## _wp15210

## Source details

**Canonical URL:** [_wp15210](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2015/_wp15210.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2015/_wp15210.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2015/_wp15210.pdf.json)

---

### Introduction: research questions and scope
- Central questions:
  - How does bank competition affect financial stability at the institution level?
  - How may policies increasing bank competition affect stability of individual institutions and the system?
  - Does the competition–stability relationship differ across banking systems (conventional vs. Islamic)?
- Focus: Middle East and North Africa (MENA) banking sector; dimensions analyzed: solvency risk, liquidity risk (funding liquidity), and credit risk (asset quality).

### Stylized facts and data
- Sample and data:
  - 367 banks in MENA: commercial banks (258), cooperative banks (2), real estate and mortgage banks (8), Islamic banks (99).
  - Annual panel: 1999 to 2013 (Bankscope); country indicators from World Bank Survey on Bank Regulation and Supervision; macro controls from IMF World Economic Outlook.
- Key sample summary statistics (selected):
  - Five-bank asset concentration ratio: 71.16
  - Lerner Index (average): 0.31
  - Bank Z-Score (average): 23.55
  - Liquidity ratio (liquid assets / (deposits + short-term borrowing)): 41.98
  - Non-Performing Loans ratio (NPL to Gross Loans, average): 1.86
- Table 2 selected figures (exact sample moments):
  - Bank Lerner Index: Obs 1941, Mean .3257241, Std. Dev. .349005, Min -1.80814, Max 1.792211
  - Bank Z-Score: Obs 2543, Mean 9.731529, Std. Dev. 18.8839, Min -29.59916, Max 241.604
  - Non-Performing Loans Ratio: Obs 1739, Mean .1032941, Std. Dev. .1214676, Min .0012, Max .683
  - Liquid Assets/Short-Term Borrowing Ratio: Obs 3311, Mean .4705926, Std. Dev. .4017674, Min .0364, Max 2.6928
  - Activity Restrictions: Obs 3900, Mean 7.818462, Std. Dev. 1.526501, Min 3, Max 12
  - Entry into Banking Requirements: Obs 4496, Mean 7.800203, Std. Dev. .6680363, Min 4, Max 8
  - Capital Regulation Stringency: Obs 3720, Mean 7.030538, Std. Dev. 1.714338, Min 3, Max 10
  - Supervisory Power: Obs 4044, Mean 11.71893, Std. Dev. 2.305612, Min 5, Max 15
  - Government Owned Bank Assets: Obs 2448, Mean 23.56516, Std. Dev. 29.17612, Min 0, Max 95.78

### Measurement and empirical strategy
- Competition measure: Lerner Index (bank-level; price minus marginal cost over average price), marginal cost from translog production function; market share reported separately.
- Stability metrics:
  - Solvency: Z-Score = (equity-asset ratio + RoA) / std. dev. of RoA (rolling four-year std. dev. of RoA).
  - Liquidity: Liquid Assets / Short-Term Borrowing ratio.
  - Credit risk: Non-Performing Loans ratio.
- Estimation:
  - Panel regressions with bank fixed effects (GLS random-effect when including time-invariant Islamic dummy).
  - Baseline includes one-year lagged explanatory variables, bank balance-sheet controls, country macro controls.
  - Interaction terms: Lerner × MarketEntry; Lerner × regulation/supervision subcomponents; Lerner × Islamic dummy.

### Major empirical findings — solvency (Z-Score)
- Baseline:
  - Increase in the Lerner Index (decrease in price competition) improves bank solvency (Z-Score).
  - A one-standard deviation rise in banks’ market power implies an increase in bank Z-Score by 4.68.
    - Context: bank-level Z-Score mean = 9.73 and standard deviation = 18.88.
  - A one-standard deviation increase in country-average market power is associated with a rise in bank Z-Score by 7.29.
- Regression highlights (selected coefficients):
  - Lerner_1: 0.134*** (0.0466).
  - Average Lerner_1: 0.293*** (0.0839).
  - Market Share_1: 0.706** (0.347).
  - Lerner_1 * Activity Restrictions: -0.0241** (0.0121).
  - Lerner_1 * Entry Requirements: 0.0611*** (0.0235).
  - Lerner_1 * Capital Regulation: 0.0823*** (0.0154).
  - Lerner_1 * Supervisory Independence: 0.102*** (0.0266).
  - Lerner_1 * No Deposit Insurance: -0.125** (0.0571).
- Thresholds and interactions:
  - Prudential entry requirements indicator range in sample: 4 to 8; sample average = 7.80; estimated threshold = 5.81.
    - Above 5.81, prudential entry requirements can ensure incumbent banks with market power maintain adequate solvency; below 5.81, price competition may improve incumbent solvency.
  - Capital regulation indicator range: 3 to 10; average = 7.03; threshold where overall impact of price competition on solvency is negative if capital regulation exceeds 6.17.
- Islamic banks:
  - No significant difference in solvency between conventional and Islamic banks; no evidence that competition–solvency nexus differs for Islamic banks.

### Major empirical findings — liquidity (Liquid Assets / Short-Term Borrowing)
- Baseline:
  - Decrease in the Lerner Index (increase in price competition) improves the liquidity ratio.
  - A one-standard deviation decrease in the bank-level Lerner Index induces an increase in the liquidity ratio by 3.7 percent for that institution.
  - Alternative magnitude: A one-standard deviation decrease in the bank Lerner Index raises the liquidity ratio by 7.4 percent (text reports both 3.7 percent and 7.4 percent in different contexts; average liquidity ratio = 47 percent).
- Regression highlights (selected coefficients):
  - Lerner_1: -0.106*** (0.0210).
  - Average Lerner_1: -0.299*** (0.0357).
  - Loans Assets Ratio_1: -0.643*** (0.0591).
  - Non-Performing Loans Ratio_1: 0.352*** (0.0599).
  - Lerner_1 * Capital Regulation: -0.0188** (0.00745).
  - Lerner_1 * Supervisory Independence: 0.0394*** (0.0130).
  - Lerner_1 * No Deposit Insurance: -0.0843*** (0.0319).
  - Lerner_1 * Government Owned Banks: -0.00885*** (0.00117).
- Market entry and activity restrictions:
  - Increase in Activity Restrictions from minimum (3) to maximum (12) raises the benefits of price competition for the liquidity ratio from 2.1 percent to 8.5 percent.
  - Rise in Entry Requirements from minimum (4) to maximum (8) changes marginal effect of competition on liquidity from -2.6 percent to 3.8 percent.
  - Threshold level for entry requirements to ensure a positive effect of competition on liquidity: 5.61 (sample average = 7.80).
- Islamic banks and liquidity:
  - Islamic dummy implies, on average, a liquidity ratio higher by 14 percent (Islamic: 13.55** (5.440) in regression).
  - Lerner_1 * Islamic: 0.138** (0.0705) — price competition does not have a significant positive effect on liquidity for Islamic banks; Islamic banks already hold larger buffers.

### Major empirical findings — credit risk (NPL ratio)
- Baseline:
  - A one-standard deviation decrease in the Lerner Index (increase in price competition) implies a rise in the NPL ratio by 1.6 percent (baseline).
  - Sample average NPL rate: 10.3 percent; standard deviation: 12.1 percent.
  - A one-standard deviation decrease in country-average Lerner Index reduces NPL ratio by 3.2 percent.
  - Bank market share: a percentage point increase implies a 0.26 percent rise in NPLs.
- Regression highlights (selected coefficients):
  - Lerner_1: -0.0462*** (0.0134).
  - Average Lerner_1: -0.127*** (0.0244).
  - Market Share_1: 0.262** (0.102).
  - Lerner_1 * Denied Applications: 0.110*** (0.0382).
  - Deposit Funding Ratio_1: 0.121*** (0.0349).
  - Return on Average Equity_1: -0.241*** (0.0333).
  - GDP Growth_1: -0.252*** (0.0662).
  - CPI Inflation_1: -0.377*** (0.0746).
  - Lerner_1 * Capital Regulation: -0.0153*** (0.00498).
  - Lerner_1 * Supervisory Power: 0.00944** (0.00391).
  - Lerner_1 * No Deposit Insurance: -0.0377* (0.0204).
- Thresholds and interactions:
  - Market power may have a positive impact on credit quality if capital regulation index exceeds threshold 5.27 (sample average = 7).
  - If fraction of denied license applications > 72 percent, an increase in market power induces a rise in NPL ratio (reversal).
- Islamic banks and credit risk:
  - Islamic nature of a bank reduces NPL ratio by 6.6 percent (document states: "The Islamic nature of a bank reduces this nonperforming loan ratio by 6.6 percent").
  - Regression: Islamic: -6.580* (3.996) in one specification; Lerner_1 * Islamic: -0.00745 (0.0457) — no significant interaction effect implying the competition–credit-risk nexus does not differ robustly for Islamic banks.

### Mechanisms and interpretation
- Liquidity channel:
  - Self-discipline mechanism: competition reduces profit margins and makes costly funding unaffordable, prompting banks to accumulate larger liquid buffers.
  - Capital regulation and regulatory environment can strengthen or weaken this channel; deposit insurance can reduce the incentive to hold liquidity.
- Solvency channel:
  - Competition reduces profitability (negative pressure on solvency); banks with market power can use rents to build capital cushions (positive for solvency).
  - Net effect ambiguous and conditional on bank capital management and regulatory stringency.
- Credit quality channel:
  - Competition reduces lending rates (improves borrower repayment capacity) but can encourage risk-taking ("reach for yield") increasing NPLs.
  - Market power can allow better screening and price discrimination, improving asset quality.
- Institutional context:
  - Market contestability, entry barriers, capital regulation, supervisory independence, and deposit insurance materially change sign and magnitude of competition–risk relationships.

### Policy implications and recommendations
- Coordinate competition policy with prudential regulation and supervision.
- Ensure stringent capital regulation and independent supervision to encourage market-power banks to build capital buffers and limit risk-taking.
- Where barriers to entry are high, increasing price competition among incumbents can improve solvency and asset quality by limiting exploitation of market power.
- Keep entry-limiting regulations to the minimum necessary; balance activity restrictions serving prudential purposes against their effect on contestability.
- Consider role of Islamic banks: their larger liquidity buffers and lower NPLs can reduce system-level credit risk and alter liquidity responses to competition.
- Reduce government ownership and influence to lower moral hazard and strengthen incentives for prudent liquidity and risk management.

### Summary conclusions (selected exact statements and statistics)
- Competition effects differ by risk type:
  - Competition has a positive effect on bank liquidity.
  - Competition may have a potentially negative impact on solvency and credit quality.
- Islamic banking:
  - Sample average rate of nonperforming loans: 10.3 percent.
  - The Islamic nature of a bank reduces this nonperforming loan ratio by 6.6 percent.
- Quantitative notes preserved:
  - "A one-standard deviation rise in banks’ market power implies an increase in bank Z-Score by 4.68."
  - "A one-standard deviation decrease in the bank-level Lerner Index induces an increase in the liquidity ratio by 3.7 percent."
  - "A one-standard deviation decrease in the Lerner Index implies a rise in the ratio of nonperforming loans to total credit by 1.6 percent."
  - "0.3 percent (almost no impact)."

*Source — _wp15210 (IMF working paper PDF chapter content provided).*

### References .............................................................................................................

### _wp15210 - References .............................................................................................................

### References
- References ............................................................................................................................... 30

### Tables
- 1. Sources and Description of the Variables ........................................................................... 3 2

*Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2015/_wp15210.pdf*

### 2. Summary Statistics ..................................................................................................

### _wp15210 - 2. Summary Statistics

### Introduction: research questions and scope
- Central questions:
  - How does bank competition affect financial stability at the institution level?
  - How may policies increasing bank competition affect stability of individual institutions and the system?
  - Does the competition–stability relationship differ across banking systems (conventional vs. Islamic)?
- Focus: Middle East and North Africa (MENA) banking sector, characterized by low competition and high barriers to entry, and the coexistence of conventional and Islamic banks.
- Dimensions of financial stability analyzed: solvency risk, liquidity risk (funding liquidity), and credit risk (asset quality).

### Key empirical findings (summarized from the text)
- Differential effects of price competition across risk types:
  - Price competition shows a significant and positive impact on bank liquidity.
  - Price competition may have potentially negative effects on bank solvency.
  - Price competition may increase credit risk of loan portfolios under certain conditions.
- Mechanisms described:
  - Liquidity: more competitive banks with lower profit margins tend to hold larger liquidity buffers because they cannot afford costly funding sources (self-discipline mechanism).
  - Solvency: competition reduces profit margins; banks with market power may use higher rents to build larger capital cushions. If competition reduces profitability more than capital responses, solvency declines.
  - Credit risk: competition reduces lending rates (can improve borrower repayment ability), but may also incentivize banks to reach for yield by lending to riskier borrowers; banks with market power can price-discriminate and screen borrowers better, potentially improving asset quality.
- Institutional and market-structure interactions:
  - Degree of market contestability and quality of banking regulation and supervision—when interacted with price competition—change the sign and magnitude of observed relationships.
  - Capital regulation and supervisory independence can enable market-power banks to safely manage profits and increase capital buffers.
  - In presence of high barriers to entry, increasing price competition among incumbents may improve solvency and asset quality by limiting exploitation of market power.
- Islamic vs. conventional banks:
  - Islamic banks hold larger liquidity buffers than conventional banks, partly because liability structures are less reliant on short-term funding.
  - Islamic banks exhibit lower credit risk (lower rates of nonperforming loans) and thus may reduce the system-level credit risk response to increased competition.
  - No significant differences found between conventional and Islamic banks regarding the impact of competition on solvency.

### The banking system in MENA: stylized facts and summary statistics
- Sample-level and regional descriptive statistics (from Global Financial Development Database, World Bank, 2011):
  - Five-bank asset concentration ratio: 71.16
  - Lerner Index (average): 0.31
  - Bank Z-Score (average): 23.55
  - Liquidity ratio (liquid assets / (deposits + short-term borrowing)): 41.98
  - Non-Performing Loans ratio (NPL to Gross Loans, average): 1.86
- Additional MENA features highlighted:
  - Coexistence and growth of Islamic banking; Islamic share of national banking markets in selected GCC countries:
    - Saudi Arabia: 48.9%
    - Kuwait: 44.6%
    - Bahrain: 27.7%
  - Relative resilience to the global financial crisis compared with other emerging markets, with heterogeneity across countries.
  - High market concentration and significant barriers to entry (activity restrictions, entry requirements, state-owned bank dominance).

### Literature review and hypotheses (concise)
- Competing theoretical views:
  - Competition-fragility (charter value hypothesis): more competition reduces franchise value and increases risk-taking.
  - Competition-stability: competition reduces lending rates and improves borrower repayment capacity, lowering credit risk.
  - Nonlinear and risk-type–specific theories reconcile differing viewpoints.
- Hypotheses for each risk type:
  - Solvency risk:
    - Competition reduces profitability (negative for solvency).
    - Competition may induce banks to target higher capital ratios (positive for solvency) if banks actively manage capital; net effect ambiguous and dependent on relative magnitudes.
    - Interaction effects: market entry conditions and quality of regulation/supervision can change outcomes.
  - Liquidity risk:
    - Two competing channels: competition may increase short-term wholesale funding (raising liquidity risk) or induce larger holdings of liquid assets (reducing liquidity risk).
    - Theoretical prediction (Carletti and Leonello, 2012): competition increases liquidity buffers — implies a positive effect of competition on liquidity.
  - Credit risk:
    - Competition reduces lending rates improving borrower repayment (reduces credit risk).
    - Competition may incentivize risk-taking to restore profits (increases credit risk).
    - Market power enables better price discrimination and screening, potentially improving asset quality.

### Empirical strategy and data
- Sample:
  - 367 banks in MENA.
  - Bank types in sample: commercial banks (258), cooperative banks (2), real estate and mortgage banks (8), Islamic banks (99).
  - Investment banks excluded.
  - Annual panel data from 1999 to 2013 (Bankscope balance sheet data).
- Key variables and measures:
  - Main explanatory variable: Lerner Index as an indicator of price competition / market power.
  - Solvency measure: Z-Score, constructed as (average return on assets + average equity-to-assets) / standard deviation of return on assets (see section III.B for full description).
  - Liquidity measure: ratio between liquid assets and short-term borrowing (an increase implies improved liquidity position).
  - Credit risk measure: Non-Performing Loans ratio (NPL to Gross Loans).
- Estimation approach:
  - Panel regressions with bank fixed effects.
  - Baseline specification includes one-year lagged explanatory variables, bank-level balance sheet controls, and country-specific macroeconomic controls.
  - Interaction terms included between price competition and country-specific indicators of regulation, supervision, and market entry to capture conditional effects.

### Policy implications (as presented in the text)
- Coordinate competition policy with prudential regulation and supervision because capital requirements, banking supervision, and market regulation can shape the competition–risk relationship.
- Specific recommendations derived from interaction results:
  - Ensure stringent capital regulation and independent supervision to provide incentives for market-power banks to build capital buffers and reduce risk-taking.
  - Where barriers to entry are high, increasing price competition among incumbents can help avoid negative stability effects of market power.
  - Keep entry-limiting regulations (activity restrictions, licensing barriers) to the minimum necessary to avoid unduly reducing market contestability; where activity restrictions serve prudential purposes, balance their use against their effect on contestability.
  - Note that the presence of Islamic banks, given their larger liquidity buffers and lower NPLs, can reduce system-level credit risk and alter the liquidity impact of competition.

*Italic: Source — _wp15210 - 2. Summary Statistics (IMF PDF chapter content provided).*

### 1. To analyze the effect of market power and of market entry, we run the following regression:

### _wp15210 - 1. To analyze the effect of market power and of market entry, we run the following regression:

### Regression framework: market power and market entry
- Specification includes an interaction term between a bank-level measure of market power and a country-specific MarketEntry jt-1 variable.
- MarketEntry jt-1 is a country-specific variable for market contestability, measuring the intensity of the barriers to entry and of the activity restrictions to financial intermediaries, based on the World Bank Survey on Bank Regulation and Supervision.
- The interaction term may be considered as an indicator of bank market power, adjusted for the entry conditions in a given market.
- Conceptual interpretation provided in source:
  - A bank with low current pricing market power but operating in a market that significantly restricts entry may be able to increase prices or collude with incumbents in the future.
  - A bank with relatively high market power in a widely open market may preserve pricing power only if its products and services are discernibly better than competitors’.
- Empirical implication emphasized:
  - Market power per se (the ability to profitably raise the price of a product or a service over the marginal cost) may not necessarily imply negative effects on stability if the market is relatively contestable and subject to low barriers to entry.
  - If high entry barriers limit access of other banks, a credit institution with market power could exploit additional profits to expand activities and increase risk taking.
  - Entry restrictions can encourage risk-taking from incumbent banks in a concentrated and closed market, potentially increasing balance-sheet size and systemic relevance via higher likelihood of implicit government support.
  - Such incentives are less relevant in a contestable market where new entrants could easily start a banking business; potential excessive expansion would have to internalize higher probability of default and lower likelihood of government intervention.

### Regression framework: interaction with regulation and supervision
- Second estimated equation introduces interaction between the Lerner index and a country-specific regulation/supervision quality variable.
- The country-specific variable for the quality of the regulatory and supervisory framework is drawn from the World Bank Survey on Bank Regulation and Supervision and covers:
  - capital stringency
  - depositor protection schemes
  - supervisory effectiveness
  - supervisory independence
  - presence of government-owned banks
- The interaction of the Lerner index with subcomponents of the World Bank Survey is analyzed depending on the specification for the dependent variable.

### Regression framework: role of Islamic banking
- Third regression includes a bank-specific dummy variable for Islamic banks.
- The Islamic dummy is defined based on specialization classified in Bankscope.10
- Purpose:
  - Explore ex ante whether Islamic banks may behave differently in terms of solvency, liquidity, and asset quality.
  - Investigate whether the business model of Islamic banks may affect the sign or magnitude of the relationship between bank competition and stability.
- Estimation detail:
  - Because the Islamic dummy is time-invariant for a given bank during the considered time period and may play a role similar to a fixed effect, the panel regression is estimated using a GLS random-effect specification.

### Data sources and description of variables
- Bank-level balance sheet variables: Bankscope.
- Country-level information on market entry conditions and quality of the regulatory framework: World Bank Survey on Bank Regulation and Supervision.
- Macroeconomic controls at the country level: IMF World Economic Outlook database.

### Bank competition measures used
- The degree of competition in the banking market is analyzed through measures related to three concepts: price competition, market contestability, and market concentration.
- Key explanatory variables include:
  - Bank-specific measures of price competition (in particular the Lerner index).

*Source: _wp15210 - 1. To analyze the effect of market power and of market entry, we run the following regression.*

### 2. Country-level indicators of regulatory restrictions (based on the World Bank indicators for

### 2. Country-level indicators of regulatory restrictions (based on the World Bank indicators for barriers to entry and activity restrictions)

### Measurement of competition (bank level)
- Competition is measured by the Lerner Index, defined as the ratio between the markup (price minus marginal cost) and the average price of bank activities.
- Price is proxied by the ratio of total revenues to total earning assets, where total revenues include interest income and non-interest operating income and equity-accounted profit/loss operating income.
- Marginal cost is derived from estimation of a translog production function at the bank level for each country, including bank and time fixed effects.
- Input price definitions:
  - Price of labor (cLabor): ratio of personnel expenses over total assets.
  - Price of funding (cFunds): ratio of total interest expenses to the total amount of deposits, money market, and short-term funding.
  - Price of fixed assets (cFixed): ratio of other operating expenses to total assets.
- Market share of a given bank is introduced as an indicator of position in terms of market concentration; market concentration and market power are structurally distinct.

### Financial stability metrics (bank-level dimensions)
- Three dimensions considered:
  - Solvency risk: measured using the Z-Score.
    - Z-Score = (equity-asset ratio (E/A) + Return on Assets (RoA)) / standard deviation of RoA.
    - Return on Average Assets is used to reduce impact of asset changes during the year.
    - Standard deviation of Return on Average Assets computed on a rolling four-year interval.
    - Higher Z-Score indicates higher solvency.
  - Funding liquidity risk:
    - Liquidity Ratio = ratio between liquid assets and short-term borrowing.
    - This explains the size of the liquid assets buffer a bank has at its disposal.
  - Credit risk:
    - Measured by Non-Performing Loans (NPLs) ratio (major component of on-balance-sheet assets is loans).

### Bank regulation and supervision indicators (World Bank Survey)
- Market entry and activity indicators:
  - Activity Restrictions.
  - Entry into Banking Requirements.
  - Fraction of Denied Applications.
- Prudential regulation and supervision indicators:
  - Capital Regulation Stringency.
  - Supervisory Power.
  - Supervisory Independence.
  - Deposit Insurance Scheme.
  - Fraction of Government-Owned Bank Assets at the country level.
- Survey waves: 2001, 2003, 2007, 2011 (covers 143 jurisdictions).

### V. Empirical results — overview
- Empirical analysis distinguishes solvency, liquidity, and credit risk.
- Examines:
  - Effect of market power and entry barriers on bank stability.
  - Interaction between market competition and banking regulation and supervision.
  - Potential role of Islamic banking in the competition–stability nexus.

### A. Solvency Risk — key findings
- Baseline result:
  - An increase in the Lerner Index (decrease in price competition) improves bank solvency (Z-Score).
  - A one-standard deviation rise in banks’ market power implies an increase in bank Z-Score by 4.68.
    - Context: bank-level Z-Score mean = 9.73 and standard deviation = 18.88.
- Country-average market power:
  - A one-standard deviation increase in the average market power of banks in a country is associated with a rise in bank Z-Score by 7.29 (56 percent larger than the bank-level effect).
- Bank balance sheet controls and their associations with Z-Score:
  - Larger share of government bond exposures / total assets → higher solvency.
  - Faster growth rate of total assets → higher solvency risk (excessive expansion can worsen solvency).
  - Higher ratio of non-performing loans / total assets → worsening of solvency.
  - Higher ratio of non-interest income / total revenues → positive impact on solvency (diversification reduces profit volatility).
- Market entry conditions can alter the competition–solvency relationship:
  - If supervisory authorities reject more than 85 percent of banking license submissions, an increase in incumbent market power would reduce solvency (i.e., more price competition would be beneficial).
  - If Activity Restrictions take their maximum value of 12, the positive effect of market power on solvency would be completely offset.
- Prudential entry requirements threshold:
  - World Bank prudential entry requirements indicator ranges between 4 and 8 in the sample; average value = 7.80.
  - Estimated threshold based on interaction coefficients = 5.81.
  - Interpretation: above 5.81, prudential entry requirements can ensure incumbent banks with market power maintain adequate solvency; below 5.81, price competition may improve incumbent solvency.
- Interaction with capital regulation:
  - In the absence of capital regulation, a one-standard deviation increase in price competition would imply an improvement in bank solvency by almost one-standard deviation.
    - Note: a one-standard deviation decrease in bank market power (Lerner Index) increases bank Z-Score by 17.73, given Z-Score standard deviation = 18.88.
  - Interaction term for capital regulation is positive and increasing with capital stringency.
  - Overall impact of price competition on solvency is negative (i.e., market power increases solvency) if capital regulation indicator exceeds threshold = 6.17.
    - Sample range for capital stringency = 3 to 10; average = 7.03.
  - Mechanism interpretation:
    - Without capital regulation: competition reduces profits but induces banks to increase capital buffers (endogenous capital), potentially improving solvency.
    - With binding capital regulation: banks keep required capital passively; competition reduces profitability and thus reduces solvency (market power raises solvency).
- Supervisory independence:
  - Positive effect of market power on solvency is increasing in the degree of supervisory independence.
  - Effective and independent supervision induces banks with market power to manage capital more prudently.
- Deposit insurance:
  - When no explicit deposit insurance exists, solvency differences between high market power and low market power banks are small.
  - In the presence of deposit insurance, banks with market power tend to show higher solvency than banks with competitive pricing.
  - Interpretation: absence of deposit insurance raises opportunity cost of insolvency, incentivizing low-profit (low market power) banks to raise capital; deposit insurance reduces that incentive.
- Islamic banking:
  - No significant difference in bank solvency between conventional and Islamic banks.
  - No evidence that the competition–solvency nexus differs for Islamic banks.

### B. Liquidity Risk — key findings
- Baseline result:
  - A decrease in the Lerner Index (increase in price competition) improves the liquidity ratio.
  - A one-standard deviation decrease in the bank-level Lerner Index induces an increase in the liquidity ratio by 3.7 percent for that institution.
    - Context: for banks in the sample, ratio of liquid assets / short-term borrowing mean = 47 percent and standard deviation = 4 percent.
- The positive effect of competition on liquidity is robust across specifications and accounting for market entry and regulatory factors.
- Country-average competition:
  - A one-standard deviation decrease in the country-average Lerner Index (increase in competition at country level) implies an improvement in the bank-level liquidity ratio (estimate discussed but numeric value not provided in the excerpt).

*Source: IMF working paper content unit "_wp15210 - 2. Country-level indicators of regulatory restrictions (based on the World Bank indicators for"*

### 7.4 percent, which is—in terms of magnitude—twice as large as the effect of a corresponding

### _wp15210 - 7.4 percent, which is—in terms of magnitude—twice as large as the effect of a corresponding

### Price competition and bank liquidity
- A one-standard deviation decrease in the bank Lerner Index (increase in price competition) raises the liquidity ratio by 7.4 percent, which is—in terms of magnitude—twice as large as the effect of a corresponding change in the bank-level Lerner Index.
- The positive effect of price competition on bank liquidity suggests banks respond to reduced profit margins by increasing liquid asset buffers to face future cash outflows.
- Average liquidity ratio in the sample: 47 percent.
- A one-standard deviation decrease in the Lerner Index implies an improvement of 3.5 percent in the liquidity ratio (used for magnitude comparison with Islamic banks).

### Bank-level balance sheet factors affecting liquidity
- Banks with larger lending activity with respect to total assets present a wider liquidity mismatch and higher liquidity risk.
- Higher nonperforming loans (NPLs) lead banks to hold more liquid assets as buffers.
- A higher fraction of non-interest income over total revenues (more fee-based revenues) may allow banks to hold less liquidity.
- Higher bank profitability improves liquidity by increasing available cash revenues.
- Evidence supports active liquidity management: banks adjust liquid buffers relative to effective cash-flow risks.
- Banks in MENA adopted prudent liquidity management on average during the analysis period; the region did not experience liquidity shocks comparable to advanced economies during the crisis.

### Market entry conditions, activity restrictions, and liquidity
- Interaction effects: the positive effect of price competition on liquidity is larger when regulators impose activity restrictions or establish legal entry requirements.
- Increase in activity restrictions from the minimum (3) to the maximum value (12) raises the benefits of price competition for the liquidity ratio from 2.1 percent to 8.5 percent.
- Rise in banking entry requirements from the minimum (4) to the maximum (8) changes the marginal effect of competition on liquidity from -2.6 percent to 3.8 percent.
- Threshold level for entry requirements to ensure a positive effect of competition on liquidity: 5.61.
- Sample average value of entry requirements: 7.80.
- Interpretation: activity restrictions and entry requirements can ensure prudent management and allow competition to improve liquidity; without these, competition may encourage recourse to short-term funding and reduce liquidity.

### Prudential regulation, supervision, and liquidity
- Capital regulation stringency increases the liquidity-enhancing effect of price competition.
- Increase in capital regulation from the minimum (3) to the maximum (10) increases the positive effect of higher competition on the bank liquidity ratio from 2 percent to 6.6 percent.
- Supervisory power interaction: nonsignificant coefficient.
- Supervisory independence may reduce the positive effect of competition on liquidity (or reduce the negative impact of market power), by reinforcing incentives to hold appropriate liquid assets.
- Deposit insurance effect: explicit deposit insurance or full compensation of depositors in past failures can eliminate the liquidity-enhancing effect of competition due to moral hazard; absence of protection induces more prudent liquidity management under competitive pressures.
- Government ownership: average 24 percent of total banking system assets are government-owned in MENA.
  - Larger fraction of government-owned banks increases the negative effect of bank market power on liquidity position (greater moral hazard and weaker incentives for prudent liquidity management).

### Islamic banking and liquidity
- Islamic banks tend to have larger liquidity buffers compared with conventional banks due to liability structures less reliant on short-term funding.
- Islamic dummy implies that, on average, a bank would have a liquidity ratio higher by 14 percent.
- For comparison: a one-standard deviation decrease in the Lerner Index implies a 3.5 percent improvement in the liquidity ratio.
- Price competition does not have a significant positive effect on liquidity for Islamic banks; the overall effect of an increase in the Lerner Index for Islamic banks may be positive when summing the Lerner coefficient and the interaction term.
- Interpretation: Islamic banks rely on cash-flow-driven (passive) liquidity accumulation and limited access to interest-paying wholesale funding; more profitable Islamic banks (through market power) may hold larger liquidity buffers.

### Price competition and credit risk (nonperforming loans)
- A one-standard deviation decrease in the Lerner Index (increase in price competition) implies a rise in the ratio of nonperforming loans to total credit by 1.6 percent (baseline specification).
- Sample average rate of nonperforming loans: 10.3 percent; standard deviation: 12.1 percent.
- Country-average Lerner Index effect: a one-standard deviation decrease in the country-average Lerner Index reduces the ratio of nonperforming loans by 3.2 percent.
- Bank market share effect: a percentage point increase in bank market share implies a 0.26 percent rise in the rate of nonperforming loans.
- If the fraction of denied license applications is higher than 72 percent, an increase in market power induces a rise in the ratio of nonperforming loans (reversed effect).
- Funding composition effects:
  - Higher fraction of deposits over total funding increases the rate of nonperforming loans (deposit-funded banks have weaker market monitoring incentives).
  - Larger share of government bond exposures to total assets is associated with a lower rate of nonperforming loans (more conservative risk profile).
  - Higher return on average equity is associated with a lower rate of nonperforming loans (more profitable banks lend to less risky borrowers).
- Macroeconomic effects:
  - GDP growth rate and CPI inflation rate are negatively related to the rate of nonperforming loans (favorable macro conditions improve asset quality).

### Regulation, supervision, and credit risk
- Capital regulation can change the sign and magnitude of the competition–credit-quality nexus.
  - Market power may have a positive impact on credit quality if the country-level indicator of capital regulation exceeds threshold 5.27.
  - Sample average value of the capital regulation index: 7.
- Supervisory power:
  - Increasing prerogatives of supervisory authorities reduces the negative effect of price competition on credit quality.
  - Rise in supervisory power from the minimum (5) to the maximum value (15) decreases the effect of price competition on the nonperforming loans ratio from 3.6 percent to [text truncated in source].

*Source: Excerpt from IMF working paper content provided in file _wp15210 - 7.4 percent, which is—in terms of magnitude—twice as large as the effect of a corresponding*

### 0.3 percent (almost no impact).

### VI. CONCLUSIONS

### Overview and motivation
- The paper analyzes the relationship between bank competition and financial stability for banks in the Middle East and North Africa (MENA).
- MENA banking systems generally show low levels of competition and relatively high market concentration, but satisfactory financial stability during the global financial crisis, including a high level of bank solvency, large buffers of liquid assets with respect to deposits and short-term borrowings, and relatively limited risk taking as indicated by markedly lower rate of nonperforming loans compared with financial sectors elsewhere.

### Main contributions
- Three primary contributions:
  - Explore how competition affects three types of bank risk at the institution level—solvency, liquidity, and credit risk—and show heterogeneous effects can be explained by different types of risk.
  - Examine how market entry, bank regulation, and bank supervision shape the impact of competition on different sources of bank risk.
  - Study whether competition has different effects for Islamic banks; find a potential difference for funding liquidity due to Islamic banks being less reliant on wholesale funding.

### Empirical findings (key statistics preserved)
- Competition effects differ by risk type:
  - Competition has a positive effect on bank liquidity.
  - Competition may have a potentially negative impact on solvency and credit quality.
- Islamic banking:
  - Sample average rate of nonperforming loans: 10.3 percent.
  - The Islamic nature of a bank reduces this nonperforming loan ratio by 6.6 percent.
  - No significant evidence that the relationship between price competition and credit risk differs for Islamic banks; higher credit quality is attributed to Islamic banks’ specific business model rather than to the competition–stability nexus.
- Deposit insurance and moral hazard:
  - Presence of deposit insurance may induce moral hazard in banks’ lending decisions; if depositors are fully reimbursed, banks are incentivized to take risk to such an extent that the relationship between price competition and credit risk is not significant anymore.
- Earlier mentioned quantitative note in the document: "0.3 percent (almost no impact)."

### Mechanisms by risk type
- Liquidity:
  - Price competition improves liquidity by inducing a self-discipline mechanism on funding choices: lower profit margins make costly funding unaffordable, so banks prefer larger buffers of liquid assets.
  - Capital regulation may strengthen the liquidity-enhancing effect of price competition.
  - Deposit insurance may reduce incentives to hold larger liquidity buffers.
- Solvency:
  - Price competition may reduce bank profits and potentially harm solvency.
  - If banks actively manage capital, they may increase capital in response to lower profitability; competition can have a positive solvency effect if capital increases are sufficiently large to offset profit reductions.
  - Deposit insurance reduces the opportunity cost of potential insolvency for competitive banks and may lessen incentives to increase capital.
  - Prudential requirements may enable banks with market power to manage additional profit margins and increase capital buffers.
- Credit risk:
  - Price competition may increase credit risk if banks take additional risks to restore profitability; lender-side risk-taking effects may dominate borrower-side improvements.
  - Capital regulation may incentivize prudent management by banks with market power.

### Policy implications
- Market contestability:
  - Reducing competition can increase bank risk, particularly solvency and credit risk, when markets are heavily regulated or access is restricted (e.g., frequent denial of license applications).
  - Authorities should consider reforms that promote price competition among incumbent banks in regulated markets and modify or reduce regulations that create unnecessary restrictions to contestability, when not justified by other major policy objectives.
- Supervision and government influence:
  - Improve quality and independence of prudential supervision to limit distortionary incentives on bank behavior.
  - Strong supervisory power can counteract competition-induced incentives for risky lending by encouraging appropriate risk management.
  - Reducing government stake and control in the financial sector may lower moral hazard incentives for banks with large market power that rely on public support and may otherwise manage liquidity imprudently.

* _wp15210 - 0.3 percent (almost no impact)._

### REFERENCES

### _wp15210 - REFERENCES

### References (selected)
- Abedifar P., Molyneux P. and Tarazi A. (2013), Risk in Islamic Banking, “Review of Finance”, Vol. 17, No. 6, pp.2035-2096
- Allen F., Carletti E. and Marquez R. (2011), Credit Market Competition and Capital Regulation, “Review of Financial Studies”, Vol. 24, Issue 4, pp.983-1018
- Anginer D., Demirguc-Kunt A. and Zhu M. (2014), How Does Competition Affect Bank Systemic Risk?, “Journal of Financial Intermediation”, Vol. 23, pp.1–26.
- Baele L., Farooq M. and Ongena S. (2014), Of Religion and Redemption: Evidence from Default on Islamic Loans, “Journal of Banking and Finance”, Vol. 44, pp. 141–159.
- Barth J., Caprio G. and Levine R. (2006), Rethinking Bank Regulation: Till Angels Govern, (New York: Cambridge University Press).
- Beck T., De Jonghe O. and Schepens G. (2013), Bank Competition and Stability: Cross-Country Heterogeneity, “Journal of Financial Intermediation”, Vol. 22, pp. 218–244.
- Beck T., Demirguc-Kunt A. and Merrouche O. (2013), Islamic vs. Conventional Banking: Business Model, Efficiency and Stability, “Journal of Banking and Finance”, Vol. 37, pp. 433–447.
- Boyd J. and De Nicolo’ G. (2005), The Theory of Bank Risk Taking Revisited, “Journal of Finance”, Vol. 60, pp. 1329–1343.
- Diamond D. (1984), Financial Intermediation and Delegated Monitoring, “Review of Financial Studies”, Vol. 51, No. 3, pp. 393–414.
- Ernst & Young (2014), “World Islamic Banking Competitiveness Report” (Dubai, UAE).
- Freixas X. and Rochet J. C. (2008), Microeconomics of Banking, Second Edition (Cambridge, MA: MIT Press).
- Hasan M. and Dridi J. (2010), “The Effects of the Global Financial Crisis on Islamic and Conventional Banks: A Comparative Study”, IMF Working Papers Series No. 10-2010 (Washington: International Monetary Fund).
- Turk-Ariss R. (2010), Competitive Conditions in Islamic and Conventional Banking: A Global Perspective, “Review of Financial Economics”, Vol. 19, pp.101-108.
- Vives X. (2014), Strategic Complementarity, Fragility and Regulation, “Review of Financial Studies”, Vol. 27, Issue 12, pp. 3547–3592.
(References list continues in source.)

### Table: Sources and Description of the Variables
- Market Power and Competition
  - Bank-level Lerner Index: measure of market power; computed as the ratio between the bank mark-up (price – marginal cost) and the average price of bank assets. Calculated from BANKSCOPE.
  - Country-average Lerner Index: average value, at the country level, of the Lerner Index of all the banks in that country.
  - Market Share: ratio of total assets of a bank to total assets of all banks in a given country.
- Bank Solvency
  - Z-Score: sum of the equity-asset ratio and RoA divided by the standard deviation of RoA. Calculated from BANKSCOPE.
  - Equity-Asset Ratio: ratio between total equity and total assets (BANKSCOPE).
  - Return on Average Assets: ratio of bank net income to the average value of assets.
- Bank Liquidity
  - Liquid Assets/Short-Term Borrowing Ratio: ratio between liquid assets and short-term borrowing (BANKSCOPE).
- Credit Risk
  - Non-Performing Loans Ratio: ratio between non-performing loans and total loans (BANKSCOPE).
- Bank Balance Sheet Controls
  - Non Interest Income Ratio: ratio of non-interest income to total revenues (BANKSCOPE).
  - Return on Average Equity: ratio of bank net income to average equity.
  - Government Bonds Ratio: bank exposures to government bonds over total assets.
  - Loans to Total Assets Ratio: ratio of bank net loans to total loans.
  - Deposits to Total Funding Ratio: ratio of total deposits to total funding.
  - Growth of Total Assets: growth rate of total assets.
- Macro Variables
  - GDP Growth Rate: annual growth rate of real GDP (World Economic Outlook (IMF)).
  - CPI Inflation Rate: inflation rate based on the consumer price index.
- Market Entry Conditions (Banking Regulation and Supervision (World Bank))
  - Activity Restrictions: extent banks may engage in securities, insurance and real estate, range between 3 and 12.
  - Entry into Banking Requirements: range between 0 and 8.
  - Fraction of Denied Applications: percentage of applications to enter banking which are denied.
- Bank Regulation and Supervision (World Bank)
  - Capital Regulation: stringency of capital requirements, range between 0 and 10.
  - Supervisory Power: power of supervisory authorities, range between 0 and 15.
  - Independence of Supervisors: range between 0 and 3.
  - Fraction of Government Owned Bank Assets: percentage of system assets in government-controlled banks.
  - Deposit Insurance Scheme: equal to 0 in presence of insurance and to 1 in absence.

### Table 2: Summary Statistics (selected figures)
- BANK COMPETITION
  - Bank Lerner Index: Obs 1941, Mean .3257241, Std. Dev. .349005, Min -1.80814, Max 1.792211
  - Country Average Lerner Index: Obs 2706, Mean .3216213, Std. Dev. .2488896, Min -.5921693, Max 1.149032
  - Bank Market Share: Obs 3390, Mean .0917404, Std. Dev. .1346114, Min .000185, Max 1
- BANK SOLVENCY
  - Bank Z-Score: Obs 2543, Mean 9.731529, Std. Dev. 18.8839, Min -29.59916, Max 241.604
  - Country Average Z-Score: Obs 4677, Mean 1.753135, Std. Dev. .5518894, Min -1.167987, Max 3.238249
  - Tot Regulatory Capital Ratio: Obs 1744, Mean 22.04591, Std. Dev. 13.56645, Min 8.05, Max 92
  - Tier 1 Capital Ratio: Obs 1178, Mean 19.64578, Std. Dev. 12.31565, Min 7.2, Max 83
- BANK PROFITABILITY
  - Return on Average Equity: Obs 3374, Mean 11.05414, Std. Dev. 13.34129, Min -51.66, Max 50.04
  - Return on Average Assets: Obs 3375, Mean 1.495665, Std. Dev. 2.636938, Min -10.27, Max 13.2
  - Net Interest Margin: Obs 3332, Mean 3.613487, Std. Dev. 3.015037, Min -3.82, Max 18.05
- BANK CREDIT RISK
  - Non-Performing Loans Ratio: Obs 1739, Mean .1032941, Std. Dev. .1214676, Min .0012, Max .683
  - Loan Loss Provisions to Interest Revenues: Obs 2761, Mean .2716322, Std. Dev. .4864896, Min -.6169, Max 3.1066
- BANK LIQUIDITY
  - Deposit to Total Funding Ratio: Obs 3243, Mean .8104914, Std. Dev. .2161244, Min .0416, Max 1
  - Liquid Assets/Short-Term Borrowing Ratio: Obs 3311, Mean .4705926, Std. Dev. .4017674, Min .0364, Max 2.6928
- MACROECONOMIC VARIABLES
  - GDP Growth Rate: Obs 5653, Mean 4.838977, Std. Dev. 7.295436, Min -62.07599, Max 104.4833
  - CPI Inflation Rate: Obs 5634, Mean 5.62938, Std. Dev. 7.314547, Min -9.86305, Max 53.24779
- MARKET ENTRY
  - Activity Restrictions: Obs 3900, Mean 7.818462, Std. Dev. 1.526501, Min 3, Max 12
  - Entry into Banking Requirements: Obs 4496, Mean 7.800203, Std. Dev. .6680363, Min 4, Max 8
  - Fraction of Denied Applications: Obs 2548, Mean .2619675, Std. Dev. .3426484, Min 0, Max 1
- BANK REGULATION AND SUPERVISION
  - Capital Regulation Stringency: Obs 3720, Mean 7.030538, Std. Dev. 1.714338, Min 3, Max 10
  - Supervisory Power: Obs 4044, Mean 11.71893, Std. Dev. 2.305612, Min 5, Max 15
  - Supervisory Independence: Obs 3368, Mean 1.5962, Std. Dev. .7955234, Min 0, Max 3
  - Government Owned Bank Assets: Obs 2448, Mean 23.56516, Std. Dev. 29.17612, Min 0, Max 95.78
  - Deposit Insurance Coverage: Obs 1092, Mean .0628674, Std. Dev. .1769852, Min .002, Max .7
  - No Deposit Insurance: Obs 3692, Mean .5016251, Std. Dev. .5000651, Min 0, Max 1

### Regression Tables — Key statistically significant coefficients (selected highlights and exact estimates)
- Table I.A — Effects on Z-Score (bank fixed effects; standard errors in parentheses; significance: *** p<0.01, ** p<0.05, * p<0.1)
  - Lerner_1: 0.134*** (0.0466) in column (1).
  - Average Lerner_1: 0.293*** (0.0839) in column (2).
  - Market Share_1: 0.706** (0.347) in column (3).
  - Lerner_1 * Activity Restrictions: -0.0241** (0.0121) in column (4).
  - Lerner_1 * Entry Requirements: 0.0611*** (0.0235) in column (5).
  - Growth Assets_1: -0.125** (0.0580) in column (1).
  - Non-Performing Loans Ratio_1: -0.251** (0.125) in column (1).
- Table I.B — Interaction with regulation (bank fixed effects)
  - Lerner_1: 0.134*** (0.0466) in column (1); -0.508*** (0.129) in column (2); 0.313** (0.141) in column (3).
  - Lerner_1 * Capital Regulation: 0.0823*** (0.0154) in column (2).
  - Lerner_1 * Supervisory Independence: 0.102*** (0.0266) in column (4).
  - Lerner_1 * No Deposit Insurance: -0.125** (0.0571) in column (5).
  - Non-Performing Loans Ratio_1: -0.503*** (0.133) in column (6).
- Table I.C — Islamic banking interactions (GLS random-effect)
  - Lerner_1: 0.0851*** (0.0281) in column (1).
  - Lerner_1 * Islamic: -0.0222 (0.0921) in column (2).
  - Deposit Funding Ratio_1: 0.152** (0.0610) in column (1).
  - Non-Performing Loans Ratio_1: -0.362*** (0.0940) in column (1).
- Table II.A — Effects on Liquidity Ratio (bank fixed effects)
  - Lerner_1: -0.106*** (0.0210) in column (1).
  - Average Lerner_1: -0.299*** (0.0357) in column (2).
  - Loans Assets Ratio_1: -0.643*** (0.0591) in column (1).
  - Non-Performing Loans Ratio_1: 0.352*** (0.0599) in column (1).
  - Return on Average Equity_1: 0.205*** (0.0551) in column (1).
- Table II.B — Regulation interactions on liquidity
  - Lerner_1: -0.106*** (0.0210) in column (1); 0.0411 (0.0615) in column (2); -0.171** (0.0744) in column (3).
  - Lerner_1 * Capital Regulation: -0.0188** (0.00745) in column (2).
  - Lerner_1 * Supervisory Independence: 0.0394*** (0.0130) in column (4).
  - Lerner_1 * No Deposit Insurance: -0.0843*** (0.0319) in column (5).
  - Lerner_1 * Government Owned Banks: -0.00885*** (0.00117) in column (6).
- Table II.C — Islamic banking and liquidity (GLS random-effect)
  - Lerner_1: -0.0991*** (0.0203) in column (1).
  - Lerner_1 * Islamic: 0.138** (0.0705) in column (2).
  - Islamic: 13.55** (5.440) in column (2).
  - Loans Assets Ratio_1: -0.542*** (0.0520) in column (1).
  - Non-Performing Loans Ratio_1: 0.355*** (0.0585) in column (1).
- Table III.A — Effects on NPL Ratio (bank fixed effects)
  - Lerner_1: -0.0462*** (0.0134) in column (1).
  - Average Lerner_1: -0.127*** (0.0244) in column (2).
  - Market Share_1: 0.262** (0.102) in column (3).
  - Lerner_1 * Denied Applications: 0.110*** (0.0382) in column (6).
  - Deposit Funding Ratio_1: 0.121*** (0.0349) in column (1).
  - Return on Average Equity_1: -0.241*** (0.0333) in column (1).
  - GDP Growth_1: -0.252*** (0.0662) in column (1).
  - CPI Inflation_1: -0.377*** (0.0746) in column (1).
- Table III.B — Regulation interactions on credit risk
  - Lerner_1: -0.0462*** (0.0134) in column (1); 0.0807* (0.0433) in column (2); -0.151*** (0.0463) in column (3).
  - Lerner_1 * Capital Regulation: -0.0153*** (0.00498) in column (2).
  - Lerner_1 * Supervisory Power: 0.00944** (0.00391) in column (3).
  - Lerner_1 * No Deposit Insurance: -0.0377* (0.0204) in column (5).
  - Return on Average Equity_1: -0.276*** (0.0363) in column (2).
- Table III.C — Islamic banking and credit risk (GLS random-effect)
  - Lerner_1: -0.0478*** (0.0119) in column (1).
  - Lerner_1 * Islamic: -0.00745 (0.0457) in column (2).
  - Islamic: -3.139 (2.608) in column (2); -6.580* (3.996) in column (4).
  - Deposit Funding Ratio_1: 0.104*** (0.0278) in column (1).
  - Return on Average Equity_1: -0.259*** (0.0312) in column (1).
  - GDP Growth_1: -0.251*** (0.0646) in column (1).
  - CPI Inflation_1: -0.319*** (0.0683) in column (1).

### Notes on estimation and inference
- Bank Fixed Effects: many regressions report Bank Fixed Effects = YES.
- Significance notation: *** p<0.01, ** p<0.05, * p<0.1.
- For regressions involving the Islamic dummy, a GLS random-effect specification is used because the Islamic dummy is time-invariant for a given bank and may be equivalent to bank fixed effects.

*Source: _wp15210 - REFERENCES (IMF PDF)._

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2015/_wp15210.pdf_