## 2.1  Bank Ownership and Overall Lending

## Source details

**Canonical URL:** [2.1  Bank Ownership and Overall Lending](https://www.imf.org/-/media/files/publications/wp/2022/english/wpiea2022188-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2022/english/wpiea2022188-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2022/english/wpiea2022188-print-pdf.pdf.json)

---

### Methodology and Empirical Specifications
- Baseline fixed effects specification (annual data 2015-2020):
  - ∆Loans_i,c,t = β1 Foreign_i,t + β2 Covid_c,t + β3 Covid_c,t * Foreign_i,t + δ X_i,t−1 + α_c + u_i,c,t
  - Loans_i,c,t: growth rate of total gross loans for bank i in jurisdiction c at period t.
  - Foreign_i,t: dummy = 1 if bank is foreign owned (>50 percent asset share held by foreign entities), 0 otherwise.
  - Covid_c,t: dummy = 1 for 2020, 0 otherwise.
  - X_i,t−1: one period lagged bank characteristics (log of assets; capital = ratio of bank equity to total assets; liquidity = ratio of liquid assets to total assets; profitability = return on assets; funding structure = customer deposits as share of total liabilities).
  - α_c: country fixed effects; additional specifications include year and bank fixed effects.

- Cross-sectional specification for 2020 using mobility as severity proxy:
  - ∆Loans_i,c = β1 Foreign_i + β2 HighMobilityChange_c + β3 Foreign_i * HighMobilityChange_c + δ X_i,2019 + u_i,c
  - HighMobilityChange_c: dummy = 1 if mobility change to retail and recreational venues in jurisdiction > median, 0 otherwise.
  - Mobility change proxied by changes in visits to retail and recreational venues (Google COVID-19 Community Mobility Reports).

- Cross-sectional specifications for loan portfolio shares in 2020:
  - ∆ConsumerShare_i,c = β1 Foreign_i + β2 HighMobilityChange_c + β3 Foreign_i * HighMobilityChange_c + δ X_i,2019 + u_i,c
  - ∆FirmShare_i,c = β1 Foreign_i + β2 HighMobilityChange_c + β3 Foreign_i * HighMobilityChange_c + δ X_i,2019 + u_i,c
  - ConsumerShare_i,c: share of consumer loans in total loans for bank i in jurisdiction c in 2020.
  - FirmShare_i,c: share of corporate loans in total loans for bank i in jurisdiction c in 2020.

- Policy interaction specifications for 2020:
  - Equation (5) (number of policy interventions):
    - ∆Loans_i,c = β1 Foreign_i + β2 HighNo.ofPolicies_c + β3 HighMobilityChange_c + β4 Foreign_i * HighNo.ofPolicies_c + β5 HighNo.ofPolicies_c * HighMobilityChange_c + β6 Foreign_i * HighMobilityChange_c + β7 Foreign_i * HighNo.ofPolicies_c * HighMobilityChange_c + δ X_i,2019 + u_i,c
    - HighNo.ofPolicies_c: dummy = 1 if number of policy interventions (monetary, fiscal, prudential, debt moratoria) > median, 0 otherwise.
  - Equation (6) (prudential measures):
    - ∆Loans_i,c = β1 Foreign_i + β2 PrudentialMeasures_c + β3 HighMobilityChange_c + β4 Foreign_i * PrudentialMeasures_c + β5 PrudentialMeasures_c * HighMobilityChange_c + β6 Foreign_i * HighMobilityChange_c + β7 Foreign_i * PrudentialMeasures_c * HighMobilityChange_c + δ X_i,2019 + u_i,c
    - PrudentialMeasures_c: dummy = 1 if jurisdiction introduced prudential policies related to capital or liquidity requirements, 0 otherwise.
  - Triple interaction terms capture whether policy interventions in areas more severely affected by COVID-19 had heterogeneous effects by bank ownership.

### Data
- Bank-level data source: Fitch Connect (ownership and gross loans 2015-2020; bank size, capitalization, liquidity, profitability, funding structure, consumer and firm loans).
- Sample: 1,482 banks across 62 jurisdictions.
- Foreign bank presence in sample:
  - Emerging markets: 32 percent
  - Advanced economies: 8 percent
- Mobility data: Google COVID-19 Community Mobility Reports (change in visits to retail and recreational venues).
- Policy announcements dataset: Kirti et al. (2022) (monetary, fiscal, prudential, others such as debt moratoria). Prudential measures in sample primarily involved capital-related requirements.

### Key Results — Impact of Bank Ownership on Overall Lending
- Baseline fixed effects estimates (equation (1), Table 4):
  - Credit growth during the pandemic was almost 6 percentage points lower for foreign banks relative to domestic banks.
- Heterogeneity by severity of mobility restrictions (equation (2), Table 5):
  - Accounting for mobility changes, foreign banks issued credit at a higher rate than domestic banks in jurisdictions with more stringent mobility restrictions.
  - Robust to alternate thresholds for foreign bank ownership and to different cutoff definitions for HighMobilityChange (median and 70th percentile).
- Income-group heterogeneity (Table 6):
  - Advanced economies: foreign banks had lower rates of credit growth relative to domestic banks irrespective of mobility restriction degree.
  - Emerging markets: foreign banks, while issuing credit at a lower rate overall, acted as shock absorbers in jurisdictions with more stringent mobility restrictions by maintaining more stable rates of lending.
- Cross-border heterogeneity (Table 7):
  - Foreign banks lent more than domestic banks when the host jurisdiction had more stringent mobility restrictions relative to the home jurisdiction — consistent with a shock-absorbing role.

### Key Results — Impact on Loan Portfolios
- Loan portfolio composition in 2020 (equations (3)-(4), Tables 8 and 9):
  - In jurisdictions with high mobility changes, foreign banks had:
    - A lower share of consumer loans in total loans relative to domestic banks.
    - A larger share of firm (corporate) credit in total loans relative to domestic banks.
  - Interpretation: In areas with more stringent lockdowns, firms were more likely to demand credit and draw down lines of credit; foreign banks’ larger exposure to the corporate sector helps explain higher lending rates in those jurisdictions.

### Key Results — Role of Policies
- Policy types analyzed: monetary, fiscal, prudential, debt moratoria; among prudential policies, capital and liquidity measures examined (equations (5)-(6), Tables 10 and 11).
- Main patterns:
  - Foreign bank lending rates were lower than domestic banks in jurisdictions with a larger number of policy interventions (baseline).
  - In jurisdictions with both high mobility restrictions and more extensive policy and regulatory support, foreign banks lent at a higher rate than locally owned banks.
    - Effect particularly strong for monetary policies and prudential stimulus.
    - Fiscal policies and debt moratoria had more muted impacts.
  - Prudential measures:
    - Foreign banks lent more than domestic banks in jurisdictions most affected by lockdowns and where capital requirements were relaxed.
    - Lending rates were not significantly different between foreign and domestic banks in jurisdictions with lower liquidity requirements.
  - Implication: Differences in capital structures of foreign and domestic banks likely played a role in heterogeneous lending responses during the COVID-19 crisis.

### Robustness Checks
- Robustness exercises included:
  - Replacing HighMobilityChange dummy with actual values of mobility changes in equations (2)-(4).
  - Replacing HighNo.ofPolicies dummy with the actual number of policy measures in equation (5).
  - Replacing prudential presence dummy with the actual number of prudential policies (capital and liquidity) in equation (6).
- All results remain broadly unchanged.

---

### 5.1  Actual Changes in Mobility

### Effect of mobility changes on credit growth (Table 12)
- Coefficients and statistics (actual values):
  - Foreign = −15.652 ∗∗∗ (column 1)
  - Foreign = −11.655 ∗∗∗ (column 2)
  - Change in Mobility = −0.378 ∗∗∗ (column 1)
  - Change in Mobility = −0.253 ∗∗∗ (column 2)
  - Foreign*Change in Mobility = 0.382 ∗∗∗ (column 1)
  - Foreign*Change in Mobility = 0.297 ∗∗∗ (column 2)
  - Bank Size = −0.537 ∗∗∗ (reported)
  - Capital = −35.656 ∗∗ (reported)
  - Profitability = 77.947 ∗∗ (reported)
  - Funding Structure = 5.283 ∗ (reported)
  - Liquidity = −0.120 ∗∗ (reported)
  - Observations = 1,409 (column 1) and 1,327 (column 2)
  - Adj. R-squared = 0.05 (column 1) and 0.08 (column 2)

- Interpretation:
  - The positive coefficient on Foreign*Change in Mobility indicates foreign banks increased lending rates in response to a rise in mobility restrictions.
  - In jurisdictions with stringent lockdown measures (i.e., above the median), foreign banks were likely to have provided more stable access to credit relative to domestic banks.

### Effect of mobility changes on credit allocation (consumer vs firm)
- Consumer lending (Table 13; Share of Consumer Loans in Total Loans; actual values):
  - Foreign = 8.589 (column 1) and 11.921 ∗ (column 2)
  - Change in Mobility = 0.187 ∗ (column 1) and 0.230 ∗ (column 2)
  - Foreign*Change in Mobility = −0.467 ∗∗ (column 1) and −0.537 ∗∗ (column 2)
  - Capital = −62.856 ∗∗∗ (reported)
  - Liquidity = −0.208 ∗∗ (reported)
  - Observations = 1,220 (column 1) and 1,186 (column 2)
  - Adj. R-squared = 0.004 (column 1) and 0.03 (column 2)

- Firm lending (Table 14; Share of Firm Loans in Total Loans; actual values):
  - Foreign = −16.285 ∗∗∗ (column 1) and −24.265 ∗∗∗ (column 2)
  - Change in Mobility = −0.400 ∗∗∗ (column 1) and −0.546 ∗∗∗ (column 2)
  - Foreign*Change in Mobility = 0.470 ∗∗ (column 1) and 0.742 ∗∗ (column 2)
  - Capital = 86.661 ∗∗∗ (reported)
  - Observations = 1,208 (column 1) and 1,168 (column 2)
  - Adj. R-squared = 0.02 (column 1) and 0.06 (column 2)

- Interpretation:
  - As mobility restrictions increased, foreign banks tended to reduce the share of consumer credit in their loan portfolios and shift exposure towards firm credit.

### Interaction with policy measures and mobility (Tables 15–16)
- Aggregate patterns:
  - A higher number of policy interventions did not benefit foreign banks overall.
  - For foreign banks in jurisdictions with stringent mobility restrictions, an increase in the number of policy interventions had a positive effect on their lending rate.
  - This positive policy effect holds for both monetary and prudential policies; fiscal policies and debt moratoria had more muted impacts.
  - Among prudential policies, higher usage of capital measures increased foreign bank lending rates in jurisdictions with high mobility restrictions.

- Selected numeric results (actual number of policies/measures):
  - Table 15 (Growth of Gross Loans; examples):
    - Foreign*High Change in Mobility*High No. of Policies = 14.782 ∗∗∗ (Monetary column)
    - High No. of Policies = 6.818 ∗∗ (Monetary column)
    - Observations = 1,269 (all columns)
    - Adj. R-squared ≈ 0.07–0.09 (by column)
  - Table 16 (Growth of Gross Loans; capital and liquidity measures actual number):
    - Foreign = −6.546 ∗∗∗ (Liquidity column) and −2.240 (Capital column)
    - No. of Prudential Measures = −2.888 ∗∗∗ (Liquidity column) and 0.230 (Capital column)
    - Foreign*No. of Prudential Measures*High Change in Mobility = 1.037 (Liquidity column) and 1.727 ∗ (Capital column)
    - Observations = 1,269 (both columns)
    - Adj. R-squared = 0.08 (Liquidity) and 0.07 (Capital)

### Key Analytical Conclusions (relevant to mobility findings)
- Overall findings:
  - Foreign banks had lower rates of credit growth relative to domestic banks during the COVID-19 pandemic (robust to controls and fixed effects).
  - Jurisdictions with higher foreign bank presence may have faced a more severe credit crunch at an overall level.

- Nuanced findings conditional on mobility and demand:
  - Conditioning on severity of pandemic effects on economic activity (mobility and lockdowns) reverses the baseline result: in jurisdictions with more stringent lockdown measures, foreign banks had higher growth rates in credit compared to domestic banks.
  - In such jurisdictions, foreign banks had larger exposure to the corporate sector; businesses facing tougher lockdowns demanded more credit, which may explain higher foreign bank lending rates there.
  - Foreign banks acted as shock absorbers in jurisdictions hit hardest by the pandemic, ensuring steady supply of liquidity.

- Role of timely policy intervention:
  - Timely and extensive monetary and prudential interventions, particularly changes to capital requirements, enhanced the shock-absorbing role of foreign banks in jurisdictions with high mobility restrictions.
  - It is possible foreign banks in these regions had larger capital buffers and thus benefited more from relaxation of prudential requirements.

### Caveat and Further Research
- Data limitation:
  - Unable to fully separate demand-side factors from supply-side ones.
  - Mobility change data partially account for demand conditions.
  - Further research using bank-borrower matched data is needed to better understand bank ownership effects on lending decisions.

*Source: Shock Absorbers or Transmitters? The Role of Foreign Banks during COVID -19, Working Paper No. WP/2022/188 (sections 2.1 and 5.1 summarized from the provided content).*

### 2.1  Bank Ownership and Overall Lending

### 2.1 Bank Ownership and Overall Lending

### Methodology and Empirical Specifications
- Baseline fixed effects specification (annual data 2015-2020):
  - ∆Loans_i,c,t = β1 Foreign_i,t + β2 Covid_c,t + β3 Covid_c,t * Foreign_i,t + δ X_i,t−1 + α_c + u_i,c,t
  - Loans_i,c,t: growth rate of total gross loans for bank i in jurisdiction c at period t.
  - Foreign_i,t: dummy = 1 if bank is foreign owned (>50 percent asset share held by foreign entities), 0 otherwise.
  - Covid_c,t: dummy = 1 for 2020, 0 otherwise.
  - X_i,t−1: one period lagged bank characteristics (log of assets; capital = ratio of bank equity to total assets; liquidity = ratio of liquid assets to total assets; profitability = return on assets; funding structure = customer deposits as share of total liabilities).
  - α_c: country fixed effects; additional specifications include year and bank fixed effects.

- Cross-sectional specification for 2020 using mobility as severity proxy:
  - ∆Loans_i,c = β1 Foreign_i + β2 HighMobilityChange_c + β3 Foreign_i * HighMobilityChange_c + δ X_i,2019 + u_i,c
  - HighMobilityChange_c: dummy = 1 if mobility change to retail and recreational venues in jurisdiction > median, 0 otherwise.
  - Mobility change proxied by changes in visits to retail and recreational venues (Google COVID-19 Community Mobility Reports).

- Cross-sectional specifications for loan portfolio shares in 2020:
  - ∆ConsumerShare_i,c = β1 Foreign_i + β2 HighMobilityChange_c + β3 Foreign_i * HighMobilityChange_c + δ X_i,2019 + u_i,c
  - ∆FirmShare_i,c = β1 Foreign_i + β2 HighMobilityChange_c + β3 Foreign_i * HighMobilityChange_c + δ X_i,2019 + u_i,c
  - ConsumerShare_i,c: share of consumer loans in total loans for bank i in jurisdiction c in 2020.
  - FirmShare_i,c: share of corporate loans in total loans for bank i in jurisdiction c in 2020.

- Policy interaction specifications for 2020:
  - Equation (5) studies number of policy interventions:
    - ∆Loans_i,c = β1 Foreign_i + β2 HighNo.ofPolicies_c + β3 HighMobilityChange_c + β4 Foreign_i * HighNo.ofPolicies_c + β5 HighNo.ofPolicies_c * HighMobilityChange_c + β6 Foreign_i * HighMobilityChange_c + β7 Foreign_i * HighNo.ofPolicies_c * HighMobilityChange_c + δ X_i,2019 + u_i,c
    - HighNo.ofPolicies_c: dummy = 1 if number of policy interventions (monetary, fiscal, prudential, debt moratoria) > median, 0 otherwise.
  - Equation (6) studies prudential measures:
    - ∆Loans_i,c = β1 Foreign_i + β2 PrudentialMeasures_c + β3 HighMobilityChange_c + β4 Foreign_i * PrudentialMeasures_c + β5 PrudentialMeasures_c * HighMobilityChange_c + β6 Foreign_i * HighMobilityChange_c + β7 Foreign_i * PrudentialMeasures_c * HighMobilityChange_c + δ X_i,2019 + u_i,c
    - PrudentialMeasures_c: dummy = 1 if jurisdiction introduced prudential policies related to capital or liquidity requirements, 0 otherwise.
  - Triple interaction terms capture whether policy interventions in areas more severely affected by COVID-19 had heterogeneous effects by bank ownership.

### Data
- Bank-level data source: Fitch Connect (ownership and gross loans 2015-2020; bank size, capitalization, liquidity, profitability, funding structure, consumer and firm loans).
- Sample: 1,482 banks across 62 jurisdictions.
- Foreign bank presence in sample:
  - Emerging markets: 32 percent
  - Advanced economies: 8 percent
- Mobility data: Google COVID-19 Community Mobility Reports (change in visits to retail and recreational venues).
- Policy announcements dataset: Kirti et al. (2022) (monetary, fiscal, prudential, others such as debt moratoria). Prudential measures in sample primarily involved capital-related requirements.

### Key Results — Impact of Bank Ownership on Overall Lending
- Baseline fixed effects estimates (equation (1), Table 4):
  - Credit growth during the pandemic was almost 6 percentage points lower for foreign banks relative to domestic banks.
- Heterogeneity by severity of mobility restrictions (equation (2), Table 5):
  - Accounting for mobility changes, foreign banks issued credit at a higher rate than domestic banks in jurisdictions with more stringent mobility restrictions.
  - Robust to alternate thresholds for foreign bank ownership and to different cutoff definitions for HighMobilityChange (median and 70th percentile).
- Income-group heterogeneity (Table 6):
  - Advanced economies: foreign banks had lower rates of credit growth relative to domestic banks irrespective of mobility restriction degree.
  - Emerging markets: foreign banks, while issuing credit at a lower rate overall, acted as shock absorbers in jurisdictions with more stringent mobility restrictions by maintaining more stable rates of lending.
- Cross-border heterogeneity (Table 7):
  - Foreign banks lent more than domestic banks when the host jurisdiction had more stringent mobility restrictions relative to the home jurisdiction — consistent with a shock-absorbing role.

### Key Results — Impact on Loan Portfolios
- Loan portfolio composition in 2020 (equations (3)-(4), Tables 8 and 9):
  - In jurisdictions with high mobility changes, foreign banks had:
    - A lower share of consumer loans in total loans relative to domestic banks.
    - A larger share of firm (corporate) credit in total loans relative to domestic banks.
  - Interpretation: In areas with more stringent lockdowns, firms were more likely to demand credit and draw down lines of credit; foreign banks’ larger exposure to the corporate sector helps explain higher lending rates in those jurisdictions.

### Key Results — Role of Policies
- Policy types analyzed: monetary, fiscal, prudential, debt moratoria; among prudential policies, capital and liquidity measures examined (equations (5)-(6), Tables 10 and 11).
- Main patterns:
  - Foreign bank lending rates were lower than domestic banks in jurisdictions with a larger number of policy interventions (baseline).
  - In jurisdictions with both high mobility restrictions and more extensive policy and regulatory support, foreign banks lent at a higher rate than locally owned banks.
    - Effect particularly strong for monetary policies and prudential stimulus.
    - Fiscal policies and debt moratoria had more muted impacts.
  - Prudential measures:
    - Foreign banks lent more than domestic banks in jurisdictions most affected by lockdowns and where capital requirements were relaxed.
    - Lending rates were not significantly different between foreign and domestic banks in jurisdictions with lower liquidity requirements.
  - Implication: Differences in capital structures of foreign and domestic banks likely played a role in heterogeneous lending responses during the COVID-19 crisis.

### Robustness Checks
- Robustness exercises included:
  - Replacing HighMobilityChange dummy with actual values of mobility changes in equations (2)-(4).
  - Replacing HighNo.ofPolicies dummy with the actual number of policy measures in equation (5).
  - Replacing prudential presence dummy with the actual number of prudential policies (capital and liquidity) in equation (6).
- All results remain broadly unchanged.

*Source: IMF staff analysis (chapter 2.1–3, data and results summarized from the provided content).*

### 5.1  Actual Changes in Mobility

### 5.1  Actual Changes in Mobility

### Effect of mobility changes on credit growth
- Table 12 reports results for Growth of Gross Loans after accounting for mobility (actual values):
  - Foreign = −15.652 ∗∗∗ (column 1)
  - Foreign = −11.655 ∗∗∗ (column 2)
  - Change in Mobility = −0.378 ∗∗∗ (column 1)
  - Change in Mobility = −0.253 ∗∗∗ (column 2)
  - Foreign*Change in Mobility = 0.382 ∗∗∗ (column 1)
  - Foreign*Change in Mobility = 0.297 ∗∗∗ (column 2)
  - Bank Size = −0.537 ∗∗∗ (reported)
  - Capital = −35.656 ∗∗ (reported)
  - Profitability = 77.947 ∗∗ (reported)
  - Funding Structure = 5.283 ∗ (reported)
  - Liquidity = −0.120 ∗∗ (reported)
  - Observations = 1,409 (column 1) and 1,327 (column 2)
  - Adj. R-squared = 0.05 (column 1) and 0.08 (column 2)

- Interpretation from text:
  - The coefficient of the interaction term between foreign ownership and change in mobility is positive, indicating foreign banks increased lending rates in response to a rise in mobility restrictions.
  - In jurisdictions with stringent lockdown measures (i.e., above the median), foreign banks were likely to have provided more stable access to credit relative to domestic banks.

### Effect of mobility changes on credit allocation (consumers vs firms)
- Consumer lending (Table 13, actual values; Share of Consumer Loans in Total Loans):
  - Foreign = 8.589 (column 1) and 11.921 ∗ (column 2)
  - Change in Mobility = 0.187 ∗ (column 1) and 0.230 ∗ (column 2)
  - Foreign*Change in Mobility = −0.467 ∗∗ (column 1) and −0.537 ∗∗ (column 2)
  - Capital = −62.856 ∗∗∗ (reported)
  - Liquidity = −0.208 ∗∗ (reported)
  - Observations = 1,220 (column 1) and 1,186 (column 2)
  - Adj. R-squared = 0.004 (column 1) and 0.03 (column 2)

- Firm lending (Table 14, actual values; Share of Firm Loans in Total Loans):
  - Foreign = −16.285 ∗∗∗ (column 1) and −24.265 ∗∗∗ (column 2)
  - Change in Mobility = −0.400 ∗∗∗ (column 1) and −0.546 ∗∗∗ (column 2)
  - Foreign*Change in Mobility = 0.470 ∗∗ (column 1) and 0.742 ∗∗ (column 2)
  - Capital = 86.661 ∗∗∗ (reported)
  - Observations = 1,208 (column 1) and 1,168 (column 2)
  - Adj. R-squared = 0.02 (column 1) and 0.06 (column 2)

- Interpretation from text:
  - As mobility restrictions increased, foreign banks tended to reduce the share of consumer credit in their loan portfolios and shift exposure towards firm credit.

### Interaction with policy measures and mobility
- Summary from text and tables 15–16:
  - At an overall level, a higher number of policy interventions did not benefit foreign banks.
  - For foreign banks based in jurisdictions with stringent mobility restrictions, an increase in the number of policy interventions had a positive effect on their lending rate.
  - This holds for both monetary and prudential policies.
  - In jurisdictions with the most stringent mobility restrictions and extensive use of policy/regulatory support, foreign bank lending rates were higher than domestic banks.
  - Among prudential policies, higher usage of capital measures led to an increase in lending rates of foreign banks in jurisdictions with high mobility restrictions.

- Selected numeric results (Tables 15 and 16, actual number of policies/measures):
  - Table 15 (Growth of Gross Loans; examples):
    - Foreign*High Change in Mobility*High No. of Policies = 14.782 ∗∗∗ (Monetary column)
    - High No. of Policies = 6.818 ∗∗ (Monetary column)
    - Observations = 1,269 (all columns)
    - Adj. R-squared ≈ 0.07–0.09 (by column)
  - Table 16 (Growth of Gross Loans; capital and liquidity measures actual number):
    - Foreign = −6.546 ∗∗∗ (Liquidity column) and −2.240 (Capital column)
    - No. of Prudential Measures = −2.888 ∗∗∗ (Liquidity column) and 0.230 (Capital column)
    - Foreign*No. of Prudential Measures*High Change in Mobility = 1.037 (Liquidity column) and 1.727 ∗ (Capital column)
    - Observations = 1,269 (both columns)
    - Adj. R-squared = 0.08 (Liquidity) and 0.07 (Capital)

### Key analytical conclusions (from section 6 Conclusion, as relevant to mobility findings)
- Overall findings:
  - Foreign banks had lower rates of credit growth relative to domestic banks during the COVID-19 pandemic (robust to controls and fixed effects).
  - Jurisdictions with higher foreign bank presence may have faced a more severe credit crunch at an overall level, consistent with previous findings.

- Nuanced findings conditional on mobility and demand:
  - Conditioning on severity of pandemic effects on economic activity (mobility and lockdowns) reverses the baseline result: in jurisdictions with more stringent lockdown measures, foreign banks had higher growth rates in credit compared to domestic banks.
  - In such jurisdictions, foreign banks had larger exposure to the corporate sector; businesses facing tougher lockdowns demanded more credit, which may explain higher foreign bank lending rates there.
  - Foreign banks acted as shock absorbers in jurisdictions hit hardest by the pandemic, ensuring steady supply of liquidity.

- Role of timely policy intervention:
  - Timely and extensive monetary and prudential interventions, particularly changes to capital requirements, enhanced the shock-absorbing role of foreign banks in jurisdictions with high mobility restrictions.
  - It is possible foreign banks in these regions had larger capital buffers and thus benefited more from relaxation of prudential requirements.

### Caveat and further research
- Data limitation noted:
  - Unable to fully separate demand-side factors from supply-side ones.
  - Mobility change data partially account for demand conditions.
  - Further research using bank-borrower matched data is needed to better understand bank ownership effects on lending decisions.

*Source: Shock Absorbers or Transmitters? The Role of Foreign Banks during COVID -19, Working Paper No. WP/2022/188 (sections 5.1, 5.2 and Conclusion).*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2022/english/wpiea2022188-print-pdf.pdf_