrieved 10.12.2022. 
IMF (2022b) Emerging market and developing economies. Datasets. 
<https://www.imf.org/external/datamapper/profile/NGDP_RPCH@WEO/OEM
DC/ADVEC/WEOWORLD>, retrieved 15.11.2022. 
Ingves, S. (2014) Banking on leverage. BCBS speech. 26.2.2014. 
<https://www.bis.org/speeches/sp140226.pdf>, retrieved 2.1.2022. 
Insider (2017) The 1000-year bond: A Danish energy company just issued debt with a 
maturity date of 3017. 17.11.2017. <https://www.businessinsider.com/danish-
energy-company-1000-year-bond-2017-11?r=US&IR=T>, retrieved 9.12.2022. 
IPCC (2022) Climate Change 2022: Impacts, Adaptation and Vulnerability. 28.2.2022. 
<https://report.ipcc.ch/ar6wg2/pdf/IPCC_AR6_WGII_FinalDraft_FullReport.pd
f>, retrieved 14.3.2022 
ISS (2022) The Depth & Breadth of Sustainable Finance Regulatory Initiatives: Global 
Developments in 2022. 3.10.2022. <https://www.issgovernance.com/library/the-
depth-and-breadth-of-sustainable-finance-regulatory-initiatives-global-
developments-in-2022/>, retrieved 11.12.2022. 
Kharas, H. – McArthur, J. (2019) Building the SDG economy: Needs, spending, and 
financing for universal achievement of the Sustainable Development Goals, 
Global Economy & Development. Brookings Working paper 131. 
Kothari, S. P. – Warner, J. B. (2007) Econometrics of Event Studies. In: Handbook of 
Corporate Finance: Empirical Corporate Finance. North Holland, Amsterdam. 
KPMG (2022) Big shifts, small steps: Survey of sustainability reporting 2022. October 
2022. <https://assets.kpmg/content/dam/kpmg/xx/pdf/2022/10/ssr-small-steps-
big-shifts.pdf>, retrieved 11.12.2022. 
Lebelle, M. – Lajili Jarjir, S. – Sassi, S. (2020) Corporate Green Bond Issuances: An 
International Evidence. Journal of Risk and Financial Management, Vol. 13 (2), 
25. 
Luo, W. – Tian, Z. – Zhong, S. – Lyu, Q. – Deng, M. (2022) Global Evolution of Research 
on Sustainable Finance from 2000 to 2021: A Bibliometric Analysis on WoS 
Database. Sustainability, Vol. 14 (15), 9435. 
68 
MacKinlay, A. C. (1997) Event studies in economics and finance. Journal of Economic 
Literature, Vol. 35 (1), 13–39. 
Markowitz, H. (1952) Portfolio selection. The Journal of Finance, Vol. 7 (1), 77–91. 
McKinsey & Company (2013) Infrastructure productivity: How to save $1 trillion a year. 
January 2013. 
<https://www.mckinsey.com/~/media/mckinsey/business%20functions/operatio
ns/our%20insights/infrastructure%20productivity/mgi%20infrastructure_full%2
0report_jan%202013.pdf>, retrieved 9.12.2022. 
McKinsey & Company (2022) The economic transformation: What would change in the 
net-zero transition. 25.1.2022. 
<https://www.mckinsey.com/capabilities/sustainability/our-insights/the-
economic-transformation-what-would-change-in-the-net-zero-transition>, 
retrieved 9.12.2022. 
Miller, M. H. (2000) The history of finance: an eyewitness account. Journal of Applied 
Corporate Finance, Vol. 13 (2), 8–14. 
MSCI (2022) The MSCI Net-Zero Tracker. October 2022. 
<https://www.msci.com/documents/1296102/26195050/MSCI-Net-ZeroTracker-
October.pdf>, retrieved 9.12.2022. 
Myers, S. C. – Majluf, N. S. (1984) Corporate financing and investment decisions when 
firms have information that investors do not have. Journal of Financial 
Economics, Vol. 13 (2), 187–221. 
Myers, S. C. (1984) The capital structure puzzle. Journal of Finance, Vol. 39, 575–592. 
New Climate Institute (2022) Corporate Climate Responsibility Monitor 2022. February 
2022. <https://carbonmarketwatch.org/wp-
content/uploads/2022/02/CMW_CCRM2022_v08_FinalStretch2.pdf>, retrieved 
9.12.2022. 
OECD (2015) Green Bonds: Mobilising the Debt Capital Markets a Low-Carbon 
Transition. December 2015. 
<https://www.oecd.org/environment/cc/Green%20bonds%20PP%20%5Bf3%5D
%20%5Blr%5D.pdf>, retrieved 11.4.2022. 
PBOC (2021) Green Bond Endorsed Projects Catalogue (2021 Edition). 21.4.2021. 
<http://www.pbc.gov.cn/goutongjiaoliu/113456/113469/4342400/202109161718
0089879.pdf>, retrieved 19.12.2022. 
69 
 
Pilbeam, K. (2018) Finance & Financial Markets, 4th Edition. Macmillan Education, 
London. 
PwC (2021) PwC’s Global investor survey: The economic realities of ESG. December 
2021. <https://www.pwc.com/gx/en/corporate-reporting/assets/pwc-global-
investor-survey-2021.pdf>, retrieved 9.12.2022. 
Refinitiv (2021) ESG data and solutions from Refinitiv: Research with confidence | 
Integrate with ease. 
<https://www.refinitiv.com/content/dam/marketing/en_us/documents/brochures/
esg-research-brochure.pdf>, retrieved 7.3.2022. 
Refinitiv (2022a) ICMA updates rules for green ABS, covered bonds. 15.7.2022. 
<https://www.refinitiv.com/perspectives/market-insights/icma-updates-rules-for-
green-abs-covered-bonds/>, retrieved 14.12.2022. 
Refinitiv (2022b) TRBC Sector Classification. <https://www.refinitiv.com/en/financial-
data/indices/trbc-business-classification>, retrieved 6.12.2022. 
Reuters (2022) EXCLUSIVE China tightens green bond rules to align them with global 
norms. 24.8.2022. <https://www.reuters.com/markets/asia/exclusive-china-
tightens-green-bond-rules-align-them-with-global-norms-2022-08-24/>, 
retrieved 19.12.2022. 
S&P Global (2022a) Europe's dominance in green bond market fades amid record growth 
in China. 27.7.2022. <https://www.spglobal.com/marketintelligence/en/news-
insights/latest-news-headlines/europe-s-dominance-in-green-bond-market-fades-
amid-record-growth-in-china-71321575>, retrieved 10.12.2022. 
S&P Global (2022b) Germany, Italy lead Q3 green bond issuance in Europe. 3.11.2022. 
<https://www.spglobal.com/marketintelligence/en/news-insights/latest-news-
headlines/germany-italy-lead-q3-green-bond-issuance-in-europe-72471450>, 
retrieved 10.12.2022. 
Schoenmaker, D. – Schramade, W. (2018) Principles of Sustainable Finance. Oxford 
University Press, Oxford. 
SIFMA (2022) 2022 Capital Markets Fact Book. July 2022. <https://www.sifma.org/wp-
content/uploads/2022/07/CM-Fact-Book-2022-SIFMA.pdf>, retrieved 
11.12.2022. 
Stock, J. H. – Watson, M. W. (2020) Introduction to Econometrics. 4th Edition. Pearson 
Education, London. 
70 
Strong, N. (1992) Modelling abnormal returns: A review article. Journal of Business 
Finance & Accounting, Vol. 19 (4), 533–553. 
Sustainable Fitch (2022) China’s New Principles for Green-Bond Issuance. 16.8.2022. 
<https://www.sustainablefitch.com/corporate-finance/chinas-new-principles-for-
green-bond-issuance-16-08-2022>, retrieved 19.12.2022. 
Taghizadeh-Hesary, F. – Yoshino, N. – Phoumin, H. (2021) Analyzing the characteristics 
of green bond markets to facilitate green finance in the post-COVID-19 world. 
Sustainability, Vol. 13 (10), 5719. 
Taking Stock (2021) A global assessment of net zero targets. March 2021. <https://ca1-
eci.edcdn.com/reports/ECIU-
Oxford_Taking_Stock.pdf?mtime=20210323005817&focal=none>, retrieved 
26.9.2022. 
Tang, D. Y. – Zhang, Y. (2020) Do Shareholders Benefit from Green Bonds? Journal of 
Corporate Finance, Vol. 61, 101427. 
TEG (2020) Taxonomy: Final report of the Technical Expert Group on Sustainable 
Finance. March 2020. <https://finance.ec.europa.eu/system/files/2020-
03/200309-sustainable-finance-teg-final-report-taxonomy_en.pdf>, retrieved 
19.12.2022. 
The Eco Experts (2022) Top 7 Most Polluting Industries in 2022. 31.10.2022. 
<https://www.theecoexperts.co.uk/blog/top-7-most-polluting-industries>, 
retrieved 14.12.2022. 
The New Climate Economy (2014) The New Climate Economy Report: Better Growth 
Better Climate. September 2014. <https://newclimateeconomy.report/2016/wp-
content/uploads/sites/2/2014/08/NCE-Global-Report_web.pdf>, retrieved 
13.4.2022. 
Thompson, S. (2021) Green and Sustainable Finance: Principles and Practice. Kogan 
Page Publishers, London. 
Tobin, J. (1969) A general equilibrium approach to monetary theory. Journal of Money, 
Credit and Banking, Vol. 1 (1), 15–29. 
Tobin, J. (1978) A proposal for international monetary reform. Eastern Economic 
Journal, Vol. 4 (3/4), 153–159. 
Tolliver, C. – Keeley, A. R. – Managi, S. (2020) Drivers of green bond market growth: 
The importance of Nationally Determined Contributions to the Paris Agreement 
71 
 
and implications for sustainability. Journal of Cleaner Production, Vol. 244, 
118643. 
UN General Assembly (2015) Transforming our World: the 2030 Agenda for Sustainable 
Development. Report No. A/RES/70/1. <https://documents-dds-
ny.un.org/doc/UNDOC/GEN/N15/291/89/PDF/N1529189.pdf?OpenElement>, 
retrieved 28.3.2022. 
UN High-Level Expert Group (2022) Integrity Matters: Net Zero Commitments by 
Businesses, Financial Institutions, Cities and Regions. 8.11.2022. 
<https://www.un.org/sites/un2.un.org/files/high-
level_expert_group_n7b.pdf?_gl=1*1dt7sfl*_ga*OTE1MTc4MTU4LjE2NjI1Nj
YyNzY.*_ga_TK9BQL5X7Z*MTY2Nzk5Mzg0NC4xLjAuMTY2Nzk5Mzg0N
C4wLjAuMA>, retrieved 9.12.2022. 
UNFCCC (2022) The Paris Agreement. <https://unfccc.int/process-and-meetings/the-
paris-agreement/the-paris-agreement>, retrieved 20.12.2022. 
Wang, J. – Chen, X. – Li, X. – Yu, J. – Zhong, R. (2020) The market reaction to green 
bond issuance: Evidence from China. Pacific-Basin Finance Journal, Vol. 60, 
101294. 
White, H. (1980) A Heteroskedasticity-Consistent Covariance Matrix Estimator and a 
Direct Test for Heteroskedasticity. Econometrica, Vol. 48 (4), 817–838. 
World Economic Forum (2019) Unlocking Capital Markets to Finance the SDGs. 
September 2019. 
<https://www3.weforum.org/docs/WEF_Unlocking_Capital_Markets_to_Financ
e_the_SDGs_2019.pdf>, retrieved 9.12.2022. 
72 
Appendices 
Appendix 1 Used market indices  
Country Used Market Index 
Australia Standard and Poor's / Australian Stock Exchange 300 
Austria Austrian Traded Index 
Belgium Belgium 20 
Brazil Brazil Bovespa 
Canada Standard and Poor's / Toronto Stock Exchange Composite Index 
Chile FTSE Chile 
China (Mainland) Shanghai Shenzhen CSI 300 
Denmark OMX Copenhagen (OMXC) 
Finland OMX Helsinki (OMXH) 
France France CAC 40 
Germany Prime All Share (Xetra) 
Greece Athex Composite 
Hong Kong Hang Seng 
India Nifty 500 
Ireland Iseq All Share Index 
Italy FTSE MIB Index 
Japan TOPIX 
Netherlands AEX All Share 
New Zealand Standard and Poor's / NZX 50 
Norway Oslo Exchange All Share 
Poland FTSE Poland 
Portugal Portugal PSI All-Share 
South Africa FTSE / JSE All Share 
Spain IBEX 35 
Sweden OMX Stockholm (OMXS) 
Switzerland Swiss Market (SMI) 
Taiwan Taiwan Stock Exchange Weighed TAIEX 
Turkey FTSE Turkey 
UAE FTSE United Arab Emirates 
United Kingdom FTSE All Share 
US Standard and Poor's 500 Composite 
 
 
 
 
73 
 
Appendix 2 Results from event study using USD as common currency 
Daily abnormal returns 
Day AAR Median Stdev Min  Max Pos:Neg t-test Sign test 
–5 0.119 0.084 1.551 –9.791 8.915 297:267 1.826* 1.643 
–4 –0.095 –0.082 1.374 –5.378 6.749 261:303 –1.644 –1.389 
–3 0.056 –0.040 1.455 –5.292 9.020 267:297 0.920 –0.884 
–2 0.017 0.026 1.341 –6.960 6.764 289:275 0.306 0.969 
–1 –0.007 0.060 1.470 –8.343 5.929 289:275 –0.109 0.969 
0 –0.069 –0.064 1.468 –6.779 6.729 270:294 –1.123 –0.631 
+1 0.022 0.050 1.382 –5.672 9.495 294:270 0.374 1.390 
+2 0.045 0.067 1.526 –7.050 13.732 301:263 0.698 1.980** 
+3 –0.053 0.002 1.475 –10.264 7.700 282:282 –0.852 0.380 
+4 0.070 –0.030 1.409 –5.279 9.318 275:289 1.187 –0.210 
+5 –0.084 –0.069 1.388 –7.602 6.933 267:297 –1.439 –0.884 
Significance levels: * p < 10%  ** p < 5%  *** p < 1% 
 
Cumulative average abnormal returns 
Window CAAR Median Stdev Min  Max Pos:Neg t-test Sign test 
[–10, –1] 0.258 0.165 4.679 –22.434 25.993 291:273 1.311 1.138 
[–20, +20] 0.043 0.178 9.476 –52.436 34.074 283:281 0.108 0.464 
[–10, +10] 0.076 0.189 6.975 –49.714 38.345 290:274 0.259 1.053 
[–5, +5] 0.022 0.093 4.661 –22.275 27.636 289:275 0.111 0.969 
[–1, +1] –0.054 –0.041 2.337 –12.040 12.755 275:289 –0.553 –0.210 
[+10, +20] 0.055 –0.221 4.352 –23.091 25.854 266:298 0.303 –0.968 
Significance levels: * p < 10%  ** p < 5%  *** p < 1%