## _cr1665

## Source details

**Canonical URL:** [_cr1665](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2016/_cr1665.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2016/_cr1665.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2016/_cr1665.pdf.json)

---

### Executive summary and context
- Argentina’s financial system is very small, bank-dominated, and mostly transactional; it has ample liquidity, solid capitalization, and strong asset quality despite limited depth.
- FSAP stress testing objectives:
  - assess capacity to withstand extreme but plausible macroeconomic shocks;
  - identify transmission channels, weaknesses, and policy priorities.
- Modeling and horizon:
  - Tests use macroeconomic and satellite models estimated on historical data; model uncertainty is potentially severe given institutional and structural changes over the last two decades.
  - Adverse external shocks assessed over a two-year horizon (2013–14) based on data available through September 2012.
- Coverage and approach:
  - 22 largest banks covering 90 percent of the system by assets.
  - Top-down macroeconomic scenarios and sensitivity analysis; satellite models map macro shocks to bank profitability and capitalization.

### Scenarios and macroeconomic projections
- Adverse calibration:
  - Adverse external shocks calibrated to trigger a cumulative decline of real GDP growth equivalent to 13.3 percentage points, or 2 standard deviations.
  - One standard deviation of two-year cumulative real GDP growth (2001–2012) = 6.6 percentage points.
- Macroeconomic projected paths (selected entries, real GDP growth in percent):
  - Baseline Authorities: 2012 1.9; 2013 4.6; 2014 4.0
  - Adverse Authorities: 2012 1.9; 2013 -0.8; 2014 -1.4
  - Baseline WEO1: 2012 1.9; 2013 2.8; 2014 3.5
  - Adverse (U): 2012 1.9; 2013 -3.4; 2014 -3.6
  - Adverse (V): 2012 1.9; 2013 -6.9; 2014 0.0
- Other macro variables (selected):
  - Nominal exchange rate, peso per US dollar (end-of-year) Adverse (V): 2013 26.4; 2014 12.0
  - Nominal annual interest rate (BADLAR, percent) Adverse (U): 2013 20.6; 2014 17.3
- Sensitivity scenarios:
  - 900 basis point increase in real interest rates.
  - 30 percent peso depreciation with 0.2 pass-through to domestic prices.
  - Failure of the five largest borrowers (concentration).
  - Reverse liquidity test with 100 percent run-off on maturing wholesale deposits.

### Solvency and credit risk findings
- Credit risk is the primary vulnerability:
  - Loan loss ratio in the banking system rises from 1.5 percent to 4.4 percent (U-shaped) and 6.2 percent (V-shaped) by 2014 in Fund-staff adverse scenarios.
- Capitalization outcomes:
  - Two banks are undercapitalized at the starting point.
  - In the U-shaped adverse scenario, 4 of the 22 largest banks would fall below the 8.0 percent total capital minimum.
  - In the V-shaped adverse scenario, 6 of the 22 largest banks would be undercapitalized.
  - In the most extreme adverse scenario, capitalization in 6 of the 22 largest banks would fall below the required minimum of 8.0 percent.
  - Estimated system-wide capital shortfall in the V-shaped adverse scenario is about 0.2 percent of GDP (noted elsewhere as less than 0.2 percent of GDP).
- Sensitivity shock outcomes:
  - 900 basis point increase in real interest rates → 5 of the 22 largest banks undercapitalized.
  - 30 percent peso depreciation (0.2 pass-through) → NPL ratios up about ½ percentage point on average; with no pre-impairment profits, capitalization of five banks could fall below the required minimum.
  - Failure of the five largest borrowers → undercapitalization in 8 of the 22 banks; failure of the 10 largest counterparts → undercapitalization of 12 banks.

### Market, sovereign, and exchange rate risk
- Market risk:
  - Exposure to corporate bonds, equity, commodities, foreign securities is negligible.
  - Banks hold highly liquid bonds and BCRA money market instruments (LEBACs, NOBACs).
- Sovereign risk:
  - Adverse scenarios produce higher interest rates and inverted yield curves causing haircut-related losses on sovereign securities.
  - Simulated yield increases (V-shaped size):
    - U.S. dollar denominated bonds: about 1,500 basis points average yield increase.
    - Peso denominated bonds adjusted by inflation: about 2,000 basis points.
    - Bonds linked to BADLAR: about 2,800 basis points.
    - Peso-denominated nominal instruments (fixed coupon and discount money market): about 450 basis points.
  - Impact on CAR: increase in sovereign yields lowers system CAR by 1.3 percentage points, from 11.7 percent to 10.4 percent; bank-level impacts range from 0.1 to 3.2 percentage points of CAR.
- Exchange rate risk:
  - Banks have positive net open foreign currency exposures; peso depreciation in adverse scenarios tends to boost bank profitability.
  - Market-effect of a 30 percent depreciation: system CAR would increase by 1.4 percentage points, from 11.7 percent to 13.1 percent; 21 of 22 banks obtain benefits ranging from 0.4 to 5.2 percentage points of CAR; only one bank would suffer a loss.

### Liquidity stress test results
- Funding composition (excluding own capital):
  - 65 percent funding from legal entities (non-financial public sector including FGS 27 percent; SMEs 17 percent; large enterprises 17 percent; others stated).
  - 32 percent funding due with individual depositors (physical persons).
  - 3 percent other liabilities/own issuances/secured funding.
- Run-off calibration based on 1995 and 2001 crises with detailed run-off rates by counterparty and maturity (examples preserved in source tables).
- Short-run resilience:
  - All banks could confront persistent and sizable deposit withdrawals for 30 days without BCRA assistance.
  - BCRA tools assumed: waive reserve requirements for up to 30 days; inject liquidity through standing facilities.
- After 30 days:
  - Only 2 of the 22 largest banks would need BCRA assistance in pesos and 1 in dollars; extension of reserve requirement waiver would suffice.
  - No bank would need emergency liquidity assistance from the BCRA for two years, besides access to own minimum required reserves (under assumptions).
- Reverse liquidity test (wholesale deposits):
  - Assuming 100 percent run-off on maturing wholesale deposits, all banks have liquidity to absorb losses of 33 percent or more of wholesale deposits without BCRA assistance.
  - 15 of the 22 largest banks could confront 100 percent withdrawals of maturing wholesale deposits without a shortfall of liquid assets at any time during the two-year assessment period.

### Contagion, interconnectedness, and FGS risks
- Interbank contagion:
  - Direct bilateral interbank contagion is limited: interbank exposures are very small relative to capitalization.
  - As of September 2012, only one of the 22 large banks had total interbank exposure larger than its excess capital over the required minimum; for that bank, five other institutions would have to fail for its capital to fall below the minimum.
- Sustainability Guarantee Fund (FGS) risks:
  - FGS accounts for 20 percent of the financial system and about 11 percent of GDP.
  - About four-fifths of FGS investment portfolio concentrated in fixed-term instruments with a duration of five years on average.
  - Simple estimate: for every 100 bps increase in the interest rate, the economic value of FGS investment portfolio would decline by 3.4 percent.
  - FGS has 6 percent of its portfolio in fixed-term deposits with very short average maturities of 35–40 days (another section cites average maturities of about 45 days for fixed-term deposits).
  - For some banks, FGS deposits represent more than 4 percent of total deposits.
  - Allocation mechanism lacks transparent criteria; many deposits (as of September 2012) were invested below market prices in certain banks where FGS has significant stakes.
  - Risks: sudden withdrawal of FGS deposits may generate liquidity pressures for smaller banks; large FGS equity stakes in private banks raise governance concerns.

### Insurance sector vulnerabilities and pension observations
- Profitability and solvency:
  - Nominal ROE rose from 8 percent in 2006 to 22 percent in 2012.
  - Under current solvency regime: 4 companies have solvency ratios below 100 percent and 57 have ratios below 125 percent.
  - Industry actual solvency margin has amounted to 175 percent of the minimum solvency margin for the past several years.
- Asset quality and liquidity:
  - Assets bearing significant credit risk (unpaid premia and other receivables as reported in 2012) amount to 100 percent of free capital available to the sector (ARG$18.34 billion).
  - A large percentage of available capital is illiquid and not fully suitable to protect companies in adverse events.
- Non-life sector:
  - Weaker than life or retirement sectors; credit and liquidity risks.
  - Among 20 largest non-life firms in 2012: 7 had ROEs >25 percent; 9 had ROEs <15 percent.
  - Outstanding premia equivalent to about one-fifth of total assets and one-third of total annual nonlife premia.
  - Liquid assets for 20 largest nonlife insurers in 2012: 106 percent of claims paid on average; 6 firms had liquid assets below 75 percent of claims paid.
- Policy implication:
  - Recommendation to lift recently introduced mandatory investment requirements which risk accentuating vulnerabilities; allow risk-transfer reinsurance.
  - New regulation disallowing more than two months of unpaid premia as admissible assets would reduce credit risk.
  - Recommendation: adapt investment guidelines to avoid significant maturity mismatches between assets and liabilities.

### Satellite models, methodology, and modeling recommendations
- Satellite models used to map macro shocks to bank profits and capital cover:
  - Credit losses (dynamic panel models), net interest income (maturity ladder and projected rates), net fee and service income (AR(1) bank-level), operating and administrative expenses, sovereign yield curve and valuation effects.
- Credit risk model (Fund-staff baseline):
  - Dynamic panel logistic transformation of loan loss rate with explanatory variables including lagged LLR, real GDP growth and its square, (ex-post) real interest rate rr, growth of GDP deflator π, capital-assets ratio CR and its polynomials, bank fixed effects μi.
  - Estimation: unbalanced panel of 193 institutions, annual observations 1994–2011; two-step Arellano-Bond.
  - Key estimated coefficients (selected): Loan loss rate (lagged) 0.19; Real GDP growth -9.77; Real GDP growth squared 73.70; Real interest rate 1.52; Growth of GDP deflator 1.86; Capital to assets ratio (lagged) -2.89 (with squared and cubic terms).
- Sovereign risk valuation:
  - Two-step: regress yield shifts on short-term deposit rate; re-price instruments by discounting cash flows at new yields; separate curves for nominal pesos, inflation-indexed pesos, BADLAR-linked, and U.S. dollars.
  - Regression estimated using weekly 2006–12 data.
- Recommendations to advance methodology (three-pronged credit approach and other actions):
  - Data: reduce uncertainty; use quarterly and monthly bank-level data.
  - Credit risk:
    - Enhance dynamic panel models using quarterly data. Timeframe: Immediate.
    - Develop more granular models by sector, loan type, borrower type. Timeframe: Near-term.
    - Develop PD models based on individual borrower data (Central de Deudores del Sistema Financiero). Timeframe: Near-term.
  - Sovereign risk:
    - Construct zero coupon yield curves for nominal and inflation-indexed instruments; obtain smooth reference yield curves (Nelson-Siegel or similar); price fixed income instruments based on shifts in zero-coupon yield curves. Timeframe: Medium-term.
  - Net fee/service income and operating expenses:
    - Exploit quarterly data; develop quarterly panel and bank-specific regression models. Timeframe: Near-term.
  - Use the bank-level panel supervisory dataset with detailed balance-sheet and P&L information for modeling. Timeframe: Immediate.

### Key policy recommendations and implementation timeframes (selected)
- Banking sector (BCRA):
  - Use bank-level panel supervisory dataset for modeling. Timeframe: Immediate.
  - Enhance satellite models for credit risk, sovereign risk, net fee and service income, operating and administration expenses (Immediate to Medium-term as specified above).
- Sustainability Guarantee Fund (FGS):
  - Auction fixed-term deposits in banks on a market price return base; increase maturity to provide stable funding. Timeframe: Near-Term.
  - Discourage unsupervised direct credit activities; register all credit and debtor information from FGS loan programs in the credit bureau. Timeframe: Near-Term.
  - Establish a limit of 5 percent for FGS exposure to banks in equity. Timeframe: Near-Term.
- Insurance sector (SSN):
  - Apply technical premiums and disallow more than two months of unpaid premia. Timeframe: Immediate.
  - Lift recent investment guidelines for insurances and allow risk-transfer reinsurance. Timeframe: Immediate.

*Source: EXECUTIVE SUMMARY and selected sections from _cr1665 (IMF staff and Banco Central de la República Argentina materials, Argentina FSAP stress tests; content as provided in the supplied PDF excerpt).*

### EXECUTIVE SUMMARY ___________________________________________________________________________ 5

### _cr1665 - EXECUTIVE SUMMARY ___________________________________________________________________________ 5

### Introduction and context
- Argentina’s financial system is very small, bank-dominated, and mostly transactional; it has ample liquidity, solid capitalization, and strong asset quality despite limited depth.
- FSAP stress testing objectives: assess capacity to withstand extreme but plausible macroeconomic shocks; identify transmission channels, weaknesses, and policy priorities.
- Tests use macroeconomic and satellite models estimated on historical data; model uncertainty is potentially severe given institutional and structural changes over the last two decades.
- Stress testing horizon and data: adverse external shocks assessed over a two-year horizon (2013–14) based on data available through September 2012.

### Design of stress tests and scenarios
- Coverage: 22 largest banks, which account for 90 percent of the system by assets.
- Methods: top-down macroeconomic scenarios and sensitivity analysis; satellite models used to map macro shocks to bank profitability and capitalization.
- Adverse external shocks calibrated to trigger a cumulative decline of real GDP growth equivalent to 13.3 percentage points, or 2 standard deviations.
- Sensitivity scenarios included: a 900 basis point increase in real interest rates; a 30 percent peso depreciation partially transmitted to inflation; concentration shocks (failure of the five largest borrowers); reverse liquidity tests with 100 percent run-off on maturing wholesale deposits.

### Solvency and credit risk findings
- Credit risk is the primary vulnerability: declines in capital ratios in 2013 and 2014 are largely driven by deterioration in credit quality.
- Nonperforming loan (NPL) dynamics: loan loss ratio in the banking system rises from 1.5 percent to more than 6 percent in the most extreme adverse scenario.
- Capitalization outcomes:
  - In the most extreme adverse scenario, capitalization in 6 of the 22 largest banks would fall below the required minimum of 8 percent.
  - Two banks are undercapitalized at the starting point of the exercise.
  - Estimated system-wide capital shortfall in the most extreme scenario is less than 0.2 percent of GDP.
- Sensitivity shocks:
  - A 900 basis point increase in real interest rates would lead to 5 of the 22 largest banks being undercapitalized.
  - A 30 percent peso depreciation partially transmitted to inflation would increase NPL ratios and could reduce capitalization of 5 banks below the required minimum.
  - Failure of the five largest borrowers would cause undercapitalization in 8 of the 22 banks (noted as a low probability event).

### Market, sovereign, and exchange rate risk
- Market risk exposure is limited: banks hold highly liquid bonds and BCRA money market instruments; exposure to corporate bonds, equity, commodities, foreign securities is negligible.
- Sovereign risk: adverse scenarios produce higher interest rates and inverted yield curves causing sizable losses on sovereign securities; losses are partly offset over time by price appreciation when yields decline and are limited by short portfolio durations and instruments indexed to inflation, the U.S. dollar, or the BADLAR rate.
- Exchange rate risk: banks have positive net open foreign currency exposures; a peso depreciation in adverse scenarios tends to boost bank profitability.

### Liquidity stress test results
- Cash-flow based liquidity tests use liability run-off rates and asset haircuts calibrated on Argentine historical data.
- Short-run resilience:
  - All banks could confront persistent and sizable deposit withdrawals for 30 days without BCRA assistance.
  - BCRA tools: waive reserve requirements for up to 30 days or inject liquidity through standing facilities.
- After 30 days:
  - Only 2 of the 22 largest banks would need BCRA assistance in pesos and 1 in dollars; extension of reserve requirement waiver would suffice.
- Reverse liquidity test (wholesale deposits):
  - Assumed 100 percent run-off on maturing wholesale deposits with full rollover on other lines.
  - All banks have sufficient liquidity to absorb losses of 33 percent or more of wholesale deposits without BCRA assistance.
  - 15 of the 22 largest banks could confront 100 percent withdrawals of maturing wholesale deposits without a shortfall of liquid assets at any time during the two-year assessment period.

### Contagion, interconnectedness, and the Sustainability Guarantee Fund (FGS)
- Direct bilateral interbank contagion is limited: interbank exposures are very small relative to capitalization; as of September 2012, only one of the 22 large banks had total interbank exposure larger than its excess capital over the required minimum—and for that bank, five other institutions would have to fail for its capital to fall below the minimum.
- FGS-related risks:
  - FGS has 6 percent of its portfolio invested in fixed-term deposits with very short average maturities of 35–40 days.
  - For some banks, FGS deposits represent more than 4 percent of total deposits.
  - Allocation mechanism for these deposits lacks transparent criteria; many deposits (as of September 2012) were invested below market prices in certain banks where FGS has significant stakes.
  - Risks: sudden withdrawal of FGS deposits may generate liquidity pressures for smaller banks; large FGS equity stakes in private banks raise governance concerns.

### Insurance sector vulnerabilities
- A large percentage of available capital for solvency is illiquid and not well suited to protect against adverse events.
- Non-life sector is weaker than life or retirement sectors, with notable credit and liquidity risks.
- Policy recommendation: lift recently introduced mandatory investment requirements which risk accentuating vulnerabilities; allow risk-transfer reinsurance.

### Key recommendations and implementation timelines (as presented)
- Banking sector (BCRA)
  - Use the bank-level panel supervisory dataset with detailed balance-sheet and P&L information for modeling. Timeframe: Immediate.
  - Enhance satellite models for credit risk:
    - estimate dynamic panel models using quarterly data; Timeframe: Immediate.
    - develop more granular models (loan loss rates or default probabilities by sector, loan or borrower type); Timeframe: Near-term.
    - explore alternative specifications and time-variant bank-specific variables; Timeframe: Near-term.
    - develop default probability models (PD) based on individual borrower data. Timeframe: Near-term.
  - Enhance satellite models for sovereign risk:
    - construct zero coupon yield curves for nominal and inflation-indexed instruments from coupon yields; Timeframe: Medium-term.
    - obtain smooth reference yield curves (nominal and real) using Nelson-Siegel or similar approaches; Timeframe: Medium-term.
    - price fixed income instruments based on shifts in estimated zero-coupon yield curves in stress tests. Timeframe: Medium-term.
  - Enhance satellite models for net fee and service income and operating and administration expenses:
    - exploit quarterly data; Timeframe: Near-term.
    - develop quarterly panel and bank-specific regression models to assess robustness; Timeframe: Near-term.
- Sustainability Guarantee Fund (FGS)
  - Auction fixed-term deposits in banks on a market price return base; increase maturity to provide stable funding. Timeframe: Near-Term.
  - Discourage unsupervised direct credit activities; register all credit and debtor information from FGS loan programs in the credit bureau. Timeframe: Near-Term.
  - Establish a limit of 5 percent for FGS exposure to banks in equity. Timeframe: Near-Term.
- Insurance sector (SSN)
  - Apply technical premiums and disallow more than two months of unpaid premia. Timeframe: Immediate.
  - Lift recent investment guidelines for insurances and allow risk-transfer reinsurance. Timeframe: Immediate.

*Source: EXECUTIVE SUMMARY (IMF Technical Note, FSAP); content as provided in the supplied PDF excerpt.*

### 7.      The stress tests examined the resilience of the Argentine banking system to solvency,

### 7.      The stress tests examined the resilience of the Argentine banking system to solvency, liquidity, and contagion risks

### Overview
- The tests examined solvency, liquidity, and contagion risks (Figure 1).
- TD solvency stress tests: macroeconomic scenarios approach and sensitivity analysis by the authorities (BCRA) and the FSAP team.
- Time horizon: two-year horizon (2013–14), based on data available through September 2012.
- Transmission to individual banks’ profitability and capitalization assessed using satellite models developed by the authorities and validated by Fund staff.
- TD liquidity stress tests: maturity ladder analysis and bank-specific information provided to the BCRA to assess capacity to confront large funding withdrawals.
- Contagion tests also covered nonbank financial institutions.

### Coverage
- Tests covered the 22 largest banks, which in terms of asset size account for 90 percent of the system.
- Composition of the tested banks:
  - Private domestic: 10 banks (23 percent of assets)
  - Public: 5 banks (41 percent of assets)
  - Private foreign: 7 banks (26 percent of assets)
  - Other (not tested): remainder

### A. Solvency Stress Tests — Approach
- Solvency stress tests covered credit, market, and sovereign risks and used:
  - Macroeconomic scenarios (dynamic, multi-factor transmission).
  - Sensitivity analysis (one risk factor at a time, holding others constant).
- Macroeconomic scenarios simulated adverse external shocks (declines in trading partners’ growth, adverse terms of trade shocks, rise in global risk aversion causing capital outflows).
- Tests required assumptions on banks’ behavior, balance sheet growth, and off-balance sheet evolution.
- Two baseline scenarios: authorities’ baseline and Fund-staff (WEO) baseline; authorities’ models used BCRA methodology, Fund-staff used VAR analysis.
- Balance sheet treatment differences:
  - Authorities’ approach: allows credit and deposits to differ from nominal GDP growth.
  - Fund-staff approach: assumed “constant” balance sheet (grow in line with nominal GDP).

### Macroeconomic Tests — Scenarios and Sensitivities
- Macroeconomic adverse scenarios assessed:
  - (i) Adverse scenario based on authorities’ baseline: cumulative decline of GDP equivalent to 1.7 standard deviations over two years.
  - (ii) U-shaped adverse scenario relative to Fund-staff baseline.
  - (iii) V-shaped adverse scenario relative to Fund-staff baseline.
- The U- and V-shaped adverse scenarios result in a cumulative decline of real GDP growth equivalent to 2 standard deviations (13.3 percentage points) over two years.
- Standard deviation detail: one standard deviation of two-year cumulative real GDP growth (2001–2012) = 6.6 percentage points.
- Sensitivity results from VAR analysis:
  - A one-standard deviation decline in trading partners’ growth (equivalent to a 4 percentage point deviation from baseline) reduces Argentina real GDP by nearly 3 percentage points after four quarters (elasticity ≈ 0.75).
  - A 10 percent decline in terms of trade sustained for 8 quarters reduces domestic output by about 2 percentage points.
  - A one standard deviation shock in the VIX reduces domestic output by 4.5 percentage points relative to the baseline.

### Macroeconomic Tests — Projected Paths (Table 2)
- Real GDP growth (in percent):
  - 2012 2013 2014
  - Baseline Authorities: 1.9 4.6 4.0
  - Adverse Authorities: 1.9 -0.8 -1.4
  - Baseline WEO1: 1.9 2.8 3.5
  - Adverse (U): 1.9 -3.4 -3.6
  - Adverse (V): 1.9 -6.9 0.0
- GDP deflator growth (in percent):
  - Baseline Authorities: 15.3 15.5 16.8
  - Adverse Authorities: 15.3 10.8 15.1
  - Baseline WEO1: 15.3 17.1 15.0
  - Adverse (U): 15.3 15.3 13.1
  - Adverse (V): 15.3 14.6 12.0
- Unemployment rate (end-of-year, percent):
  - Baseline Authorities: 6.9 6.6 6.2
  - Adverse Authorities: 6.9 8.2 8.0
  - Baseline WEO: 6.9 6.7 6.4
  - Adverse (U): 6.9 10.5 10.6
  - Adverse (V): 6.9 12.6 8.4
- Nominal exchange rate, peso per US dollar (end-of-year):
  - Baseline Authorities: 13.8 11.2 11.0
  - Adverse Authorities: 13.8 18.6 11.9
  - Baseline WEO: 13.8 17.8 16.0
  - Adverse (U): 13.8 21.5 15.0
  - Adverse (V): 13.8 26.4 12.0
- Nominal annual interest rate (BADLAR, percent):
  - Baseline: 15.4 16.0 15.4
  - Adverse Authorities: 15.4 17.7 16.6
  - Baseline WEO: 15.4 16.7 16.4
  - Adverse (U): 15.4 20.6 17.3
  - Adverse (V): 15.4 21.4 16.4
- Note: All the WEO numbers correspond to the January update. Unemployment rates in Fund-staff baseline scenarios obtained by applying Okun’s law; Okun coefficient estimates for Argentina range 1.7-2.3.

### Solvency Thresholds (Table 3)
- Minimum Total Capital (In percent):
  - Year 2012 2013 2014: 8.0 8.0 8.0
- Minimum Tier 1 Capital (In percent):
  - Year 2012 2013 2014: 4.0 4.5 5.5

### Satellite Models — Methodology
- Satellite models developed by BCRA, validated by Fund-staff, used to map macroeconomic shocks to bank profits and capital.
- Satellite models cover:
  - Credit losses, yield curve movements for bonds and money market instruments, net interest income, net fee and service income, operational and administrative expenses, and other profit sources.
- Credit risk model: dynamic panel data where loan loss rates depend on real GDP growth, real interest rates, inflation, and bank-specific characteristics—approach similar to some G20 countries.
- Interest income: maturity ladder approach; yield curve shifts modeled separately by instrument type (pesos, inflation-linked, BADLAR-linked, U.S. dollars).
- Data and estimation:
  - Models estimated using long annual bank-specific data spanning almost two decades; market risk models (yield curves) use daily data.
  - Ongoing BCRA work to estimate refined models exploiting monthly and quarterly bank-specific data back to 1994.
- Modeling caveats:
  - Long series span institutional and regime changes (convertibility to post-convertibility) and banking crises (1995, 2001), which may create structural breaks but provide evidence for extreme-condition calibration.
- Note: Satellite models used in authorities’ baseline differ from models used in Fund-staff baseline; Fund-staff models revised per FSAP team comments.

### Results — Key Findings
- Credit risk is the most important vulnerability.
- Fund-staff baseline macroeconomic stress test results:
  - Declines in capital ratios in 2013 and 2014 largely driven by deterioration in credit quality.
  - NPL rates rise sharply under adverse scenarios: loan loss ratio in the banking system increases from 1.5 percent to 4.4 percent (U-shaped) and 6.2 percent (V-shaped) by 2014.
  - In the U-shaped adverse scenario, 4 of the 22 largest banks would fall below the 8 percent total capital minimum.
  - In the V-shaped adverse scenario, 6 banks would be undercapitalized.
  - Note: two banks accounting for 9 percent of banking assets were undercapitalized at the starting point and are under plans to restore solvency.
- Capital calculation: conforming to the new standard under Pillar 1 of Basel II adopted in January 2013; most capital consists of common equity and retained earnings (part of Tier 1 capital), so total capital hurdles more restrictive than Tier 1 for most banks.

### Results — Market and Sovereign Risk
- Banks hold liquid bonds and money market instruments, mostly issued by the BCRA (LEBACs and NOBACs) and to a much lesser extent government securities (peso and U.S. dollar denominated).
- Exposure to foreign sovereigns is negligible.
- Sovereign risk vulnerability is mild:
  - Adverse scenarios cause significantly higher interest rates and inverted yield curves, producing haircut-related losses on sovereign paper.
  - Losses partially offset by price appreciation when yields decline over time and gains from quasi-fiscal instruments linked to inflation, U.S. dollar, or BADLAR that produce gains for banks.
  - Short duration of bank sovereign portfolios limits exposure.
- Simulated yield changes (very large in V-shaped adverse scenario):
  - Average yield increases about 1,500 basis points for U.S. dollar denominated bonds.
  - About 2,000 basis points increase for peso denominated bonds adjusted by inflation.
  - About 2,800 basis points increase for bonds linked to the BADLAR interest rate.
  - Yield changes of peso-denominated nominal instruments (fixed coupon and discount money market) about 450 basis points.

*Source: IMF staff and Banco Central de la República Argentina materials, Argentina FSAP stress tests (2013–14 scenarios).*

### 26.      Banks appeared resilient to non-sovereign market risk. The exposure of banks to

### _cr1665 - 26.      Banks appeared resilient to non-sovereign market risk. The exposure of banks to

### Resilience to non-sovereign market risk
- Banks’ exposure to corporate bonds, equity, commodities, non-sovereign foreign securities, and other sources of market risk are negligible.
- Banks hold positive net open foreign currency exposures—including balance sheet exposures and positions in foreign currency forward and future contracts—so a depreciation of the peso in the adverse scenarios has a positive impact on profits.

### System capital shortfall under adverse scenarios
- In all adverse scenarios the capital shortfall in the banking system would be small relative to the size of the economy.
- V-shaped adverse scenario system shortfall estimated at about 0.2 percent of GDP.
- The small shortfall partly reflects the small size of the banking system relative to the size of the economy.
- Although the overall shortfall is small, a number of banks would be under-capitalized in adverse scenarios.

### Recommendations on advancing the stress tests methodology
- Data
  - Reducing the uncertainty on the measurement of economic variables is recommended.
- Satellite models for credit risk (three-pronged approach)
  - First: enhance current model (dynamic panel techniques) by using available quarterly data (instead of annual data) and explore inclusion of additional bank-specific and time varying explanatory variables.
  - Second: develop more granular models of credit risk, including models of loan loss rates or default probabilities by economic sector (e.g., agriculture, commerce, manufacturing), type of loan (e.g., consumer, mortgage, corporate), or type of borrower (e.g., individuals, SMEs, large corporations).
  - Third: develop models of probability of default (PD) estimated based on individual borrower data. Information from the Central de Deudores del Sistema Financiero could be useful. Development of sufficiently granular models may require collection of new data.
- Satellite models for sovereign risk
  - Estimate the term structure of nominal and real interest rates using standard methods.
  - Suggested agenda:
    - Step 1: From observed yields of nominal coupon bonds obtain zero-coupon yield curves (replication of zeros; bootstrapping; recursive or simultaneous equation approaches; or regression methods). Apply similar approach to inflation-indexed bonds to construct “real” term structure, with careful analysis for instruments adjusted to CER.
    - Step 2: Obtain smooth reference curves (nominal and real) using interpolation methods (e.g., Nelson-Siegel).
    - Step 3: Use projected shifts in zero-coupon yield curves to conduct stress tests; include separate projections for a real and a nominal interest rate and measure response of zero-coupon curves to projected rate changes using regressions.
- Satellite models for net fee and service income and for operating and administration expenses
  - Use available quarterly data to allow more powerful stationarity tests and reassessment of model specification; consider additional macroeconomic control variables and explore panel data or bank-specific regressions in a richer quarterly data environment.

### Sensitivity tests: scope and methodology
- Sensitivity tests assess instantaneous impact on banks’ balance sheets and off-balance sheet positions as of September 2012; tests are static with risk-weighted assets assumed constant after shocks.
- Key domestic shocks assessed:
  - Increase in domestic real interest rates (tightening of domestic monetary conditions).
  - Nominal depreciation of the peso (loss of confidence, capital outflows, wider parallel-official exchange rate gap).
  - Failure of a number of large corporate exposures (concentration).
  - Sharp decline in prices of sovereign securities.
- The credit risk model used for some sensitivity tests is dynamic and estimated on annual data; interest rate or depreciation effects are fully transmitted to loan loss rates only after two years, though balance sheets are held as observed in September 2012.

### Increase in real interest rates
- A 900 basis point increase in real interest rates:
  - Sensitivity tests suggest 5 of the 22 largest banks would be undercapitalized after a 900 basis point increase in real interest rates.
  - Test assumes banks earn no pre-impairment profits under stress, the increase is sustained for two years, and output is assumed constant.
  - The test ignores second-round effects (e.g., effects on inflation expectations, deposit or capital outflows) and therefore shows only a partial impact on credit quality and capitalization.

### Nominal peso depreciation (loss of confidence)
- Under a 30 percent peso depreciation with 0.2 pass-through to domestic prices:
  - Inflation would increase by 6 percentage points.
  - NPL ratios would increase by about ½ percentage point on average, with variation across banks.
  - If banks earned no pre-impairment profits, capitalization of five banks could fall below the required minimum (partial impact only).
- Market risk effect of a 30 percent peso depreciation against the U.S. dollar:
  - System CAR would increase by 1.4 percentage points, from 11.7 percent to 13.1 percent.
  - Across banks: only one bank would suffer a loss; 21 of the 22 banks would obtain benefits ranging from 0.4 to 5.2 percentage points of CAR.
- Tests ignore second-round effects of a depreciation-inflation spiral that could reduce real money demand and cause a real credit contraction, adversely affecting output and credit losses.

### Concentration risk: failure of large corporate exposures
- Failure of the five largest borrowers would cause undercapitalization in 8 of the 22 banks.
- Failure of the 10 largest counterparts would result in undercapitalization of 12 banks.
- Many large counterparties are large counterparts across multiple banks, compounding systemic risk.
- Some large counterparties operate in the agro-industrial sector; a decline in commodity prices could trigger common-name concentration risk.

### Sovereign securities price declines
- Banks hold highly liquid bonds and money market instruments issued by the BCRA and the government, including:
  - LEBACs (zero-coupon, nominal, peso, maturities up to 270 days).
  - NOBACs (coupon bonds up to three years, principal and coupons either fixed or floating linked to BADLAR).
  - Government bonds denominated in pesos and U.S. dollars.
- Simulated one-year yield changes (50th and 90th percentiles from daily 2007–12 series) were similar in magnitude to V-shaped adverse scenario shocks:
  - Average yields increased by about 1,500 basis points for U.S. dollar denominated bonds.
  - Average yields increased by about 2,000 basis points for peso denominated bonds adjusted by inflation.
  - Average yields increased by about 2,800 basis points for bonds linked to the BADLAR interest rate.
  - Yield changes of peso-denominated nominal instruments (fixed coupon bonds and discount money market instruments) were about 450 basis points.
- Results
  - Increase in sovereign yields, combined with short duration of banks’ portfolios, lowers system CAR by 1.3 percentage points, from 11.7 percent to 10.4 percent.
  - Impact across banks varies from 0.1 to 3.2 percentage points of CAR.
  - Haircuts applied to adjusted (marked-to-market) balance sheet values; previous economic valuation losses or gains were recognized before haircuts.

### Liquidity stress tests
- Liquidity stress tests used a maturity ladder (TD) approach with data on maturity structures collected from banks.
- The exercise captured:
  - (i) liquidity needs derived from outflows;
  - (ii) available standby liquidity from inflows;
  - (iii) liquidity buffers available to counterbalance liquidity gaps.
- Additional analyses included:
  - A reverse liquidity stress test assessing capacity to withstand maximum withdrawal of wholesale deposits.
  - A TD liquidity test linked to macroeconomic scenarios based on the authorities’ baseline.
- Note: Common FSAP practice is to assume an environment where funding pressures are sizable but limited to a number of banks (not systemic).

*Source: IMF staff and authorities’ calculations as presented in the source PDF.*

### 41.      Banks’ funding consists mainly of deposits from individuals, SMEs, and large

### _cr1665 - 41.      Banks’ funding consists mainly of deposits from individuals, SMEs, and large

### Funding structure of the banking system (excluding own capital)
- 65 percent is non-collateralized funding provided by legal entities:
  - non-financial public sector (FGS included) (27 percent)
  - SMEs (17 percent)
  - large enterprises (17 percent)
  - other institutions regulated by the BCRA (0.4 percent)
  - unregulated institutions with financial activities (3.2 percent)
  - others (0.4 percent)
- 32 percent is funding due with individual depositors (physical persons)
- 3 percent includes own issuances; secured funding; liabilities related to derivatives transactions and committed credit lines; and other liabilities

### Run-off rate assumptions and calibration
- General principle: more informed and sophisticated depositors withdraw funding more rapidly than less informed depositors
- Sight deposits:
  - large enterprises: 15 percent of the initial balance in the first day, increasing to 25 percent within one month
  - SMEs and individuals: 7.5 percent of the initial outstanding balance in the first day, increasing to 10 percent within a two month horizon
- Time and other deposits (defined as fraction maturing in a period withdrawn by depositors):
  - large corporations: constant 40 percent
  - SMEs: gradually increase from 15 to 25 percent
  - individuals: gradually increase from 10 to 25 percent
- Nonfinancial public sector: 20 percent run-off rate, constant across maturities
- Others:
  - own issuances: 60 percent
  - secured funding: 100 percent
  - committed credit lines: 10-40 percent
- Calibration basis:
  - Run-off rates calibrated based on Argentina’s historical experience (convertibility period crises of 1995 and 2001)
  - Run-off rates on individual banks calibrated to be higher than those experienced by a typical bank during the crises of 1995 and 2001
  - Noted caveats: systemic panics raise run-off rates and lower asset haircuts; BCRA’s current ability to act as lender of last resort and lower dollarization reduce likelihood of historical run-off levels materializing

### Liquidity inflows assumptions (standby inflows)
- Assets generating cash proceeds:
  - maturing loans: 76 percent
  - debt securities: 19 percent
  - reverse repos: 4 percent
  - others: 1 percent
- Conversion rates to cash proceeds:
  - maturing loans to households and enterprises: 50 percent
  - maturing loans to financial institutions: 100 percent
  - cash flows from debt securities and reverse repos: 100 percent
- Assumption: going concern—actions do not compromise relations with important borrowers nor cause premature liquidation

### BCRA assistance and market haircuts
- BCRA emergency tools assumed:
  - waive reserve requirements for a maximum period of 30 days
  - inject liquidity through standing facilities
- Assumption: market haircuts are lower than BCRA haircuts (banks pay a price if relying on BCRA emergency liquidity)

### Liquidity stress test results (pesos)
- All banks would be able to confront persistent and sizable withdrawals of funding for 30 days without any assistance from the BCRA
- After 30 days:
  - only two of the 22 largest banks would need BCRA assistance in pesos
  - 1 bank would need assistance in dollars
  - an extension of the reserve requirement waiver would suffice to render them liquid
- No bank would need emergency liquidity assistance from the BCRA for two years, besides access to own minimum required reserves
- Pilot program: BCRA developing framework for LCR and NFSR; three banks tested so far show ratios above 100 percent for both indicators
- Table 4 summary (pesos liquidity tests):
  - Number of banks meeting contractual obligations without BCRA support: entries show meeting obligations up to various horizons (implied all banks up to 30 days; beyond one month two banks fail under full reserve enforcement)
  - Number of banks meeting contractual obligations with BCRA support: entries show banks meet obligations across all horizons with BCRA support
- Note: assumed BCRA waives compliance with minimum reserve requirement for up to one month; two banks need liquidity support to comply with fully enforced minimum reserve requirement after one month but would need no other liquidity assistance if waiver extended

### Reverse liquidity stress test and wholesale deposit run-off
- Assumption: 100 percent run-off rates on maturing wholesale deposits and full rollover rates in other funding lines
- Results:
  - all banks have liquidity to deal with losses of 33 percent or more of total wholesale deposits without resorting to BCRA facilities
  - 15 of the 22 largest banks would be able to confront withdrawals of 100 percent of maturing wholesale deposits without experiencing a shortfall of liquid assets at any time in the two-year assessment period

### Contagion via interbank exposures and nonbank interconnectedness
- Direct contagion via bilateral interbank exposures is limited:
  - as of September 2012, only one of the 22 largest banks had total interbank exposure larger than its excess capital over the required minimum; in that case five other institutions would have to fail for that bank’s capital to fall below the required minimum
- Interconnectedness with the FGS (FGS risks):
  - FGS has 6 percent of its portfolio invested in fixed-term deposits, very short term (on average 35–40 days)
  - For some banks, FGS deposits represent more than 4 percent of total deposits
  - Allocation mechanism for such deposits appears not transparent; as of September 2012 most were invested below market prices (in the largest public bank, or in a private bank where the FGS is a significant shareholder)
  - Lack of clear criteria and risk of sudden withdrawal may generate liquidity pressures for smaller banks
  - Large equity stakes of the FGS in some private banks raise governance issues

### The FGS (Sustainability Guarantee Fund) — size, portfolio, and risks
- Institutional role and scale:
  - FGS is a public entity under ANSES control; main institutional investor in Argentina
  - Accounts for 20 percent of the financial system and about 11 percent of GDP
  - Created in 2007; received proceeds from nationalization of private pension funds in late 2008
  - Main provider of long-term financing for productive and infrastructure projects
- Portfolio characteristics and market risk:
  - About four-fifths of the investment portfolio concentrated in fixed term instruments with a duration of five years on average
  - Simple estimates: for every 100 bps increase in the interest rate, the economic value of its investment portfolio would decline by 3.4 percent
  - A fraction of this debt is probably bonds indexed to official inflation
- Lending and regulatory perimeter risk:
  - FGS has begun to extend credit; falls outside BCRA regulatory perimeter and loans are not registered in the credit registry—this could be a source of risk
- Interconnections with banks:
  - Over 6 percent of the portfolio invested in fixed-term deposits, majority in public banks, with average maturities of about 45 days
  - Cash is shared between three custodian banks
  - FGS has mostly rolled-over fixed-term deposits so far, but lack of allocation clarity may create pressures
  - FGS owns 31 percent of the third largest private bank (about 6 percent deposit market share) and has stakes in 4 other private banks

### Insurance sector vulnerabilities and pension scheme observations
- Industry profitability and composition:
  - Nominal ROE rose from 8 percent in 2006 to 22 percent in 2012
  - Nominal ROE in 2012 still below many indicators of inflation, but somewhat above the rate of increase in the GDP deflator
  - ROE supported by returns on investment, not underwriting income
- Solvency and asset quality:
  - Under current solvency regime: 4 companies have solvency ratios below 100 percent and 57 have ratios below 125 percent
  - For the industry as a whole actual solvency margin has amounted to 175 percent of the minimum solvency margin for the past several years
  - Assets bearing significant credit risk (unpaid premia and other receivables as reported in 2012) amount to 100 percent of the free capital available to the sector (ARG$18.34 billion)
  - Large percentage of available capital is illiquid and not fully suitable to protect companies in adverse events
- Non-life sector specifics:
  - Weaker footing than life or retirement sectors; vulnerabilities to credit and liquidity risks
  - Among 20 largest non-life firms in 2012: 7 had ROEs in excess of 25 percent, 9 had ROEs below 15 percent
  - Poor underwriting results: claims paid plus expenses have exceeded premium income for several years
  - Outstanding premia equivalent to about one-fifth of total assets and one-third of total annual nonlife premia
  - New regulation disallowing more than two months of unpaid premia as admissible assets would reduce credit risk
  - For the 20 largest nonlife insurers in 2012, liquid assets amounted to 106 percent of claims paid on average; 6 firms had liquid assets below 75 percent of claims paid
  - High number of court claims could indicate dissatisfaction with claim settlements and potential liquidity strains
- Life insurance and retirement products:
  - Life insurance sector declining since 2001; now accounts for less than 30 percent of whole insurance sector
  - Life insurance relies mostly on annual renewal of group life insurance—liabilities have much shorter duration
  - Mandatory investments of up to 30 percent of assets in economic projects and infrastructure could lengthen asset duration and potentially accentuate maturity mismatch
  - Recommendation: adapt investment guidelines to avoid significant maturity mismatches between assets and liabilities
- Contagion to banks:
  - Limited: banks’ exposures to insurance sector via deposits and some bonds are of very limited significance
  - Only two banks hold equity in insurance company subsidiaries

*Source: IMF staff summary of content from _cr1665 - 41.      Banks’ funding consists mainly of deposits from individuals, SMEs, and large*

### Appendix I. Argentina: Methodology for the Construction of

### Appendix I. Argentina: Methodology for the Construction of Macroeconomic Scenarios

### Macroeconomic scenarios based on the authorities’ baseline
- Step 1. Projection of key macroeconomic variables using models and expert judgement.
  - The BCRA uses a VAR model and a small dynamic macroeconomic model in the New Keynesian tradition to project real GDP, a short-term interest rate (overdraft lending rate, in pesos), the inflation rate (based on the GDP deflator), and the nominal exchange rate.
  - Projections take as given the foreseen paths of the sovereign spread and exogenous external variables—including the terms of trade and an international interest rate (the U.S. Federal Funds rate).
  - The open economy model incorporates four standard dynamic equations—a Phillips curve; an IS curve; an uncovered interest parity condition; and a Taylor rule—and is estimated using Bayesian techniques. Variants can accommodate pure fixed, pure floating, dirty floating regimes.

- Step 2. Projection of a broader set of interest rates.
  - Future paths of key deposit rates are projected based on regression models of the form (as shown in source):
    - 30-60 day time deposits, pesosoverdraft, pesos
    - BADLAR private banks, pesosoverdraft, pesos
    - Spread ()(macro variables),
    - Spread ()(macro variables).
    - tt tt iif iif  
  - These models take as input the macroeconomic projections from Step 1.
  - Other lending interest rates (consumer loans, mortgages) are projected using regressions on the rate corresponding to 30–60 day time deposits.
  - A similar approach is used for interest rates on U.S. dollar transactions.

- Step 3. Projection of aggregate money demand (M2).
  - Real M2 aggregates that include dollar denominated deposits are regressed on interest rates.
  - Estimated coefficients are used to project real M2 using inputs: real GDP and the nominal interest rates from Steps 1 and 2.

- Step 4. Projection of elements of the monetary survey, including aggregate credit and deposits.
  - Future paths of credit and deposit aggregates (in both pesos and U.S. dollars) are projected using standard financial programming techniques.
  - The balance-sheet identity for the consolidated banking system (including the BCRA and depository institutions) is given in the source as:
    - M2CD NFADCOIN   
    - where C=currency in circulation; D=demand and time deposits; NFA=Net Foreign Assets; DC=Domestic Credit; OIN=Other items net.
  - To project DC and D, assumptions are required on international reserves (NFA), other items net (OIN) and the ratio of currency-in-circulation to deposits.
  - The size of the BCRA balance sheet (the monetary base) is projected by imposing assumptions on the money multiplier.
  - Assumptions on accumulation of net foreign assets by individual banks and the BCRA allow disentangling balance sheet projections for the BCRA and consolidated depository institutions; NFA projections must be consistent with external sector financial programming.
  - Domestic credit can be channeled to public or private sectors; projections are guided by expert judgment and must be consistent with fiscal sector financial programming (deficit projections and treasury’s debt issuance).
  - This approach yields projections of deposit and credit aggregates of the consolidated banking system that will not necessarily grow at the same rate as nominal GDP—a key difference from the Fund-staff baseline.

- Step 5. Projection of bank-specific deposits and credit.
  - Before stress-related losses, individual bank balance sheets are assumed to grow at the same rate as the aggregate balance sheet; the relative size of each bank (in terms of assets) is assumed constant.

### Macroeconomic scenarios based on the Fund-staff baseline
- Data.
  - All series are monthly and taken from the IMF’s International Financial Statistics (IFS), Haver, and CEIC; complemented with BCRA website data when needed.
  - Series are sampled quarterly; average values are used for financial market data (interest rates, exchange rate and VIX).
  - Sample period: 1993–2012.
  - Variables used: U.S. real GDP, Brazil real GDP, U.S. federal funds rate, U.S. VIX, Argentina’s terms of trade (the price of exports index used in robustness exercises), the benchmark short-term 30-day deposit interest rate, the nominal effective exchange rate, and Argentina’s real GDP.
  - GDP and GDP deflator series are seasonally adjusted by X-12; annualized log-growth rates are calculated.

- Model.
  - A VAR model is used where yt is the n x 1 data vector containing external economic growth (Brazil and United States), U.S. VIX, U.S. interest rate, terms of trade growth, domestic inflation, domestic economic growth, domestic interest rate, and nominal effective exchange rate; k is a vector of constants, Bi is an n x n matrix of coefficients (i = 1, ..., p), and ut is a white-noise vector of “structural” shocks with diagonal variance-covariance matrix D.
  - The reduced form can be rewritten (notation as in source) and identification involves imposing restrictions on B0 to orthogonalize reduced form errors. A standard identification is the recursive (Cholesky) ordering of Sims (1980).

- Estimation and identification.
  - The reduced form model is estimated by least-squares (MLE); lag structure determined by Bayesian information criteria (BIC).
  - Standard errors for impulse response functions are calculated according to Sims and Zha (1999).
  - According to multivariate BIC, generally one or two lags suffice; for baseline results a 1-lag structure with one crisis dummy was preserved.
  - Residual tests reveal mild forms of autocorrelation (but not of heteroskedasticity); adding more lags or dummies can reduce autocorrelation.

- Contemporaneous restrictions for the non-recursive structural model (as imposed in source).
  - Trading partners’ growth is exogenous with respect to all system variables.
  - Global risk aversion shocks respond to trading partners’ growth shocks (U.S. GDP growth included).
  - U.S. interest rate responds to both U.S. GDP growth (modified Taylor rule) and global risk aversion.
  - Domestic terms of trade shocks respond contemporaneously to trading partners’ growth.
  - Domestic inflation reacts contemporaneously to terms of trade and domestic activity (modified Phillips curve); the coefficient on inflation in the domestic activity equation is assumed zero.
  - Domestic activity responds to trading partners’ activity (through exports) and to global risk aversion (through investment).
  - Domestic interest rates react to global uncertainty, foreign interest rates, inflation (Fisher effect), and domestic activity (money demand relationship).
  - The nominal effective exchange rate reacts contemporaneously to all variables in the system.
  - The over-identifying restrictions are generally not rejected at the 5 percent level in the baseline VAR.

- Impulse response analysis.
  - Accumulated responses of real GDP to structural one-standard-deviation shocks in trading partners’ growth, terms of trade, and global risk aversion were analyzed (figures shown in source).

- Balance sheet and capital assumptions under Fund-staff baseline.
  - Individual bank balance sheets were assumed to stay constant as a share of nominal GDP (since nominal growth was not negative in any scenario).
  - Assets:
    - The value (in pesos) of total assets and every asset sub-category (credit, fixed assets, cash holdings, bond holdings, etc.) grow at the same rate as nominal GDP.
    - The relative structure of assets remains unchanged; RWA grows at the same rate as nominal GDP.
  - Liabilities:
    - Capital is projected based on outstanding capital at the end of the previous period, and current period’s profits net of dividends distributed and taxes paid. The formula shown in the source is:
      - 1111
      - CapitalCapitalProfitsDividendsTaxes   .
      - ttttt     
    - When losses materialize, the term Profits has a negative sign.
    - Current BCRA regulation on dividend distribution—whereby a bank can distribute dividends only when its CAR exceeds the minimum regulatory requirement by 75 percent—was assumed to remain in place.
    - For banks qualified to distribute dividends, the payout policy was assumed consistent with past behavior.
    - No injection of external capital was assumed during the stress testing period.
  - Remaining liabilities are calculated as the residual (assets minus capital) and their relative structure (deposit types, bonds issued, other liabilities) is assumed invariant.
  - Implication of the “constant” balance sheet assumption:
    - Banks that exhibit larger declines in capital have larger gains in deposits and other liabilities—counterintuitive and counterfactual, but convenient to avoid two stress-testing difficulties:
      - If a bank’s balance sheet declines sharply, higher loan loss rates and haircuts applied to a smaller asset base can make losses decline rather than increase.
      - If RWA declines sharply, the CAR could rise in adverse scenarios.
    - The constant balance sheet assumption provides a common anchor that prevents tests from being affected by optimistic projections of total assets- and credit-to-GDP.

*Prepared by IMF staff as presented in the source PDF.*

### Appendix II. Argentina: Satellite Models

### Appendix II. Argentina: Satellite Models

### Satellite Model for Net Interest Income
- Projection approach:
  - A broad set of lending and deposit rates is projected using regression methods.
  - Separate trajectories for all relevant interest rates are obtained for the baseline and adverse macroeconomic scenarios.
  - For each bank, interest income and expenses are projected based on bank-specific balance sheet, flow-of-funds, and loan maturity data (available as of September 2012), imposing roll-over assumptions on loans and deposits, and using the interest rates’ paths projected for the corresponding macroeconomic scenario.

### Satellite Model for Credit Risk
- Model specification (Fund-staff baseline):
  - Dynamic panel model with logistic transformation of loan loss rate (LLR), where LLR = ln(loan loss rate / (1 - loan loss rate)).
  - Right-hand side variables include:
    - Lagged LLR terms (up to two lags in displayed equation).
    - Real GDP growth (g) and real GDP growth squared terms (non-linear effect).
    - (Ex-post) real interest rate rr = (lagged short-term nominal lending rate) - (contemporaneous inflation measured by GDP deflator).
    - Growth of GDP deflator π.
    - Bank-specific capital-assets ratio CR (time-varying) and polynomial terms of CR.
    - Bank-specific fixed effects μi.
- Estimation sample and method:
  - Unbalanced panel including 193 banking institutions.
  - Annual observations for the period 1994–2011.
  - Two-step Arellano-Bond approach; results robust to sample period changes.
- Key estimated coefficients (Fund-staff baseline table):
  - Loan loss rate (lagged): 0.19
  - Real GDP growth: -9.77
  - Real GDP growth squared: 73.70
  - Capital to assets ratio (lagged): -2.89
  - Capital to assets ratio squared (lagged): 2.25
  - Capital to assets ratio cubic (lagged): -0.53
  - Real interest rate: nominal lending rate (peso loans, short term) (lagged) - growth of GDP deflator: 1.52
  - Growth of GDP deflator: 1.86
  - Wald chi2(8) = 574.50, Prob > chi2 = 0.0000
  - Arellano-Bond test for zero autocorrelation in first-differenced errors: Order 1 z = -5.0859, Prob > z = 0.0000; Order 2 z = .89139, Prob > z = 0.3727
- Model properties and interpretation:
  - Non-linear in real GDP growth: loan loss rate increases at an accelerated rate as real GDP growth declines further.
  - Effects of real interest rate and inflation are captured in separate terms for sensitivity testing (e.g., exchange rate depreciation pass-through to inflation without affecting rr).
  - Higher capital-to-assets ratio reduces moral hazard and lowers incentives to take excessive credit risk.
  - Robustness: parameters stable when estimated with data through 2008 and 2012.
- Numerical example (illustrative):
  - Assumptions: intercept = average fixed effect = -3.1; rr = 0.03; CR = 0.09; initial loss rate = 1.7 percent (similar to private banks in 2012).
  - Using the model, loss rates implied by baseline and adverse macro scenarios would rise significantly (specific paths are described in source; numeric paths not reproduced verbatim beyond assumptions above).

- Model used for tests based on the authorities’ baseline (alternative specification):
  - Includes additional terms: nominal market interest rate i (overdraft loans, end of previous period) and spreadi (bank i’s average nominal lending rate minus system lending rate).
  - Estimated coefficients (authorities’ baseline table):
    - Loan loss rate (lagged): 0.29
    - Real GDP growth: -6.57
    - Real GDP growth squared: 60.94
    - Real GDP growth cubic: -472.73
    - Nominal interest rate (lagged): 1.13
    - Capital to assets ratio (lagged): -2.08
    - Capital to assets ratio squared (lagged): 2.17
    - Capital to assets ratio cubic (lagged): -0.51
    - Spread (lagged): 1.21

### Satellite Model for Net Fee and Service Income
- Fund-staff baseline model (bank-level):
  - Simple autoregressive model of order 1:
    - NSIit = α + β NSIit-1 + εit
  - NSIA is net fee and service income-to-assets ratio (component of ROA).
  - Estimated for each bank using annual data for the period 1994-2012.
  - IMF and BCRA staff favored the simple AR(1) given annual data limitations.
  - Recommendation for future: use quarterly data and a more general specification NSIit = α + β NSIit-1 + γ Xit + εit where X includes macro and bank-specific time-varying controls.
- Authorities’ baseline model (system-level, quarterly):
  - NSIt = C + α1 NSIt-1 + α2 NSIt-4 + α3 UNEMPt-2 + α4 (DEFLATORt - DEFLATORt-1) + α5 UNEMPt-2 * D0201t
  - Estimation period: 2003-Q1 to 2011-Q4.
  - Estimated coefficients:
    - C: 171.8
    - NSIt-1: 0.761
    - NSIt-4: 0.327
    - UNEMPt-2: -0.147
    - DEFLATORt - DEFLATORt-1: 4.001
    - UNEMPt-2 * D0201t: 0.000
  - Limitations noted:
    - Non-stationarity and likely spurious regression.
    - Loss of precision from using aggregate instead of bank-specific data.

### Satellite Model for Operating and Administrative Expenses
- Fund-staff baseline model (bank-level):
  - Functional form used to project operating and administrative expenses (AE) for each bank:
    - AEit / NSIit = α + β AEit-1 / NSIit-1 + γ Real GDP growth + εit
    - (expressed in source as a model where Real GDP growth and NSI ratios determine AE dynamics)
  - Rationale: wage payroll stickiness vs. NSI flexibility implies AE-to-NSI ratio increases in recessions; excluding this effect biases stress tests.
  - Estimation: one model per bank using annual data for 1994-2012.
  - Comparable structure to UK RAMSI model (RAMSI uses gross fee income in denominator).
  - Recommendation: use quarterly data in future to permit stationarity tests and re-specification.
- Authorities’ baseline model (system-level, quarterly):
  - AE t = C + α1 AE t-1 + α2 AE t-3 + α3 (BANKEMP t-2 - BANKEMP t-3) + α4 (DEFLATOR t - DEFLATOR t-1) + α5 (DEFLATOR t-2 - DEFLATOR t-3)
  - Estimation period: 2003:Q3 to 2011:Q4.
  - Estimated coefficients:
    - C: -133.5
    - AE t-1: 0.734
    - AE t-3: 0.322
    - BANKEMP t-2 - BANKEMP t-3: 0.064
    - DEFLATOR t - DEFLATOR t-1: 19.255
    - DEFLATOR t-2 - DEFLATOR t-3: 12.401
  - Model limitations: non-stationarity, likely spurious, and loss of precision from aggregate data.

### Satellite Model for Sovereign Risk
- Two-step valuation procedure for bond and money market instruments under stress:
  - Step 1 — Shifts in yield curves:
    - Yield y_{m,M,t} for maturity range m and instrument type M is regressed on the 30–60 day time deposit rate i_{t-30,t-60}:
      - y_{m,M,t} = α + C i_{t-30,t-60} + ε
    - Separate yield curves constructed and evaluated for instrument types: nominal in pesos; inflation-adjusted in pesos; linked to the BADLAR rate; and denominated in US dollars.
    - Regression estimated using weekly data during 2006–12.
    - Floating rate instruments linked to the BADLAR rate are treated separately using BADLAR projections.
  - Step 2 — Valuation effects:
    - Re-price instruments by setting new prices equal to present value of future cash flows discounted at new yields.
    - Non-publicly traded instruments (e.g., some LEBACs and NOBACs) priced using yield curves of similar traded instruments.

- Parameters by instrument/maturity (excerpt from source table — In pesos / inflation-adjusted / U.S. dollars):
  - In pesos (maturity ranges 0-2 yrs., 2-6 yrs., >6 yrs.): parameters a2.99, 3.32, 2.56, 1.17, 1.68, 1.28, 0.61 and C: -0.08, -0.21, -0.12, -0.01, -0.06, 0.000, 0.05 (table formatting in source indicates parameter values across maturity bands and instrument types).

### Other Satellite Models
- Less relevant models cover asset categories with low bank exposures.
- Methodology reference: BCRA, Informe al Directorio 736/45/13.

### Appendix III. Argentina: Contributions to the Changes in CAR — Adverse Scenario (V)
- CAR definition and decomposition:
  - CAR = Capital / RWA
  - Change in CAR expressed as:
    - ΔCAR = (1 / RWA_{t+1}) * ΔCapital - (Capital_{t} / RWA_{t+1} * RWA_{t}) * ΔRWA + joint term.
  - For chart construction, the joint effect is added to the contribution of changes in capital.
- Decomposition of capital evolution (assuming no capital injections):
  - ΔCapital = Profit before losses - Losses due to stress - Dividends - Taxes
  - Note: profits before losses include net interest income; net fee and service income; and operating and administrative expenses.
- Sources of change in capital (enumerated for contribution analysis):
  - Profit before losses (excl. govt and BCRA securities)
  - Income from govt. and BCRA securities
  - Credit losses
  - Losses/gains due to market risk (non-sovereign)
  - Losses/gains due to market risk (sovereign)
  - Other (taxes, dividends, ...)

### Appendix IV. Argentina: Liquidity Stress Tests
- Definitions:
  - Run-off rate: fraction of liability maturing in a period that is withdrawn by claim holders.
  - Sight deposits: run-off rate is the fraction of initial outstanding balance withdrawn in a given period due to instantaneous maturity.
- Run-off rates for cash outflows (In percent) — selected entries (columns denote time bands: Up to one day; >1 day–1 week; >1 week–1 month; >1 month–2 months; >2 months–3 months; >3 months–6 months; >6 months–1 year; >1 year–2 years):
  - Own issuances due, short term (initial maturity less than 1 year) row: 60, 60, 40, 40, 40, 40, 40, 20
  - Own issuances due, long term (initial maturity more than 1 year) row: 80, 80, 80, 60, 60, 60, 60, 30
  - Unsecured funding due with legal entities with revenue less than $ 10.000.000:
    - sight deposits: 7.5, 10.5, 10, 0, 0, 0, 0, 0
    - others: 15, 15, 15, 20, 25, 25, 25, 3
  - Unsecured funding due with legal entities with revenue more than $ 10.000.000:
    - sight deposits: 15, 5, 5, 100, 0, 0, 0, 0
    - others: 40, 40, 40, 40, 40, 40, 40, 3
  - Financial institutions regulated by the BCRA: 75, 75, 75, 75, 30, 30, 30, 15
  - Other institutions with financial activities: 75, 75, 75, 75, 30, 30, 30, 15
  - Non-financial public sector: 20, 20, 20, 20, 20, 15, 12, 12
  - Secured funding due to legal entities (all maturity bands): 100 across maturity bands for collateral types listed (bonds and money market instruments issued by the BCRA; government debt instruments with/without publicly quoted prices; secured by other instruments).
  - Repos due with the BCRA: 100 across all maturity bands.
  - Funding due with depositors (physical persons):
    - sight deposits: 7.5, 10.5, 10, 0, 0, 0, 0, 0
    - other: 10, 10, 10, 15, 20, 25, 25, 3
  - Outflows from derivatives: 100 across all maturity bands.
  - Undrawn committed credit/liquidity lines to financial institutions regulated by the BCRA: 10, 20, 50, 75, 100, 100, 100, 100
  - Undrawn lines to fideicomisos: 40, 40, 40, 40, 40, 40, 40, 20
  - Undrawn lines to other financial institutions: 10, 15, 20, 25, 30, 30, 30, 30
  - Undrawn lines to physical persons and legal entities of the non-financial public sector: 1, 2, 3, 4, 5, 5, 5, 5 (table shows digits "12345555" across bands)
  - Others rows include zeros in source table.
- Run-off rates for cash inflows (In percent) — selected entries:
  - New own issuances (already contracted): 100 across all maturity bands.
  - Loans maturing to financial institutions regulated by the BCRA: 100, 100, 90, 70, 50, 50, 50, 25
  - Loans to other institutions with financial activities: 100, 100, 90, 70, 50, 50, 50, 25
  - Loans to others (credit cards): 50, 50, 50, 30, 30, 10, 10, 10
  - Other loans to others: 50, 50, 50, 30, 30, 10, 10, 10
  - Inflows from derivatives: 100 across all maturity bands.
  - Paper in own portfolio maturing: 100 across all maturity bands.
  - Reverse repos secured by BCRA instruments and government debt (with/without publicly quoted prices) and secured by other instruments: 100 across all maturity bands.
  - Volume of available credit/liquidity lines from financial institutions regulated by the BCRA: 10, 20, 50, 75, 100, 100, 100, 100
  - Volume of available lines from other financial institutions: 10, 15, 20, 25, 30, 30, 30, 30
  - Others: 100 across all maturity bands.
- Counterbalancing Capacity (haircuts and availability) — selected entries (In percent and maturity bands):
  - Cash and reserves in the BCRA: 0 (Up to one day), 0 (>1 day–1 week), 100 (remaining bands).
  - Unencumbered BCRA eligible collateral — bonds and money market instruments issued by the BCRA: 5, 15, 30, 50, 100, 100, 100, 100
  - Government (treasury) debt instruments with publicly quoted prices: 15, 25, 30, 50, 100, 100, 100, 100
  - Government debt instruments without publicly quoted prices (not included in VELI): n.a, 30, 30, 50, 100, 100, 100, 100
  - Instruments eligible for VELI — Collateralized loans: n.a, 35, 30, 50, 100, 100, 100, 100
  - Equity: 40 (Up to one day), n.a, 5, 12, 50, 100, 100, 100, 100
  - Other assets accepted by the BCRA as collateral: n.a, 35, 5, 12, 50, 100, 100, 100, 100

*Source: _cr1665 - Appendix II. Argentina: Satellite Models*

### Appendix Table 1. Argentina: Stress Test Matrix

### Appendix Table 1. Argentina: Stress Test Matrix

### Banking sector: Solvency risk — Institutional perimeter and data
- Institutions included:
  - 22 major banks : Banco de la Nación Argentina; Banco de la Provincia de Buenos Aires; Banco de Galicia y Buenos Aires S.A.; Banco Santander Río S.A.; Banco Macro S.A.; BBVA Banco Frances S.A.; HSBC Bank Argentina S.A.; Banco de la Ciudad de Buenos Aires; Banco Credicoop Cooperativo Limitado; Banco Patagonia S.A.; Standard Bank Argentina S.A.; Citibank N.A.; Banco Hipotecario S.A.; Banco de la Provincia de Cordoba S.A.; Banco Supervielle S.A.; Nuevo Banco de Santa Fe S.A.; Banco Itaú Argentina S.A.; Banco de San Juan S.A.; Banco Comafi S.A.; Banco de la Pampa Sociedad de Economía Mixta; Nuevo Banco de Entre Rios S.A.; and Banco Industrial S.A.
- Market share:
  - Ninety percent of total sector assets
- Data and baseline date:
  - September 2012.
  - Supervisory data.
  - Scope of consolidation: Solo.
  - Coverage of sovereign exposures: Trading and banking books including own sovereign, valuated at MTM values.
  - Publicly available data: balance sheet and income statements.
  - Data provided by authorities.

### Banking sector: Solvency risk — Methodology and models
- Methodology:
  - Authorities’ macroeconomic and satellite models, with FSAP team guidance.
  - FSAP team VAR model.
  - Breaking Point Method by Ong, Maino, and Duma (2010).
- Satellite models for macro-financial linkages:
  - Models for credit losses
  - Methodology to calculate net interest income
  - Methodology to calculate losses from bonds and money market instruments (sovereign and other issuers)
  - Models for market risk
  - Expert judgment
  - Model for credit losses

### Banking sector: Solvency risk — Horizon, scenarios, and sensitivities
- Stress test horizon:
  - 2013–14.
- Scenario analysis:
  - Two baseline scenarios: Baseline (authorities) and Adjusted Baseline (WEO).
  - Three adverse scenarios incorporate the external shocks listed in the RAM, including a recession in Brazil caused by a sudden stop of capital inflows to Emerging Markets; and a sharp decline in commodity prices caused by a global recession:
    - (i) a U-shaped adverse scenario based on authorities’ baseline based on a cumulative decline of GDP of 1.7 standard deviations over two years;
    - (ii) a U-shaped adverse scenario based on the WEO baseline; and
    - (iii) a V-shaped adverse scenario based on the WEO baseline.
  - The latter two scenarios result in a cumulative decline of GDP equivalent to 2 standard deviations over two years.
- Sensitivity analysis:
  - Sensitivity analysis to evaluate domestic shocks: (i) tightening of domestic monetary conditions aimed at containing inflation pressures and (ii) a loss of confidence in the monetary and financial system that triggers capital outflows and widens the gap between the parallel and official exchange rates;
  - A failure of a number of large corporate exposures;
  - Sovereign shock based on the 90th percentile of the changes in bond yields.

### Banking sector: Solvency risk — Risks assessed and behavioral adjustments
- Risks/factors assessed:
  - Credit losses.
  - Losses from bonds and money market instruments (sovereign and other issuers) in the banking and trading books.
  - Market risk, including foreign exchange risk.
  - Losses from fixed income holdings/ sovereign.
- Behavioral adjustments:
  - Balance sheet growth in line with nominal GDP
  - Dividends can only be paid out by banks that remain adequately capitalized and subject to BCRA regulation.

### Banking sector: Solvency risk — Calibration, standards, and outputs
- Calibration of risk parameters:
  - Backward-looking rates for specific provisioning.
  - Fixed risk-weights for RWAs.
- Regulatory/Accounting and Market-Based Standards:
  - Regulatory Capital and Tier 1 Capital
  - Standard approach
  - Basel III hurdle rates
- Output presentation:
  - Capital shortfall system wide.
  - For each hurdle rate (or range), percentage of assets that fail.

### Banking sector: Liquidity risk — Institutional perimeter and data
- Institutions included:
  - 22 major banks : Banco de la Nación Argentina; Banco de la Provincia de Buenos Aires; Banco de Galicia y Buenos Aires S.A.; Banco Santander Río S.A.; Banco Macro S.A.; BBVA Banco Frances S.A.; HSBC Bank Argentina S.A.; Banco de la Ciudad de Buenos Aires; Banco Credicoop Cooperativo Limitado; Banco Patagonia S.A.; Standard Bank Argentina S.A.; Citibank N.A.; Banco Hipotecario S.A.; Banco de la Provincia de Cordoba S.A.; Banco Supervielle S.A.; Nuevo Banco de Santa Fe S.A.; Banco Itaú Argentina S.A.; Banco de San Juan S.A.; Banco Comafi S.A.; Banco de la Pampa Sociedad de Economía Mixta; Nuevo Banco de Entre Rios S.A.; and Banco Industrial S.A.
- Market share:
  - 90 percent of total sector assets.
- Data and baseline date:
  - September 2012.
  - Source: Granular data provided by banks for this purpose and supervisory data.
  - Scope of consolidation: solo.

### Banking sector: Liquidity risk — Methodologies and buffers
- Methodology:
  - Three methods:
    - Cash flow-based liquidity stress test using maturity buckets.
    - Macro stress tests: authorities’ macroeconomic and satellite models with FSAP team guidance.
    - Reserve liquidity test as sensitivity tests.
- Risks:
  - Funding liquidity.
  - Market liquidity.
- Buffers:
  - Counterbalancing capacity.
  - Central bank facilities.

### Banking sector: Liquidity risk — Tail shocks, standards, and reporting
- Size of the shock:
  - Bank run and dry up of wholesale funding markets, taking into account haircuts to liquid assets.
  - Run-off rates calculated following historical events and LCR rates.
- Regulatory standards:
  - Liquidity gap, survival period.
  - Basel III draft standards (LCR and NSFR).
- Output presentation:
  - Liquidity gap by bank and currency, aggregated.
  - Survival period in days by bank, number of banks that still can meet their obligations.

### Banking sector: Contagion risk — Perimeter, models, and shocks
- Institutions included:
  - 22 major banks : Banco de la Nación Argentina; Banco de la Provincia de Buenos Aires; Banco de Galicia y Buenos Aires S.A.; Banco Santander Río S.A.; Banco Macro S.A.; BBVA Banco Frances S.A.; HSBC Bank Argentina S.A.; Banco de la Ciudad de Buenos Aires; Banco Credicoop Cooperativo Limitado; Banco Patagonia S.A.; Standard Bank Argentina S.A.; Citibank N.A.; Banco Hipotecario S.A.; Banco de la Provincia de Cordoba S.A.; Banco Supervielle S.A.; Nuevo Banco de Santa Fe S.A.; Banco Itaú Argentina S.A.; Banco de San Juan S.A.; Banco Comafi S.A.; Banco de la Pampa Sociedad de Economía Mixta; Nuevo Banco de Entre Rios S.A.; and Banco Industrial S.A.
  - FGS; mutual funds; insurance companies, retirement funds, brokerages.
- Market share:
  - 90 percent of total sector assets.
- Data and baseline date:
  - December 2012.
  - Source: institutions’ own, supervisory.
  - Scope of consolidation: Solo.

- Methodology for contagion:
  - Network model map using centrality measures for contagion between banks and nonbanks financial institutions.
  - Pure contagion interbank model by Cihak (2007).
- Size of the shock:
  - Pure contagion: default of institutions; market closure; or retrenchment of claims.
- Output presentation:
  - Network Model Map.

### References (selected works cited in the matrix)
- Burrows, Oliver; Learmonth, David, and Jack McKeown, 2012, ”RAMSI: a Top-down Stress-testing Model,” Bank of England, Financial Stability Paper No. 17, September.
- Cihak, Martin, 2007, “Introduction to Applied Stress Testing,” available at http://www.imf.org/external/pubs/ft/wp/2007/wp0759.pdf.
- Collyns, C. and G. Russell Kincaid , 2003, "Managing Financial Crises: Recent Experience and Lessons for Latin America," IMF Occasional Paper No. 217.
- Foglia, Antonella (2009), “Stress Testing Credit Risk: A Survey of Authorities’ Approaches,” International Journal of Central Banking, Vol. 5, No. 3, September.
- Nelson, Charles, and Andrew Siegel, 1987, “Parsimonious Modeling of Yield Curves,” The Journal of Business, Vol. 60, No. 4, October, pp. 473–489.
- Ong, Li Lian; Rodolfo Maino; and Nombulelo Duma (2010), “Into the Great Unknown: Stress Testing with Weak Data,” IMF Working Paper No. 282.
- World Bank, 2012, "The Labor Market Story behind Latin America's Transformation," Office of the Chief Economist for the Latin America and the Caribbean Region, October.

*Source: Appendix Table 1. Argentina: Stress Test Matrix (_cr1665).*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/scr/2016/_cr1665.pdf_