*2.1. An Introduction to Green Finance and Green Bonds*

The concept of green finance emerged in the 2010s and can be defined as 'a type of future-oriented finance that simultaneously pursues the development of financial industry, improvement of the environment, and economic growth' [4]. Green finance is a broad concept that includes sustainable finance for socially inclusive green projects, environmental finance to promote environmental protection, carbon finance, targeting a reduction in GHG emissions, and climate finance, focusing on climate change adaptation and mitigation [4]. The term 'green finance' also covers a wide range of instruments, from private loans to insurance, and includes equity, derivatives, and fiscal or investment funds [4].

Increasing green finance, climate finance, and low-carbon investments are directly and indirectly related to various SDGs. Investments in green energy projects are crucial to achieve the SDGs and meet the Paris Agreements [1]. Fiscal policy has an essential role in assuring the sustainable use of resources and keeping the environment for meeting the related SDGs. This applied to both sides of the government budget. In the revenue side, carbon taxation adjusted with greenness efforts and green bonds are two essential tools, and their importance is increasing. Various fiscal measures could help green-specific priority sectors. Green-adjusted tax on polluting gases can help generate revenue for environmental purposes and redirect the flow of investments from brown to green and low-carbon sectors by introducing green floating rate bonds [5,13]. Another fiscal measure is global taxation on CO2, NOX, and other pollutions, for bringing back optimal portfolio allocation in green investments.

In this research, we focus on green bonds. Since their creation in 2007, USD754 billion worth of green bonds have been issued—primarily in the United States, China, and France in compliance with the Green Bond Principles [7]. Green bonds can be issued by central and local governments, banks, or corporations, and include any debt format [6]. Since 2014, Asia-Pacific's bond issuance has been growing at 35%, placing the region second in terms of green bond volume [6]. Figure 1 shows the evolution of the amount issued for green bonds, per region of issuance. The graph clearly shows that green bonds are a relatively new form of financial instrument, as their issuance started timidly in the early 2010s and skyrocketed after 2015. Europe is the leading issuer of green bonds, although Asia-Pacific has witnessed steady growth in recent years.

**Figure 1.** Evolution of the amount issued for green bonds per region. Note: 'North America' includes issuers from other regions of the world, apart from Europe and Asia-Pacific. However, the majority of the issuance in this category is from the United States and Canada. Source: Authors' compilation, using data from [7].

> Increasing awareness of climate change could be the reason behind the surge in popularity of this instrument. Typically, green bonds are a form of fixed-income finance which can be applied to many debt formats, such as private placements, securitizations, and covered bonds, as well as green loans [7]. The particularity of this form of finance is their target, as the term only encompasses finance for climate change solutions whose proceeds go to green assets [7]. To clarify which bonds could be qualified as such, a consortium of investment banks established the Green Bonds Principles in 2014, based on four main components: (i) the use of the proceeds, (ii) the process for project evaluation and selection, (iii) the management of the proceeds, and (iv) reporting [19]. The principles do not define what is 'green' about the bonds, but merely list target sectors in which green bonds are considered valid. However, these principles simply have an indicative value, and were only agreed on by the investment banks that created them. To date, there is no general taxonomy for green bonds, although the European Union has proposed including one in the upcoming European Green Deal [7].

## *2.2. Characteristics and Challenges of Green Bonds*

The increasing popularity of this instrument has attracted the attention of academic researchers. Studies have provided some empirical proof that green bonds can be useful in fighting climate change [20]. The main academic debate regarding green bonds is the existence of the 'green premium', also called 'greenium', defined as 'a discount that makes green bonds funded cheaper than other bonds from the same issuer' [8,21]. Many recent studies have attempted to compare the yields of green bonds with those of conventional bonds, and the results vary depending on the methodology used. The authors of [11] conducted a global study, matching green bonds with similar conventional bonds and applying a two-step regression method, and concluded that green bonds had lower yields, on average. This effect was especially pronounced for bonds issued by the financial sector and low-rated bonds [11]. This conclusion is shared by recent studies such as [8,10,11].

Other studies, however, tend to have mixed results. For instance, [22] showed that the green premium was actually positive, meaning that matched green bonds had higher yields than their closest brown counterparts. The authors explain their results by arguing that the sign of the green premium depends on the issuer, and that privately issued bonds generally have a positive premium [22]. Similarly, [23] found that the sign of the green premium was not obvious, and depended on the rating achieved by the bond. In particular, highly rated green bonds consistently showed higher returns, which, the authors argued, could make up for the external costs of issuance [23,24]. Finally, [25] could not find statistically significant evidence of the existence of the green premium, even though they used several methodologies such as matching with difference-in-differences and traditional panel techniques (fixed effect). Due to the lack of consensus regarding the green premium, [26] provided a comprehensive literature review on the topic, detailing the methodology of each paper. The authors concluded that the majority of the studies on the topic prove the existence of a green premium in secondary markets.

Interestingly, there does not appear to be a consensus on the riskiness of green bonds either. While [22] found that green bonds had lower variance than conventional bonds, the results of [16], who studied the volatility of the green bond market using a multivariate GARCH approach, contradict this theory. The authors of [16] proved that the market of labeled green bonds was highly volatile—far more so than the unlabeled market of conventional bonds. There is a close link between green bonds and fixed-income and currency markets, with the latter's green bonds receiving price spillover from the latter [27]. Generally, green bonds are strongly affected by changes in stock, changes in energy, and high-yield corporate bond markets [28], as well as the liquidity risk of the bond market [29].

Apart from their generally low returns and high risks, green bonds also represent a challenge for their issuers. Both [23] and [27] highlighted that issuing green bonds tends to be more expensive than issuing a conventional bond due to additional costs arising from the certification, reporting, and administrative burden of the proceeds. The authors of [27] also pointed out the need to bridge the informational gap between issuers and investors and offer clear and unified green criteria to provide assurance of the green nature of the investment [27]. The major issue faced by green bonds is generally the lack of uniform definition and labeling. While the Green Bonds Principles are a major step towards this direction, they remain an informal form of labeling that was only generated by a handful of private actors. Hence, it does not have global legitimacy.

A review of the literature has revealed the evolution and contribution of green bonds. As fixed-income instruments, green bonds can be useful in fighting climate change and bridging the investment gap for green projects. At the same time, these bonds are characterized by lower returns and higher risks than their conventional counterparts. Administrative costs arising from certification and lack of uniform taxonomy have added to their relative lack of attractiveness. Nevertheless, there is ongoing debate regarding the characteristics of green bonds, particularly the existence of a green premium, while results tend to vary depending on the bond rating and issuer [22,23].

A literature review shows that green bonds are essential financial instruments for financing ecological and green projects, and that their importance is increasing. However, we could not find any study that provides a comparative study of the characteristics of green bonds, based on the region. In particular, we could not find any study that determines whether the domination of traditional banking impacts the return of green bonds issued

in Asia and the Pacific. Hence, from this aspect, this study is novel and contributes to the literature.

## **3. Methodology and Data Description**

In this section, we detail the approach taken in this study to determine the regional characteristics of green bonds, with a specific focus on those issued in Asia and the Pacific.

#### *3.1. Data and Description of Variables*

The study combined two datasets from Bloomberg New Energy Finance (BNEF) and the Climate Bonds Initiative (CBI). The BNEF database only provides bonds with an issued amount of at least \$100. Both sources are considered authorities on data related to green finance and have been employed in many recent studies (e.g., [9,11,22–24]). In this research, we only focus on green bonds with a minimum of \$100 in size, issued from 2017 to 2020. Hence, this study presents an analysis of unbalanced panel data of 1014 bonds, from 2017 to 2020, for a total of 1174 observations. To be precise, since we are missing many observations of the rate of return of bonds in 2017 and 2018, the length of the panel is about two time periods. A description of the variables used in the study is provided in Table 1.


Source: Authors' compilation.

## *3.2. Methodology*

To determine the characteristics of green bonds, we propose several methods, each assessing different dimensions of bonds. First, an analysis of the distribution of issuers, maturity, and issued size is proposed, to determine whether green bonds issued in Asia present an inherent difference in their nature. We then move on to a mean-variance analysis, distinguishing between regions and sectors of issuance, to discuss how Asian green bonds compare with their counterparts in terms of risks and returns. Finally, the latter part of the empirical analysis is devoted to investigating the impact of the sector of issuance on the performance of green bonds, as measured by the rate of return, depending on the region.

To this end, we develop an econometric model, which is given by the following equation:

$$Return\_{i,l} = a + \sum\_{l=1}^{4} \beta\_l \text{Sector}\_l + \sum\_{l=2018}^{2020} \gamma\_l \text{Year}\_l + \chi\_1 \text{Coupon}\_l + \chi\_2 \text{Maturity}\_{l,l} + \varepsilon\_{i,l} + u\_{i,l}, \tag{1}$$

where *Returni*,*<sup>t</sup>* denotes the rate of return of bond *i* at year *t*, *Sectori* is a set of dummy variables denoting the bond *i*'s issuing sector, *Yeart* is a set of dummy variables for time fixed-effects, *Couponi* is the bond *i*'s coupon rate, *Maturityi*,*<sup>t</sup>* denotes the number of days until the bond *i* reaches maturity at year *t*, and *i*,*<sup>t</sup>* and *ui*,*<sup>t</sup>* are idiosyncratic and timevarying error terms, respectively.

While many studies use yield as a dependent variable [11,22,23], we decided to use the rate of return of the bonds as our dependent variable, as an approximation of the bond's performance, due to limitations on data availability. Since this study aims to determine the impact of the type of issuer on the bond's performance, we also include a set of four dummy variables, representing the issuer's sector, constructed based on the issuer name

provided by BNEF. Sectors analyzed in the study are grouped into five categories: public, banking and finance, manufacturing, power and utilities, and other issuers. Public issuers are generally state and regional development banks and international organizations, but we do not include state-owned enterprises in this category. Banking and finance are essentially composed of national and local banking institutions, but investment banks and insurance are also considered. Finally, manufacturing in our sample is mostly composed of information technology and paper companies, while other issuers are dominated by companies belonging to real estate and construction.

The choice of remaining control variables is based on existing literature on the topic. The coupon rate, issued size, and maturity are often used in studies tackling the existence of the green premium, as they are essential components for matching green and brown bonds [11,22] or as control variables in regression [8,23,24]. Furthermore, we decided to include year fixed-effects to control for variation over time since our other control variables describe fixed characteristics of bonds. The variable *Maturity* is time-dependent, but its variation is fixed over time so it cannot fully capture changes in time periods. We aim to capture the effects of changes in the financial market and economic policy uncertainty through these dummy variables, as these macroeconomic variables were shown to have a significant impact on green bonds' returns. The authors of [18,25] took a similar approach by including year fixed-effects as a control variable in their regression.
