Green banks versus non-green banks: A financial stability comparative analysis in terms of CAMEL ratios
Abstract
This study examines green and non-green-banks from a financial stability point of view and specifically whether there are any discernible performance differences between the two groups. Using the supervisory ratios namely CAMEL variables, and employing panel data techniques (random effects model) and a global panel data set of 165 banks from 38 countries for the period 1999 to 2021, we adopt the Differences-In-Differences approach to examine whether green (“treatment” group) and non-green (“control” group) banks exhibit differential behaviour, using the outbreak of the financial crisis (2008) as the time of intervention. Our results mainly show that green banks differ (and specifically perform better than their non-green counterparts) only in terms of Total Capital, Tier 1 Capital, and NPLs/Reserve for Loan Losses ratios during and after the financial crisis. As for the rest of the CAMEL factors, it seems that both groups exhibit the same behaviour, especially in the post-crisis period. Thus, green banks are not stronger in total than their non-green counterparts in terms of financial stability. We also find that the financial crisis had either a positive or a negative effect on most of the CAMEL factors of both bank types, except for the Leverage Ratio (a capital adequacy proxy) and Operational Expenses/Operational Income ratios (a management quality proxy), which proved crisis-insensitive.
1 INTRODUCTION
Global warming is the main element of climate change and is mainly attributed to the extended use of fossil fuels and Greenhouse Gas emissions (GHG). The IPCC (2021), p. 18) mentions that an increase in the average global temperature by 2°C is quite possible between 2041 and 2060 in the intermediate scenario. The severity of climate change is reflected in climate-driven natural disasters affecting humans and the economy. Eckstein et al. (2020) mention that, between 1999 and 2018, more than 495,000 people died as a direct result of more than 12,000 extreme weather events, while the economic losses were approximately 3.54 trillion USD (in Purchasing Power Parities). So, it becomes obvious that limiting the risk of an environmental disaster is rather impossible without a green transition, that is, a substantial transformation of production and consumption patterns, which implies a radical change in firms' production characteristics and households' consumption habits (Besley & Persson, 2023).
Although banks are firms that are not directly engaged in environmental degradation, they lend to other firms (which are either goods producers or services providers) and to households. To this extent, the banking sector can have a significant indirect contribution to reducing the risk of an environmental disaster by lending enterprises with increased environmental conscience and households with increased perception of “green values”.
In addition, climate change affects both countries' economies and financial sectors. Natural disasters, as a result of climate change, amplify the fragility of banking systems as they increase banks' probability of default (Danmarks Nationalbank, 2019; Klomp, 2014) and the probability of banking crises (from 26% to 248% by 2100 under the extreme scenario according to Lamperti et al. (2019)). The need for the protection of the environment and of the financial system from the negative effects of climate change has led to the creation and launch of “green banking” (i.e., credit facilities to low-carbon investments, renewables, etc.) and of the so-called “green banks” (fully public and quasi-public). Additionally, many small, medium, and large-sized—financial institutions around the world have introduced “green credit lines,” aiming to achieve the above-mentioned climate goal, and also to mitigate the negative impacts on the banking sector stemming from climate change.
As pure green banks are very few worldwide, for the purposes of our analysis we employ data on commercial banks (i.e., the typical banks that accept deposits, grant loans, etc.) and we characterize them as green or not by developing certain criteria, the basic criterion being whether a bank is a member of certain international organizations focusing on sustainability and environmental protection (see Subsection 3.3). During the last twenty years, the number of environmentally friendly banks exhibited a significant increase, showing that the banking sector takes seriously the issue of climate change and its adverse effects on humans and economies. For instance, membership of the Equator Principles rose from 13 in 2003 to 123 in 2021, while, according to Weber (2012), out of 61 banks examined (from 17 countries) 28 conducted systematic environmental evaluations since 2006.
In this study, we compare green and conventional (non-green hereafter) banks in terms of financial strength and soundness indicators known as CAMEL/CAMELS ratios (solvency, asset quality, management quality, profitability, and liquidity) which are employed as proxies of financial stability. For the purposes of our analysis, we use an unbalanced panel data set of 165 green and non-green banks (either global or non-global) from 38 countries worldwide covering the period 1999 to 2021, and we apply a random effects model and Differences-In-Differences (DID) method to examine if there are discernible performance differences between the two groups. Moreover, we compare green and non-green banks by considering that the implementation of a new “green” lending logic has to be a gradual process for the banking sector from a financial stability perspective, as a sudden transition could pose a threat to the stability of the financial system, taking into account that the adoption of lower risk weights for green assets could reduce the financial sector's capital adequacy without reducing its actual risks (Manninen & Tiililä, 2020; Nieto, 2017).
To the best of our knowledge, this is the first study examining the differences between green and non-green banks from a financial stability perspective in terms of CAMEL variables, both in the pre- and post-financial crisis periods and across countries, using a panel data set for a 23-year period.
Our analysis provides the following key results. First, the global financial crisis significantly impacts most CAMEL factors, except the leverage ratio and operational expenses to operational income ratio, proving crisis-insensitive across all bank types. Second, green banks demonstrate consistent behaviour across various fundamental aspects, except for risk-weighted capital ratios, indicating that they do not outperform non-green banks in terms of risk-adjusted capital adequacy ratios, especially before the crisis. Last, the crisis benefits green banks only in terms of two risk-adjusted capital ratios and partially in NPLs coverage ratio, with other CAMEL factors (leverage ratio, NPLs and provision ratios, management quality, earning ability and liquidity ratios) remaining unaffected, providing empirical evidence that green banks do not substantially differ from non-green banks in terms of financial stability.
The remainder of the paper is organized as follows. Section 2 presents related literature. Section 3 describes data, variables, and methodology. Section 4 contains results (including descriptive statistics and statistical tests) and discusses empirical findings. Finally, Section 5 concludes and presents policy implications.
2 A BRIEF LITERATURE REVIEW
2.1 CAMEL/CAMELS rating system and financial sector stability
The CAMEL/CAMELS rating system comprises financial criteria used by regulators to gauge banks' financial health, serving as a metric for banking sector stability, soundness, and risk, thereby measuring financial stability (Hirtle et al., 2020; Restoy, 2017). These ratings and factors are also utilized in predictive models for analysing and forecasting financial crises and bank failures (Demyanyk & Hasan, 2010). CAMEL/CAMELS ratings, combine ratio values and specific questionnaire outcomes (not publicly available), ranging from 1 (strongest performance and risk management) to 5 (weakest performance and highest supervisory concern) (FDIC, 2018–2023).
CAMEL/CAMELS proxies cover six main segments: capital adequacy (C), asset quality (A), management quality (M), earning ability (E), liquidity (L), and sensitivity to market risk (S) 1 (Board of Governors of the Federal Reserve System, 2018, 2023, p. 9). Each segment includes specific variables such as Total Capital and Leverage Ratio (C), Non-performing Loans to Total Loans (A), Operational Expenses to Operational Income (M), Return on Assets (E), Loan-to-Deposit and Liquidity Coverage Ratio (L), and Regulatory Capital for market risk/Regulatory Capital (S). These variables are well-documented in numerous banking sector studies (Cole & White, 2017; Papanikolaou, 2018). An analytical bibliographic review of CAMEL/CAMELS variables is available in Online Appendix A.
CAMEL/CAMELS ratings and factors also measure banks' risk and financial health (Chernykh & Kotomin, 2022; Martinez-Peria & Schmuckler, 2001), banks' performance from a supervisory view, and financial stability perspective (Gropp et al., 2022; Hirtle et al., 2020). They are used in studies of various countries' banking systems (see e.g., Afroj, 2022; Nguyen et al., 2020) and in creating new indexes for assessing banks' financial strength (see e.g., Doumpos et al., 2017).
2.2 Green banking and green banks
Green banking encompasses: (a) banking commitments (implementation of green finance principles), (b) financial flows (volume and distribution of loans to green investments), (c) financial risk (impact on NPLs and ROE), and (d) environmental and social outcomes (avoidance of negative E&S impacts, GHG emissions reduction, economic growth, job creation, etc.) (IFC & SBN, 2017; Park & Kim, 2020). Related concepts include green and sustainable finance (see Online Appendix B).
- “Pure Green Banks”: Dedicated public or non-profit entities focused on increasing and accelerating investments in clean energy and services using financial tools to mitigate climate change (CGC, 2019).
- “Green Investment Banks (GIBs)”: Public or quasi-public institutions financing renewable energy, energy efficiency, and other clean energy projects in partnership with private lenders to advance public objectives (CGC, 2017; OECD, 2015).
- “Social Banks”: Institutions that consider profit, environment, and people; in some cases, green and social banks coincide (Benedikter, 2011).
Pure green banks and GIBs are limited in number, with no more than 12 by the end of 2021, and currently only 9 (Green Bank Network, 2018, 2023).
3 DATA, VARIABLES, AND METHODOLOGIES
3.1 Data
We use data from Thomson Reuters Eikon for the CAMEL factors, and from: (a) supranational organizations (Federal Stability Board, Basel Committee on Banking Supervision) and the “Banks around the World” website, and (b) non-governmental organizations and institutions (BEI, EP, GABV, and UNEP FI) for classifying banks as global and green, respectively (see Subsection 3.3 and Online Appendix C).
From our initial sample, observations for the CAMEL variables NPLs/Reserve for Loan Losses % (asset quality proxy) and Operating Expenses/Operating Income % (management quality proxy) with extreme values (above 1.5 times the 99.5th percentile or below 0.5 times the 0.5th percentile) are characterized as outliers and excluded from the analysis. Our final unbalanced panel dataset comprises a maximum of 3795 observations from 1999 to 2021 on an annual basis, 2 including specific CAMEL variables for 165 banks from 38 countries worldwide.
3.2 Variables
Following studies on banks' financial strength and sector stability, we use CAMEL proxies for the first five segments (excluding the S segment due to data unavailability). All CAMEL variables employed are included in both the FDIC (2018–2023) and the Board of Governors of the Federal Reserve System (2018, 2023) examination manuals.
Specifically, we use 13 quantitative bank-specific variables (12 CAMEL factors plus bank size) and 4 dummy variables (global bank, green bank, financial crisis period, and the interaction between crisis and green banks) (see Table 1). Country dummies are also included. The global bank, crisis, and green bank variables are primary (or main) variables, with the interaction variable being the key variable of interest.
| CAMEL segment | Abbreviation/acronym | Variable definition and calculation | Data source |
|---|---|---|---|
| Panel A: Explanatory—control—variables | |||
| Capital adequacy | TCR | Total Capital Ratio (%) (=Total Capital to total risk-weighted assets *100) | Thomson Reuters Eikon |
| CRTIER1 | Tier 1 Capital Ratio (%) (=Tier 1 Capital to total risk-weighted assets *100) | Thomson Reuters Eikon | |
| LR | Leverage Ratio (%) (=Total Capital/(Total Assets − Customer Liabilities on Acceptances) * 100) | Thomson Reuters Eikon | |
| Asset quality | NPLS | Non-Performing Loans/Total Loans (%) (NPLs) | Thomson Reuters Eikon |
| PROVLLLOANS | Provision for Loan Losses/Total Loans (%) (=Provision for Loan Losses/(Total Loans−Interbank Loans) * 100) | Thomson Reuters Eikon | |
| NPLSRESERLL | Non-Performing Loans/Reserve for Loan Losses (%) | Thomson Reuters Eikon | |
| RESERLLLOANS | Reserve For Loan Losses/Total Loans (%) (=Reserve for Loan Losses/(Total Loans−Interbank Loans) * 100 | Thomson Reuters Eikon | |
| Management quality | OPEXPENSOPINCOME | Operating Expenses/Operating Income (%) (Operating Expenses = sum of all expenses related to operations; Operating Income = sales−total operating expenses) | Thomson Reuters Eikon; authors' calculation |
| OPEXPENSTA | Operating Expenses/Total Assets (%) | Thomson Reuters Eikon; authors' calculation | |
| Earning ability | ROA | Return on Assets (%) (ROA) (=Net Income before Preferred Dividends + ((Interest Expense on Debt-Interest Capitalized) * (1-Tax Rate)))/Average of Last Year's (Total Assets−Customer Liabilities on Acceptances) and Current Year's (Total Assets−Customer Liabilities on Acceptances) * 100 | Thomson Reuters Eikon |
| ROE | Return on Equity (%) (ROE) (=Net Income before Preferred Dividends−Preferred Dividend Requirement)/Average of Last Year's and Current Year's Common Equity * 100. | Thomson Reuters Eikon | |
| Liquidity | LTD | Total Loans/Total Deposits (%) | Thomson Reuters Eikon |
| – | SIZE | Bank size = log of Total Assets | Thomson Reuters Eikon; authors' calculation |
| Panel B: key or main explanatory variables (dummy variables and interaction term) | |||
| – | GLOBAL (= global) | 0 if non-global and 1 if global | FSB, BCBS, “Banks around the World” (http://www.relbanks.com), own research; authors' construction |
| – | NEWGREEN (= green) | 0 if non-green and 1 if green | ΒΕΙ, EP, GABV, UNEP FI, own research; authors' construction |
| – | CRISIS (= crisis) | 0 if time ≤ 2007Q4 and 1 if time ≥ 2008Q1 | Authors' construction |
| – | INTERACTION | Crisis × green | Authors' construction |
- Note: (a) The term “calculation” in the third column header refers to the Thomson Reuters Eikon (formerly Data Stream) calculation methodology described in Thomson Reuters Financial (2007). (b) The OPEXPENSOPINCOME, OPEXPENSTA, and SIZE variables are our own calculations based on Thomson Reuters Eikon relevant data. (c) Operating Expenses, Operating Income, and Total Assets amounts are all in 000's USD. (d) Total Loans = Loans−Interbank Loans. (e) “Customer Liabilities on Acceptances” are only subtracted when included in Total Assets (Thomson Reuters Financial, 2007). (f) Total Assets = the sum of cash and due from banks, total investments, net loans, customer liability on acceptances (if included in total assets), investment in unconsolidated subsidiaries, real estate assets, net property, plant and equipment, and other assets (Thomson Reuters Financial, 2007). (g) The characterization of a bank as global and green, respectively, is in accordance with the developed classification criteria (see Subsection 3.3). (h) Cut-off dates for the financial crisis period (pre- and post-crisis) were set according to the BIS (2009) timelines.
- Abbreviations: BCBS, Basel Committee on Banking Supervision; BEI, Banking Environment Initiative; EP, Equator Principles; FSB, Financial Stability Board; GABV, Global Alliance for Banking on Values; UNEP FI, United Nations Environment Programme- Finance Initiative.
The financial crisis variable is defined with the crisis starting in the last quarter of 2007. The pre-crisis and post-crisis periods are before and after 2007Q4, respectively, based on BIS (2009) and Reinhart and Rogoff (2014). 3 The dataset is not split into more sub-periods due to the varying duration and effects of the crisis across regions (Baur, 2012; De Bondt & Vermeulen, 2021; Laeven & Valencia, 2012; Reinhart & Rogoff, 2014). “During and after crisis” refers to 2008–2009, the peak of the global financial crisis (Reinhart & Rogoff, 2011), used as the interaction point for green banks with the crisis.
3.3 Classification criteria
We classify a bank as “green” if it was a member, by the end of 2021, of at least one of the following organizations: Banking Environment Initiative (BEI), Equator Principles (EP), United Nations Environment Programme—Finance Initiative (UNEP FI), and Global Alliance on Banking on Values (GABV) (see Online Appendix C for details). This aligns with Benedikter (2011), pp. 43–45) analysis for social and green banks. Additional criteria (e.g., sustainability policy soundness, green loan types, amounts, and share in total loan portfolio) were not employed due to data unavailability. Confirmation of green bank status involved reviewing annual and sustainability reports (covering the vast majority of banks characterized as green).
A bank is classified as “global” if it: (a) operates in more than one region via subsidiaries or branches, including significant representative offices worldwide, and (b) has an average asset size of $100 billion USD or more during the entire sample period. The first criterion is necessary, the second is not. These criteria follow BCBS (2017), De Haas and Van Lelyveld (2014), and Jeucken (2001), p. 186). Most global banks are included in the FSB (2022) list of Global Systemically Important Banks (G-SIBs) and the “Banks around the World” website. 4 Confirmation of global status was done through banks' official websites and reports.
All banks that are characterized as green and non-green (either global or not) are presented in Appendix Table A1.
According to our criteria, from 165 banks, 84 are identified as green (53 global, 31 non-global) and 81 as non-green (19 global, 62 non-global) as of December 31, 2021. Care was taken to ensure a balanced number of green and non-green banks per region and country. Figure 1 and Table 2 summarize green and non-green banks per region and country.
| Region | Country | Banks number (total) | Of which: Green banks | % | Of which: Global banks | % |
|---|---|---|---|---|---|---|
| Africa | Kenya | 2 | 1 | 50% | 0 | 0% |
| Mauritius | 2 | 1 | 50% | 0 | 0% | |
| Morocco | 2 | 1 | 50% | 0 | 0% | |
| Nigeria | 12 | 6 | 50% | 3 | 50% | |
| South Africa | 8 | 4 | 50% | 4 | 100% | |
| Total | 26 | 13 | 50% | 7 | 54% | |
| Asia | India | 2 | 1 | 50% | 0 | 0% |
| Total | 2 | 1 | 50% | 0 | 0% | |
| Asia Pacific | Bangladesh | 2 | 1 | 50% | 0 | 0% |
| China | 8 | 4 | 50% | 2 | 50% | |
| Indonesia | 2 | 1 | 50% | 0 | 0% | |
| Japan | 10 | 4 | 40% | 3 | 75% | |
| Malaysia | 2 | 0 | 0% | 0 | – | |
| South Korea | 6 | 3 | 50% | 1 | 33% | |
| Taiwan R.O.C. | 2 | 1 | 50% | 0 | 0% | |
| Thailand | 2 | 1 | 50% | 0 | 0% | |
| Total | 34 | 15 | 44% | 6 | 40% | |
| Europe | Belgium | 2 | 1 | 50% | 0 | 0% |
| Denmark | 2 | 1 | 50% | 0 | 0% | |
| France | 6 | 4 | 67% | 4 | 100% | |
| Germany | 4 | 1 | 25% | 0 | 0% | |
| Greece | 2 | 0 | 0% | 0 | – | |
| Italy | 4 | 2 | 50% | 2 | 100% | |
| Netherland | 3 | 2 | 67% | 2 | 100% | |
| Norway | 2 | 1 | 50% | 1 | 100% | |
| Spain | 8 | 5 | 63% | 5 | 100% | |
| Sweden | 4 | 4 | 100% | 4 | 100% | |
| Switzerland | 6 | 2 | 33% | 2 | 100% | |
| Turkey | 6 | 3 | 50% | 0 | 0% | |
| UK | 5 | 5 | 100% | 5 | 100% | |
| Total | 54 | 31 | 57% | 25 | 81% | |
| Latin America | Argentina | 2 | 1 | 50% | 0 | 0% |
| Brazil | 6 | 3 | 50% | 3 | 100% | |
| Chile | 2 | 1 | 50% | 0 | 0% | |
| Colombia | 2 | 1 | 50% | 0 | 0% | |
| Ecuador | 2 | 1 | 50% | 0 | 0% | |
| Mexico | 2 | 1 | 50% | 0 | 0% | |
| Peru | 2 | 1 | 50% | 0 | 0% | |
| Total | 18 | 9 | 50% | 3 | 33% | |
| Middle East | Oman | 2 | 1 | 50% | 0 | 0% |
| Total | 2 | 1 | 50% | 0 | 0% | |
| North America | Canada | 10 | 5 | 50% | 5 | 100% |
| USA | 12 | 5 | 42% | 3 | 60% | |
| Total | 22 | 10 | 45% | 8 | 80% | |
| Oceania | Australia | 7 | 4 | 57% | 4 | 100% |
| Total | 7 | 4 | 57% | 4 | 100% | |
| Grand total | 165 | 84 | 51% | 53 | 63% |
- Note: (a) Banks are classified as green and global according to the developed classification criteria (see Subsection 3.3). (b) Data sources: for green banks: BEI, EP, UNEP FI, GABV, and our own research; for global banks: Financial Stability Board (FSB), “Banks around the World” (http://www.relbanks.com), and our own research.
3.4 Methodology
Our analysis (apart from descriptive statistics and the necessary statistical tests) involves: (a) hypothesis testing (using a t-test) to examine whether there are statistically significant differences in the mean values of CAMEL factors between green and non-green banks, and (b) the use of panel data regression techniques to investigate whether there are statistically significant differences between the two basic groups, allowing for the appropriate dynamics to capture potential persistence.
We estimate a series of models using random effects, ending with a Difference-in-Differences (DID) approach to assess differential behaviour between green and non-green banks. Following Imbens and Wooldridge (2007) and Berger and Raluca (2020), we define green banks as the “treatment” group and non-green banks as the “control” group, using the financial crisis outbreak as the time of “intervention” (see Subsections 3.2 and 3.4.1).
3.4.1 Panel data regression models & estimation methodology
The model (Equation (2)) assumes that each bank's intercept a varies over time (i.e., is random) and the slope coefficients β are constant across banks (RE model) (Gujarati, 2004; p.642, p.647). Here, u it is where μ i denotes the unobservable individual-specific effect and ν it denotes the remainder disturbance. μ i is time-invariant (e.g., country) and accounts for any individual-specific effect not included in the regression. ν it varies with individuals (e.g., banks) and time and can be thought of as the usual disturbance in the regression.
We choose to cluster on the bank level rather than country level since some countries in our sample have significantly more banks than others (see Figure 1 and Table 2).
By adding CR t × GR ijt we test for significant differences between green and non-green banks during and after the global financial crisis. Equation (3), the fully expanded version of our RE model, represents the DID approach applied to assess differences between green and non-green banks concerning CAMEL factors, where non-green bank is the control group, green bank is the treatment group, and the financial crisis outbreak (after 2007 Q4) is the time of intervention.
We consider the RE model most appropriate since it allows for time invariant variables (Wooldridge, 2002, p. 288) like country of origin and bank type (see e.g., Doumpos et al., 2017), while in the FE model, such variables' effects are absorbed by the constant term. Moreover, if differences across entities (i.e. banks) may affect the dependent variable(s), then the RE model is presumably most suitable (Clark & Linzer, 2015).
The final number of variables employed (after correlation analysis—see Subsection 4.4) is 15: (a) 11 control variables (10 CAMEL variables and bank size), (b) two main variables (dummy variables for crisis (CR) and global bank (GL)) and one key variable (dummy variable for green bank (GR)), and (c) the effect of the crisis on green banks (CR×GR), the main variable of interest. This interaction term captures the effect of the financial crisis on green banks. Specifically, we focus on this interaction term, where being a green bank during the crisis can become critical. In this case γ 5 , the coefficient of the interaction variable (see Equation 3), captures the impact of the financial crisis (CR) on the dependent variable Y i when GR equals one (and vice versa). Moreover, we have a two-way interaction describing the effect of a joint increase of CR and GR on Y i (Brambor et al., 2006; Braumoeller, 2004).
3.4.2 Main hypotheses
We test three main hypotheses:
Hypothesis a. (CAMEL variables and financial crisis): H0: γ 3 = 0, that is, the crisis has not affected CAMEL factors, against the alternative Ha: γ 3 ≠ 0, that is, the crisis has affected CAMEL factors, concerning both bank types. If the crisis (CR) estimate is not statistically significant, we accept H0 and reject Ha.
Hypothesis b. (CAMEL variables and green banks): H0: γ 4 = 0, that is, bank type does not affect CAMEL factors, against the alternative Ha: γ 4 ≠ 0, i.e., bank type affects CAMEL factors. If the green bank (GR) estimate is not statistically significant, we accept H0 and we reject Ha.
Hypothesis c. (Green banks and financial crisis): H0: γ 5 = 0, that is, the crisis has not affected green banks with respect to CAMEL factors, against the alternative Ha: γ 5 ≠ 0, i.e., the crisis has affected green banks with respect to CAMEL factors; if the interaction term (CR×GR) estimate is not statistically significant, we accept H0 and reject Ha.
The documentation for these hypotheses is presented in Online Appendix D.
4 RESULTS AND DISCUSSION
4.1 Descriptive statistics
Table 3 presents descriptive statistics for bank-specific CAMEL factors and bank size.
| Panel 1. All banks (green & non-green) | Panel 2: All banks (green & non-green): Pre- vs post-crisis period | Panel 3. Green vs non-green banks: Whole period | |||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Whole period | Pre-crisis period only | Post-crisis period only | Green banks only | Non-green banks only | |||||||||||
| Variables | Obs. | Mean | SD | Obs. | Mean | SD | Obs. | Mean | SD | Obs. | Mean | SD | Obs. | Mean | SD |
| TCR | 2435 | 15.42 | 6.28 | 567 | 13.38 | 9.84 | 1868 | 16.03 | 4.53 | 1140 | 15.52 | 3.95 | 1295 | 15.32 | 7.78 |
| CRTIER1 | 2414 | 12.54 | 4.81 | 592 | 9.98 | 5.49 | 1822 | 13.38 | 4.24 | 1154 | 12.55 | 3.83 | 1260 | 12.54 | 5.55 |
| LR | 3043 | 19.19 | 12.99 | 1050 | 18.52 | 14.46 | 1993 | 19.54 | 15.13 | 1220 | 18.52 | 10.07 | 1823 | 19.64 | 14.60 |
| NPLS | 2762 | 3.51 | 5.35 | 893 | 2.72 | 4.16 | 1869 | 3.88 | 5.79 | 1166 | 2.97 | 3.15 | 1596 | 3.90 | 6.47 |
| PROVLLLOANS | 3153 | 1.17 | 2.25 | 1093 | 1.14 | 2.30 | 2060 | 1.18 | 2.22 | 1287 | 0.93 | 1.29 | 1866 | 1.33 | 2.71 |
| NPLSRESERLL | 2356 | 129.69 | 75.58 | 762 | 105.36 | 68.29 | 1594 | 141.32 | 76.15 | 1012 | 134.90 | 70.99 | 1344 | 125.77 | 78.66 |
| RESERLLLOANS | 3006 | 3.04 | 4.52 | 1047 | 3.37 | 4.09 | 1959 | 2.86 | 4.73 | 1224 | 2.50 | 2.15 | 1782 | 3.41 | 5.57 |
| OPEXPENSOPINCOME | 2872 | 510.35 | 369.80 | 1006 | 552.31 | 378.54 | 1866 | 487.72 | 363.10 | 1195 | 515.83 | 2.15 | 1677 | 506.44 | 378.65 |
| OPEXPENSTA | 3191 | 6.20 | 6.24 | 1099 | 7.68 | 8.88 | 2092 | 5.42 | 4.03 | 1301 | 5.27 | 3.40 | 1890 | 6.84 | 7.54 |
| ROA | 2888 | 1.72 | 4.62 | 972 | 1.80 | 2.51 | 1916 | 1.67 | 5.39 | 1171 | 1.32 | 1.11 | 1717 | 1.98 | 5.91 |
| ROE | 3241 | 9.51 | 80.78 | 1087 | 14.36 | 30.19 | 2154 | 7.07 | 96.66 | 1333 | 11.65 | 21.21 | 1908 | 8.03 | 103.77 |
| LTD | 3161 | 128.06 | 366.97 | 1089 | 147.71 | 560.21 | 2072 | 117.73 | 200.75 | 1286 | 124.69 | 202.93 | 1875 | 130.37 | 445.91 |
| SIZE | 3355 | 18.05 | 2.16 | 1156 | 17.55 | 2.18 | 2199 | 18.31 | 2.10 | 1342 | 19.17 | 1.97 | 2013 | 17.30 | 1.95 |
| Panel 4: Green banks pre- vs post-crisis (green banks only) | Panel 5: Non-green banks pre- vs post-crisis (non-green banks only) | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Pre-crisis period (green banks only) | Post-crisis period (green banks only) | Pre-crisis period (non-green banks only) | Post-crisis period (non-green banks only) | |||||||||
| Variables | Obs. | Mean | SD | Obs. | Mean | SD | Obs. | Mean | SD | Obs. | Mean | SD |
| TCR | 208 | 12.08 | 2.22 | 932 | 16.29 | 3.85 | 359 | 14.14 | 12.20 | 936 | 15.77 | 5.10 |
| CRTIER1 | 231 | 8.91 | 2.40 | 923 | 13.46 | 3.57 | 361 | 10.66 | 6.68 | 899 | 13.29 | 4.83 |
| LR | 276 | 16.46 | 7.69 | 944 | 19.13 | 10.60 | 774 | 19.26 | 16.14 | 1049 | 19.92 | 13.35 |
| NPLS | 259 | 1.52 | 1.58 | 907 | 3.38 | 3.36 | 634 | 3.21 | 4.75 | 962 | 4.36 | 7.35 |
| PROVLLLOANS | 291 | 0.67 | 1.35 | 996 | 1.01 | 1.27 | 802 | 1.31 | 2.54 | 1064 | 1.34 | 2.83 |
| NPLSRESERLL | 224 | 97.40 | 64.36 | 788 | 145.52 | 69.18 | 538 | 108.68 | 69.64 | 806 | 137.18 | 88.23 |
| RESERLLLOANS | 277 | 2.13 | 1.82 | 947 | 2.60 | 2.23 | 770 | 3.81 | 4.57 | 1012 | 3.12 | 6.21 |
| OPEXPENSOPINCOME | 281 | 547.96 | 311.91 | 914 | 505.95 | 369.48 | 725 | 554.00 | 401.58 | 952 | 470.22 | 356.19 |
| OPEXPENSTA | 293 | 5.81 | 2.97 | 1008 | 5.11 | 3.50 | 806 | 8.36 | 10.13 | 1084 | 5.71 | 4.46 |
| ROA | 266 | 1.38 | 0.94 | 905 | 1.31 | 1.16 | 706 | 1.96 | 2.87 | 1011 | 2.00 | 7.32 |
| ROE | 299 | 15.19 | 39.88 | 1034 | 10.62 | 10.80 | 788 | 14.04 | 25.59 | 1120 | 3.79 | 133.59 |
| LTD | 292 | 167.20 | 414.70 | 994 | 112.21 | 46.96 | 797 | 140.57 | 604.90 | 1078 | 122.83 | 274.60 |
| SIZE | 302 | 19.11 | 1.86 | 1040 | 19.18 | 2.00 | 854 | 17.00 | 2.02 | 1159 | 17.53 | 1.86 |
- Note: All variables listed in this table are defined in Table 1. The sample data period is between December 1999 and December 2021 (yearly observations).
An analysis of these statistics, including relevant graphs comparing average CAMEL variable values for both bank types in the pre- and post-crisis periods, is in Online Appendix E.
4.2 Unit root test
We examine our variables for unit roots using the Phillips-Peron (PP) and Augmented Dickey-Fuller (ADF) tests (Choi, 2001). The null hypothesis for both tests is that the panel contains a unit root. The results are reported in Table 4.
| Variable | Fisher-type test: PP | Fisher-type test: ADF | |||
|---|---|---|---|---|---|
| Inverse χ 2 P | Inverse χ 2 P | ||||
| Statistic | p-Value | Statistic | p-Value | ||
| TCR | 576.3223 | 0.0000 | 432.7961 | 0.0000 | |
| CRTIER1 | 456.2322 | 0.0000 | 421.1506 | 0.0000 | |
| LR | 842.8484 | 0.0000 | 491.0263 | 0.0000 | |
| NPLS | 1032.3208 | 0.0000 | 829.3101 | 0.0000 | |
| PROVLLLOANS | 1655.8885 | 0.0000 | 1044.3339 | 0.0000 | |
| NPLSRESERLL | 580.8081 | 0.0000 | 349.0031 | 0.0123 | |
| RESERLLLOANS | 1121.9818 | 0.0000 | 922.4793 | 0.0000 | |
| OPEXPENSOPINCOME | 1747.6865 | 0.0000 | 799.2404 | 0.0000 | |
| OPEXPENSTA | 1131.2831 | 0.0000 | 721.6166 | 0.0000 | |
| ROA | 1395.8630 | 0.0000 | 953.2277 | 0.0000 | |
| ROE | 1533.9488 | 0.0000 | 963.4409 | 0.0000 | |
| LTD | 802.2495 | 0.0000 | 674.4124 | 0.0000 | |
| SIZE | 886.6411 | 0.0000 | 571.8224 | 0.0000 | |
- Νote: (a) The Fisher-type unit root test refers to the Phillips-Perron (PP) and Augmented Dickey-Fuller (ADF) tests, respectively; the PP test was performed with a constant term, lag (1), and no trend, and the ADF test with a constant term, lag (1), and no trend. (b) For these tests, the null hypothesis of a unit root is tested against the alternative of stationarity.
The results reject the null hypothesis, indicating that all series-variables are stationary.
4.3 t-Test
t-Test results are shown in Table 5. We form six baseline hypotheses, each decomposed into 13 null hypotheses (the number of variables under investigation), to examine if the mean value of the i th CAMEL variable equals the mean value of the ith CAMEL variable of the hypothesis (H0) under investigation. The hypotheses tested are depicted in Table 5.
| Variable | 1. Ho: Green vs. non-green banks: Whole period | 2. All banks: Pre- vs. post-crisis period | 3. Pre vs. post-crisis: Green banks only | 4. Pre vs. post-crisis: Non-green banks only | 5. Green vs. non-green banks: Only pre-crisis period | 6. Green vs. non-green banks: Only post-crisis period | ||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Obs. (combined) | Unequal Variances | Obs. (combined) | Unequal variances | Obs. (combined) | Unequal variances | Obs. (combined) | Unequal variances | Obs. (combined) | Unequal variances | Obs. (combined) | Unequal Variances | |
| TCR | 2435 |
−0.204 (0.83; 0.407) |
2435 |
−2.649*** (6.212; 0.000) |
1140 |
−4.211*** (21.194; 0.000) |
1295 |
−1.635** (2.459; 0.014) |
567 |
2.052*** (3.108; 0.002) |
1868 |
−0.519** (2.484; 0.013) |
| CRTIER1 | 2414 |
−0.010 (0.05; 0.957) |
2414 |
−3.398*** (13.775; 0.000) |
1154 |
−4.552*** (23.099; 0.000) |
1260 |
2.626*** (6.799; 0.000) |
592 |
1.757*** (4.557; 0.000) |
1822 |
−0.169 (0.848; 0.397) |
| LR | 3043 |
1.118** (2.50; 0.01) |
3043 |
−1.026* (1.964; 0.049) |
1220 |
−2.669*** (4.626; 0.000) |
1283 |
−0.668 (0.939; 0.348) |
1050 |
2.799*** (3.772; 0.000) |
1993 |
0.798 (1.485; 0.138) |
| NPLS | 2762 |
0.931*** (4.99; 0.000) |
2762 |
−1.166*** (6.034; 0.000) |
1166 |
−1.863*** (12.537; 0.000) |
1596 |
−1.148*** (3.790; 0.000) |
893 |
1.688*** (7.941; 0.000) |
1869 |
0.973*** (3.714; 0.000) |
| PROVLLLOANS | 3153 |
0.396*** (5.47; 0.000) |
3153 |
−0.043 (0.503; 0.615) |
1287 |
−0.346*** (3.893; 0.000) |
1866 |
−0.031 (0.249; 0.804) |
1093 |
0.646*** (5.394; 0.000) |
2060 |
0.331*** (3.467; 0.000) |
| NPLSRESERLL | 2536 |
−9.129*** (2.95; 0.003) |
2536 |
−35.960*** (11.512; 0.000) |
1012 |
−48.163*** (9.717; 0.000) |
1344 |
−28.499*** (6.832; 0.000) |
762 |
11.283** (2.151; 0.032) |
1594 |
−8.381** (2.204; 0.028) |
| RESERLLLOANS | 3006 |
0.917*** (6.30; 0.000) |
3006 |
0.505*** (3.048; 0.002) |
1224 |
−0.471*** (−3.595; 0.000) |
1782 |
0.705*** (2.762; 0.006) |
1047 |
1.682*** (8.513; 0.000) |
1959 |
0.506** (2.429; 0.015) |
| OPEXPENSOPINCOME | 2872 |
−9.385 (0.68; 0.499) |
2872 |
64.589*** (4.425; 0.000) |
1195 |
42.013* (1.887; 0.059) |
1677 |
83.774*** (4.442; 0.000) |
1006 |
6.038 (0.253; 0.800) |
1866 |
−35.723** (2.751; 0.034) |
| OPEXPENSTA | 3191 |
1.570*** (7.96; 0.000) |
3191 |
6.197*** (8.019; 0.000) |
1301 |
0.697*** (3.389; 0.000) |
1890 |
2.655*** (6.958; 0.000) |
1099 |
2.553*** (6.433; 0.000) |
2092 |
0.595*** (3.409; 0.000) |
| ROA | 2888 |
0.661*** (4.519; 0.000) |
2888 |
0.126 (0.858; 0.391) |
1171 |
0.075 (1.077; 0.282) |
1717 |
−0.045 (0.178; 0.859) |
972 |
0.577*** (4.710; 0.000) |
1916 |
0.697*** (2.984; 0.003) |
| ROE | 3241 |
−3.619 (1.480; 0.139) |
3241 |
7.288*** (3.203; 0.001) |
1333 |
4.570* (1.961; 0.051) |
1908 |
10.249** (2.503; 0.012) |
1087 |
- 1.148 (0.463; 0.644) |
2154 |
−6.827* (1.704; 0.089) |
| LTD | 3161 |
5.674 (0.483; 0.629) |
3161 |
29.977* (1.709; 0.088) |
1286 |
54.989** (2.262; 0.025) |
1875 |
17.743 (0.771; 0.441) |
1089 |
−26.628 (0.823; 0.411) |
2072 |
10.618 (1.250; 0.212) |
| SIZE | 3355 |
−1.861*** (26.919; 0.000) |
3355 |
−0.763*** (9.752; 0.000) |
1342 |
−0.077 (0.621; 0.535) |
2013 |
−0.533*** (6.045; 0.000) |
1156 |
−2.109*** (16.547; 0.000) |
2199 |
−1.653*** (19.963; 0.000) |
- Note: (a) Testable hypotheses: 1. H0: The mean value of the i th CAMEL variable of green banks equals the mean value of the i th CAMEL variable of non-green banks. 2. H0: The mean value of the i th CAMEL variable of both bank types in the pre-crisis period equals the mean value of the i th CAMEL variable of both bank types in the post-crisis period. 3. H0: The mean value of the i th CAMEL variable of green banks in the pre-crisis period equals the mean value of the i th CAMEL variable of green banks in the post-crisis period. 4. H0: The mean value of the i th CAMEL variable of non-green banks in the pre-crisis period equals the mean value of the i th CAMEL variable of non-green banks in the post-crisis period. 5. H0: The mean value of the i th CAMEL variable of green banks in the pre-crisis period equals the mean value of the i th CAMEL variable of non-green banks in the pre-crisis period. 6. H0: The mean value of the i th CAMEL variable of green banks in the post-crisis period equals the mean value of the i th CAMEL variable of non-green banks in the post-crisis period. (b) First value denotes difference in means; absolute values in parentheses denote t and p values, respectively. (c) t-test critical values: t = 1.96, p = 0.05: if t > 1.96 and p < 0.05, there is a statistically significant difference at the 5% level; otherwise, there is no statistically significant difference (see: Verbeek, 2000; p. 23). (d) Number of asterisks denotes significance level: *** at the 1% level (p-value <0.01), ** at the 5% level (0.01 < p-value <0.05), * at the 10% level (0.05 < p-value <0.1). (e) Statistically significant cases are indicated by bold characters.
The t-test results show statistically significant differences in the mean values of CAMEL variables in 59 out of 78 cases examined. This indicates that the majority of the CAMEL variables are good candidates for distinguishing between pre- and post-crisis periods as well as between green and non-green banks in both time periods. However, these differences might be driven by other bank characteristics or factors. Additionally, the results do not necessarily imply that green banks outperform non-green banks or vice versa. This necessitates further analysis to investigate if green (“treatment” group) and non-green banks (“control” group) exhibit differential behaviour.
4.4 Correlation analysis
We employ Pearson pairwise correlation (r) to check for potential multicollinearity among variables. We performed two separate correlation analyses: (a) one with all CAMEL variables and bank size, and (b) one with the remaining variables after examining the size of the correlation coefficients. Correlation matrices can be found in Tables 6 and 7. Specifically, to avoid multicollinearity problems, we excluded all variables with a correlation coefficient r ≥ 0.50, which can be considered a threshold between comparatively low and moderate correlations (Gujarati, 2004, p. 359).
| Variables | TCR | CRTIER1 | LR | NPLS | PROVLL LOANS | NPLS RESERLL | RESERL LLOANS | OPEXPENS OPINCOME | OPEXP ENSTA | ROA | ROE | LTD | SIZE |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| TCR | 1.000 | ||||||||||||
| CRTIER1 | 0.949 | 1.000 | |||||||||||
| LR | 0.292 | 0.281 | 1.000 | ||||||||||
| NPLS | 0.176 | 0.195 | 0.052 | 1.000 | |||||||||
| PROVLLLOANS | 0.425 | 0.378 | 0.169 | 0.365 | 1.000 | ||||||||
| NPLSRESERLL | 0.077 | 0.108 | −0.034 | 0.370 | −0.142 | 1.000 | |||||||
| RESERLLLOANS | 0.297 | 0.284 | 0.105 | 0.770 | 0.655 | −0.096 | 1.000 | ||||||
| OPEXPENSOPINCOME | −0.103 | −0.124 | −0.009 | 0.199 | 0.109 | 0.158 | 0.135 | 1.000 | |||||
| OPEXPENSTA | 0.381 | 0.340 | 0.203 | 0.272 | 0.803 | −0.193 | 0.560 | 0.169 | 1.000 | ||||
| ROA | 0.413 | 0.399 | 0.305 | 0.016 | 0.376 | −0.246 | 0.282 | −0.192 | 0.598 | 1.000 | |||
| ROE | 0.116 | 0.106 | 0.045 | −0.167 | −0.184 | −0.040 | −0.117 | 0.030 | −0.044 | 0.360 | 1.000 | ||
| LTD | −0.029 | −0.026 | 0.521 | −0.004 | −0.035 | 0.059 | −0.045 | 0.178 | 0.013 | 0.043 | −0.036 | 1.000 | |
| SIZE | −0.277 | −0.309 | −0.210 | −0.173 | −0.230 | 0.073 | −0.272 | 0.005 | −0.414 | −0.386 | 0.003 | −0.059 | 1.000 |
- Note: (a) All variables are defined in Table 1. (b) Data sources: Thomson Reuters Eikon (ex-DataStream) and own calculations. (c) Pairwise correlations above r = 0.50 are marked with red bold characters.
| Variables | TCR | CRTIER1 | LR | NPLS | PROVLL LOANS | NPLS RESERLL | OPEXPENS OPINCOME | ROA | ROE | LTD | SIZE |
|---|---|---|---|---|---|---|---|---|---|---|---|
| TCR | 1.000 | ||||||||||
| CRTIER1 | 0.949 | 1.000 | |||||||||
| LR | 0.292 | 0.281 | 1.000 | ||||||||
| NPLS | 0.176 | 0.195 | 0.052 | 1.000 | |||||||
| PROVLLLOANS | 0.425 | 0.378 | 0.169 | 0.365 | 1.000 | ||||||
| NPLSRESERLL | 0.077 | 0.108 | −0.034 | 0.370 | −0.142 | 1.000 | |||||
| OPEXPENSOPINCOME | −0.103 | −0.124 | −0.009 | 0.199 | 0.109 | 0.158 | 1.0000 | ||||
| ROA | 0.413 | 0.399 | 0.305 | 0.016 | 0.376 | −0.246 | −0.192 | 1.000 | |||
| ROE | 0.116 | 0.106 | 0.045 | −0.167 | −0.184 | −0.040 | 0.030 | 0.360 | 1.000 | ||
| LTD | −0.029 | −0.026 | 0.521 | −0.004 | −0.035 | 0.059 | 0.178 | 0.043 | −0.036 | 1.000 | |
| SIZE | −0.277 | 0.309 | −0.211 | −0.173 | −0.230 | 0.073 | 0.005 | −0.386 | 0.003 | −0.059 | 1.0000 |
- Note: (a) All variables are defined in Table 1. (b) Data sources: Thomson Reuters Eikon (ex-DataStream) and own calculations. (c) Pairwise correlations above r = 0.50 are marked with red bold characters.
The correlation coefficients of the remaining variables are all below 0.50 (see Table 7), 5 suggesting no evidence of multicollinearity problems.
4.5 Main results: Regressions results and discussion
In Tables 8–17, we present the estimation results of the RE model.
| Depended variable: TCR | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Method of estimation: RE GLS | [1] | [2] | [3] | [4] | [5] | [6] | [7] | [8] | [9] |
| TCR (T-1) |
0.805*** (26.31) |
0.463*** (2.67) |
0.459*** (2.65) |
0.459*** (2.65) |
0.431** (2.57) |
0.391** (2.42) |
0.389** (2.40) |
0.383** (2.39) |
|
| CRTIER1(T-1) |
0.637*** (8.96) |
||||||||
| LR(T-1) |
0.553*** (4.25) |
0.096** (2.48) |
0.097** (2.51) |
0.098** (2.48) |
0.107*** (2.61) |
0.991*** (2.68) |
0.100*** (2.68) |
0.099*** (2.73) |
|
| NPLS(T-1) |
−0.007 (0.25) |
0.057 (1.50) |
0.060 (1.59) |
0.059 (1.57) |
0.080* (1.77) |
0.095** (2.05) |
0.094** (2.05) |
0.098** (2.16) |
|
| PROVLLLOANS(T-1) |
0.095 (1.33) |
0.130 (0.72) |
0.126 (0.70) |
0.132 (0.72) |
0.190 (0.82) |
0.152 (0.73) |
0.154 (0.74) |
0.156 (0.75) |
|
| NPLSRESERLL(T-1) |
0.001 (1.24) |
0.003* (1.92) |
0.003* (1.89) |
0.003* (1.90) |
−9.86e-06 (0.01) |
0.003* (1.65) |
−0.003* (1.66) |
−0.004* (1.92) |
|
| OPEXPENSOPINCOME(T-1) |
−0.0003 (1.19) |
−0.0005 (1.47) |
−0.0005 (1.46) |
−0.0006 (1.47) |
−0.0007* (1.84) |
−0.0008** (2.15) |
−0.0009** (2.17) |
−0.0009** (2.26) |
|
| ROA(T-1) |
−0.084 (0.64) |
0.029 (0.17) |
0.351 (0.21) |
0.023 (0.14) |
−0.174 (0.86) |
−0.080 (0.40) |
−0.084 (0.42) |
−0.096 (0.48) |
|
| ROE(T-1) |
−0.002*** (4.18) |
−0.001 (1.49) |
−0.001 (1.48) |
−0.001 (1.42) |
−0.0006 (0.55) |
−0.0005 (0.48) |
−0.0005 (0.45) |
−0.0004 (0.39) |
|
| LTD(T-1) |
−0.003 (0.98) |
−0.008 (1.53) |
−0.008 (1.54) |
−0.008 (1.54) |
−0.009 (1.47) |
−0.008 (1.44) |
−0.008 (1.44) |
−0.008 (1.36) |
|
| SIZE(T-1) |
0.034 (0.48) |
−0.006 (0.06) |
0.214* (1.92) |
−0.155 (1.40) |
−0.198 (1.57) |
−0.187 (1351) |
|||
| GLOBAL |
0.276 (0.61) |
−0.346 (0.84) |
0.422 (1.13) |
0.315 (0385) |
0.252 (0.67) |
||||
| COUNTRY | YES | YES | YES | YES | |||||
| CRISIS |
2.146*** (4.477) |
2.135*** (4.83) |
1.580*** (4.17) |
||||||
| NEWGREEN |
0.350 (1.18) |
−0.592** (2.05) |
|||||||
| INTERACTION (= CRISIS × NEWGREEN) |
1.219*** (3.01) |
||||||||
| CONSTANT |
3.176*** (7.10) |
6.646*** (7.92) |
7.071*** (3.30) |
6.492** (2.48) |
7.099** (2.31) |
3.387 (1.24) |
9.489*** (2.92) |
10.223*** (2.83) |
10.574*** (2.92) |
| Sigma u i | 0 | 0.766 | 1.022 | 1.038 | 1.026 | 0.715 | 0.710 | 0.717 | 0.735 |
| Sigma e i | 2.895 | 2.043 | 2.288 | 2.239 | 2.239 | 2.239 | 2.180 | 2.180 | 2.156 |
| Rho | 0 | 0.123 | 0.166 | 0.177 | 0.174 | 0.093 | 0.096 | 0.098 | 0.104 |
| Observations | 2252 | 1404 | 1461 | 1461 | 1461 | 1461 | 1461 | 1461 | 1461 |
| No. of countries | 38 | 38 | 38 | 38 | 38 | 33 | 33 | 33 | 33 |
| Diagnostics: | |||||||||
| R 2: within | 0.614 | 0.412 | 0.315 | 0.318 | 0.316 | 0.331 | 0.383 | 0.384 | 0.394 |
| R 2: between | 0.969 | 0.890 | 0.837 | 0.837 | 0.842 | 0.861 | 0.877 | 0.876 | 0.873 |
| R 2: overall | 0.755 | 0.714 | 0.626 | 0.624 | 0.627 | 0.686 | 0.712 | 0.713 | 0.714 |
| Wald chi2(1) test: p-value | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 |
- Note: Please refer footnote of Table 17.
| Depended variable: CRTIER1 | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Method of estimation: RE GLS | [1] | [2] | [3] | [4] | [5] | [6] | [7] | [8] | [9] |
| TCR (T-1) |
0.367** (2.15) |
||||||||
| CRTIER1(T-1) |
0.869*** (47.11) |
0.828*** (14.95) |
0.829*** (14.28) |
0.828*** (14.26) |
0.762*** (10.43) |
0.705*** (8.83) |
0.704*** (8.83) |
0.700*** (8.80) |
|
| LR(T-1) |
0.124*** (3.09) |
0.023*** (3.12) |
0.024*** (3.13) |
0.024*** (3.18) |
0.038*** (3.04) |
0.033*** (2.75) |
0.034*** (2.76) |
0.034*** (2.84) |
|
| NPLS(T-1) |
0.085** (2.23) |
0.005 (0.18) |
0.006 (0.21) |
0.005 (0.19) |
0.036 (0.95) |
0.052 (1.44) |
0.052 (1.44) |
0.053 (1.47) |
|
| PROVLLLOANS(T-1) |
0.014 (0.08) |
−0.064 (1.33) |
−0.066 (1.41) |
0.061 (1.25) |
−0.029 (0.51) |
−0.028 (0.49) |
−0.027 (0.48) |
−0.014 (0.25) |
|
| NPLSRESERLL(T-1) |
0.008*** (4.08) |
0002*** (2.83) |
0.003*** (2.72) |
0.002*** (2.73) |
0.002 (1.53) |
−0.0007 (0.58) |
−0.0007 (0.59) |
−0.0009 (0.73) |
|
| OPEXPENSOPINCOME(T-1) |
−0.001*** (2.62) |
−0.0003* (1.79) |
−0.0003* (1.69) |
−0.0003* (1.74) |
−0.0005* (1.86) |
−0.0006** (2.20) |
−0.0006** (2.25) |
−0.0006** (2.36) |
|
| ROA(T-1) |
−0.007 (0.03) |
−0.051 (0.41) |
−0.046 (0.41) |
−0.058 (0.50) |
−0.151 (1.23) |
−0.033 (0.25) |
−0.036 (0.26) |
−0.034 (0.25) |
|
| ROE(T-1) |
−0.002 (1.53) |
−0.003*** (9.11) |
−0.003*** (8.74) |
−0.003*** (8.51) |
−0.003*** (6.65) |
−0.003*** (6.67) |
−0.003*** (6.70) |
−0.003*** (6.79) |
|
| LTD(T-1) |
−0.013** (2.09) |
−0.0004 (0.34) |
−0.0004 (0.34) |
−0.0005 (0.40) |
−0.003 (1.07) |
−0.003 (1.17) |
−0.003 (1.17) |
−0.003 (1.10) |
|
| SIZE(T-1) |
0.008 (0.17) |
−0.014 (0.25) |
0.056 (0.80) |
−0.122 (1.63) |
−0.132* (1.67) |
−0.127 (1.61) |
|||
| GLOBAL |
0.126 (0.74) |
0.467 (0.21) |
0.403** (2.19) |
0.378** (2.05) |
0.364** (1.96) |
||||
| COUNTRY | YES | YES | YES | YES | |||||
| CRISIS |
1.597*** (4.56) |
1.594*** (4.55) |
1.288*** (3.73) |
||||||
| NEWGREEN |
0.081 (0.59) |
−0.465** (2.29) |
|||||||
| INTERACTION (= CRISIS × NEWGREEN) |
0.693*** (3.23) |
||||||||
| CONSTANT |
1.862*** (7.96) |
5.690*** (2.61) |
1.953*** (3.29) |
1.796 (1.29) |
2.162 (1.42) |
1.638 (0.95) |
4.662** (2.39) |
4.846** (2.42) |
5.034** (2.50) |
| Sigma u i | 0 | 1.688 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| Sigma e i | 2.152 | 2.317 | 1.879 | 1.837 | 1.837 | 1.837 | 1.802 | 1.800 | 1.789 |
| Rho | 0 | 0.347 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| Observations | 2233 | 1398 | 1438 | 1438 | 1438 | 1438 | 1438 | 1438 | 1438 |
| No. of countries | 38 | 38 | 38 | 38 | 38 | 34 | 34 | 34 | 34 |
| Diagnostics: | |||||||||
| R 2: within | 0.580 | 0.341 | 0.565 | 0.565 | 0.564 | 0.573 | 0.603 | 0.603 | 0.607 |
| R 2: between | 0.982 | 0.627 | 0.972 | 0.973 | 0.973 | 0.969 | 0.971 | 0.972 | 0.971 |
| R 2: overall | 0.789 | 0.538 | 0.813 | 0.813 | 0.813 | 0.82 | 0.838 | 0.838 | 0.839 |
| Wald chi2(1) test: p-value | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 |
- Note: Please refer footnote of Table 17.
| Depended variable: LR | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Method of estimation: RE GLS | [1] | [2] | [3] | [4] | [5] | [6] | [7] | [8] | [9] |
| TCR (T-1) |
0.298** (2.39) |
0.001 (0.05) |
−0.002 (0.07) |
0.0006 (0.02) |
0.036 (1.39) |
0.036 (1.41) |
0.038 (1.47) |
0.037 (1.45) |
|
| CRTIER1(T-1) | |||||||||
| LR(T-1) |
0.921*** (34.25) |
0.890*** (32.39) |
0.889*** (33.10) |
0.887*** (32.55) |
0.799*** (25.86) |
0.799*** (25.60) |
0.799*** (25.68) |
0.799*** (25.61) |
|
| NPLS(T-1) |
−0.071 (0.57) |
0.032 (0.97) |
0. 028 (0.83) |
0.030 (0.90) |
0.007 (0.14) |
0.007 (0.13) |
0.006 (0.13) |
0.007 (0.13) |
|
| PROVLLLOANS(T-1) |
0 0.059 (0.98) |
−0.299* (1.87) |
−0.290* (1.76) |
−0.303* (1.83) |
−0.260 (1.29) |
−0.259 (1.28) |
−0.262 (1.30) |
−0.260 (1.28) |
|
| NPLSRESERLL(T-1) |
0.004 (0.98) |
−0.002* (1.80) |
−0.002* (1.80) |
−0.002* (1.83) |
−0.002 (1.11) |
−0.002 (1.08) |
−0.002 (1.06) |
−0.002 (1.09) |
|
| OPEXPENSOPINCOME(T-1) |
−0.0009 (1.10) |
0.0002 (0.36) |
0.0001 (0.27) |
0.0002 (0.37) |
0.0004 (0.72) |
0.0004 (0.72) |
0.0004 (0.77) |
0.0004 (0.76) |
|
| ROA(T-1) |
1.055** (2.39) |
0.210 (0.99) |
0.166 (0.68) |
0.200 (0.80) |
0.353 (1.23) |
0.353 (1.21) |
0.353 (1.22) |
0.354 (1.22) |
|
| ROE(T-1) |
−0.0012 (0.60) |
0.00003 (0.04) |
0.0002 (0.23) |
0.0001 (0.12) |
−0.0004 (0.35) |
−0.0004 (0.35) |
−0.0005 (0.37) |
−0.0005 (0.37) |
|
| LTD(T-1) |
0.071*** (3.06) |
0.008* (1.73) |
0.008* (1.74) |
0.009* (1.76) |
0. 008 (1.23) |
0. 008 (1.23) |
0.007 (1.24) |
0.008 (1.25) |
|
| SIZE(T-1) |
−0.060 (1.01) |
−0.005 (0.07) |
0.085 (1.27) |
0.087 (1.34) |
0.112 (1.59) |
0.114 (1.60) |
|||
| GLOBAL |
−0.314 (1.63) |
−0.492** (2.03) |
−0.495** (2.13) |
−0.429* (1.93) |
−0.436* (1.95) |
||||
| COUNTRY | YES | YES | YES | YES | |||||
| CRISIS |
−0.017 (0.06) |
−0.010 (0.04) |
−0.118 (0.36) |
||||||
| NEWGREEN |
−0.209 (1.10) |
−0.404 (0.87) |
|||||||
| INTERACTION (= CRISIS × NEWGREEN) |
0.244 (0.49) |
||||||||
| CONSTANT |
1.588*** (3.49) |
3.847 (1.20) |
1.173** (2.19) |
2.411 (1.63) |
1.470 (0.86) |
1.031 (0.64) |
1.004 (0.66) |
0.554 (0.35) |
0.614 (0.39) |
| Sigma u i | 0 | 5.285 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| Sigma e i | 4.460 | 4.627 | 3.472 | 3.467 | 3.467 | 3.467 | 3.465 | 3.467 | 3.468 |
| Rho | 0 | 0.566 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| Observations | 2808 | 1447 | 1447 | 1447 | 1447 | 1447 | 1447 | 1447 | 1447 |
| No. of countries | 38 | 38 | 38 | 38 | 38 | 33 | 33 | 33 | 33 |
| Diagnostics: | |||||||||
| R 2 : within | 0.552 | 0.196 | 0.543 | 0.543 | 0.543 | 0.544 | 0.544 | 0.544 | 0.545 |
| R 2 : between | 0.995 | 0.376 | 0.980 | 0.980 | 0.980 | 0.980 | 0.980 | 0.980 | 0.980 |
| R 2: overall | 0.869 | 0.350 | 0.841 | 0.841 | 0.841 | 0.850 | 0.850 | 0.850 | 0.850 |
| Wald chi2(1) test: p-value | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 |
- Note: Please refer footnote of Table 17.
| Depended variable: NPLS | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Method of estimation: RE GLS | [1] | [2] | [3] | [4] | [5] | [6] | [7] | [8] | [9] |
| TCR (T-1) |
−0.008 (0.50) |
−0.317** (2.16) |
−0.030** (2.09) |
−0.030** (2.09) |
−0.025* (1.82) |
−0.037** (2.12) |
−0.037** (2.13) |
−0.036** (2.13) |
|
| CRTIER1(T-1) | |||||||||
| LR(T-1) |
−0.002 (0.12) |
0.007 (1.10) |
0.007 (1.20) |
0.007 (1.20) |
0.005 (0.61) |
0.0007 (0.09) |
0.0006 (0.07) |
0.0006 (0.07) |
|
| NPLS(T-1) |
0.819*** (12.26) |
0.848*** (13.15) |
0.851*** (13.15) |
0.851*** (13.24) |
0.777*** (9.35) |
0.781*** (9.40) |
0.781*** (9.39) |
0.781*** (9.39) |
|
| PROVLLLOANS(T-1) |
0.722*** (2.84) |
0.116** (2.35) |
0.109** (2.19) |
0.110** (2.22) |
0.163*** (2.67) |
0.160** (2.56) |
0.158** (2.55) |
0.159** (2.57) |
|
| NPLSRESERLL(T-1) |
0.016*** (8.64) |
0.0008 (0.65) |
0.0008 (0.62) |
0.0008 (0.63) |
0.001 (0.87) |
0.0005 (0.28) |
0.0005 (0.30) |
0.0005 (0.29) |
|
| OPEXPENSOPINCOME(T-1) |
0.0009 (1.11) |
0.0003 (0.98) |
0.0003 (1.06) |
0.0003 (1.02) |
0.0003 (0.76) |
0.0002 (0.71) |
0.0003 (0.75) |
0.0005 (0.75) |
|
| ROA(T-1) |
−0.100 (0.32) |
0.171* (1.69) |
0.202** (1.98) |
0.199* (1.85) |
0.121 (1.08) |
0.151 (1.44) |
0.152 (1.44) |
0.152 (1.44) |
|
| ROE(T-1) |
−0.004** (2.52) |
−0.003*** (4.63) |
−0.003*** (4.79) |
−0.003*** (4.66) |
−0.003*** (3.86) |
−0.003*** (3.88) |
−0.003*** (3.89) |
−0.003*** (3.89) |
|
| LTD(T-1) |
0.004 (0.73) |
0.002 (1.05) |
0.001 (1.01) |
0.001 (1.02) |
0.002 (1.06) |
0.002 (1.07) |
0.002 (1.09) |
0.003 (1.11) |
|
| SIZE(T-1) |
0.042* (1.91) |
0.038 (1.29) |
0.107*** (3.30) |
0.050 (1.42) |
0.062* (1.65) |
0.062* (1.66) |
|||
| GLOBAL |
0.023 (0.21) |
−0.061 (0.52) |
0.059 (0.47) |
0.089 (0.70) |
0.088 (0.69) |
||||
| COUNTRY | YES | YES | YES | YES | |||||
| CRISIS |
0.584*** (4.16) |
0.587*** (4.16) |
0.587*** (2.94) |
||||||
| NEWGREEN |
−0.096 (1.21) |
−0.129 (0.98) |
|||||||
| INTERACTION (= CRISIS × NEWGREEN) |
0.042 (0.24) |
||||||||
| CONSTANT |
0.588*** (3.06) |
−0.398 (0.67) |
0.065 (0.27) |
−0.810 (1.60) |
−0.742 (1.14) |
−2.025*** (2.67) |
−1.108 (1.37) |
−1.323 (1.57) |
−1.312 (1.56) |
| Sigma u i | 0 | 1.753 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| Sigma e i | 2.683 | 2.152 | 1.463 | 1.448 | 1.448 | 1.448 | 1.443 | 1.444 | 1.444 |
| Rho | 0 | 0.399 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| Observations | 2562 | 1456 | 1456 | 1456 | 1456 | 1456 | 1456 | 1456 | 1456 |
| No. of countries | 38 | 38 | 38 | 38 | 38 | 33 | 33 | 33 | 33 |
| Diagnostics: | |||||||||
| R 2: within | 0.472 | 0.255 | 0.655 | 0.656 | 0.656 | 0.660 | 0.666 | 0.666 | 0.666 |
| R 2: between | 0.982 | 0.380 | 0.954 | 0.955 | 0.955 | 0.959 | 0.960 | 0.960 | 0.961 |
| R 2: overall | 0.706 | 0.302 | 0.803 | 0.803 | 0.803 | 0.816 | 0.820 | 0.820 | 0.820 |
| Wald chi2(1) test: p-value | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 |
- Note: Please refer footnote of Table 17.
| Depended variable: PROVLLLOANS | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Method of estimation: RE GLS | [1] | [2] | [3] | [4] | [5] | [6] | [7] | [8] | [9] |
| TCR (T-1) |
0.002 (0.39) |
−0.011** (1.98) |
−0.010* (1.78) |
−0.008 (1.57) |
−0.002 (0.41) |
−0.005 (0.80) |
−0.004 (0.77) |
−0.004 (0.75) |
|
| CRTIER1(T-1) | |||||||||
| LR(T-1) |
0.004 (0.47) |
−0.004 (1.34) |
−0.003 (1.23) |
−0.005* (1.67) |
0.0002 (0.04) |
−0.0005 (0.11) |
−0.0005 (0.12) |
−0.0005 (0.12) |
|
| NPLS(T-1) |
0.050** (2.15) |
0.017 (1.38) |
0.019 (1.54) |
0.020 (1.63) |
0.210 (1.29) |
0.022 (1.38) |
0.022 (1.38) |
0.022 (1.38) |
|
| PROVLLLOANS(T-1) |
0.277** (2.57) |
0.737*** (12.97) |
0.734*** (12.72) |
0.724*** (12.53) |
0.622*** (9.37) |
0.621*** (9.43) |
0.621*** (9.39) |
0.620*** (9.35) |
|
| NPLSRESERLL(T-1) |
−0.0004 (0.60) |
0.00004 (0.12) |
0.00002 (0.06) |
0.000 (0.03) |
0.00002 (0.06) |
−0.0002 (0.37) |
−0.0002 (0.36) |
−0.0001 (0.32) |
|
| OPEXPENSOPINCOME(T-1) |
0.0006*** (3.16) |
0.0003 (1.56) |
0.0003 (1.59) |
0.0003* (1.71) |
0.0003* (1.65) |
0.0003 (1.64) |
0.0003 (1.63) |
0.0003 (1.63) |
|
| ROA(T-1) |
0.310*** (3.00) |
0.278*** (3.54) |
0.295*** (3.50) |
0.321*** (3.66) |
0.350*** (3.30) |
0.357*** (3.35) |
0.358*** (3.35) |
0.358*** (3.35) |
|
| ROE(T-1) |
−0.002*** (4.06) |
0.0002 (0.71) |
0.0002 (0.46) |
0.00007 (0.22) |
−0.0003 (0.69) |
−0.0003 (0.67) |
−0.0003 (0.67) |
−0.0003 (0.67) |
|
| LTD(T-1) |
0.003 (1.48) |
0.0009 (1.29) |
0.0009 (1.23) |
0.001 (1.53) |
0.002* (1.67) |
0.002* (1.66) |
0.002* (1.67) |
0.002 (1.64) |
|
| SIZE(T-1) |
0.022* (1.73) |
0.064*** (2.72) |
0.039** (1.96) |
0.026 (1.22) |
0.028 (1.17) |
0.028 (1.15) |
|||
| GLOBAL |
−0.237*** (2.99) |
−0.223** (2.48) |
−0.196** (2.24) |
−0.190** (2.38) |
−0.188** (2.38) |
||||
| COUNTRY | YES | YES | YES | YES | |||||
| CRISIS |
0.130** (2.28) |
0.131** (2.29) |
0.159** (2.14) |
||||||
| NEWGREEN |
−0.019 (0.40) |
0.033 (0.57) |
|||||||
| INTERACTION (= CRISIS × NEWGREEN) |
−0.065 (0.91) |
||||||||
| CONSTANT |
0.860*** (6.16) |
−0.257 (0.95) |
−0.208 (1.34) |
−0.668* (1.78) |
−1.384** (2.53) |
−1.200** (2.41) |
−0.998** (1.94) |
−1.039* (1.83) |
−1.056* (1.85) |
| Sigma u i | 0.625 | 0.832 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| Sigma e i | 1.564 | 0.710 | 0.690 | 0.688 | 0.688 | 0.688 | 0.688 | 0.688 | 0.688 |
| Rho | 0.138 | 0.578 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| Observations | 2978 | 1477 | 1474 | 1474 | 1474 | 1474 | 1474 | 1474 | 1474 |
| No. of countries | 38 | 38 | 38 | 38 | 38 | 33 | 33 | 33 | 33 |
| Diagnostics: | |||||||||
| R 2 : within | 0.025 | 0.078 | 0.099 | 0.102 | 0.105 | 0.113 | 0.116 | 0.116 | 0.116 |
| R 2 : between | 0.834 | 0.440 | 0.950 | 0.949 | 0.952 | 0.955 | 0.955 | 0.956 | 0.956 |
| R 2: overall | 0.273 | 0.335 | 0.722 | 0.723 | 0.727 | 0.745 | 0.745 | 0.746 | 0.746 |
| Wald chi2(1) test: p-value | 0.010 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 |
- Note: Please refer footnote of Table 17.
| Depended variable: NPLSRESERLL | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Method of estimation: RE GLS | [1] | [2] | [3] | [4] | [5] | [6] | [7] | [8] | [9] |
| TCR (T-1) |
1.232** (2.24) |
−0.108 (0.40) |
−0.050 (0.19) |
−0.028 (0.10) |
−0.180 (0.73) |
−0.494** (2.02) |
−0.513** (2.11) |
−0.531** (2.20) |
|
| CRTIER1(T-1) | |||||||||
| LR(T-1) |
0.687 (1.54) |
−0.065 (0.42) |
−0.055 (0.36) |
−0.077 (0.48) |
−0.218 (1.08) |
−0.310 (1.51) |
−0.307 (1.49) |
−0.304 (1.46) |
|
| NPLS(T-1) |
7.371*** (4.15) |
−0.034 (0.08) |
0.061 (0.14) |
0.076 (0.17) |
0.092 (0.21) |
0.281 (0.65) |
0.286 (0.66) |
0.304 (0.70) |
|
| PROVLLLOANS(T-1) |
−10.778*** (3.15) |
−2.847*** (3.00) |
−3.122*** (3.26) |
−3.266*** (3.43) |
−3.071*** (2.83) |
−2.962*** (2.75) |
−2.930*** (2.72) |
−2.854*** (2.64) |
|
| NPLSRESERLL(T-1) |
0.840*** (49.71) |
0.840*** (37.23) |
0.839*** (36.30) |
0.838*** (37.19) |
0.786*** (33.08) |
0.761*** (29.27) |
0.761*** (29.37) |
0.758*** (29.20) |
|
| OPEXPENSOPINCOME(T-1) |
0.011 (1.22) |
0.006 (1.46) |
0.006* (1.66) |
0.007* (1.82) |
0.006 (1.40) |
0.006 (1.21) |
0.005 (1.13) |
0.005 (1.11) |
|
| ROA(T-1) |
−3.847 (0.96) |
3.177** (2.24) |
4.295*** (2.84) |
4.706*** (3.09) |
5.760*** (3.42) |
6.143*** (3.68) |
6.124*** (3.68) |
6.131*** (3.70) |
|
| ROE(T-1) |
−0.215* (1.76) |
−0.156*** (3.39) |
−0.176*** (3.66) |
−0.176*** (3.85) |
−0.182*** (3.64) |
−0.154*** (3.87) |
−0.154*** (3.87) |
−0.154*** (3.85) |
|
| LTD(T-1) |
0.110* (1.73) |
0.071** (2.58) |
0.069*** (2.24) |
0.073*** (2.69) |
0.071** (2.45) |
0.074** (2.31) |
0.073** (2.26) |
0.077** (2.33) |
|
| SIZE(T-1) |
1.286* (1.96) |
1.921** (2.67) |
2.408** (2.46) |
0.932 (0.86) |
0.574 (0.54) |
0.626 (0.58) |
|||
| GLOBAL |
−3.674 (1.38) |
−9.325*** (2.85) |
−6.212* (1.88) |
−7.115** (2.11) |
−7.308** (2.15) |
||||
| COUNTRY | YES | YES | YES | YES | |||||
| CRISIS |
15.263*** (4.65) |
15.156*** (4.64) |
11.570*** (2.76) |
||||||
| NEWGREEN |
2.881 (1.33) |
−3.684 (0.81) |
|||||||
| INTERACTION (= CRISIS × NEWGREEN) |
8.148* (1.67) |
||||||||
| CONSTANT |
20.733*** (9.48) |
81.834*** (5.13) |
13.434*** (2.61) |
−13.165 (0.89) |
−24.248 (1.64) |
−20.827 (1.02) |
3.188 (0.14) |
9.644 (0.43) |
11.592 (0.52) |
| Sigma u i | 0 | 45.474 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| Sigma e i | 38.515 | 51.654 | 37.349 | 37.150 | 37.150 | 37.151 | 36.749 | 36.748 | 36.672 |
| Rho | 0 | 0.437 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| Observations | 2087 | 1469 | 1398 | 1398 | 1398 | 1398 | 1398 | 1398 | 1398 |
| No. of countries | 38 | 38 | 38 | 38 | 38 | 33 | 33 | 33 | 33 |
| Diagnostics: | |||||||||
| R 2: within | 0.513 | 0.143 | 0.506 | 0.508 | 0.508 | 0.508 | 0.520 | 0.520 | 0.522 |
| R 2: between | 0.965 | 0.241 | 0.950 | 0.949 | 0.950 | 0.947 | 0.950 | 0.950 | 0.949 |
| R 2: overall | 0.713 | 0.185 | 0.729 | 0.730 | 0.731 | 0.743 | 0.748 | 0.748 | 0.749 |
| Wald chi2(1) test: p-value | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 |
- Note: Please refer footnote of Table 17.
| Depended variable: OPEXPENSOPINCOME | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Method of estimation: RE GLS | [1] | [2] | [3] | [4] | [5] | [6] | [7] | [8] | [9] |
| TCR (T-1) |
−3.791 (0.71) |
0.211 (0.06) |
0.263 (0.07) |
0.255 (0.07) |
0.528 (0.11) |
0.261 (0.05) |
0.238 (0.05) |
0.185 (0.04) |
|
| CRTIER1(T-1) | |||||||||
| LR(T-1) |
−5.511** (2.18) |
−2.444* (1.69) |
−2.439* (1.69) |
−2.434* (1.68) |
−2.085 (1.10) |
−2.152 (1.14) |
−2.147 (1.14) |
−2.154 (1.14) |
|
| NPLS(T-1) |
−3.701 (0.53) |
−0 0.627 (0.14) |
−0.533 (0.12) |
−0.539 (0.12) |
−5.381 (1.19) |
−5.146 (1.11) |
−5.139 (1.10) |
−5.121 (1.11) |
|
| PROVLLLOANS(T-1) |
45.690*** (3.57) |
7.582 (0.80) |
7.301 (0.76) |
7.345 (0.74) |
−3.804 (0.34) |
−4.011 (0.36) |
−3.963 (0.35) |
−3.682 (0.33) |
|
| NPLSRESERLL(T-1) |
−0.019 0.10) |
−0.049 (0.42) |
−0.051 (0.43) |
−0.051 (0.43) |
−0.027 (0.20) |
−0.051 (0.36) |
−0.051 (0.35) |
−0.058 (0.41) |
|
| OPEXPENSOPINCOME(T-1) |
0.561*** (20.38) |
0.577*** (13.14) |
0.578*** (12.95) |
0.578*** (12.57) |
0.497*** (8.70) |
0.498*** (8.76) |
0.498*** (8.68) |
0.497*** (8.68) |
|
| ROA(T-1) |
−5.091 (0.31) |
2.632 (0.25) |
3.889 (0.37) |
3.805 (0.36) |
3.572 (0.25) |
3.683 (0.26) |
3.620 (0.25) |
3.626 (0.25) |
|
| ROE(T-1) |
−6.231*** (3.33) |
−0.558 (0.55) |
−0.604 (0.62) |
−0.610 (0.60) |
−0.857 (0.90) |
−0.591 (0.65) |
−0.584 (0.64) |
0.603 (0.66) |
|
| LTD(T-1) |
1.644*** (3.99) |
0.672*** (2.75) |
0.671*** (2.76) |
0.670*** (2.75) |
0.343 (1.31) |
0.348 (1.33) |
0.347 (1.33) |
0.355 (1.38) |
|
| SIZE(T-1) |
1.336 (0.25) |
1.170 (0.18) |
−6.956 (0.93) |
−8.712 (1.12) |
−9.195 (1.13) |
−9.072 (1.12) |
|||
| GLOBAL |
0.954 (0.05) |
−10.175 (0.41) |
−6.904 (0.27) |
−8.000 (0.30) |
−8.490 (0.32) |
||||
| COUNTRY | YES | YES | YES | YES | |||||
| CRISIS |
17.429 (0.78) |
17.388 (0.78) |
7.909 (0.30) |
||||||
| NEWGREEN |
3.737 (0.21) |
−13.283 (0.53) |
|||||||
| INTERACTION (= CRISIS × NEWGREEN) |
21.388 (0.74) |
||||||||
| CONSTANT |
204.821*** (13.81) |
536.212*** (5.70) |
159.360*** (2.62) |
132.253 (1.01) |
135.267 (0.87) |
313.802* (1.79) |
336.889* (1.84) |
345.438* (1.81) |
351.828* (1.83) |
| Sigma u i | 30.163 | 166.020 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| Sigma e i | 239.126 | 267.778 | 233.734 | 233.802 | 233.802 | 233.802 | 233.754 | 233.843 | 233.710 |
| Rho | 0.016 | 0.278 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| Observations | 2558 | 1488 | 1404 | 1404 | 1404 | 1404 | 1404 | 1404 | |
| No. of countries | 38 | 38 | 38 | 38 | 38 | 33 | 33 | 33 | |
| Diagnostics: | |||||||||
| R 2: within | 0.138 | 0.053 | 0.100 | 0.100 | 0.100 | 0.102 | 0.103 | 0.103 | 0.104 |
| R 2: between | 0.900 | 0.331 | 0.879 | 0.879 | 0.879 | 0.840 | 0.841 | 0.840 | 0.839 |
| R 2: overall | 0.381 | 0161 | 0.397 | 0.397 | 0.397 | 0.435 | 0.435 | 0.435 | 0.435 |
| Wald chi2(1) test: p-value | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 |
- Note: Please refer footnote of Table 17.
| Depended variable: ROA | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Method of estimation: RE GLS | [1] | [2] | [3] | [4] | [5] | [6] | [7] | [8] | [9] |
| TCR (T-1) |
0.003 (0.40) |
0.002 (0.74) |
0.0006 (0.22) |
−0.0003 (0.10) |
−0.003 (1.07) |
0.003 (0.81) |
0.002 (0.78) |
0.002 (0.77) |
|
| CRTIER1(T-1) | |||||||||
| LR(T-1) |
0.013* (1.96) |
0.004* (1.84) |
0.003 (1.43) |
0.004* (1.82) |
0.004* (1.71) |
0.006** (2.33) |
0.006** (2.33) |
0.006** (2.32) |
|
| NPLS(T-1) |
−0.029** (1.97) |
−0.004 (0.48) |
−0.007 (0.80) |
−0.008 (0.87) |
−0.008 (0.69) |
−0.010 (0.85) |
−0.010 (0.85) |
−0.010 (0.85) |
|
| PROVLLLOANS(T-1) |
0.110 (1.48) |
0.078*** (2.67) |
0. 084*** (2.92) |
0.090*** (3.12) |
0.098*** (3.10) |
0.101*** (3.25) |
0.101*** (3.25) |
0.101*** (3.27) |
|
| NPLSRESERLL(T-1) |
−0.002*** (5.05) |
−0.0008*** (4.10) |
−0.0008*** (3.94) |
−0.001*** (3.72) |
−0.001*** (2.86) |
−0.001* (1.70) |
−0.001* (1.70) |
−0.001* (1.66) |
|
| OPEXPENSOPINCOME(T-1) |
−0.0004*** (4.64) |
−0.0001 (1.39) |
−0.0002 (1.62) |
−0.0002* (1.77) |
−0.0002** (2.08) |
−0.0002** (2.09) |
−0.0002** (2.09) |
−0.0002** (2.05) |
|
| ROA(T-1) |
0.746*** (3.20) |
0.802*** (18.56) |
0.773*** (17.03) |
0.759*** (16.16) |
0.655*** (13.33) |
0.640*** (13.14) |
0.640*** (13.14) |
0.640*** (13.14) |
|
| ROE(T-1) |
0.0007*** (3.13) |
−0.002*** (10.04) |
−0.002*** (9.31) |
−0.002*** (8.96) |
−0.002*** (7.33) |
−0.002*** (7.38) |
−0.002*** (7.37) |
−0.002*** (7.38) |
|
| LTD(T-1) |
0.003* (1.71) |
0.002 (0.60) |
0.003 (0.74) |
0.0002 (0.43) |
0.001 (1.34) |
0.001 (1.35) |
0.001 (1.34) |
0.001 (1.36) |
|
| SIZE(T-1) |
−0.038*** (3.96) |
−0.062*** (3.71) |
−0.027 (1.61) |
−0.001 (0.06) |
−0.003 (0.14) |
−0.003 (0.14) |
|||
| GLOBAL |
0.138** (2.38) |
0.083 (1.47) |
0.028 (0.49) |
0.023 (0.45) |
0.023 (0.45) |
||||
| COUNTRY | YES | YES | YES | YES | |||||
| CRISIS |
−0.264*** (5.74) |
−0.264*** (5.73) |
−0.265*** (3.93) |
||||||
| NEWGREEN |
0.014 (0.37) |
0.011 (0.19) |
|||||||
| INTERACTION (= CRISIS × NEWGREEN) |
0.003 (0.04) |
||||||||
| CONSTANT |
0.419 (1.05) |
1.254*** (4.24) |
0.263*** (3.25) |
1.052*** (4.43) |
1.466*** (4.14) |
0.949** (2.50) |
0.542 (1.55) |
0.571 (1.47) |
0.572 (1.44) |
| Sigma u i | 0 | 0.636 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| Sigma e i | 2.301 | 0.609 | 0.562 | 0.556 | 0.556 | 0.556 | 0.552 | 0.552 | 0.552 |
| Rho | 0 | 0.522 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| Observations | 2655 | 1479 | 1441 | 1441 | 1441 | 1441 | 1441 | 1441 | 1441 |
| No. of countries | 38 | 38 | 38 | 38 | 38 | 33 | 33 | 33 | 33 |
| Diagnostics: | |||||||||
| R 2: within | 0.004 | 0.101 | 0.212 | 0.219 | 0.222 | 0.224 | 0.247 | 0.247 | 0.247 |
| R 2: between | 0.997 | 0.384 | 0.952 | 0.947 | 0.947 | 0.963 | 0.962 | 0.962 | 0.962 |
| R 2: overall | 0.505 | 0.275 | 0.739 | 0.742 | 0.745 | 0.763 | 0.769 | 0.769 | 0.769 |
| Wald chi2(1) test: p-value | 0.001 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 |
- Note: Please refer footnote of Table 17.
| Depended variable: ROE | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Method of estimation: RE GLS | [1] | [2] | [3] | [4] | [5] | [6] | [7] | [8] | [9] |
| TCR (T-1) |
−0.036 (0.34) |
−0.037 (0.37) |
−0.048 (0.51) |
−0.083 (0.98) |
−0.133* (1.76) |
−0.026 (0.45) |
−0.026 (0.45) |
−0.030 (0.51) |
|
| CRTIER1(T-1) | |||||||||
| LR(T-1) |
−0.254** (2.37) |
−0.238* (1.91) |
−0.240* (1.90) |
−0.209* (1.84) |
−0.192* (1.65) |
−0.160 (1.39) |
−0.160 (1.39) |
−0.160 (1.39) |
|
| NPLS(T-1) |
−0.424** (2.26) |
−0.410** (2.05) |
−0.427** (2.07) |
−0.458** (2.17) |
−0.286 (0.93) |
−0.334 (1.07) |
−0.334 (1.07) |
−0.332 (1.06) |
|
| PROVLLLOANS(T-1) |
0.644 (1.12) |
0.817 (1.61) |
0.867 (1.62) |
1.076* (1.92) |
0.808 (1.18) |
0.832 (1.20) |
0.832 (1.20) |
0.852 (1.23) |
|
| NPLSRESERLL(T-1) |
−0.021*** (3.35) |
−0.021*** (3.26) |
−0.021*** (3.25) |
−0.020*** (3.16) |
−0.020** (2.33) |
−0.012 (1.41) |
−0.012 (1.41) |
−0.013 (1.39) |
|
| OPEXPENSOPINCOME(T-1) |
−0.009** (2.20) |
−0.009** (2.24) |
−0.009** (2.21) |
−0.010** (2.24) |
−0.011** (1.98) |
−0.011* (1.95) |
−0.011* (1.95) |
−0.011* (1.94) |
|
| ROA(T-1) |
2.828*** (4.86) |
2.460*** (4.69) |
2.258*** (3.95) |
1.708*** (2.60) |
0.921 (1.11) |
0.669 (0.79) |
0.669 (0.79) |
0.661 (0.78)) |
|
| ROE(T-1) |
0.289*** (2.92) |
0.034 (0.58) |
0.038 (0.66) |
0.038 (0.74) |
0.036 (0.68) |
0.026 (0.62) |
0.026 (0.62) |
0.027 (0.63) |
|
| LTD(T-1) |
0.014 (1.24) |
0.014 (1.22) |
0.014 (1.22) |
0.010 (0.91) |
0.001 (0.07) |
−0.001 (0.07) |
−0.001 (0.09) |
−0.001 (0.05) |
|
| SIZE(T-1) |
−0.236 (0.97) |
−1.133* (1.87) |
−0.947 (1.10) |
−0.347 (0.41) |
−0.572 (0.56) |
−0.562 (0.56) |
|||
| GLOBAL |
5.169** (2.19) |
4.346 (1.29) |
3.079 (0.92) |
2.514 (0.86) |
2.475 (0.85) |
||||
| COUNTRY | YES | YES | YES | YES | |||||
| CRISIS |
−6.030*** (6.14) |
−6.088*** (6.24) |
−6.720*** (4.21) |
||||||
| NEWGREEN |
1.813 (1.19) |
0.661 (0.44) |
|||||||
| INTERACTION (= CRISIS × NEWGREEN) |
1.434 (0.69) |
||||||||
| CONSTANT |
6.183*** (2.89) |
19.021*** (8.43) |
18.672*** (6.47) |
23.523*** (3.24) |
39.106*** (2.87) |
37.501* (1.87) |
28.196 (1.44) |
32.219 (1.43) |
32.582 (1.43) |
| Sigma u i | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| Sigma e i | 81.125 | 19.082 | 19.065 | 18.934 | 19.934 | 18.934 | 18.894 | 18.901 | 18.908 |
| Rho | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| Observations | 3067 | 1491 | 1489 | 1489 | 1489 | 1489 | 1489 | 1489 | 1489 |
| No. of countries | 38 | 38 | 38 | 38 | 38 | 33 | 33 | 33 | 33 |
| Diagnostics: | |||||||||
| R 2: within | 0.003 | 0.032 | 0.029 | 0.031 | 0.038 | 0.044 | 0.056 | 0.056 | 0.056 |
| R 2: between | 0.204 | 0.322 | 0.349 | 0.343 | 0.374 | 0.458 | 0.450 | 0.453 | 0.453 |
| R 2: overall | 0.008 | 0.088 | 0.088 | 0.089 | 0.099 | 0.122 | 0.132 | 0.133 | 0.133 |
| Wald chi2(1) test: p-value | 0.004 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 |
- Note: Please refer footnote of Table 17.
| Depended variable: LTD | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Method of estimation: RE GLS | [1] | [2] | [3] | [4] | [5] | [6] | [7] | [8] | [9] |
| TCR (T-1) |
−0.972* (1.91) |
0.114 (0.47) |
0.035 (0.15) |
0.030 (0.12) |
0.091 (0.27) |
0.410 (0.89) |
0.401 (0.89) |
0.414 0.93 |
|
| CRTIER1(T-1) | |||||||||
| LR(T-1) |
3.840*** (2.40) |
0.611* (1.95) |
0.589* (1.89) |
0.593* (1.90) |
1.255* (1.77) |
1.345* (1.83) |
1.347* (1.83) |
1.348* (1.83) |
|
| NPLS(T-1) |
0.293 (0.21) |
0.513 (0.53) |
0. 382 (0.43) |
0.378 (0.44) |
0.429 (0.38) |
0.291 (0.28) |
0.293 (0.28) |
0.284 (0.27) |
|
| PROVLLLOANS(T-1) |
−6.498 (1.61) |
−2.656 (1.30) |
−2.381 (1.19) |
−2.352 (1.16) |
−3.504 (1.29) |
−3.352 (1.33) |
−3.333 (1.33) |
−3.403 (1.37) |
|
| NPLSRESERLL(T-1) |
0.041 (0.57) |
0.011 (0.26) |
0.012 (0.29) |
0.013 (0.30) |
0.028 (0.63) |
0.052 (1.04) |
0.052 (1.04) |
0.053 (1.11) |
|
| OPEXPENSOPINCOME(T-1) |
0.018** (2.50) |
−0.003 (0.69) |
−0.004 (0.89) |
−0.004 (0.86) |
0.0001 (0.03) |
0.001 (0.14) |
0.0005 (0.10) |
0.0006 (0.13) |
|
| ROA(T-1) |
9.883 (1.33) |
3.755 (1.00) |
2.481 (0.62) |
2.407 (0.57) |
3.564 (0.84) |
2.640 (0.69) |
2.624 (0.68) |
2.642 (0.69) |
|
| ROE(T-1) |
−0.060 (1.44) |
−0.007 (0.39) |
−0.002 (0.09) |
−0.002 (0.08) |
−0.011 (0.51) |
−0.012 (0.58) |
−0.012 (0.57) |
−0.012 (0.58) |
|
| LTD(T-1) |
0.388*** (10.10) |
0.996*** (10.27) |
0. 998*** (10.38) |
0. 998*** (10.20) |
0.962*** (4.81) |
0.957*** (4.90) |
0.0.956*** (4.89) |
0.954*** (4.78) |
|
| SIZE(T-1) |
−1.718 (1.55) |
−1.838 (1.23) |
−3.776* (1.95) |
−2.145* (1.73) |
−2.329 (1.55) |
−2.368 (1.50) |
|||
| GLOBAL |
0.689 (0.23) |
8.631 (1.64) |
5.154 (1.51) |
4.695* (1.65) |
4.838* (1.67) |
||||
| COUNTRY | YES | YES | YES | YES | |||||
| CRISIS |
−16.351* (1.89) |
−16.400* (1.88) |
−13.995 (1.52) |
||||||
| NEWGREEN |
1.485 (0.59) |
5.854 (0.47) |
|||||||
| INTERACTION (= CRISIS × NEWGREEN) |
−5.454 (0.41) |
||||||||
| CONSTANT |
72.586*** (9.45) |
37.418 (0.91) |
−15.750 (0.66) |
19.854 (0.55) |
21.933 (0.51) |
39.956 (0.83) |
14.506 (0.31) |
17.783 (0.36) |
16.384 (0.36) |
| Sigma u i | 12.487 | 31.042 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| Sigma e i | 158.711 | 53.717 | 48.436 | 47.740 | 47.740 | 47.740 | 47.728 | 47.718 | 47.720 |
| Rho | 0.006 | 0.250 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| Observations | 2995 | 1478 | 1478 | 1478 | 1478 | 1478 | 1478 | 1478 | 1478 |
| No. of countries | 38 | 38 | 38 | 38 | 38 | 33 | 33 | 33 | 33 |
| Diagnostics: | |||||||||
| R 2: within | 0.397 | 0.144 | 0.303 | 0.305 | 0.306 | 0.318 | 0.326 | 0.326 | 0.326 |
| R 2: between | 0.846 | 0.305 | 0.945 | 0.945 | 0.945 | 0.977 | 0.978 | 0.977 | 0.977 |
| R 2: overall | 0.456 | 0.238 | 0.574 | 0.576 | 0.576 | 0.600 | 0.606 | 0.606 | 0.606 |
| Wald chi2(1) test: p-value | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 |
- Note: Tables 8–17: (a) Each table shows the results of each CAMEL factor as the dependent variable against the rest of the CAMEL factors, bank size, and dummy variables (global bank, country, financial crisis, green bank, and interaction = crisis × green bank) as independent variables. All right-hand side variables (excluding dummies) are lagged by one period to avoid possible endogeneity issues. Note that the number of observations varies across the regressions depending on the missing values for the bank-specific variables. In all specifications, case [1] is the basic specification of our model, which is expanded gradually by adding successive control variables (CAMEL factors). From specification (or column hereafter) [5] onwards, we include successive dummy variables in the regressions, adding them one by one to control for the effect on the dependent variable of global bank (column 5), country (column 6), crisis (column 7), green bank (column 8), and interaction = crisis × green bank (column 9). (b) Specification 9, which is the fully expanded version of our model as we include all variables simultaneously (see Equation (3) in Subsection 3.4.1), represents the DID approach implemented to test for differences between the control and the treatment group. (c) Method of estimation: Random Effects (RE) Generalized Least Squares (GLS). (d) Data period: 1999–2021. (e) Group variable: bank. (f) All variables are defined in Table 1. (g) Absolute values in parentheses denote heteroskedasticity-robust z-statistics (two-tailed test): the superscripts ***, **, and * denote coefficients statistically different from zero at the 1% (p-value <0.01), 5% (0.01 < p-value <0.05), and 10% (0.05 < p-value <0.1) levels, respectively. (h) In Table 9—columns 6, 7, 8, and 9, four countries (country IDs: 3 = Bangladesh, 11 = Ecuador, 15 = India, and 38 = USA) were omitted because of collinearity. In Table 8 and in Tables 10–17 (columns 6, 7, 8, and 9), five countries (country IDs: 1 = Argentina, 3 = Bangladesh, 11 = Ecuador, 15 = India, and 38 = USA) were omitted because of collinearity. (i) rho is the fraction of variance due to u i and is considered the Intraclass Correlation Coefficient (ICC). (j) Bold characters indicate the statistically significant values of the key variables and the main variable of interest.
- a: “Crisis has not affected CAMEL factors”.
The crisis dummy is statistically significant in eight out of 10 cases.
The crisis contributed to the increase in risk-adjusted capital ratios (TCR and CRTIER1) for both bank types, a result that may be attributed to the capital injections by the state and the limitation of high risk-weighted loans during and after the crisis. However, the leverage ratio (LR) remained unaffected, supporting earlier studies that found LR to be less sensitive to systemic shocks and financial crises. This suggests that LR is a more objective measure of capital adequacy, and that a higher LR increases the soundness of banks during crisis periods and reduces the incentives for regulatory arbitrage (Ahnert et al., 2021; Drakos & Malandrakis, 2021; Haq & Heaney, 2012; Hogan, 2015; Smith et al., 2017).
Regarding asset quality, the crisis negatively impacted NPLs, increasing the NPLs ratio for both green and non-green banks. This is expected given the financial crisis's severity, which led to increased NPLs and reduced lending. Conversely, it positively affected provisions (PROVLLLOANS) and reserves for loan losses (NPLSRESERLL), indicating lower credit risk and better capitalization (se e.g. Alessi et al., 2021).
The management quality proxy (OPEXPENSOPINCOME) was unaffected by the crisis, 6 aligning with t-test results for this ratio (see Table 5).
The crisis had a highly significant negative effect on earning ability, as both ROA and ROE ratios deteriorated during and after the crisis, a result that is consistent with previous studies findings (see e.g., Baselga-Pascual et al., 2015).
For liquidity (LTD), the crisis had a negative impact at the 10% level, but this effect vanished with the introduction of the interaction term, possibly due to decreased green banks loans or increased deposits.
- b: “Bank type (i.e., green) does not affect CAMEL factors.”
The dummy variable for green banks is insignificant in eight out of 10 cases. Examining capital adequacy, our results show no differences between green and non-green banks in terms of the LR, excluding TCR and CRTIER1. Similarly, there are no differences in asset quality, management quality, earning ability, and liquidity proxies between green and non-green banks. This finding aligns with our prior belief that the asset quality, managerial quality, earnings, and liquidity of green banks do not differentiate them from non-green counterparts. This could be because the positive outcomes from green loans and investments are too early to reflect on green banks' fundamentals, given their relatively nascent stage in green banking. 7 These results are consistent with Scholtens and Dam (2007), who found no significant improvement in financial performance between adopters and non-adopters of the Equator Principles. Park and Kim (2020) also noted that assessing green banking effectiveness is premature due to insufficient data availability. Furthermore, the green bank estimate (NEWGREEN) is negative and statistically significant (at the 5% level) for the TCR and CRTIER1 ratios, indicating that green banks do not outperform non-green banks in terms of the risk-adjusted capital adequacy ratios, particularly in the pre-crisis period.
- c: “Crisis has not affected green banks.”
The main variable of interest, that is, the interaction between green banks and the crisis, is statistically insignificant for LR, two out of three asset quality proxies (NPLs and provisions ratios), management quality, earning ability, and liquidity proxies. These results indicate that green banks did not perform better or worse during and after the financial crisis in terms of the majority of the CAMEL factors examined.
- Higher capital increments after the crisis, possibly combined with increased capital injections by the State.
- Limitation of exposures to environmentally unfriendly economic sectors (e.g., chemical industries).
- Gradual increase in the share of green loans and investments (which seem to be less risky—EBF, 2017, p. 38) coupled with the implementation of the “green supporting factor” (GSF) (a lower risk-weighting factor for green loans), resulting in lower capital requirements for credit risk and higher capital ratios, assuming also a “green differentiated capital requirements” scheme (Dafermos & Nikolaidi, 2021).
This suggests that green banks might be less risky and more sound than non-green banks from a risk-adjusted capital perspective in the post-crisis period. However, the interaction term with LR is statistically insignificant (see Table 10), indicating no difference between green and non-green banks in terms of LR during and after the crisis.
Moreover, the interaction variable with respect to the NPLSRESERLL is statistically significant (although at the 10% level—see Table 13), indicating that green banks exhibit differentiated behaviour compared to non-green banks, particularly in presenting a better coverage ratio for NPLs.
In summary, we reject the null hypothesis for TCR, CRTIER1, and NPLSRESERLL, concluding that the crisis benefited green banks only in terms of the two risk-adjusted capital ratios and partially in the coverage ratio for NPLs. However, we do not reject the null hypothesis for the remaining CAMEL factors, indicating that green banks did not exhibit differential behaviour in the post-crisis period with respect to the rest of the CAMEL factors.
Finally, we analyse two of the important main control variables, which are bank size and global bank.
Bank size: Overall, bank size does not significantly impact the capital adequacy ratios of either bank type. However, the impact on asset quality variables is mixed. Initially, the NPLs ratio shows a positive and significant relationship with bank size, but this becomes insignificant after the introduction of the crisis dummy variable, only to become significant again (albeit at the 10% level) after the introduction of the green bank dummy variable. This suggests that the credit risk of green banks may increase with their size (see also Table 3—Panel 4). This result aligns partially with previous studies by Čihák and Hesse (2010) and Laeven et al. (2014). The other two asset quality indicators become insignificant after the introduction of the crisis variable. Additionally, size does not impact OPEXPENSOPINCOME, implying that management quality remains unaffected by size. Bank size also becomes insignificant concerning earning ability once country dummies are included in the regressions. Finally, there is no statistically significant impact of bank size on liquidity.
Global bank: The impact of a bank's global presence, whether green or non-green, on CAMEL variables is not consistent across all estimations. First, the global presence has a positive and significant impact on the risk-adjusted capital ratio CRTIER1 and a negative and significant impact on the LR after the introduction of the crisis variable. This suggests that the worldwide presence of a bank contributes positively to the risk-adjusted capital ratio but negatively to the non-risk-adjusted leverage ratio, a result that is partially in line with previous studies such as Altunbas et al. (2017). 8 Second, the global presence has a negative and significant impact on asset quality indicators, excluding the NPLs ratio. Thirdly, it has no significant impact on management quality and earning ability ratios. Finally, there is a statistically significant positive effect (although at the 10% level) on the LTD after the introduction of the green bank dummy variable. This suggests that multinational green banks may have benefited in terms of liquidity from their global activities, particularly in the post-crisis period.
“See also Online Appendix F, where the main findings obtained from the empirical analysis for a narrower sample period (1999–2015) from an earlier version of this paper are discussed.”
5 CONCLUSION AND POLICY IMPLICATIONS
Our results show that green banks, whether global or not, generally do not differ from their non-green counterparts in terms of most of the CAMEL ratios before and after the financial crisis. However, we find that the financial crisis has a statistically significant positive effect on Total Capital and Tier 1 Capital ratios (at the 1% level), as well as on the NPLs loss reserve ratio (albeit weakly significant), only during and after the financial crisis. These findings provide a first indication that green banks surpass non-green banks only in terms of the risk-adjusted capital ratios—and partially in terms of one of the asset quality ratios—in the post-crisis period. These results do not necessarily suggest that they are better capitalized and less risky than their non-green counterparts, as their better performance with respect to the aforementioned ratios may be the result of higher capital injections by the State (during and after the crisis) and perhaps of other factors such as variations in regulatory rules across countries and/or banks of different sizes. We also find that the crisis affected both bank types, positively influencing capital adequacy (except the Leverage Ratio), negatively impacting NPLs but positively affecting provisions and NPLs coverage ratios, and negatively affecting earnings ability and liquidity.
In summary, green banks do not significantly differ from non-green banks regarding leverage ratio, asset quality, management quality, earning ability, and liquidity. Thus, there is no strong evidence that green banks are more stable, and their contribution to financial stability is rather limited.
Since our main findings indicate that green banks are not necessarily better in terms of credit risk and other regulatory performance ratios, lower capital requirements and a different regulatory regime for green banks cannot be justified. This is in line with Boot and Schoenmaker (2018), who argue that banks will need not less but more capital in the new era. Moreover, green banks presumably are not green enough, given that the exposures of leading global (many of them green) banks in environmentally unfriendly sectors are still high ($742 billion USD in 2021 vs. $723 billion USD in 2016 according to the Banking on Climate Chaos (2022)). In addition, favourable measures for green loans, such as a “green-supporting factor” (GSF) and a “brown-penalizing factor” (BPF), should not yet be adopted without sufficient evidence on green loans' risk levels and their contribution to environmental protection (i.e., GHG mitigation) and green investments' returns. Such measures may result in the reduction of banks' capital and the creation of general and/or systematic financial instability (Berenguer et al., 2020; Dafermos et al., 2018), especially considering that, according to Oehmke and Opp (2022), the regulator's ability to use capital requirements to reduce dirty lending is limited. Also, a sudden implementation of a green regulatory framework may result in regulatory arbitrage. An interim measure such as the “Sustainable Finance Supporting Factor,” proposed by the EBF (2019), and a gradual movement from polluting to eco-friendly projects and sectors is advised to avoid significant direct and indirect negative effects such as a sudden increase in energy costs or an abrupt depreciation of fossil fuel reserves, which could contribute to generalized financial instability (Batten et al., 2016; Manninen & Tiililä, 2020; Nieto, 2017) as a result, inter alia, of “stranded assets” (Papandreou, 2019). Finally, it must be considered that banks face not only the ever-rising climate change risk but also other significant downside risks from the effects of inflation, rising market interest rates, slower economic growth, and geopolitical uncertainty (FDIC, 2023).
Regarding future research directions, this study can be extended by employing a larger sample size of green (commercial) and non-green banks, incorporating data for green loans/bank/country/region, adjusted for the effect of GSF and DPF and using CAMEL factors or other financial stability measures (e.g., Z-score). In the future, additional metrics such as the Green Asset Ratio (GAR) 9 can be incorporated, which can serve as an additional criterion for classifying a bank as green or not.
ACKNOWLEDGEMENTS
Special thanks are due to the two anonymous referees for their constructive and insightful remarks. The authors also would like to thank the participants of the 2021 Annual Conference of the Scottish Economic Society, and of the 10th Intl. Conference of the Financial Engineering and Banking Society (FEBS) 2021 for their helpful comments and suggestions.
CONFLICT OF INTEREST STATEMENT
The authors declare no conflicts of interest.
Endnotes
APPENDIX A
| Bank ID | Bank name | Country | Region | Global (Y/N) | Green (Y/N) | Member of: | Member since: | Member also of: | Member since: | Member also of: | Member since: | Total assets mean value (1999–2021) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 | ABN AMRO BANK N.V. | Netherlands | Europe | YES | YES | UNEP FI | 2006 | EP | 2009 | 460,000,000 | ||
| 2 | Absa Bank Limited (Absa) | South Africa | Africa | NO | NO | 80,500,000 | ||||||
| 3 | Access Bank Plc. | Nigeria | Africa | YES | YES | UNEP FI | 2009 | EP | 2009 | 10,400,000 | ||
| 5 | Akbank | Turkey | Europe | NO | NO | 56,900,000 | ||||||
| 6 | Albaraka Türk Katılım Bankası A.Ş (alBaraka) | Turkey | Europe | NO | NO | 6,314,659 | ||||||
| 7 | Attijariwafa bank | Morocco | Africa | NO | NO | 32,600,000 | ||||||
| 8 | Australia & New Zealand Banking Group Limited (ANZ) | Australia | Oceania | YES | YES | UNEP FI | 2003 | EP | 2009 | 447,000,000 | ||
| 9 | Auswide Bank Ltd | Australia | Oceania | NO | NO | 1,767,207 | ||||||
| 10 | Banca Monte dei Paschi di Siena S.p.A. (BMPS) | Italy | Europe | YES | NO | 197,000,000 | ||||||
| 11 | Banco Bradesco S.A. | Brazil | Latin America | YES | YES | EP | 2004 | UNEP FI | 2011 | 233,000,000 | ||
| 12 | Banco de Crédito | Peru | Latin America | NO | YES | EP | 2013 | 25,300,000 | ||||
| 13 | Banco de Crédito e Inversiones (BCI) | Chile | Latin America | NO | NO | 32,600,000 | ||||||
| 14 | Banco de Galicia y Buenos Aires S.A. | Argentina | Latin America | NO | YES | EP | 2007 | UNEP FI | 2011 | 10,900,000 | ||
| 15 | Banco de la Produccion S.A. Produbanco | Ecuador | Latin America | NO | NO | 2,979,787 | ||||||
| 16 | Banco do Brasil | Brazil | Latin America | YES | YES | EP | 2006 | 289,000,000 | ||||
| 17 | Banco do Estado do Rio Grande do Sul S.A. (“Banrisul”) | Brazil | Latin America | NO | NO | 13,700,000 | ||||||
| 18 | Banco do Nordeste | Brazil | Latin America | NO | NO | 9,946,335 | ||||||
| 19 | Banco Pichincha C.A. | Ecuador | Latin America | NO | YES | UNEP FI | 2011 | 7,621,492 | ||||
| 20 | Banco Popolare | Italy | Europe | NO | NO | 132,000,000 | ||||||
| 21 | Banco Popular Espaňol | Spain | Europe | YES | YES | EP | 2013 | 125,000,000 | ||||
| 22 | Banco Sabadell | Spain | Europe | YES | YES | EP | 2011 | 143,000,000 | ||||
| 23 | Banco Santander (Mexico) S.A, | Mexico | Latin America | NO | NO | 49,200,000 | ||||||
| 24 | Banco Santander S.A. | Spain | Europe | YES | YES | UNEP FI | 1992 | EP | 2009 | BEI | 2010 | 1,220,000,000 |
| 25 | Bancolombia S.A. | Colombia | Latin America | NO | YES | UNEP FI | 2011 | EP | 2008 | 37,100,000 | ||
| 26 | Bank Central Asia (BCA) | Indonesia | Asia Pacific | NO | NO | 37,100,000 | ||||||
| 27 | Bank J. Safra Sarasin Ltd | Switzerland | Europe | YES | NO | – | n/a | 18,900,000 | ||||
| 28 | Bank Muscat (SOAG)/Bank Muscat S.A.O.G. | Oman | Middle East | NO | YES | UNEP FI | 2007 | EP | 2007 | 21,200,000 | ||
| 29 | Bank of America | USA | North America | YES | YES | UNEP FI | 2001 | EP | 2004 | 1,780,000,000 | ||
| 30 | Bank of Montreal | Canada | North America | YES | YES | UNEP FI | 1993 | 406,000,000 | ||||
| 31 | Bank of Queensland Ltd | Australia | Oceania | NO | NO | 26,500,000 | ||||||
| 32 | Bankia Sau | Spain | Europe | NO | NO | 281,000,000 | ||||||
| 33 | Bankinter, S.A. | Spain | Europe | NO | NO | 64,800,000 | ||||||
| 34 | Banque Cantonale Vaudoise (BCV) | Switzerland | Europe | NO | NO | 37,700,000 | ||||||
| 35 | Barclays Africa | South Africa | Africa | NO | NO | 67,800,000 | ||||||
| 36 | Barclays Group plc | UK | Europe | YES | YES | UNEP FI | 1997 | EP | 2003 | BEI | 2010 | 1,640,000,000 |
| 37 | BBVA Continental | Peru | Latin America | NO | NO | 14,000,000 | ||||||
| 38 | BBVA Group/Banco Bilbao Vizcaya Argentaria, S.A. (BBVA) | Spain | Europe | YES | YES | UNEP FI | 1999 | EP | 2004 | 625,000,000 | ||
| 39 | Bendigo and Adelaide Bank Group | Australia | Oceania | NO | NO | 34,500,000 | ||||||
| 40 | Beneficial State Bank | USA | North America | NO | YES | GABV | 2009 | 4,568,805 | ||||
| 41 | BMCE Bank (BANQUE MAROCAINE) | Morocco | Africa | NO | YES | EP | 2009 | UNEP FI | 2000 | 19,500,000 | ||
| 42 | BNK Financial Group | South Korea | Asia Pacific | NO | NO | 45,800,000 | ||||||
| 43 | BNP Paribas | France | Europe | YES | YES | EP | 2008 | UNEP FI | 2010 | BEI | 2010 | 2,040,000,000 |
| 44 | BRAC Bank | Bangladesh | Asia Pacific | NO | YES | GABV | 2009 | 3,301,502 | ||||
| 45 | CAISSE CONTINENT/REGIONALE (F:CAI) | France | Europe | NO | NO | 41,800,000 | ||||||
| 46 | CaixaBank | Spain | Europe | YES | YES | EP | 2007 | 299,000,000 | ||||
| 47 | Canadian Imperial Bank of Commerce (CIBC) | Canada | North America | YES | YES | EP | 2003 | 339,000,000 | ||||
| 48 | Canadian Western Bank (CWB) | Canada | North America | NO | NO | 12,800,000 | ||||||
| 49 | Capitec Bank | South Africa | Africa | NO | NO | 3,046,976 | ||||||
| 51 | China CITIC Bank | China | Asia Pacific | YES | NO | 517,000,000 | ||||||
| 52 | China Construction Bank (CCB) | China | Asia Pacific | YES | NO | 1,970,000,000 | ||||||
| 53 | China Merchants Bank CO., LTD | China | Asia Pacific | YES | YES | UNEP FI | 2007 | 538,000,000 | ||||
| 54 | China Minsheng Bank | China | Asia Pacific | NO | NO | 402,000,000 | ||||||
| 55 | CIMB Group Holdings Berhad | Malaysia | Asia Pacific | YES | NO | 79,400,000 | ||||||
| 56 | Citigroup Inc. | USA | North America | YES | YES | UNEP FI | 1997 | EP | 2003 | 1,680,000,000 | ||
| 57 | Commerzbank | Germany | Europe | YES | NO | 649,000,000 | ||||||
| 58 | Commonwealth Bank of Australia | Australia | Oceania | YES | YES | UNEP FI | 2009 | EP | 2014 | 484,000,000 | ||
| 59 | Community Bank N.A. | USA | North America | NO | NO | 6,788,893 | ||||||
| 60 | Co-operative Bank of Kenya Limited | Kenya | Africa | NO | NO | 3,108,633 | ||||||
| 61 | CORPBANCA (Ita? CorpBanca) | Chile | Latin America | NO | YES | EP | 2007 | 23,200,000 | ||||
| 62 | Crédit Agricole Corporate and Investment Bank | France | Europe | YES | YES | EP | 2003 | 1,650,000,000 | ||||
| 63 | Crédit Industriel et Commercial (CIC) | France | Europe | YES | NO | 283,000,000 | ||||||
| 64 | Credit Suisse & Credit Suisse Group | Switzerland | Europe | YES | YES | UNEP FI | 1992 | EP | 2003 | 880,000,000 | ||
| 65 | Daishi Bank | Japan | Asia Pacific | NO | NO | 48,500,000 | ||||||
| 66 | Danske Bank A/S | Denmark | Europe | NO | YES | UNEP FI | 1992 | 476,000,000 | ||||
| 67 | Davivienda | Colombia | Latin America | NO | NO | 29,900,000 | ||||||
| 68 | Deutsche Bank AG | Germany | Europe | YES | YES | UNEP FI | 1992 | BEI | 2010 | 1,730,000,000 | ||
| 69 | Dexia | Belgium | Europe | YES | NO | 437,000,000 | ||||||
| 70 | DGB Financial Group | South Korea | Asia Pacific | NO | YES | UNEP FI | 2006 | 33,000,000 | ||||
| 71 | Diamond Bank Plc | Nigeria | Africa | NO | NO | 5,794,770 | ||||||
| 72 | DNB | Norway | Europe | YES | YES | UNEP FI | 1999 | EP | 2008 | 247,000,000 | ||
| 73 | E.SUN Commercial Bank, LTD | Taiwan | Asia Pacific | NO | YES | EP | 2015 | 46,500,000 | ||||
| 74 | Ecobank Transnational Inc (ETI): Ecobank Nigeria | Nigeria | Africa | NO | NO | 17,100,000 | ||||||
| 75 | FCMB Group Plc | Nigeria | Africa | YES | NO | 3,994,850 | ||||||
| 76 | Fidelity Bank plc | Nigeria | Africa | NO | YES | UNEP FI | 2008 | EP | 2012 | 4,750,591 | ||
| 77 | First Connecticut Credit Union, Inc. | USA | North America | NO | NO | 2,148,804 | ||||||
| 78 | FirstRand Group Limited | South Africa | Africa | YES | YES | EP | 2009 | UNEP FI | 2010 | 83,200,000 | ||
| 79 | Garanti Bank | Turkey | Europe | NO | YES | UNEP FI | 2011 | 60,900,000 | ||||
| 80 | Green Bank | USA | North America | NO | NO | – | n/a | 2,259,761 | ||||
| 81 | Grupo Financiero Banorte, S.A.B. de C.V. | Mexico | Latin America | NO | YES | UNEP FI | 2013 | 48,400,000 | ||||
| 82 | Guaranty Trust Bank plc. | Nigeria | Africa | YES | YES | UNEP FI | 2012 | 8,591,723 | ||||
| 83 | HSBC Bank Canada | Canada | North America | NO | NO | 67,200,000 | ||||||
| 84 | HSBC Holdings plc | UK | Europe | YES | YES | UNEP FI | 1997 | EP | 2003 | 2,040,000,000 | ||
| 85 | Hua Xia Bank | China | Asia Pacific | NO | NO | 194,000,000 | ||||||
| 86 | ICBC Turkey Bank AS | Turkey | Europe | NO | NO | 1,929,135 | ||||||
| 87 | Industrial and Commercial Bank of China Limited (ICBC) | China | Asia Pacific | YES | YES | UNEP FI | 2014 | 2,520,000,000 | ||||
| 88 | Industrial Bank Co. Ltd | China | Asia Pacific | NO | YES | UNEP FI | 2007 | EP | 2008 | 560,000,000 | ||
| 89 | Industrial Bank of Korea (IBK) | South Korea | Asia Pacific | YES | NO | 160,000,000 | ||||||
| 90 | ING Bank N.V. | Netherlands | Europe | YES | YES | EP | 2003 | UNEP FI | 2007 | 1,200,000,000 | ||
| 91 | Intesa Sanpaolo | Italy | Europe | YES | YES | UNEP FI | 2006 | EP | 2006 | 694,000,000 | ||
| 92 | Itaú Unibanco S.A. | Brazil | Latin America | YES | YES | EP | 2004 | 258,000,000 | ||||
| 93 | JPMorgan Chase & Co. | USA | North America | YES | YES | UNEP FI | 2004 | EP | 2006 | 1,930,000,000 | ||
| 94 | Julius Baer Group Ltd. (or: Julius Baer Group) | Switzerland | Europe | YES | NO | 75,900,000 | ||||||
| 95 | Jyske Bank A/S | Denmark | Europe | NO | NO | 53,000,000 | ||||||
| 96 | KB Financial Group/KB Kookmin Bank | South Korea | Asia Pacific | YES | NO | 277,000,000 | ||||||
| 97 | KBC Group N.V. | Belgium | Europe | NO | YES | EP | 2004 | 338,000,000 | ||||
| 98 | KCB | Kenya | Africa | NO | YES | UNEP FI | 2012 | 4,614,426 | ||||
| 99 | KEB Hana Bank/Hana Financial Group | South Korea | Asia Pacific | YES | YES | UNEP FI | 2007 | 247,000,000 | ||||
| 100 | Kiatnakin Bank | Thailand | Asia Pacific | NO | NO | 5,145,665 | ||||||
| 101 | Laurentian Bank of Canada (LBC) | Canada | North America | NO | NO | 23,200,000 | ||||||
| 102 | Liberbank | Spain | Europe | NO | NO | 49,900,000 | ||||||
| 103 | Lloyds Banking Group | UK | Europe | YES | YES | EP | 2008 | BEI | 2010 | 938,000,000 | ||
| 104 | Mauritius Commercial Bank Ltd. | Mauritius | Africa | NO | YES | EP | 2012 | 8,026,967 | ||||
| 105 | Merkur Bank KGaA | Germany | Europe | YES | NO | 1,087,647 | ||||||
| 106 | Mitsubishi UFJ Financial Group Inc | Japan | Asia Pacific | YES | YES | UNEP FI | 2015 | 2,020,000,000 | ||||
| 107 | Mizuho Financial Group Inc. | Japan | Asia Pacific | YES | YES | UNEP FI | 2006 | 1,660,000,000 | ||||
| 108 | National Australia Bank Limited | Australia | Oceania | YES | YES | UNEP FI | 2002 | EP | 2007 | 508,000,000 | ||
| 109 | National Bank Limited | Bangladesh | Asia Pacific | NO | NO | 3,167,526 | ||||||
| 110 | National Bank of Canada | Canada | North America | NO | NO | 138,000,000 | ||||||
| 111 | National Bank of Greece | Greece | Europe | NO | NO | 97,400,000 | ||||||
| 112 | National Bank of Oman (NBO) | Oman | Middle East | NO | NO | 6,139,921 | ||||||
| 113 | Natixis | France | Europe | YES | YES | EP | 2010 | 482,000,000 | ||||
| 114 | Nedbank Group Ltd | South Africa | Africa | YES | YES | UNEP FI | 2004 | EP | 2005 | 62,600,000 | ||
| 115 | Nedbank Ltd | South Africa | Africa | YES | YES | UNEP FI | 2004 | EP | 2005 | 71,200,000 | ||
| 116 | Nordea Bank AB | Sweden | Europe | YES | YES | UNEP FI | 1994 | EP | 2007 | 544,000,000 | ||
| 117 | Ping An Bank | China | Asia Pacific | NO | YES | UNEP FI | 2010 | 245,000,000 | ||||
| 118 | Piraeus Bank S.A | Greece | Europe | NO | NO | – | n/a | 59,500,000 | ||||
| 119 | PT Bank Negara Indonesia (Persero) Tbk | Indonesia | Asia Pacific | NO | YES | UNEP FI | 2005 | 31,200,000 | ||||
| 120 | Public Bank BHD | Malaysia | Asia Pacific | NO | NO | 63,900,000 | ||||||
| 121 | PWC Capital Inc. | Canada | North America | NO | NO | 1,007,273 | ||||||
| 122 | Resona Holdings, Inc. | Japan | Asia Pacific | NO | NO | 442,000,000 | ||||||
| 123 | RMB Holdings Limited (RMBH) | South Africa | Africa | NO | NO | 2,566,379 | ||||||
| 124 | Royal Bank of Canada (RBC) | Canada | North America | YES | YES | UNEP FI | 1992 | EP | 2003 | 670,000,000 | ||
| 125 | Royal Bank of Scotland Group/Royal Bank of Scotland | UK | Europe | YES | YES | EP | 2003 | BEI | 2010 | 1,640,000,000 | ||
| 126 | Santander Brasil | Brazil | Latin America | NO | NO | 147,000,000 | ||||||
| 127 | Santander Rio | Argentina | Latin America | NO | NO | 9,856,951 | ||||||
| 128 | Scotiabank (Bank of Nova Scotia) | Canada | North America | YES | YES | UNEP FI | 1992 | EP | 2006 | 504,000,000 | ||
| 129 | Sekerbank | Turkey | Europe | NO | YES | UNEP FI | 2015 | 5,587,042 | ||||
| 130 | Shinhan Bank/Shinhan Bank Financial Group | South Korea | Asia Pacific | NO | YES | UNEP FI | 2008 | 275,000,000 | ||||
| 131 | Shinkin Central Bank (SCB) | Japan | Asia Pacific | NO | NO | 288,000,000 | ||||||
| 132 | Shinsei Bank | Japan | Asia Pacific | NO | NO | 97,300,000 | ||||||
| 133 | SinoPac Financial Holdings Company Ltd (Bank SinoPac) | Taiwan | Asia Pacific | NO | NO | 43,800,000 | ||||||
| 134 | Skandinaviska Enskilda Banken (SEB) | Sweden | Europe | YES | YES | UNEP FI | 1995 | EP | 2007 | 278,000,000 | ||
| 135 | Skye Bank PLC | Nigeria | Africa | NO | YES | UNEP FI | 2013 | 5,303,892 | ||||
| 136 | Société Générale | France | Europe | YES | YES | UNEP FI | 2001 | EP | 2007 | 1,260,000,000 | ||
| 137 | Sparebanken Vest | Norway | Europe | NO | NO | 15,200,000 | ||||||
| 138 | Standard Bank Group Ltd. | South Africa | Africa | YES | YES | UNEP FI | 2010 | 135,000,000 | ||||
| 139 | Standard Chartered plc | UK | Europe | YES | YES | UNEP FI | 2002 | EP | 2003 | 455,000,000 | ||
| 140 | State Bank of Mauritius (SBM) | Mauritius | Africa | YES | NO | 3,840,620 | ||||||
| 141 | State Bank of Mysore | India | Asia | NO | NO | 9,481,723 | ||||||
| 142 | Sumitomo Mitsui Financial Group Inc. | Japan | Asia Pacific | YES | YES | UNEP FI | 2002 | 1,450,000,000 | ||||
| 143 | Sumitomo Mitsui Trust Bank, Limited | Japan | Asia Pacific | YES | NO | 309,000,000 | ||||||
| 144 | Svenska Handelsbanken | Sweden | Europe | YES | YES | UNEP FI | 1995 | 274,000,000 | ||||
| 145 | Swedbank AB | Sweden | Europe | YES | YES | UNEP FI | 1996 | 218,000,000 | ||||
| 146 | The Shiga Bank, Ltd. | Japan | Asia Pacific | NO | YES | UNEP FI | 2001 | 44,100,000 | ||||
| 147 | TISCO Financial Group Public Company Limited | Thailand | Asia Pacific | NO | YES | UNEP FI | 1992 | 5,444,971 | ||||
| 148 | Toronto Dominion Bank | Canada | North America | YES | YES | UNEP FI | 1994 | 627,000,000 | ||||
| 149 | Turkiye Sinai Kalkinma Bankasi (TSKB) | Turkey | Europe | NO | YES | UNEP FI | 2009 | 4,507,833 | ||||
| 150 | U.S. Bancorp | USA | North America | NO | NO | 318,000,000 | ||||||
| 151 | UBS AG | Switzerland | Europe | YES | YES | UNEP FI | 1992 | 1,200,000,000 | ||||
| 152 | UmweltBank AG | Germany | Europe | NO | YES | UNEP FI | 1999 | 2,357,480 | ||||
| 153 | UniCredit | Italy | Europe | YES | YES | UNEP FI | 1998 | 881,000,000 | ||||
| 154 | Union Bank of Nigeria Plc | Nigeria | Africa | YES | NO | 5,102,837 | ||||||
| 155 | United Bank for Africa (UBA) Plc | Nigeria | Africa | YES | NO | 10,900,000 | ||||||
| 156 | Unity Bank Plc | Nigeria | Africa | NO | NO | 1,568,062 | ||||||
| 157 | Van Lanschot NV | Netherlands | Europe | NO | NO | 19,600,000 | ||||||
| 158 | Wells Fargo Bank, N.A. | USA | North America | NO | YES | EP | 2005 | 1,160,000,000 | ||||
| 159 | WEMA BANK PLC | Nigeria | Africa | NO | YES | UNEP FI | 2015 | 1,499,768 | ||||
| 160 | Westpac Banking Corporation | Australia | Oceania | YES | YES | UNEP FI | 1992 | EP | 2003 | 439,000,000 | ||
| 161 | YES BANK Limited | India | Asia | NO | YES | UNEP FI | 2006 | 19,100,000 | ||||
| 162 | Zenith Bank plc | Nigeria | Africa | YES | YES | UNEP FI | 2008 | 14,800,000 | ||||
| 163 | Zuger Kantonalbank | Switzerland | Europe | NO | NO | 11,300,000 | ||||||
| 164 | Hokuhoku Financial Group, Inc. | Japan | Asia Pacific | NO | NO | 106,000,000 | ||||||
| 165 | Citizens Financial Group, Inc. | USA | North America | NO | NO | 148,000,000 | ||||||
| 166 | PNC Financial services Group, Inc. | USA | North America | NO | NO | 250,000,000 | ||||||
| 167 | Suntrust Banks, Inc. | USA | North America | NO | NO | 166,000,000 |
- Note: (a) Banks' ranking is by alphabetical order (excluding those with IDs 165, 166, and 167); total number of banks: 165. (b) Fourteen global banks out of 71 do not fulfil the second classification criterion (i.e., average value of total assets ≥100 billion USD), but they satisfy the first and necessary condition (i.e., being included in the FSB list of GSIBs and listed on the “Banks around the World” website), so they have been classified as global too; these are the banks with IDs: 3, 27, 55, 75, 78, 82, 94, 105, 114, 115, 140, 154, 155, and 162. (c) Total assets in thousands of USD. (d) Data sources: Total assets: Thomson Reuters Eikon (formerly Datastream); green bank classification: BEI, EP, GABV, UNEP FI according to the developed classification criteria and authors' research (see Subsection 3.3); global bank classification: FSB, www.relbanks.com, and authors' calculations and research (see Subsection 3.3).
- Abbreviations: BEI, banking environment initiative (https://www.cisl.cam.ac.uk/business-action/sustainable-finance/banking-environment-initiative); EP, equator principles (http://equator-principles.com); GABV, Global Alliance for Banking on Values (http://www.gabv.org); UNEP FI, United Nations Environment Programme—Finance Initiative (http://www.unepfi.org).
Open Research
DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available upon request from the corresponding author. The data (excluding those for banks' classification as green and global—see Appendix A) are not publicly available due to privacy or ethical restrictions.
