## Systemic Implications of Financial Inclusion ## Source details **Canonical URL:** [Systemic Implications of Financial Inclusion](https://www.imf.org/-/media/files/publications/wp/2024/english/wpiea2024203-print-pdf.pdf) ## Other formats - [Markdown version](/-/media/files/publications/wp/2024/english/wpiea2024203-print-pdf.pdf.md) - [Structured JSON version](/-/media/files/publications/wp/2024/english/wpiea2024203-print-pdf.pdf.json) --- ### Introduction — context, distinctions, and scope - Effect of financial inclusion (FI) on financial stability is critical for policymaking; policymakers aim to achieve FI and financial stability simultaneously. - Existing studies often focus on individual bank risk indicators (e.g., Z-Scores) and may not capture economy-wide repercussions of bank failures. - This study: - Differentiates FI measures by type of financial service (deposits versus loans). - Distinguishes providers: commercial banks (CB) versus non-commercial banks (NCB). - Examines multiple risk dimensions including systemic risk. - Considers macroprudential regulation at country-levels (MP) and adherence to Basel III guidance at bank-levels (Mic). ### Literature review — main gaps and mechanisms - Common individual risk metric: Z-Score (effective for insolvency risk but limited for systemic assessment). - Systemic risk measures discussed: CoVaR, SES, SRISK (SRISK combines market and balance sheet information; shown to have greater predictive capability than SES). - Heterogeneous channels of FI: - Deposits: broader deposit FI can reduce individual bank risk via cheap retail deposits and funding stability. - Loans: credit inclusion can diversify loan portfolios (reducing concentration) but can also raise idiosyncratic risk if lending standards loosen during rapid credit expansion. - Role of NCB: - NCBs advance FI by targeting underserved segments; increased NCB presence can impose competitive pressure on CBs leading to lower credit standards and higher funding rates. - Literature gap on NCB contributions to FI and effects on CB stability. - Prudential regulation: - Stricter capital requirements and regulatory compliance can mitigate FI-induced risks but may induce regulatory arbitrage and non-bank intermediation. - Methodological recommendations: - Avoid composite FI indices that mask heterogeneous effects; disaggregate by service and provider. - Need for empirical work using systemic metrics (e.g., SRISK) and examining NCB role. ### Methodology and data — sample, variables, and identification - Sample and period: - Unbalanced panel 2009–2021. - 574 commercial banks from 31 countries (12 developing or emerging economies). - Exclusions: banks without bank-level controls; banks with less than four years of observations; countries with less than four banks. - Key data sources: - Bank-level data: Bloomberg and SRISK database (V-Lab). - FI indicators: IMF Financial Access Survey (FAS). - Stock price volatility (SV): The World Bank’s Global Financial Development database. - Macroprudential policy measures: iMaPP database. - Systemic risk measure: - SRISK aggregated to annual values in billion of U$S. - SRISK formula: SRISK_{i,t} = k D_{i,t} − (1 − k) W_{i,t} (1 − LRMES_{i,t}). - LRMES_{i,t} = 1 − exp(log(1 − d) * β_{i,t}), with d default set to 40%. - V-Lab k set to 8% for Africa, Asia, Americas and 5.5% for Europe. - Idiosyncratic risk measures: - Z-Score: Z_{i,t} = (−1) × (ROA_{i,t} + EQA_{i,t}) / σ(ROA_{i,t}); conventional Z-Scores multiplied by −1 so higher Z-Scores correspond to higher individual risk. - CDS-based idiosyncratic risk: principal component on log-transformed CDS spreads and SV; first principal component explains on average 60.5 percent (range 45% to 82%). - FI indicators used: - Dep.CB: outstanding deposits with CB to GDP. - Dep.NCB: outstanding deposits with NCB to GDP. - Loan.CB: outstanding loans from CB to GDP. - Loan.NCB: outstanding loans from NCB to GDP. - Adu.Loan: percentage of adults with loan accounts (does not distinguish CB vs NCB). - Regulation measures: - MPP: macroprudential policy index aggregated from iMaPP instruments related to capital and liquidity; treated as endogenous. - MPP.cred: MPP integrated with cumulative change in Loan.CB. - Mic: bank-level microprudential compliance — simple average of deviations from four Basel III ratios (leverage, TIER1, LCR, NSFR); Mic.cred integrates Mic with cumulative credit development. - DIRQ: index combining deposit insurance coverage with cumulative changes in regulatory quality. - Estimation: - Baseline dynamic panel: yi,j,t = α yi,j,t−1 + β1 FI j,t−1 + β2 FI j,t−1 × Dev j + γ Bank i,j,t−1 + ρ Macro j,t−1 + δ1 Reg j,t + δ2 Reg.Cred j,t−1 + fe i + fe t + ε i,j,t, with yi,j,t ∈ {SRISK, Z-Scores, CDS}. - Arellano & Bond difference GMM; two-step GMM with Windmeijer (2005) SE correction. - Instruments: endogenous variables instrumented by lags; optimal lag structure uses up to three lags; diagnostic tests include J-test, VIF, Arellano & Bond p2 (no second-order autocorrelation). ### Estimation results — systemic risk (SRISK) - General: - Lagged dependent variables significant and negative — mean-reverting yearly changes in banking risks. - Control variable signs consistent across regressions when significant. - Key SRISK coefficients (selected exact estimates and interpretations): - SRISK_{t−1}: -0.1659*** (Loans, MPP); -0.1967*** (Loans, MIC); -0.1855*** (Deposits, MPP); -0.2776*** (Deposits, MIC). - Loan.CB_{t−1}: 0.1267*** (Loans, MPP); 0.2518*** (Loans, MIC) — credit expansion by CB elevates SRISK. - Dev.Loan.CB_{t−1}: -0.0606* (Loans, MPP); -0.1683*** (Loans, MIC) — effect diluted in Dev economies, possibly via loan diversification lowering LRMES. - Dep.CB_{t−1}: -0.1183*** (Deposits, MPP); -0.1707*** (Deposits, MIC) — deposits with CB mitigate SRISK generally. - Dev.Dep.CB_{t−1}: 0.1268*** (Deposits, MPP); 0.1718*** (Deposits, MIC) — in Dev countries broader deposit FI associated with higher SRISK. - MPP: -0.6456*** (Loans, MPP); -0.6204*** (Deposits, MPP) — tighter macroprudential policy reduces contemporaneous SRISK. - MPP.cred_{t−1}: -0.0102*** (Loans, MPP); -0.0154*** (Deposits, MPP) — coordinating MPP with credit expansions further reduces SRISK. - Mic: -2.0341*** (Loans, MIC); -6.4424*** (Deposits, MIC) — higher Basel compliance reduces contemporaneous SRISK. - Mic.cred: not significant for SRISK (coefficients -0.0619 and -0.0489). - Additional significant controls (examples with exact values): - SV: 0.1537***, 0.3440***, 0.1694***, 0.2100*** across columns. - Size_{t−1}: 7.0673***, 4.0952***, 7.1279***, 4.0504***. - CF_{t−1}: -0.0743***, -0.0229***, -0.0721***, -0.0192***. - Int_{t−1}: 0.5450+; 1.2845***; 0.7571+; 1.1443***. - Sample and diagnostics: - Obs: 329359 in each reported column. - J-test p-values: 0.1787, 0.1511, 0.1355, 0.1937. - Arellano-Bond p2: 0.7841, 0.8085, 0.8439, 0.5431. - Adu.Loan (13-country subsample): - Adu.Loan measures imply FI in loans decreases SRISK across the sample and particularly in Dev economies (e.g., Adu.Loan_{t−1}: -0.5750***; Dev.Adu.Loan_{t−1}: -0.8225*** in reduced-sample regressions). ### Estimation results — idiosyncratic risk (Z-Score and CDS) - General: results largely mirror SRISK findings with exceptions for deposit effects in Dev countries. - Key idiosyncratic coefficients (selected exact estimates and interpretations): - Z-Score_{t−1}: -0.0846*** (Z-Score column), -0.0645*** (CDS column). - CDS_{t−1}: -0.5162***, -0.5061***. - Loan.CB_{t−1}: 0.0200*** (Z-Score, Loans); 0.0165*** (CDS, Loans) — credit inclusion increases individual banking risks. - Dev.Loan.CB_{t−1}: -0.1127*** (Z-Score, Loans); -0.0010 (CDS, Loans) — moderated effect in Dev economies. - Dep.CB_{t−1}: -0.1039*** (Z-Score, Deposits); -0.0058** (CDS, Deposits) — broader deposit inclusion reduces idiosyncratic risk. - Dev.Dep.CB_{t−1}: -0.0351*** (Z-Score, Deposits); -0.0118* (CDS, Deposits) — deposit inclusion attenuates idiosyncratic risk in Dev economies. - Mic and Mic.cred: - Mic: -4.2921*** (Z-Score, Loans); -0.3068* (CDS, Loans); -0.1471 (Z-Score, Deposits); -0.7884*** (CDS, Deposits). - Mic.cred: -0.1631***, -0.0170***, -0.0326***, -0.0252*** — improving Basel ratios alongside credit expansions reduces idiosyncratic risks. - Adu.Loan subsample: - Adu.Loan measurement indicates FI in loans decreases idiosyncratic risk, especially in Dev economies. ### Role of non-commercial banks (NCB) - SLoan.NCB (share of NCB in loans) and SDep.NCB (share of NCB in deposits) findings: - SLoan.NCB higher share associated with higher CB systemic and idiosyncratic risks (e.g., SLoan.NCB_{t−1} 0.1139** for SRISK in appendix results). - SDep.NCB increased share also increases banking risks in many specifications. - Interpretation: - NCB loan competition can induce CBs to relax credit standards, increasing risk-taking. - NCB deposit competition can raise funding costs for CBs and reduce efficiency, prompting riskier behavior. - Interaction with prudential policies: - MPP.cred and Mic.cred show significant negative coefficients across regressions — coordinating prudential policies with credit developments mitigates banking risks. - No direct evidence found of MPP leaking into higher banking risks via non-bank intermediation; rather, NCB penetration itself elevates CB risks in some services. - Policy implication: regulation of NCB may be necessary to harmonize FI objectives with banking system stability. ### Additional empirical results and robustness - Descriptive averages and growth (selected exact figures): - Loan.CB mean 72.76% of GDP; Loan.CB increases by 0.69 pp per year. - Dep.CB mean 78.48% of GDP; Dep.CB increases by 1.82 pp per year. - In Dev economies: Loan.CB ≈ 15 pp lower than panel baseline; Dev.Dep.CB = 62.08% of GDP. - Adu.Loan mean 0.44 (44% of adults); Dev.Adu.Loan = 0.35. - Dep.CB increases: Dev.Dep.CB yearly change 1.12 pp. - SLoan.NCB mean 0.19; in Dev economies SLoan.NCB = 0.12. - SDep.NCB mean 0.17; in Dev economies SDep.NCB = 0.09. - SRISK mean 2.57 U$S bn; SRISK diff.mean -0.19 U$S bn per year. - Mic mean 2.21 (times above minimum); Mic.diff.mean -0.01 per year; Mic.cred mean -0.47. - DIRQ interaction: - Dep.CB.DIRQ_{t−1}: -0.2879*** mitigates SRISK. - Dev.Dep.CB.DIRQ_{t−1}: 0.2627*** deposit expansion in Dev countries still associated with increased SRISK. - DIRQ coefficient: -1.8099+ (DIRQ itself associated with lower SRISK at p < 0.1). - Competition and DTI interactions: - Loan.CB_{t−1} 0.1841*** (Competition specification). - Dev.Loan.CB_{t−1} -0.0828*** (Competition) and -0.2350* (DTI specification). - Loan.CB.Com_{t−1} 0.0966***; Dev.Loan.CB.Com_{t−1} 0.0747+. - Loan.CB.DTI_{t−1} 0.0323; Dev.Loan.CB.DTI_{t−1} 0.4461*. - Interpretation: more competitive systems and higher DTI amplify systemic risk when loan FI expands. - Robustness: - Reduced-sample and alternate specifications (Adu.Loan, SLoan.NCB, squared FI terms) maintain coefficient signs and significant findings; U-shape tests do not indicate inverted U driving Dev versus overall effects. ### Conclusion — synthesis and policy implications - Novel contribution: first study to analyze FI impact across multiple commercial banking risk dimensions including systemic risk (SRISK) and idiosyncratic risks (Z-Score, CDS-derived). - Distinct effects by service: - Loans: - Credit expansions generally associated with higher banking risks. - Loan diversification among many small borrowers mitigates both systemic and idiosyncratic risks, especially in Dev economies. - Deposits: - Broader deposit FI generally reduces individual bank risk via stable, cheaper funding. - Exception: in Dev countries, deposit expansion’s reduction in systemic risk is dampened (possible causes: limited asset diversification, inefficient allocation to creditworthy borrowers). - Non-commercial banks (NCB): - Higher NCB share in loans and deposits linked to higher CB systemic and idiosyncratic risks — supports calls for NCB regulation to protect banking and financial system stability. - Prudential coordination: - Tighter macroprudential policy (MPP) and higher bank-level Basel compliance (Mic) reduce SRISK and idiosyncratic risk. - Coordinating macroprudential measures with credit developments (MPP.cred) and strengthening Mic alongside credit expansions (Mic.cred) are effective mitigation channels. - Noted concern: banks tend to relax Basel scores as credit activity expands (Mic.cred mean -0.47), suggesting need for vigilance and coordination to prevent erosion of resilience. - Policy takeaway: - Promote FI while strengthening macroprudential frameworks and considering regulation of non-bank providers to avoid undermining banking system stability. - Align macroprudential policy and bank-level microprudential compliance with credit developments to contain systemic and idiosyncratic risks. *Source: wpiea2024203-print-pdf* ### 1. Introduction ........................................................................................................ ### 1. Introduction ### Context and motivation - The effect of financial inclusion (FI) on financial stability has gained substantial attention and is critical for effective policymaking. - Policymakers aim to achieve FI and financial stability goals simultaneously; overlooking trade-offs can expose economies to costly systemic crises or hinder FI progress (Čihák et al., 2016). - Recent research has begun investigating FI’s effect on the banking system, the main provider of formal finance (Feghali et al., 2021), but has focused mainly on individual banking risk indicators like Z-Scores. ### Literature gap addressed - Existing studies emphasizing idiosyncratic/individual risk may not capture broader repercussions of bank failures on the entire economy, particularly where the financial sector lacks sufficient capitalization. - This paper examines FI’s effects across multiple risk dimensions, including systemic risk, which has received limited attention. - Prior literature reveals diverse implications of FI depending on service type: deposit inclusion can bolster resilience via more stable retail deposits (Ahamed & Mallick, 2019), while credit inclusion can promote loan diversification (Morgan & Pontines, 2014) but may elevate idiosyncratic risk in competitive systems with relaxed lending standards (Čihák et al., 2021; Feghali et al., 2021). ### Distinctions and scope - The analysis differentiates FI measures by type of financial service and further distinguishes sources of FI: commercial banks (CB) versus non-commercial banks (NCB). - There is a noted literature gap on the impact of non-banking financial institutions on banking sector stability despite their role in advancing FI (Lopez & Winkler, 2018). - NCB entry creates competitive pressures on CB that can: - Lower credit standards (Feghali et al., 2021). - Raise funding rates (Craig & Dinger, 2013). - Potentially undermine CB soundness. ### Role of prudential regulation - Regulators must balance promoting FI with enhancing systemic stability, focusing on collective risk-taking behavior (Čihák et al., 2016). - The paper examines macroprudential regulation at country-levels and adherence to Basel III guidelines at bank-levels (Anginer et al., 2014; Beck et al., 2013; Sahay et al., 2015). - While such policies enhance banking sector resilience by curbing individual risk-taking during rapid FI expansion, they may induce regulatory arbitrage and entry of less regulated institutions (Claessens et al., 2021; Irani et al., 2021). ### Focus on developing economies - Financial exclusion is significant in developing and emerging (Dev) countries due to a larger unbanked population gap (Demirgüç-Kunt et al., 2020). - The study captures FI’s impact on banking risks in Dev economies while starting from a broad panel including both advanced (Ad) and Dev countries to reflect rapid FI advancements across income levels and Dev-specific challenges in fund allocation and asset diversification (Čihák et al., 2021; Gennaioli et al., 2018). ### Key empirical findings (as reported in this chapter) - In Dev economies, credit expansions are associated with lower banking risks, highlighting the role of loan diversification despite relatively rapid FI in credit. - FI in deposits generally reduces individual banking risks, but this effect on systemic risk is dampened in Dev countries, possibly due to limited capacity for financial asset diversification. - NCBs tend to: - Increase systemic and idiosyncratic risks for CB through competitive pressure in the loan segment. - Enhance overall CB resilience through involvement in deposit services. - Tighter macroprudential policies effectively reduce banking systemic risks (e.g., Altunbas et al., 2018). - Aligning macroprudential policies with credit developments further mitigates systemic risk; banks with stronger Basel capital ratios exhibit reduced idiosyncratic risks. - Evidence suggests banks may relax Basel ratios to meet lending demands, underlining the need to synchronize macroprudential policies with credit developments. - The findings support the need for NCB regulation to bolster both banking sector and financial system stability. ### Paper structure (as outlined) - Section 2: Literature Review - Section 3: Methodology and Data (variables and estimation model) - Section 4: Estimation Results (systemic and idiosyncratic risk, role of NCB) - Section 5: Conclusion (key findings and implications) - Appendix: Robustness checks and descriptive statistics *Source: wpiea2024203-print-pdf - 1. Introduction* ### 2. Literature Review ### 2. Literature Review ### Research focus and gaps - Research to date has focused on the effect of financial inclusion (FI) on the well-being of consumers and overall development and growth. - There is much less international evidence on the effect of FI on financial stability, in particular the banking system, which is the main provider of formal finance (Feghali et al., 2021). - The effect of FI on banking systemic risk remains largely unexplored; existing studies often do not distinguish between types of financial services or providers. ### Systemic risk metrics and their properties - Commonly used bank-level risk metric: - Z-Score: a prevalent indicator for assessing individual banks’ risk-taking choices (Čihák et al., 2021). Effective for insolvency risk at the individual bank level but limited in capturing broader economy-wide repercussions of bank failure, especially when the financial sector lacks sufficient capitalization. Brownlees & Engle (2017) highlight firms’ risk-taking decisions overlooking negative external costs during crises. - Alternative systemic metrics recommended for broader assessment: - CoVaR: financial system’s value at risk conditional on institutions being under distress (Tobias & Brunnermeier, 2016). - SES: systemic expected shortfall measures (Acharya et al., 2017). - SRISK: quantifies the capital shortfall of a financial entity in the event of a prolonged market downturn; merges market and balance sheet information and explicitly considers leverage. SRISK has demonstrated greater predictive capability than SES (Brownlees & Engle, 2017; Acharya et al., 2017). ### Empirical findings on FI and banking stability (mixed evidence) - Some studies find FI enhances banking stability: - Examples: Ahamed & Mallick, 2019; Danisman & Demirel, 2019; Ghosh, 2022; Neaime & Gaysset, 2018; Vo et al., 2021. - Other studies find FI may increase banking risks: - Examples: Čihák et al., 2016; Feghali et al., 2021; Le et al., 2019. - Causes of mixed results: - Use of composite FI indicators combining different FI measures can produce divergent findings (Feghali et al., 2021). Footnote: Some studies construct a composite FI development index for 115 countries using principal component analysis (PCA), combining indicators related to financial services provided by commercial banks, credit unions, and microfinance institutions without distinction.1 - Recent studies thus use single FI indicators distinguishing financial services (e.g., deposits versus loans) (Feghali et al., 2021; Ghosh, 2022). ### Differentiation by type of financial service - Deposits: - Low-income savers’ behaviour typically does not change significantly through the business cycle; during systemic distress, low-income clients continue to provide deposits (Han & Melecky, 2017). - Small customers are posited as a stable source of funding; broader FI in deposits reduces individual bank risk by collecting cheap retail deposits and reducing cost of funding, enhancing profitability and resilience (Ahamed & Mallick, 2019). - Loans (credit inclusion): - Positive channel: diversification of loan portfolios; small loans associated with lower systemic risk than large, infrequent losses (Adasme et al., 2006). Small loans reduce nonperforming loans and probability of default via loan diversification (Morgan & Pontines, 2014). - Negative channel: extending credit to previously constrained borrowers (limited collateral or incomplete information) can increase banking risks. Fast credit growth may be accompanied by reduction of lending standards, increasing standalone and potentially systemic risk through contagion (Mehrotra & Yetman, 2014). - Feghali et al. (2021) show credit inclusion can lead to elevated idiosyncratic risk (measured with Z-Scores) in competitive banking systems where banks compete to offer loans to less creditworthy individuals. ### Role of non-commercial banks (NCB) and competition - NCBs play a significant role in advancing FI by targeting low-income communities and underserved regions (Lopez & Winkler, 2018). - Increased reliance on NCBs may reduce commercial banks’ (CB) market share and create competitive pressure prompting CBs to relax credit standards (Feghali et al., 2021). - Competitive pressure can increase incentives to invest in risky projects by raising banks’ funding costs (Craig & Dinger, 2013). There is a gap in the literature on the specific impact of NCB contributions to FI on traditional banking sector stability. - Some studies focus exclusively on services provided by CBs to assess banking risks (Feghali et al., 2021; Le et al., 2019; Marcelin et al., 2022). ### Cross-country context: advanced vs developing economies - Financial exclusion is generally considered more of an issue in developing and emerging countries due to larger gaps in unbanked adults (Demirgüç-Kunt et al., 2020); however, significant FI work remains in advanced economies (Deku et al., 2016; Danisman & Demirel, 2019). - Most empirical work focuses on developing countries, specific regions, or broad panels spanning diverse countries; the present research aims to capture specific effects of broader FI in developing and emerging economies on banking risks. - Some evidence: - Ahamed & Mallick (2019) suggest FI might be more beneficial for developing countries, significantly enhancing bank stability.4 - Wang & Luo (2022) find that higher FI in developing economies is associated with significantly decreased nonperforming loans, indicating stability enhancement via loan portfolio diversity. However, these economies may have limited capacity for efficient allocation of funding to creditworthy firms and scarce asset diversification opportunities, potentially affecting banking risks (Čihák et al., 2021; Gennaioli et al., 2018). ### Prudential regulation and mitigation channels - Prudential regulations can mitigate adverse effects of FI on banking risks: - Certain regulation types may limit banks’ engagement in riskier activities amid expanded FI (Beck et al., 2013). - Stricter capital requirements help minimize contagion and may incentivize investors to control systemic risk-taking, particularly during rapid expansion of financial services (Anginer et al., 2014). - Regulators should enhance systemic stability while promoting FI because correlation in banks’ risk-taking behaviour matters more for systemic risk than absolute risk levels of individual institutions (Čihák et al., 2021). - Empirical evidence on regulation: - Sahay et al. (2015): constructed Basel Core Principles (BCP) scores (based on Basel Committee guidance, 2012). Broad FI raises individual bank risks due to increasing nonperforming loans, especially where loan losses are under-provisioned; however, where broader FI is accompanied by enhanced compliance to Basel regulation (BCP scores interacted with FI indicators), the negative impact on standalone risks is mitigated.6 - Wang & Luo (2022): using four indicators from Barth et al. (2013)—Activity, Capital, Supervisory power, and Market discipline—find that stricter banking regulation mitigates elevated idiosyncratic risk stemming from FI. - Regulatory trade-offs and leakage: - Increasing capital requirements can reduce banks’ propensity to make riskier loans (Darst et al., 2020; Agénor & Bayraktar, 2023) but may reduce banks’ capacity to provide financial services, potentially conflicting with FI objectives (Anarfo et al., 2020). - Tighter bank capital regulation can lead to non-bank financial institutions expanding their role (Aiyar et al., 2014; Irani et al., 2021; Claessens et al., 2021), potentially creating regulatory leakage where less-regulated institutions increase lending in response to tighter bank regulation. ### Methodological notes and recommendations from the literature - Caution against composite FI indices: - Composite indicators often combine different FI measures and can mask heterogeneous effects; authors recommend caution in their use (Feghali et al., 2021).2 - Importance of disaggregation: - Contemporary research emphasizes differentiating types of financial services and providers when assessing FI’s effect on banking stability (Feghali et al., 2021; Ghosh, 2022; Vo et al., 2021). - Identified literature gaps: - Limited evidence on FI’s effect on systemic banking risk using systemic metrics (e.g., SRISK, CoVaR, SES). - Limited research on the role of NCB contributions to FI and the consequent impacts on traditional banking stability. - Sparse literature integrating the interplay of FI expansion, the mix of providers (CB vs NCB), and the stringency of prudential regulation. *Source: wpiea2024203-print-pdf - 2. Literature Review* ### 3. Methodology and Data ### 3. Methodology and Data ### 3.1. Data and Variables - Sample: - Unbalanced panel covering the years 2009 to 2021. - 574 commercial banks (CB) from 31 countries, of which 12 are developing or emerging economies. - Exclusions: banks without bank-level control variables; banks with less than four years of observations across the selected dependent variables; countries with less than four banks. - Rationale for period selection: - Commences in 2009 to align with the availability of FI data and to capture the post-global financial crisis regulatory shift. - Data sources: - Bank-level dependent and control variables: Bloomberg and the SRISK database available on the V-Lab website. - FI indicators: IMF Financial Access Survey (FAS). - Stock price volatility (SV): The World Bank’s Global Financial Development database. - Macroprudential policy measures: iMaPP database. - Notation and cross-references: - Variable notations, definitions, and data sources summarized in Tables 4, 5, and 6 (in the source). ### 3.1.1. Banking Risks - Conceptual framing: - Distinction emphasized between idiosyncratic banking risks and systemic risks; systemic risk defined as widespread failures or freezing of capital markets that impair financial intermediation (payments system and lending to corporations or households). - Contagion drivers: interconnectedness, shared risk exposure, financial imbalances. - Systemic risk measure: - SRISK (Brownlees & Engle, 2017), monthly data from V-Lab aggregated to annual SRISK values in billion of U$S. - SRISK defined as the expected capital shortfall of a financial entity conditional on a prolonged market decline; depends on firm size, leverage, and expected equity loss conditional on the market decline. - SRISK formula (for firm i in period t): - SRISK_{i,t} = k D_{i,t} − (1 − k) W_{i,t} (1 − LRMES_{i,t}), where D_{i,t} is the book value of debt, W_{i,t} is the market value of equity, k is the prudential capital factor, and LRMES_{i,t} is the Long Run Marginal Expected Loss. - LRMES_{i,t} computed as 1 − exp(log(1 − d) * β_{i,t}), where d is the six-month crisis threshold for the market index decline (default value 40%), and β_{i,t} is the firm’s beta coefficient with respect to the country’s market return. - V-Lab set k to be 8% for firms in Africa, Asia, and Americas and 5.5% for firms in Europe. - Idiosyncratic risk measures: - Z-Score: - Z_{i,t} = (−1) × (ROA_{i,t} + EQA_{i,t}) / σ(ROA_{i,t}), where σ(ROA_{i,t}) is the standard deviation of returns on total assets (ROA) computed using a three-year rolling time window, and EQA_{i,t} is equity to total assets; both retrieved from Bloomberg. - Conventional Z-Scores multiplied by −1 so higher Z-Scores correspond to higher individual risk. - CDS-based idiosyncratic risk: - Use principal component approach on log-transformed CDS spreads of banks within a country and country-level stock equity volatility (SV), first principal component captures common factor. - Average proportion of total variation explained by first principal component: 60.5 percent (range across countries: 45% to 82%). - The portion of risk unexplained by the first principal component is treated as idiosyncratic CDS risk (denoted CDS); log-transformed CDS spreads with both common and individual risks denoted CDS.lt. - Distinction: Z-Score relies on balance sheet metrics; CDS-derived idiosyncratic component derives from market prices and reflects market expectations. ### 3.1.2. Financial Inclusion - Definition and focus: - FI defined as access to and use of formal financial services; focus on the market-clearing outcome — the use of financial services. - Data choice: - IMF Financial Access Survey (FAS) used due to annual coverage (2004–2021) and granularity; some FAS indicators have incomplete data for specific countries. - WBGF (Global Findex) available only for 2011, 2014, 2017, and 2021 and thus less suitable for annual panel analysis. - FI indicators selected (see Table 5 in the source): - Ratios of outstanding deposits with CB to GDP (Dep.CB). - Outstanding deposits with NCB to GDP (Dep.NCB). - Outstanding loans from CB to GDP (Loan.CB). - Outstanding loans from NCB to GDP (Loan.NCB). - Percentage of adults with loan accounts (Adu.Loan). - Note: Adu.Loan does not distinguish between CB and NCB providers. - Rationale for indicator choice: - Deposits and loans services less susceptible to technological advancement distortions compared to branches, ATMs, or mobile transactions. - Use of both deposit and loan measures helps capture distinct dynamics: outstanding loans to GDP as financial deepening versus borrower counts as breadth of inclusion. - Concerns highlighted: - Using only borrowers per 1000 adults may overstate inclusivity. - Using only loan-to-GDP may hide concentration of loans among high-income clients. - Outstanding loans to GDP can reflect credit booms that have different implications for stability. - NCB share measures: - Proposed ratio: outstanding loans (deposits) provided by NCB divided by total outstanding loans (deposits) provided by CB — to quantify NCB penetration and its implications for CB risk. ### 3.1.3. Control Variables - Strategy: - Parsimonious set of bank-specific and country-specific controls to limit endogeneity; regressors tested for multicollinearity via VIF. - Controls used in differences (not levels), affecting interpretation. - Bank-level controls (from Bloomberg): - Logarithm of total assets (Size) — size linked to risk via leverage and concentration. - Return on total assets (ROA) — lower profitability reduces loss-absorbing capacity and may alter market betas. - Country-level controls: - Stock price volatility (SV) — captures country-level market volatility; SV standardized and used in CDS PCA. - GDP per capita (GDP) — business cycle fluctuations associated with credit risk. - Net portfolio equity inflows to GDP (CF) — used instead of Chinn and Ito index to capture capital flow dynamics relevant to liquidity and stability. - Deposit interest rates — proxy for monetary policy rate. - Endogeneity handling for controls: - In SRISK estimations: SV treated as endogenous; Size and ROA first-lagged. - In Z-Score estimations: Size and ROA treated as predetermined. - In CDS estimations: ROA predetermined; Size first-lagged; SV removed because CDS idiosyncratic component is net of common variation with SV. ### 3.1.4. Bank Regulation - Regulatory dimensions analyzed: - Macroprudential regulation (MP), bank-level compliance with Basel III guidance (Mic), and deposit insurance coverage with regulatory quality developments (DIRQ). - Macroprudential policy measures: - iMaPP database monthly dummy indicators: tightening (+1) and loosening (−1). - MPP index constructed by aggregating instruments related to bank capital requirements: (i) countercyclical capital buffers, (ii) capital conservation buffers, (iii) other capital requirements, (iv) limits on leverage, (v) liquidity requirements. - MPP is treated as endogenous because macroprudential actions respond to contemporaneous banking risks. - MPP.cred: integrates cumulative change in outstanding loans from CB to GDP into the one-year lagged cumulative sum of MPP; captures regulatory reactions coordinated with credit developments. - Basel III bank-level measures: - Considered ratios: leverage ratio; TIER1 capital ratio; liquidity coverage ratio; net stable funding ratio. - For each bank-year, deviation from Basel III guidance calculated; ratios > 1 indicate compliance above minimum regulatory thresholds. - Mic: simple average of the four Basel III ratio deviations — interpreted as bank-level microprudential compliance / risk-taking indicator. - Mic.cred: integrates cumulative change in Basel scores with cumulative credit development, capturing bank-level reactions to credit expansions. - Deposit insurance and regulatory quality: - DIRQ: index combining deposit insurance coverage with cumulative changes in regulatory quality to investigate interactions between deposit inclusion and regulatory environment. ### 3.2. Estimation Model - Baseline panel model: - yi,j,t = α yi,j,t−1 + β1 FI j,t−1 + β2 FI j,t−1 × Dev j + γ Bank i,j,t−1 + ρ Macro j,t−1 + δ1 Reg j,t + δ2 Reg.Cred j,t−1 + fe i + fe t + ε i,j,t, where yi,j,t ∈ {SRISK, Z-Scores, CDS} for CB i in country j at year t; FI denotes a financial inclusion indicator; Dev is a dummy for developing and emerging economies; fei and fet are bank and time fixed effects. - Estimation approach: - Arellano & Bond (1991) difference GMM applied: equation first-differenced to remove bank fixed effects; use Holtz-Eakin et al. (1988) style instrumentation with linear moment conditions. - Exogenous variables in first differences instrumented by themselves; endogenous regressors in first differences instrumented by their lags in levels. - Two-step GMM with initial weighting matrix as in Arellano & Bond (1991); standard errors corrected as in Windmeijer (2005). - Treatment of endogeneity: - MPP and Mic treated as endogenous (prudent regulators and banks react to contemporaneous risk). - Specific endogenous/predetermined choices by dependent variable: - SRISK model: SV endogenous; Size and ROA first-lagged. - Z-Score model: Size and ROA predetermined. - CDS model: ROA predetermined; Size first-lagged; SV removed from regressors. - Instrument construction: - Optimal number of lags for instruments found to be three across dependent variables. - For endogenous variables: use two lags; for predetermined variables: use three lags. - Diagnostic tests: - J-test (Hansen, 1982) used for overidentifying restrictions — test did not indicate invalid instruments. - Variance inflation factors (VIF) computed among control variables using Pearson’s r; multicollinearity not a concern. - Wald test: population parameters of slopes and time dummies jointly different from zero. - Arellano & Bond test: no second-order autocorrelation (p2) in idiosyncratic errors present in estimated models. *Italic: IMF Working Paper — Section 3. Methodology and Data (content as provided).* ### 4. Estimation Results ### 4. Estimation Results ### Systemic Risk (Section 4.1) - Two FI metrics used: - outstanding loans and deposits with CB as a percentage of GDP (Loan.CB, Dep.CB); - borrowers as a percentage of adults (Adu.Loan). - General model behavior: - First lagged dependent variables are significant and negative, indicating mean-reverting yearly changes in banking risks. - Estimated coefficients for control variables, when significant, maintain the same sign across regressions. - Key findings from Table 1: Systemic Risk — Loans and Deposits to GDP - SRISKt−1: -0.1659*** (Loans, MPP), -0.1967*** (Loans, MIC), -0.1855*** (Deposits, MPP), -0.2776*** (Deposits, MIC). - Loan.CBt−1: 0.1267*** (Loans, MPP), 0.2518*** (Loans, MIC). Interpretation: credit expansion by CB elevates SRISK. - Dev.Loan.CBt−1: -0.0606* (Loans, MPP), -0.1683*** (Loans, MIC). Interpretation: effect diluted in Dev economies, possibly via loan diversification lowering LRMES. - Dep.CBt−1: -0.1183*** (Deposits, MPP), -0.1707*** (Deposits, MIC). Interpretation: deposits with CB mitigate SRISK generally. - Dev.Dep.CBt−1: 0.1268*** (Deposits, MPP), 0.1718*** (Deposits, MIC). Interpretation: in Dev countries broader deposit FI is associated with higher SRISK. - MPP: -0.6456*** (Loans, MPP), -0.6204*** (Deposits, MPP). Interpretation: macroprudential policy (MPP) reduces contemporaneous SRISK. - MPP.credt−1: -0.0102*** (Loans, MPP), -0.0154*** (Deposits, MPP). Interpretation: tighter macroprudential actions associated with credit expansions further reduce SRISK. - Mic: -2.0341*** (Loans, MIC), -6.4424*** (Deposits, MIC). Interpretation: Basel scores above minimum requirements (Mic) reduce contemporaneous SRISK. - Mic.credt−1: -0.0619 (Loans, MIC), -0.0489 (Deposits, MIC) — not significant for SRISK. - Other significant controls (examples): - SV: 0.1537***, 0.3440***, 0.1694***, 0.2100*** (four columns). - Size t−1: 7.0673***, 4.0952***, 7.1279***, 4.0504***. - CFt−1: -0.0743***, -0.0229***, -0.0721***, -0.0192***. - Intt−1: 0.5450+; 1.2845***; 0.7571+; 1.1443***. - Sample and diagnostics: - Obs: 329359 (all four columns). - J-test p-values: 0.1787, 0.1511, 0.1355, 0.1937. - Wald slope/time/all: *** in all columns. - Arellano-Bond p2: 0.7841, 0.8085, 0.8439, 0.5431. - Significance notation: + p < 0.1, * p < 0.05, ** p < 0.01, *** p < 0.001. - FI measured as percentage of adults with loan accounts (Adu.Loan) — summary (Section 4.1.2) - Adu.Loan available for 13 countries; comparative analysis limited to these countries. - When FI is measured by percentage of adults with loan accounts, FI in loans appears to decrease SRISK across the entire sample and particularly in Dev economies. - Interpretation: distributing loans among numerous small borrowers disperses risk and reduces loan-portfolio concentration, lowering systemic risk. ### Idiosyncratic Risk (Section 4.2) - General: - Estimated coefficients, when significant, show consistent direction across Z-Score and idiosyncratic component of CDS spreads. - Findings largely mirror SRISK results, with exceptions for deposit effects in Dev countries. - Key findings from Table 2: Idiosyncratic Risk — Loans and Deposits to GDP - Z-Score t−1: -0.0846*** (Z-Score column), -0.0645*** (CDS column). - CDS t−1: -0.5162***, -0.5061***. - Loan.CBt−1: 0.0200*** (Z-Score, Loans), 0.0165*** (CDS, Loans). Interpretation: credit inclusion increases individual banking risks. - Dev.Loan.CBt−1: -0.1127*** (Z-Score, Loans), -0.0010 (CDS, Loans). Interpretation: effect moderated in Dev economies. - Dep.CBt−1: -0.1039*** (Z-Score, Deposits), -0.0058** (CDS, Deposits). Interpretation: broader deposit inclusion reduces idiosyncratic risk. - Dev.Dep.CBt−1: -0.0351*** (Z-Score, Deposits), -0.0118* (CDS, Deposits). Interpretation: deposit inclusion attenuates idiosyncratic risk in Dev economies. - Mic and Mic.cred effects: - Mic: -4.2921*** (Z-Score, Loans), -0.3068* (CDS, Loans), -0.1471 (Z-Score, Deposits), -0.7884*** (CDS, Deposits). - Mic.credt−1: -0.1631***, -0.0170***, -0.0326***, -0.0252*** (all four columns). Interpretation: Mic and Mic.cred reduce idiosyncratic risks; improving Basel ratios alongside credit expansions mitigates individual risk. - Other controls (examples): - SV: 0.0813*** (Z-Score, Loans), 0.0024 (CDS, Loans). - GDP t−1: -0.1796***, -0.0197***, -0.0978***, -0.0143*. - CFt−1: -0.0039***, -0.0040***, -0.0176***, -0.0043**. - Intt−1: 0.8013***, 0.3435***, 0.5044***, 0.2752***. - Size t−1 and Size: positive and significant coefficients where reported. - ROA: -0.5237***, -0.0062+, -0.6131***, -0.0033. - Sample and diagnostics: - Obs: 329359, 329360, 329359, 329360. - J-test p-values: 0.2127, 0.2445, 0.1925, 0.1742. - Arellano-Bond p2: 0.9895, 0.4942, 0.9954, 0.7026. - FI measured as percentage of adults with loan accounts (Adu.Loan) — summary (Section 4.2.2) - Limited to 13 countries; signs of significant coefficients remain consistent. - When measured by percentage of adults with loan accounts, FI in loans appears to decrease idiosyncratic risk, especially in Dev economies. - Interpretation: loan distribution among numerous small borrowers reduces portfolio concentration and individual bank risk. ### The Role of Non-Commercial Banks (NCB) (Section 4.3) - Share of NCB in provision of financial services affects CB risks differently by service type. - Loans provided by NCB: - Higher share of loan services provided by NCB is associated with higher CB systemic and idiosyncratic risks (Table 9, Appendix). - Interpretation: competition from NCB may induce CB to relax credit standards, leading to riskier lending (e.g., reduced collateral requirements). - Deposits provided by NCB: - Increased share of NCB in deposits increases banking risks (Table 10, Appendix). - Interpretation: deposit competition raises deposit rates, increases funding costs, reduces efficiency, and prompts risk-taking in the traditional banking sector. - Interaction with prudential policies: - MPP.cred and Mic.cred show significant and negative coefficients across regressions, indicating that when credit extensions are accompanied by sound macroprudential policy and improved Basel metrics, banking risks are mitigated. - The study finds no direct evidence of macroprudential policy leaking into higher banking risks in the manner documented by Aiyar et al. (2014); instead, NCB penetration in certain services (notably loans) can elevate banking risks. - Policy implication: effective regulatory frameworks, including potential regulation of NCB, are crucial to harmonize FI objectives with banking stability. ### Conclusion (Section 5) — summary of main implications - Novelty: First study to analyze FI impact across multiple commercial banking risk dimensions including systemic risk. - Distinct effects by FI service: - Loan services: - Rapid credit expansions generally associated with higher banking risks. - Credit diversification among low-income borrowers mitigates these risks. - In Dev economies, loan diversification appears particularly important; higher proportion of adults with loan accounts reduces systemic and idiosyncratic risk, with the effect enhanced in Dev economies. - Deposit services: - FI in deposits generally favorable for individual banking risks due to cheaper, stable funding. - Exception: systemic risk in Dev countries — reduction in systemic risk from deposit expansion is dampened. - Possible reasons: limited capacity to allocate funding efficiently to creditworthy firms/individuals; limited availability of financial assets for diversification. - NCB effects: - Higher share of loan services by NCB associated with higher systemic and idiosyncratic risks for CB. - Increased NCB share in deposits elevates CB risks via higher funding costs and reduced efficiency. - Supports calls for regulation of NCB to enhance banking and financial system stability. - Prudential coordination: - Aligning macroprudential regulation with credit developments and aligning stronger bank-level Basel scores with credit growth reduce banks’ risk exposure. - Noted trend: prevalent tendency to reduce Basel scores to facilitate greater credit extensions — potential area of concern. ### Appendix — Descriptive Analysis excerpt (Section 6.1.1) - Average values and growth: - On average, outstanding loans from commercial banks account for 72.76% of GDP (Loan.CB). - Annually, Loan.CB increases by an average of 0.69 percentage points (pp). - In developing and emerging economies (Dev), Loan.CB proportion of GDP is lower by approximately 15 pp compared to the full sample baseline, but Dev economies exhibit a higher annual growth rate in Loan.CB. *IMF Working Paper — "Systemic Implications of Financial Inclusion", Section 4: Estimation Results* ### 1.01 pp per year (Figure 1). Similarly, outstanding deposits with CB average 78.48% of GDP (Dep.CB). On an ### Systemic Implications of Financial Inclusion ### Financial inclusion trends (deposits, loans, account ownership) - Outstanding deposits with commercial banks (Dep.CB) average 78.48% of GDP; Dep.CB increases by an average of 1.82 pp per year. - In Dev economies, Dep.CB is 62.08% of GDP (i.e., approximately 16 pp lower than the panel average); average annual change in Dev.Dep.CB is 1.12 pp. - Outstanding loans from commercial banks (Loan.CB) average 72.76% of GDP; Loan.CB increases by 0.69 pp per year. - Borrowers represent, on average, 44% of total adult population (Adu.Loan = 0.44); in Dev economies borrowers constitute 35% (Dev.Adu.Loan = 0.35). - Average annual change in borrowers: 2 pp in Dev economies vs 1 pp for the overall sample. - Depositors represent 98% of total adults (Adu.Dep = 0.98); in Dev economies depositors represent 80% (Dev.Adu.Dep = 0.80). - Yearly increase in depositors: averaging 3 pp in Dev economies vs 2 pp for total sample. - Figures show Ad economies increase the proportion of borrowers and depositors at a slower pace compared to Dev countries. ### Non-commercial banks (NCB) participation - Share of NCB in loans (SLoan.NCB) averages 19% of outstanding loans from commercial banks; in Dev economies SLoan.NCB = 12%. - SLoan.NCB shows a slight annual increase of 0.1 pp across the panel. - Trend driven primarily by higher NCB participation in Ad economies. - Share of NCB in deposits (SDep.NCB) averages 17% of outstanding deposits with commercial banks; in Dev economies SDep.NCB = 9%. - SDep.NCB slightly decreases annually by 0.2 pp in Dev economies. - Dev economies exhibit more fluctuations in SLoan.NCB and SDep.NCB; e.g., SLoan.NCB experienced a significant decline during the COVID-19 crisis in Dev economies while remaining relatively stable in Ad economies. ### Banking risks and macroprudential policy (SRISK, MPP, idiosyncratic risk) - SRISK average is 2.57 U$S bn; SRISK declines by 0.19 U$S bn per year on average. - Since the global financial crisis, macroprudential authorities have tended to reinforce capital buffers, coinciding with a gradual decrease in SRISK; during COVID-19 SRISK increased and authorities loosened capital requirements on average. - MPP index average is 1.05, indicating a tendency to tighten capital requirements over the sample period. - Positive MPP.Cred reflects accumulation of capital during credit expansions. - Idiosyncratic risks: - In Ad economies, Z-Score and CDS decline as of 2012 with an increase in 2020. - In Dev economies, Z-Scores seem to gradually increase over time. - Mic (average of six selected Basel scores expressed relative to minimum requirements) mean is 2.21 times above minimum requirements; average yearly change is negative, suggesting gradual relaxation of Basel scores by banks. - Mic increased in Ad economies until the end of the global financial crisis, then gradually declined; in Dev economies Mic peaked in 2014 before declining. - Mic.cred is predominantly negative, indicating banks tend to relax Basel ratios as credit activity expands. ### Key descriptive statistics (selected exact figures from Table 3) - MPP: mean 1.05, sd 2.36, min -13, max 10, diff.mean 0.08, diff.sd 3.23, obs 403, obs (C&B) 31. - MPP.cred: mean 41.01, sd 85.85, min -810, max 1,518, diff.mean 12.60, diff.sd 77.92, obs 327, obs (C&B) 31. - Dep.CB: mean 78.48, sd 20.02, min 20, max 300, diff.mean 1.82, diff.sd 8.76, obs 403, obs (C&B) 31. - Dep.NCB: mean 11.79, sd 8.97, min 1, max 39, diff.mean 0.13, diff.sd 1.45, obs 246, obs (C&B) 22. - Loan.CB: mean 72.76, sd 29.51, min 16, max 247, diff.mean 0.69, diff.sd 7.52, obs 403, obs (C&B) 31. - Loan.NCB: mean 16.79, sd 19.66, min 1, max 85, diff.mean 0.30, diff.sd 1.83, obs 312, obs (C&B) 25. - SDep.NCB: mean 0.17, sd 0.08, min 0.01, max 0.60, diff.mean -0.002, diff.sd 0.02, obs 246, obs (C&B) 22. - SLoan.NCB: mean 0.19, sd 0.11, min 0.01, max 0.75, diff.mean 0.001, diff.sd 0.02, obs 312, obs (C&B) 25. - SRISK (bank-level): mean 2.57, sd 17.00, min -56.09, max 206.06, diff.mean -0.19, diff.sd 4.14, obs 6,935, obs (C&B) 574. - Z-Score: mean -2.32, sd 4.74, min -53.37, max 40.93, diff.mean 0.01, diff.sd 3.24, obs 6,597, obs (C&B) 549. - Mic: mean 2.21, sd 0.82, min -5.27, max 15.64, diff.mean -0.01, diff.sd 0.62, obs 3,194, obs (C&B) 273. - Mic.cred: mean -0.47, sd 11.09, min -308.23, max 135.21, diff.mean -0.06, diff.sd 6.70, obs 5,709, obs (C&B) 574. ### Regulatory quality, deposit insurance, and systemic risk - Constructed DIRQ index combines static deposit insurance coverage (Demirgüç-Kunt et al., 2015 moral hazard index) with cumulative changes in regulatory quality (World Bank RQ index scaled and log-transformed). - Interaction variables: Dep.CB.DIRQ (Dep.CB multiplied by DIRQ) and Dev.Dep.CB.DIRQ (Dev subset). - Regression results (Table 11): - Dep.CB.DIRQ(t−1) coefficient: -0.2879*** (mitigating effect on SRISK). - Dev.Dep.CB.DIRQ(t−1) coefficient: 0.2627*** (deposit expansion in Dev countries associated with increased systemic risk even after accounting for DIRQ). - DIRQ coefficient: -1.8099+ (DIRQ itself associated with lower SRISK at p < 0.1). - Interpretation: Deposit insurance coverage combined with improvements in regulatory quality mitigates systemic risk, but the expansion of deposits in Dev countries remains associated with increased systemic risk—limited asset diversification in Dev countries appears predominant. ### Credit standards, competition, DTI, and banking risk - Interaction of Loan.CB with Competition and Debt Service to Income (DTI) shows: - Loan.CB(t−1) coefficient 0.1841*** (Competition specification). - Dev.Loan.CB(t−1) coefficient -0.0828*** (Competition) and -0.2350* (DTI specification). - Loan.CB.Com(t−1) coefficient 0.0966***; Dev.Loan.CB.Com(t−1) 0.0747+. - Loan.CB.DTI(t−1) coefficient 0.0323; Dev.Loan.CB.DTI(t−1) 0.4461*. - Findings: - More competitive banking systems are associated with higher systemic risk when FI in loans expands. - In Dev economies, increased competition also heightens systemic risk. - Higher DTI amplifies the systemic risk associated with FI in loans. ### Robustness checks, sample reductions, and U-shape tests - Adults with loan accounts (Adu.Loan) available for 13 countries: reduced-sample regressions maintain coefficient signs consistent with benchmark results; significant coefficients remain significant with same sign (Tables 7–8). - Example coefficients: Adu.Loan(t−1) -0.5750*** (SRISK column), Dev.Adu.Loan(t−1) -0.8225***; Loan.CB(t−1) 0.4804**. - Share of NCB regressions (Tables 9–10) show positive associations of SLoan.NCB and SDep.NCB with SRISK, Z-Score, and CDS in many specifications (e.g., SLoan.NCB(t−1) 0.1139** for SRISK). - U-shape analysis (Tables 13): - Squared FI terms included; signs of linear and squared terms preserve directions from benchmark. - Results imply that differing signs for overall FI effects and Dev-specific effects are not driven by an inverted U-shaped relationship between FI and SRISK. *IMF WORKING PAPERS — Systemic Implications of Financial Inclusion; INTERNATIONAL MONETARY FUND.* ### References ### References ### Measures and indicators of systemic risk - Acharya, V. V., Pedersen, L. H., Philippon, T., & Richardson, M. (2017). Measuring systemic risk. The review of financial studies, 30(1), 2–47. - Brownlees, C. & Engle, R. F. (2017). Srisk: A conditional capital shortfall measure of systemic risk. The Review of Financial Studies, 30(1), 48–79. - Bhansali, V., Gingrich, R., & Longstaff, F. A. (2008). Systemic credit risk: What is the market telling us? Financial Analysts Journal, 64(4), 16–24. - Giglio, S. (2016). Credit default swap spreads and systemic financial risk. Technical report, ESRB working paper series. - Tobias, A. & Brunnermeier, M. K. (2016). Covar. The American Economic Review, 106(7), 1705. - Berndt, A. & Obreja, I. (2010). Decomposing european cds returns. Review of Finance, 14(2), 189–233. - Zhang, B. Y., Zhou, H., & Zhu, H. (2009). Explaining credit default swap spreads with the equity volatility and jump risks of individual firms. The Review of Financial Studies, 22(12), 5099–5131. ### Financial inclusion, access, and deposit behavior - Demirgüç-Kunt, A., Klapper, L., Singer, D., Ansar, S., & Hess, J. (2020). The global findex database 2017: Measuring financial inclusion and opportunities to expand access to and use of financial services. The World Bank Economic Review, 34(Supplement_1), S2–S8. - Demirgüç-Kunt, A., Klapper, L. F., Singer, D., & Van Oudheusden, P. (2015). The global findex database 2014: Measuring financial inclusion around the world. World Bank Policy Research Working Paper, (7255). - Cull, R., Demirgüç-Kunt, A., & Morduch, J. (2014). Banks and microbanks. Journal of Financial Services Research, 46, 1–53. - Han, R. & Melecky, M. (2013). Financial inclusion for financial stability: Access to bank deposits and the growth of deposits in the global financial crisis. World bank policy research working paper, (6577). - Han, R. & Melecky, M. (2017). Broader use of saving products among people can make deposit funding of the banking system more resilient. Journal of International Financial Markets, Institutions and Money, 47, 89–102. - Sahay, M. R., Cihak, M., N’Diaye, M. P., Barajas, M. A., Mitra, M. S., Kyobe, M. A., Mooi, M., & Yousefi, M. R. (2015). Financial inclusion: can it meet multiple macroeconomic goals? International Monetary Fund. - Sarma, M. & Pais, J. (2011). Financial inclusion and development. Journal of international development, 23(5), 613–628. - Klapper, L. & Singer, D. (2015). The role of informal financial services in africa. Journal of African Economies, 24(suppl_1), i12–i31. - López, T. & Winkler, A. (2018). The challenge of rural financial inclusion–evidence from microfinance. Applied Economics, 50(14), 1555–1577. - López, T. & Winkler, A. (2019). Does financial inclusion mitigate credit boom-bust cycles? Journal of Financial Stability, 43, 116–129. - Le, T.-H., Chuc, A. T., & Taghizadeh-Hesary, F. (2019). Financial inclusion and its impact on financial efficiency and sustainability: Empirical evidence from asia. Borsa Istanbul Review, 19(4), 310–322. - Zheng, C. & Zhang, J. (2021). The impact of covid-19 on the efficiency of microfinance institutions. International Review of Economics & Finance, 71, 407–423. - Schulte, M. & Winkler, A. (2019). Drivers of solvency risk–are microfinance institutions different? Journal of Banking & Finance, 106, 403–426. - Anarfo, E. B., Abor, J. Y., et al. (2020). Financial regulation and financial inclusion in subsaharan africa: Does financial stability play a moderating role? Research in International Business and Finance, 51, 101070. - Vo, D. H., Nguyen, N. T., & Van, L. T.-H. (2021). Financial inclusion and stability in the asian region using bank-level data. Borsa Istanbul Review, 21(1), 36–43. - Wang, R. & Luo, H. R. (2022). How does financial inclusion affect bank stability in emerging economies? Emerging Markets Review, 51, 100876. - Ghosh, S. (2022). Financial inclusion and banking stability: Does interest rate repression matter? Finance Research Letters, 50, 103205. - Hua, X., Bi, J., & Shi, H. (2023). The appropriate level of financial inclusion: The perspective of financial stability. China Economic Quarterly International, 3(3), 167–178. - Marcelin, I., Sun, W., Teclezion, M., & Junarsin, E. (2022). Financial inclusion and bank risk-taking: The effect of information sharing. Finance Research Letters, 50, 103182. - Feghali, K., Mora, N., & Nassif, P. (2021). Financial inclusion, bank market structure, and financial stability: International evidence. The Quarterly Review of Economics and Finance, 80, 236–257. - Ahamed, M. M. & Mallick, S. K. (2019). Is financial inclusion good for bank stability? international evidence. Journal of Economic Behavior & Organization, 157, 403–427. - Danisman, G. O. & Tarazi, A. (2020). Financial inclusion and bank stability: Evidence from europe. The European Journal of Finance, 26(18), 1842–1855. - Neaime, S. & Gaysset, I. (2018). Financial inclusion and stability in mena: Evidence from poverty and inequality. Finance Research Letters, 24, 230–237. - Han, R. & Melecky, M. (2017). Broader use of saving products among people can make deposit funding of the banking system more resilient. Journal of International Financial Markets, Institutions and Money, 47, 89–102. ### Bank risk, profitability, market structure, and funding - Albertazzi, U. & Gambacorta, L. (2009). Bank profitability and the business cycle. Journal of financial stability, 5(4), 393–409. - Lepetit, L. & Strobel, F. (2013). Bank insolvency risk and time-varying z-score measures. Journal of International Financial Markets, Institutions and Money, 25, 73–87. - Chiaramonte, L., Croci, E., & Poli, F. (2015). Should we trust the z-score? evidence from the european banking industry. Global Finance Journal, 28, 111–131. - Cull, R., Demirgüç-Kunt, A., & Morduch, J. (2014). Banks and microbanks. Journal of Financial Services Research, 46, 1–53. - Craig, B. R. & Dinger, V. (2013). Deposit market competition, wholesale funding, and bank risk. Journal of Banking & Finance, 37(9), 3605–3622. - Anginer, D., Demirguc-Kunt, A., & Zhu, M. (2014). How does competition affect bank systemic risk? Journal of financial Intermediation, 23(1), 1–26. - Beck, T. & De Jonghe, O. (2013). Lending concentration, bank performance and systemic risk: exploring cross-country variation. World Bank Policy Research Working Paper, (6604). - Beck, T., De Jonghe, O., & Schepens, G. (2013). Bank competition and stability: Crosscountry heterogeneity. Journal of financial Intermediation, 22(2), 218–244. - Danisman, G. O. & Demirel, P. (2019). Bank risk-taking in developed countries: The influence of market power and bank regulations. Journal of International Financial Markets, Institutions and Money, 59, 202–217. - Khan, M. S., Scheule, H., & Wu, E. (2017). Funding liquidity and bank risk taking. Journal of banking & finance, 82, 203–216. - Vazquez, F. & Federico, P. (2015). Bank funding structures and risk: Evidence from the global financial crisis. Journal of banking & finance, 61, 1–14. - Luck, S. & Schempp, P. (2023). Inefficient liquidity creation. Journal of financial intermediation, 53, 100996. - Xu, M. T., Hu, K., & Das, M. U. S. (2019). Bank profitability and financial stability. International Monetary Fund. ### Macroprudential policy, regulation, and non-bank intermediation - Aiyar, S., Calomiris, C. W., & Wieladek, T. (2014). Does macro-prudential regulation leak? evidence from a uk policy experiment. Journal of Money, Credit and Banking, 46(s1), 181–214. - Alam, Z., Alter, M. A., Eiseman, J., Gelos, M. R., Kang, M. H., Narita, M. M., Nier, E., & Wang, N. (2019). Digging deeper–Evidence on the effects of macroprudential policies from a new database. International Monetary Fund. - Galati, G. & Moessner, R. (2018). What do we know about the effects of macroprudential policy? Economica, 85(340), 735–770. - Claessens, S., Cornelli, G., Gambacorta, L., Manaresi, F., & Shiina, Y. (2021). Do macroprudential policies affect non-bank financial intermediation?. BIS Working Papers, (927). - Kim, S. & Mehrotra, A. (2018). Effects of monetary and macroprudential policiesâevidence from four inflation targeting economies. Journal of Money, Credit and Banking, 50(5), 967–992. - Maddaloni, A. & Peydró, J.-L. (2013). Monetary policy, macroprudential policy and banking stability: evidence from the euro area. ECB Working Paper, (1560). - Mehrotra, A. N. & Yetman, J. (2014). Financial inclusion and optimal monetary policy. BIS Working Paper, (476). - Ari, M. A., Darracq-Paries, M., Kok, C., & Zochowski, D. (2017). Shadow banking and market discipline on traditional banks. International Monetary Fund. - Irani, R. M., Iyer, R., Meisenzahl, R. R., & Peydro, J.-L. (2021). The rise of shadow banking: Evidence from capital regulation. The Review of Financial Studies, 34(5), 2181–2235. - Górnicka, L. A. (2016). Banks and shadow banks: Competitors or complements? Journal of Financial Intermediation, 27, 118–131. - Darst, M., Refayet, E., & Vardoulakis, A. (2020). Banks, non banks, and lending standards. FEDS Working Paper, (2020-86). - Jiang, E. X. (2023). Financing competitors: Shadow banksâ funding and mortgage market competition. The Review of Financial Studies, 36(10), 3861–3905. ### Monetary policy, leverage, and macro-financial linkages - Borio, C. E. (2014). Monetary policy and financial stability: what role in prevention and recovery? BIS working paper. - Drehmann, M., Borio, C. E., & Tsatsaronis, K. (2012). Characterising the financial cycle: don’t lose sight of the medium term! BIS working paper. - Laeven, M. L., Dell’Ariccia, M. G., & Marquez, M. R. (2010). Monetary policy, leverage, and bank risk taking. International Monetary Fund. - Gennaioli, N., Martin, A., & Rossi, S. (2018). Banks, government bonds, and default: What do the data say? Journal of Monetary Economics, 98, 98–113. - Khan, M. S., Scheule, H., & Wu, E. (2017). Funding liquidity and bank risk taking. Journal of banking & finance, 82, 203–216. ### Econometric methods and diagnostics - Arellano, M. & Bond, S. (1991). Some tests of specification for panel data: Monte carlo evidence and an application to employment equations. The review of economic studies, 58(2), 277–297. - Hansen, L. P. (1982). Large sample properties of generalized method of moments estimators. Econometrica: Journal of the econometric society, 1029–1054. - Holtz-Eakin, D., Newey, W., & Rosen, H. S. (1988). Estimating vector autoregressions with panel data. Econometrica: Journal of the econometric society, 1371–1395. - Windmeijer, F. (2005). A finite sample correction for the variance of linear efficient two-step gmm estimators. Journal of econometrics, 126(1), 25–50. - Dormann, C. F., Elith, J., Bacher, S., Buchmann, C., Carl, G., Carré, G., Marquéz, J. R. G., Gruber, B., Lafourcade, B., Leitão, P. J., et al. (2013). Collinearity: a review of methods to deal with it and a simulation study evaluating their performance. Ecography, 36(1), 27–46. ### Regulation, capital requirements, and growth - Agénor, P.-R. & Bayraktar, N. (2023). Capital requirements and growth in an open economy. Journal of Economic Dynamics and Control, 147, 104595. - Barth, J. R., Caprio Jr, G., & Levine, R. (2013). Bank regulation and supervision in 180 countries from 1999 to 2011. Journal of Financial Economic Policy, 5(2), 111–219. - Irani, R. M., Iyer, R., Meisenzahl, R. R., & Peydro, J.-L. (2021). The rise of shadow banking: Evidence from capital regulation. The Review of Financial Studies, 34(5), 2181–2235. - Agénor, P.-R. & Bayraktar, N. (2023). Capital requirements and growth in an open economy. Journal of Economic Dynamics and Control, 147, 104595. ### Regional and sectoral studies - Allen, F., Carletti, E., Cull, R., Qian, J. â., Senbet, L., & Valenzuela, P. (2014). The african financial development and financial inclusion gaps. Journal of African economies, 23(5), 614–642. - Adasme, O., Majnoni, G., & Uribe, M. (2006). Access and risk: friends or foes?: lessons from Chile, volume 4003. World Bank Publications. - Vo, D. H., Nguyen, N. T., & Van, L. T.-H. (2021). Financial inclusion and stability in the asian region using bank-level data. Borsa Istanbul Review, 21(1), 36–43. - Neaime, S. & Gaysset, I. (2018). Financial inclusion and stability in mena: Evidence from poverty and inequality. Finance Research Letters, 24, 230–237. - Klapper, L. & Singer, D. (2015). The role of informal financial services in africa. Journal of African Economies, 24(suppl_1), i12–i31. *Systemic Implications of Financial Inclusion Working Paper No. WP/2024/203* --- _Source: https://www.imf.org/-/media/files/publications/wp/2024/english/wpiea2024203-print-pdf.pdf_