## _wp1225 - 1. Assets of Four Major Banks for Selected Countries, 2010

## Source details

**Canonical URL:** [_wp1225 - 1. Assets of Four Major Banks for Selected Countries, 2010](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp1225.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp1225.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp1225.pdf.json)

---

### I. Introduction — overall assessment
- The Australian banking system was resilient during the global financial crisis, attributed in part to intensive supervision and sound regulation.
- The banking sector is profitable with capital above regulatory minimums and is dominated by four major banks (all Australian-owned).
- The four major banks are individually and collectively large relative to the size of the banking system and their combined assets are large relative to GDP.
- Main vulnerabilities identified:
  - Exposure to highly indebted households through residential mortgage lending.
  - Sizable short-term offshore borrowing.
- Mitigating factors:
  - Household debt is high at about 150 percent of disposable income but is held mainly by higher income households.
  - Exposure to high-risk mortgages is small.
  - Prudent lending practices and APRA’s conservative approach in implementing Basel II.
  - Reduced use of short-term offshore wholesale funding through increased deposits and lengthened tenor of funding.
- Conclusions:
  - Four major Australian banks have capital well above regulatory requirements with high quality capital.
  - Headline capital ratios are below the global average for large banks in the IMF sample, but Australia’s conservative Basel II implementation implies headline ratios underestimate capital strength.
  - The four major banks are well-positioned to meet higher Basel III capital requirements and are progressing toward Basel III liquidity standards.
  - Stress tests calibrated on the Irish crisis show banks largely able to withstand sizable residential mortgage shocks, but combining mortgage shocks with corporate losses at global financial crisis peaks would bring the banks’ average total capital ratio below the regulatory minimum.
  - Given high bank concentration and market uncertainty, merits of higher capital requirements for systemically important domestic banks should be considered alongside evolving international standards.

### II. Features of the Australian banking system and concentration
- Market shares and concentration:
  - The four major banks’ assets are around 75 percent of total banking sector assets and 80 percent of the residential mortgage market.
  - Combined assets of the four major banks in Australia are about 180 percent of GDP (end 2010).
  - Relative to the size of the total banking sector, Australia lies in the middle of the distribution for the IMF sample of countries.
- Factors increasing concentration after the crisis:
  - Slower growth of smaller banks reliant on securitization due to reduced funding access.
  - Reduced lending by foreign-owned banks.
  - Acquisitions of two medium-sized banks by larger banks in 2008 (St. George by Westpac; BankWest by Commonwealth Bank of Australia).
- Systemic implications:
  - Any distress among these banks could have a sizable impact on the financial sector and real economy in Australia and New Zealand.
  - The four major Australian banks’ subsidiaries and branches control 90 percent of the assets of New Zealand’s banking sector.
  - Potential market perception of “too big to fail” implies possible fiscal liability.
- Supervisory approach:
  - APRA uses a graduated risk-based approach: Probability and Impact Rating System (PAIRS) and Supervisory Oversight and Response System (SOARS), assigning institutions to undisclosed supervisory categories: normal; oversight; mandated improvement; and restructure.

### III. Asset quality, household exposure and liquidity vulnerabilities
- Asset composition and household exposure:
  - Nonperforming loan ratio is low compared to other advanced countries.
  - Residential mortgages comprised 56 percent of total loans at end-2010.
  - Less than 10 percent of owner-occupiers had mortgages with loan-to-value ratios higher than 80 percent and debt service ratios greater than 30 percent.
  - Household debt distribution (Figure 5): First 3%; Second 7%; Third 18%; Fourth 28%; Fifth 44% — households in the top two income quintiles hold almost three quarters of household debt.
  - Full recourse mortgage lending limits strategic loan defaults.
- Short-term external debt and funding:
  - Short‑term external debt remains sizable at 45 percent of GDP at end-September 2011.
  - The maturity profile has extended, with a greater share maturing in the six-month to one year window.
- Financial soundness indicators (selected, in percent unless stated):
  - Return on assets: ANZ 1.0; NAB 0.8; CBA 1.0; Westpac 1.1.
  - Return on equity: ANZ 16.2; NAB 15.0; CBA 19.5; Westpac 16.0.
  - Tier one capital ratio (Basel II): ANZ 10.9; NAB 9.7; CBA 10.0; Westpac 9.7.
  - Total capital ratio (Basel II): ANZ 12.1; NAB 11.3; CBA 11.7; Westpac 11.5.
  - TCE/Total Assets: ranges around 4.3–5.2 across banks and dates.
  - Past due 90 days plus/total loans: values range 0.4–0.8 across banks and dates.

### IV. Basel II implementation, LGD floor, and impact on measured capital ratios
- APRA conservative measures under Basel II:
  - A 20 percent loss given default (LGD) floor for residential mortgages (above Basel II floor of 10 percent).
  - Higher risk weights for certain residential mortgages under the standardized approach.
  - No introduction of permissible reduced risk weights for retail lending under standardized approach.
  - Until June 2011, advanced approaches subject to a 90 percent floor of the Basel I capital requirement.
  - APRA required banks using advanced approaches to hold capital against interest rate risk in the banking book.
- Consequences for measured capital:
  - Australian banks’ reported LGD rates are higher than many other countries.
  - Headline regulatory ratios (total and Tier 1) for the four major Australian banks are lower than for some other countries, but definitional and conservative risk-weighting differences imply caution in cross-country comparisons.
  - Australian banks tend to hold higher quality capital, reflected in relatively higher rankings in tangible common equity ratios compared with total and Tier 1 capital ratios.

### V. Comparative analysis with Canada: LGD, PD, and capital impacts
- Comparator rationale:
  - Nonperforming housing loan ratios in Australia and Canada have been broadly similar in recent years.
  - Eight banks studied (four Australian, four Canadian) are Fitch-rated AA or AA- and adopted advanced internal ratings-based approaches.
- LGD and capital impact (weighted averages; Table 2):
  - Using current LGD (20.2 percent): Tier 1 capital = 9.4; Total capital = 11.4.
  - Assuming LGD 10 percent: Tier 1 capital = 10.3; Total capital = 12.5.
  - Assuming LGD 15 percent: Tier 1 capital = 9.9; Total capital = 12.0.
  - Assuming average for Canadian 4 large banks' LGD (13.9 percent): Tier 1 capital = 10.0; Total capital = 12.1.
- Key numeric effects:
  - Reducing Australian LGD to the Basel II 10 percent floor would increase the four major Australian banks’ weighted average Tier 1 and total capital ratios by almost 100 basis points, respectively.
  - Lowering LGD to Canada’s four large banks’ average of 13.9 percent would increase Tier 1 and total capital ratios by about 60 basis points, respectively.
- Probability of default (PD) and mortgage insurance:
  - The weighted average PD for the Australian four major banks is 2½ times that of Canada’s three large banks.
  - Almost 70 percent of the four large Canadian banks’ residential mortgages belong to the lowest risk bucket versus 40 percent for the four major Australian banks — Canada’s CMHC mortgage insurance assigns a zero risk weight for regulatory capital purposes.
- Risk-weight comparison:
  - Australian banks’ average risk weight is "almost 2½ times" the average of the Canadian banks.
  - Applying Canadian banks' risk weight to Australian banks would raise total capital ratio by "more than 120 basis points" and Tier 1 capital ratio by "about 100 basis points".
  - Excluding APRA’s interest-rate-in-banking-book Pillar 1 requirement would raise the four large Australian banks’ average Tier 1 and total capital ratios by "about 40 and 50 basis points, respectively."

### VI. Basel III, liquidity, and funding composition in Australia
- Basel III capital proposals and APRA positions:
  - APRA proposed a minimum "4.5 percent" Common Equity Tier 1 ratio and a "6 percent" Tier 1 capital ratio from January 2013.
  - APRA proposed introducing a capital conservation buffer of "2.5 percent" from January 2016.
  - APRA has proposed revisions aligning some Australian practices to Basel III but retaining more conservative treatments in certain areas (e.g., deductions for capitalized expenses and transaction costs).
- Liquidity standards and Australian arrangements:
  - Liquidity Coverage Ratio (LCR) objective: ensure banks have adequate high-quality liquid assets to survive an acute stress scenario lasting one month.
  - In many jurisdictions the LCR will largely be met via government securities; Australia has a somewhat limited supply of government securities.
  - APRA and the Reserve Bank of Australia (RBA) designed an approach allowing banks to establish a committed secured liquidity facility with the RBA to cover any shortfall between holdings of high-quality liquid assets and the LCR requirement.
  - Collateral for the facility includes all assets normally eligible for repurchase transactions with the RBA.
  - Fee for access to this facility: "15 basis points per annum."
- Net Stable Funding Ratio (NSFR) and funding trends:
  - NSFR requires banks to have sufficient stable sources of funding over a 1 year horizon; requirement is that NSFR be above "100 percent".
  - Since the global financial crisis, Australian banks’ funding structure improved: increased retail deposits and long-term wholesale funding; reduced reliance on short-term offshore funding (original maturity basis).
  - Estimates suggest the NSFR has improved for three of the four major Australian banks over the past three years.
  - Estimated NSFRs in 2010 show most banks, including the Australian banks, lie below the "100 percent" benchmark, with Australian banks at or just below the average level.
  - Revised laws permit issuance of covered bonds (October 2011 legislation), which may increase the share of long-term funding.

### VII. Vulnerability to shocks to residential mortgages: scenarios and impacts
- Context and data disclosure:
  - Residential mortgage lending comprises "more than half" of the four major banks’ loans.
  - Pillar 3 disclosures report exposures disaggregated into seven risk categories with reported PD, LGD, and risk weights.
  - Example Westpac (as of September 30, 2011; in millions of Australian dollars; Table 3) exposures and parameters:
    - Corporate: Exposure "92,389"; PD "2.3%"; LGD "45%"; Average Risk Weight "61%"; Risk Weighted Assets "56,792".
    - Business lending: Exposure "60,254"; PD "5.8%"; LGD "32%"; Risk Weight "72%"; RWA "43,661".
    - Small business: Exposure "9,974"; PD "3.9%"; LGD "37%"; Risk Weight "42%"; RWA "4,232".
    - Residential mortgages: Exposure "376,480"; PD "1.5%"; LGD "20%"; Risk Weight "15%"; RWA "56,597".
    - Credit cards: Exposure "17,376"; PD "2.3%"; LGD "78%"; Risk Weight "28%"; RWA "4,884".
    - Other retail: Exposure "9,553"; PD "5.3%"; LGD "66%"; Risk Weight "84%"; RWA "8,029".
    - Sovereign: Exposure "35,034"; PD "0.04%"; LGD "9%"; Risk Weight "4%"; RWA "1,492".
    - Bank: Exposure "26,677"; PD "0.08%"; LGD "54%"; Risk Weight "25%"; RWA "6,627".
    - Total exposure "700,566"; Total RWA "234,857".
- Stress scenarios calibrated to Irish experience:
  - Irish shocks: unemployment rose to "13.6 percent" in 2010 from "4.6 percent" in 2007; housing prices declined "46 percent" from the peak in 2007 through November 2011; high loan-to-value ratios at origination.
  - Scenario construction: assume shares of the three riskiest categories for residential mortgages at the four Australian banks rise to those of the Irish banks in 2010; the share of the next low risk category declines accordingly.
- Scenario results (four large Australian banks combined; amounts in millions of Australian dollars unless otherwise indicated):
  - Baseline (September 2011 Actual):
    - Residential mortgages exposure: "1,204,001".
    - Total exposure: "2,603,910".
    - Residential mortgages PD: "2.0%".
    - Residential mortgages LGD: "20.2%".
    - Residential mortgages Risk weight: "17.0%".
    - Residential mortgages RWA: "205,058".
    - Total RWA: "1,182,705".
    - Tier 1 capital: "119,002".
    - Total capital: "136,074".
    - Provisions: "19,499".
    - Estimated loss: "4,411".
    - Tier 1 (%) "10.1".
    - Total (%) "11.5".
  - Scenario 1 (shares of the 3 highest risk categories at Irish banks’ 2010 levels):
    - Residential mortgages PD: "11.1%".
    - Residential mortgages LGD: "20.3%".
    - Residential mortgages Risk weight: "30.3%".
    - Residential mortgages RWA: "364,223".
    - Total RWA: "1,341,870".
    - Tier 1 capital: "114,863".
    - Total capital: "127,796".
    - Provisions: "19,499".
    - Estimated loss: "27,777".
    - Total loss to capital: "8,278".
    - Tier 1 (%) "8.6".
    - Total (%) "9.5".
    - Estimated impact: Tier 1 capital ratio declines by "1½ percentage points" from 10.1 percent baseline to 8.6 percent.
    - All four banks’ Tier 1 ratios would remain above the regulatory minimum of "4 percent".
  - Scenario 2 (Scenario 1 plus increases of LGD and risk weights by "1½ times"):
    - Residential mortgages PD: "11.1%".
    - Residential mortgages LGD: "30.4%".
    - Residential mortgages Risk weight: "45.4%".
    - Residential mortgages RWA: "546,334".
    - Total RWA: "1,523,981".
    - Tier 1 capital: "107,666".
    - Total capital: "113,908".
    - Provisions: "19,499".
    - Estimated loss: "41,665".
    - Total loss to capital: "22,166".
    - Tier 1 (%) "7.1".
    - Total (%) "7.5".
    - Under Scenario 2, one bank’s total capital ratio is projected to decline to below "6 percent"; other banks’ total capital ratios remain above "8 percent".
    - Note: Such a large increase in LGD is considered unlikely given Australia’s low loan-to-value ratios and estimated house price overvaluation of "10–15 percent".
- Additional stress considerations and tail risks:
  - The exercise does not include shocks to corporate and other lending. Irish banks suffered heavy losses from commercial property lending (commercial property lending was "31 percent" of total loans in Ireland in 2006); average haircut when transferred to NAMA was about "58 percent".
  - Four major Australian banks’ corporate exposures, including commercial property lending, are about "one-quarter" of total bank exposures; commercial property exposures are around "10 percent" of total loans.
  - Takats and Tumbarello (2009) estimated expected losses from corporate sector distress one year ahead at about "6 percent" of banks' loans to the corporate sector during the peak of the global financial crisis.
  - If these "6 percent" corporate-sector losses are applied as a tail-risk shock to corporate exposures under Scenario 1 and Scenario 2:
    - The four banks’ average total capital ratio would decline by "more than 2 percentage points" to about "7 percent" under Scenario 1 and "5¼ percent" under Scenario 2 (below the regulatory minimum).
  - Potential losses from other credit exposures (retail lending, personal loans) are not included in these calculations.
- Stress-test design implications:
  - APRA may consider a more severe downside scenario together with funding risk and a longer risk horizon when conducting future stress testing.
  - Recent FSAP stress-test risk horizons for United Kingdom, Germany, and Netherlands are "five years".
  - Future scenarios should explicitly include funding risk and disruptions, and potential large increases in longer-term real interest rates.
- Overall conclusion from scenarios:
  - The major Australian banks could withstand sizeable shocks to residential mortgages in the two scenario exercises, with Tier 1 ratios remaining above "4 percent" in these specific exercises.
  - However, inclusion of severe corporate-sector shocks and broader credit losses could push average total capital ratios below regulatory minima in severe tail events.

### VIII. Stress scenarios and variant descriptions (2.6 SD from baseline and related variants)
- Scenario labels and outcomes included in source:
  - "2.6 SD from baseline (also consider prolonged slow growh over five years)"
  - "Two SD from baseline"
  - "Two SD from baseline (also consider prolonged slow growh over five years)"
  - Outcome/path text excerpts: "-1 percent in year 1 then V-shaped recovery"; "-3 percent in year 1 then V-shaped recovery"; "- 1 percent in year 2 then 2.7 percent in year 3"; "-5 percent cumulatively over 3 years"; "-1.6 percent in year 1 then a 0.3 percent increase"; "-2.3 percent in year 1 then       V - shapped recovery"; "negative growth (yoy) for nine quarters"
- Standalone numeric entries and SD measures extracted exactly:
  - "-4.8-1.6"
  - "-2.3"
  - "(number of SD from beginning year's outturn) 2/ -2.3-2.4-3.3-3.3-0.2-0.1-2.6"
  - "(number of SD from historical mean) 2/ -2.4-3.5 -2.3 -3.0-2.0-2.3-3.5"
  - "Unemployment 3/910.8"
  - "2.6 SD from baseline 9.715.8"
  - "Two SD from baseline 9.8 12"
  - "(number of SD from beginning year's outturn) 2/ 2.43.21.10.41.81.7"
  - "(number of SD from historical mean) 2/ 0.81.9-0.36/0.91.71.6"
- House price inflation and funding risk indicators (extracted):
  - House price inflation 4/-30-2505 /-20-33-25-14
  - Funding riskYesNoYesYesYesYesNoYes
- Country labels appearing in table context:
  - "Australia"
  - "New ZealandIreland"
- Notes and footnotes preserved exactly:
  - "Sources: Various stress test reports and IMF staff calculations."
  - "1/ The lowest growth rate assumed."
  - "2/ Based on the data from 1981-2005."
  - "3/ The highest umemployment rate assumed."
  - "4/ Cumulative."
  - "5/ House prices in Germany have been flat for more than a decade."
  - "6/ Owing due to double digit unemployment rates from 1982-1997. The average umemployment rate for 2000-05 was 4.3 percent."

### IX. Policy and supervisory recommendations
- Consider a fuller analysis of variances in definitions of eligible capital across jurisdictions to facilitate international comparisons of headline capital ratios.
  - Example referenced: Australia and New Zealand Bank analysis indicating Tier 1 would rise from "10.1 percent" (Australian rules) to "13.5 percent" (UK rules); Westpac analysis showing common equity ratio of "8 percent" (March 2011) would increase to "13 percent" under Canadian rules.
- Continue implementing Basel III higher-quality capital requirements; Australia’s high share of common equity places banks in a relatively good position to meet requirements.
- Use committed RBA secured liquidity facility (fee "15 basis points per annum") as an alternative mechanism to meet LCR where government securities supply is limited.
- APRA to consider:
  - aligning certain conservative Australian practices with Basel III while retaining prudential conservatism where appropriate;
  - conducting stress tests that incorporate longer horizons (e.g., up to "five years"), funding shock scenarios, and larger downside shocks to commodity and house prices and global longer-term interest rates;
  - exploring higher capital requirements for systemically important domestic banks given large market shares and potential fiscal risk arising from "too big to fail" perceptions.
- Consider analysis of appropriate capital requirements over the next year (including stress tests) in the context of the 2012 update of the Financial Sector Stability Assessment with the IMF.

*Source: IMF staff analysis and figures/tables as presented in the chapter "Assets of Four Major Banks for Selected Countries, 2010" of the provided PDF content.*

### 1. Assets of Four Major Banks for Selected Countries, 2010 .............................................. 4

### _wp1225 - 1. Assets of Four Major Banks for Selected Countries, 2010 .............................................. 4

### I. Introduction — overall assessment
- The Australian banking system was resilient during the global financial crisis, attributed in part to intensive supervision and sound regulation.
- The banking sector is profitable with capital above regulatory minimums and is dominated by four major banks (all Australian-owned).
- The four major banks are individually and collectively large relative to the size of the banking system and their combined assets are large relative to GDP.
- Main vulnerabilities identified:
  - Exposure to highly indebted households through residential mortgage lending.
  - Sizable short-term offshore borrowing.
- Mitigating factors cited:
  - Household debt is high at about 150 percent of disposable income but is held mainly by higher income households.
  - Exposure to high-risk mortgages is small.
  - Prudent lending practices and APRA’s conservative approach in implementing Basel II.
  - Reduced use of short-term offshore wholesale funding through increased deposits and lengthened tenor of funding.
- Conclusion:
  - The four major Australian banks have capital well above regulatory requirements with high quality capital.
  - Headline capital ratios are below the global average for large banks in the IMF sample, but Australia’s conservative Basel II implementation implies headline ratios underestimate capital strength.
  - The four major banks are well-positioned to meet higher Basel III capital requirements and are progressing toward Basel III liquidity standards.
  - Stress tests calibrated on the Irish crisis show banks largely able to withstand sizable residential mortgage shocks, but combining mortgage shocks with corporate losses at global financial crisis peaks would bring the banks’ average total capital ratio below the regulatory minimum.
  - Given high bank concentration and market uncertainty, merits of higher capital requirements for systemically important domestic banks should be considered alongside evolving international standards.

### II. Features of the Australian banking system and concentration
- The four major banks’ assets are around 75 percent of total banking sector assets and 80 percent of the residential mortgage market.
- Factors increasing concentration after the crisis:
  - Slower growth of smaller banks reliant on securitization due to reduced funding access.
  - Reduced lending by foreign-owned banks.
  - Acquisitions of two medium-sized banks by larger banks in 2008 (St. George by Westpac; BankWest by Commonwealth Bank of Australia).
- International comparisons:
  - Combined assets of the four major banks in Australia are about 180 percent of GDP (end 2010).
  - Relative to the size of the total banking sector, Australia lies in the middle of the distribution for the IMF sample of countries.
- Systemic implications:
  - Any distress among these banks could have a sizable impact on the financial sector and real economy in Australia and New Zealand.
  - The four major Australian banks’ subsidiaries and branches control 90 percent of the assets of New Zealand’s banking sector.
  - Potential market perception of “too big to fail” implies possible fiscal liability.
- Supervisory approach:
  - APRA uses a graduated risk-based approach: Probability and Impact Rating System (PAIRS) and Supervisory Oversight and Response System (SOARS), assigning institutions to undisclosed supervisory categories: normal; oversight; mandated improvement; and restructure.

### III. Asset quality, household exposure and liquidity vulnerabilities
- Asset quality and household exposure:
  - Banks’ nonperforming loan ratio is low compared to other advanced countries.
  - Residential mortgages comprised 56 percent of total loans at end-2010.
  - Less than 10 percent of owner-occupiers had mortgages with loan-to-value ratios higher than 80 percent and debt service ratios greater than 30 percent.
  - Households in the top two income quintiles hold almost three quarters of household debt (Figure 5: First 3%; Second 7%; Third 18%; Fourth 28%; Fifth 44%).
  - Full recourse mortgage lending limits strategic loan defaults.
- Short-term external debt:
  - Short‑term external debt remains sizable at 45 percent of GDP at end-September 2011.
  - The maturity profile has extended, with a greater share maturing in the six-month to one year window.
- Financial soundness indicators (selected, in percent unless stated):
  - Return on assets: ANZ 1.0; NAB 0.8; CBA 1.0; Westpac 1.1 (sample points Sep-11, Sep-10, Jun-11 etc. reported in Table 1).
  - Return on equity: ANZ 16.2; NAB 15.0; CBA 19.5; Westpac 16.0 (selected reporting dates).
  - Tier one capital ratio (Basel II): ANZ 10.9; NAB 9.7; CBA 10.0; Westpac 9.7 (selected dates).
  - Total capital ratio (Basel II): ANZ 12.1; NAB 11.3; CBA 11.7; Westpac 11.5 (selected dates).
  - TCE/Total Assets: ranges reported around 4.3–5.2 across banks and dates.
  - Past due 90 days plus/total loans: values range 0.4–0.8 across banks and dates.

### IV. Basel II implementation, LGD floor, and impact on measured capital ratios
- APRA’s conservative Basel II implementation highlights:
  - A 20 percent loss given default (LGD) floor for residential mortgages, above the Basel II floor of 10 percent.
  - Higher risk weights for certain residential mortgages under the standardized approach.
  - No introduction of permissible reduced risk weights for retail lending under standardized approach.
  - Until June 2011, capital requirements under advanced approaches remained subject to a 90 percent floor of the Basel I capital requirement (instead of the 80 percent floor applicable in the second year).
  - APRA required banks using advanced approaches to hold capital against interest rate risk in the banking book.
- Consequences:
  - Australian banks’ reported LGD rates are higher than many other countries (Figure 7).
  - Headline regulatory ratios (total and Tier 1) for the four major Australian banks are lower than for some other countries, but differences in calculation and conservative risk-weighting imply caution in cross-country comparisons.
  - Australian banks tend to hold higher quality capital, reflected in relatively higher rankings in tangible common equity ratios compared with total and Tier 1 capital ratios.

### V. Comparative analysis with Canada: LGD, PD, and capital impacts
- Comparator rationale:
  - Nonperforming housing loan ratios in Australia and Canada have been broadly similar in recent years.
  - Eight banks studied (four Australian, four Canadian) are Fitch-rated AA or AA- and adopted advanced internal ratings-based approaches.
- LGD and capital impact (weighted averages; Table 2):
  - Using current LGD (20.2 percent): Tier 1 capital = 9.4; Total capital = 11.4.
  - Assuming LGD 10 percent: Tier 1 capital = 10.3; Total capital = 12.5.
  - Assuming LGD 15 percent: Tier 1 capital = 9.9; Total capital = 12.0.
  - Assuming average for Canadian 4 large banks' LGD (13.9 percent): Tier 1 capital = 10.0; Total capital = 12.1.
- Key numeric comparisons and effects:
  - Reducing Australian LGD to the Basel II 10 percent floor would increase the four major Australian banks’ weighted average Tier 1 and total capital ratios by almost 100 basis points, respectively.
  - Lowering LGD to Canada’s four large banks’ average of 13.9 percent would increase Tier 1 and total capital ratios by about 60 basis points, respectively.
- Probability of default (PD) on residential mortgages:
  - The weighted average PD for the Australian four major banks is 2½ times that of Canada’s three large banks (Figure 14).
  - Mortgage insurance in Canada (CMHC) assigns a zero risk weight for regulatory capital purposes; almost 70 percent of the four large Canadian banks’ residential mortgages belong to the lowest risk bucket versus 40 percent for the four major Australian banks (Figures 15).

### VI. Stress testing results and policy implications
- Stress-test summary:
  - Stress tests calibrated on the Irish crisis experience show banks largely able to withstand sizable shocks to residential mortgage exposures.
  - Combining residential mortgage shocks with corporate losses expected at the peak of the global financial crisis would bring the banks’ average total capital ratio below the regulatory minimum.
- Policy recommendations and considerations:
  - Given high bank concentration and market uncertainty, consider merits of higher capital requirements for systemically important domestic banks.
  - Any consideration of higher capital should be taken into account alongside evolving international standards (Basel III) and complemented by intensive supervision.

*Source: IMF staff analysis and figures/tables as presented in the chapter "Assets of Four Major Banks for Selected Countries, 2010" of the provided PDF content.*

### 0.0 to 0.2      0.2 to 0.5      0.5 to 2.0      2.0 to 10.0   10.0 to 99.9100

### _wp1225 - 0.0 to 0.2      0.2 to 0.5      0.5 to 2.0      2.0 to 10.0   10.0 to 99.9100

### PD range and composition of residential mortgages (Australia and Canada)
- PD bands shown in figures: "0.0 to 0.2", "0.2 to 0.5", "0.5 to 2.0", "2.0 to 10.0", "10.0 to 99.9100".
- Canada (October 2010) sample: four large banks (BMO, CIBC, Scotiabank, and TD Bank).
- Australia (September 2010) sample: four large banks (ANZ, CBA, NAB, and Westpac). For CBA, data for December 2010.
- Sources for PD-range composition: Banks' disclosure statements; IMF staff estimates.

### Risk weights and capital ratios: Australia versus Canada
- Australian banks’ average risk weight is "almost 2½ times" the average of the Canadian banks (Figure 18).
- Applying Canadian banks' risk weight to Australian banks would raise:
  - total capital ratio by "more than 120 basis points",
  - Tier 1 capital ratio by "about 100 basis points" (Figure 19).
- Regulatory and definitional differences affecting comparisons:
  - jurisdictions apply different approaches to definitions of eligible capital, Pillar 1 risk-weighted assets, and capital limits;
  - regulators’ supervisory review of banks’ internal capital adequacy assessment influences capital levels.
- APRA specific note:
  - APRA requires interest rate risk in the banking book to be a Pillar 1 capital requirement in addition to credit, market, and operational risks.
  - If this APRA requirement is excluded, the four large Australian banks’ average Tier 1 and total capital ratios are estimated to rise by "about 40 and 50 basis points, respectively."

### Basel III, liquidity, and funding composition in Australia
- Basel III will require banks to hold more and higher-quality capital.
- APRA observations and proposals:
  - APRA proposed a minimum "4.5 percent" Common Equity Tier 1 ratio and a "6 percent" Tier 1 capital ratio from January 2013.
  - APRA proposed introducing a capital conservation buffer of "2.5 percent" from January 2016.
  - APRA has proposed revisions aligning some Australian practices to Basel III but retaining more conservative treatments in certain areas (e.g., deductions for capitalized expenses and transaction costs).
- Basel III liquidity standards introduced:
  - Liquidity Coverage Ratio (LCR) objective: ensure banks have adequate high-quality liquid assets to survive an acute stress scenario lasting one month.
  - In many jurisdictions the LCR will largely be met via government securities; Australia has a somewhat limited supply of government securities.
- Australian approach to liquidity:
  - APRA and the Reserve Bank of Australia (RBA) designed an approach allowing banks to establish a committed secured liquidity facility with the RBA to cover any shortfall between holdings of high-quality liquid assets and the LCR requirement.
  - Collateral for the facility includes all assets normally eligible for repurchase transactions with the RBA.
  - Fee for access to this facility: "15 basis points per annum."
- Net Stable Funding Ratio (NSFR):
  - NSFR requires banks to have sufficient stable sources of funding over a 1 year horizon; differing weights are applied to balance-sheet components and the requirement is that NSFR be above "100 percent".
  - Since the global financial crisis, Australian banks’ funding structure improved: increased retail deposits and long-term wholesale funding; reduced reliance on short-term offshore funding (original maturity basis) (Figure 20).
  - Estimates suggest the NSFR has improved for three of the four major Australian banks over the past three years (Figure 21).
  - Estimated NSFRs in 2010 show most banks, including the Australian banks, lie below the "100 percent" benchmark, with Australian banks at or just below the average level (Figure 22).
  - Revised laws permit issuance of covered bonds (October 2011 legislation), which may increase the share of long-term funding.

### Vulnerability to shocks to residential mortgages: scenarios and impacts
- Context:
  - Residential mortgage lending comprises "more than half" of the four major banks’ loans.
  - Pillar 3 disclosures report exposures disaggregated into seven risk categories with reported PD, LGD, and risk weights.
  - Example Westpac (as of September 30, 2011; in millions of Australian dollars; Table 3):
    - Corporate: Exposure "92,389"; Probability of Default "2.3%"; Loss Given Default "45%"; Average Risk Weight "61%"; Risk Weighted Assets "56,792".
    - Business lending: Exposure "60,254"; PD "5.8%"; LGD "32%"; Risk Weight "72%"; RWA "43,661".
    - Small business: Exposure "9,974"; PD "3.9%"; LGD "37%"; Risk Weight "42%"; RWA "4,232".
    - Residential mortgages: Exposure "376,480"; PD "1.5%"; LGD "20%"; Risk Weight "15%"; RWA "56,597".
    - Credit cards: Exposure "17,376"; PD "2.3%"; LGD "78%"; Risk Weight "28%"; RWA "4,884".
    - Other retail: Exposure "9,553"; PD "5.3%"; LGD "66%"; Risk Weight "84%"; RWA "8,029".
    - Sovereign: Exposure "35,034"; PD "0.04%"; LGD "9%"; Risk Weight "4%"; RWA "1,492".
    - Bank: Exposure "26,677"; PD "0.08%"; LGD "54%"; Risk Weight "25%"; RWA "6,627".
    - Other: RWA "72,829"; Risk Weighted Assets total "52,543".
    - Total exposure "700,566"; Total RWA "234,857".
- Scenarios calibrated to Irish banks’ experience:
  - Irish context: unemployment rose to "13.6 percent" in 2010 from "4.6 percent" in 2007; housing prices declined "46 percent" from the peak in 2007 through November 2011; high loan-to-value ratios at origination (Figure 23).
  - Scenario construction: assume shares of the three riskiest categories for residential mortgages at the four Australian banks rise to those of the Irish banks in 2010; the share of the next low risk category declines accordingly (Table 4).
- Scenario results (Table 5; summary for four large Australian banks combined; amounts in millions of Australian dollars unless otherwise indicated):
  - Baseline (September 2011 Actual):
    - Residential mortgages exposure: "1,204,001".
    - Total exposure: "2,603,910".
    - Residential mortgages PD: "2.0%".
    - Residential mortgages LGD: "20.2%".
    - Residential mortgages Risk weight: "17.0%".
    - Residential mortgages RWA: "205,058".
    - Total RWA: "1,182,705".
    - Tier 1 capital: "119,002".
    - Total capital: "136,074".
    - Provisions: "19,499".
    - Estimated loss: "4,411".
    - Tier 1 (%) "10.1".
    - Total (%) "11.5".
  - Scenario 1 (shares of the 3 highest risk categories at Irish banks’ 2010 levels):
    - Residential mortgages PD: "11.1%".
    - Residential mortgages LGD: "20.3%".
    - Residential mortgages Risk weight: "30.3%".
    - Residential mortgages RWA: "364,223".
    - Total RWA: "1,341,870".
    - Tier 1 capital: "114,863".
    - Total capital: "127,796".
    - Provisions: "19,499".
    - Estimated loss: "27,777".
    - Total loss to capital: "8,278".
    - Tier 1 (%) "8.6".
    - Total (%) "9.5".
    - Estimated impact: Tier 1 capital ratio declines by "1½ percentage points" from 10.1 percent baseline to 8.6 percent.
    - All four banks’ Tier 1 ratios would remain above the regulatory minimum of "4 percent".
  - Scenario 2 (Scenario 1 plus increases of LGD and risk weights by "1½ times"):
    - Residential mortgages PD: "11.1%".
    - Residential mortgages LGD: "30.4%".
    - Residential mortgages Risk weight: "45.4%".
    - Residential mortgages RWA: "546,334".
    - Total RWA: "1,523,981".
    - Tier 1 capital: "107,666".
    - Total capital: "113,908".
    - Provisions: "19,499".
    - Estimated loss: "41,665".
    - Total loss to capital: "22,166".
    - Tier 1 (%) "7.1".
    - Total (%) "7.5".
    - Under Scenario 2, one bank’s total capital ratio is projected to decline to below "6 percent"; other banks’ total capital ratios remain above "8 percent".
    - Note: Such a large increase in LGD is considered unlikely given Australia’s low loan-to-value ratios and estimated house price overvaluation of "10–15 percent".
- Primary driver of capital reductions: downward internal ratings migration increasing measured RWA and capital requirements.
- Additional stress considerations:
  - The exercise does not include shocks to corporate and other lending. Irish banks suffered heavy losses from commercial property lending (commercial property lending was "31 percent" of total loans in Ireland in 2006); average haircut when transferred to NAMA was about "58 percent".
  - Four major Australian banks’ corporate exposures, including commercial property lending, are about "one-quarter" of total bank exposures; commercial property exposures are around "10 percent" of total loans (well below Irish banks’ "31 percent").
  - Takats and Tumbarello (2009) estimated expected losses from corporate sector distress one year ahead at about "6 percent" of banks' loans to the corporate sector during the peak of the global financial crisis.
  - If these "6 percent" corporate-sector losses are applied as a tail-risk shock to corporate exposures under Scenario 1 and Scenario 2:
    - The four banks’ average total capital ratio would decline by "more than 2 percentage points" to about "7 percent" under Scenario 1 and "5¼ percent" under Scenario 2 (below the regulatory minimum).
  - Potential losses from other credit exposures (retail lending, personal loans) are not included in these calculations.
- Stress-test design implications:
  - APRA may consider a more severe downside scenario together with funding risk and a longer risk horizon when conducting future stress testing.
  - Recent FSAP stress-test risk horizons for United Kingdom, Germany, and Netherlands are "five years" (Table 6).
  - The Irish crisis showed severe shocks for a longer period than the stress test assumptions of 2006 (Figure 25).
  - Future scenarios should explicitly include funding risk and disruptions, and potential large increases in longer-term real interest rates.
- Overall conclusion from scenarios:
  - The major Australian banks could withstand sizeable shocks to residential mortgages in the two scenario exercises, with Tier 1 ratios remaining above "4 percent" in these specific exercises.
  - However, inclusion of severe corporate-sector shocks and broader credit losses could push average total capital ratios below regulatory minima in severe tail events.

### Policy and supervisory recommendations
- Consider a fuller analysis of variances in definitions of eligible capital across jurisdictions to facilitate international comparisons of headline capital ratios.
  - Example results referenced: Australia and New Zealand Bank analysis indicating Tier 1 would rise from "10.1 percent" (Australian rules) to "13.5 percent" (UK rules); Westpac analysis showing common equity ratio of "8 percent" (March 2011) would increase to "13 percent" under Canadian rules.
- Continue implementing Basel III higher-quality capital requirements; Australia’s high share of common equity places banks in a relatively good position to meet requirements.
- Use committed RBA secured liquidity facility (fee "15 basis points per annum") as an alternative mechanism to meet LCR where government securities supply is limited.
- APRA to consider:
  - aligning certain conservative Australian practices with Basel III while retaining prudential conservatism where appropriate;
  - conducting stress tests that incorporate longer horizons (e.g., up to "five years"), funding shock scenarios, and larger downside shocks to commodity and house prices and global longer-term interest rates;
  - exploring higher capital requirements for systemically important domestic banks given large market shares and potential fiscal risk arising from "too big to fail" perceptions.
- Consider analysis of appropriate capital requirements over the next year (including stress tests) in the context of the 2012 update of the Financial Sector Stability Assessment with the IMF.

*Sources: Banks' disclosure statements; APRA; RBA; IMF staff estimates and calculations.*

### 2.6 SD from baseline

### 2.6 SD from baseline

### Stress scenarios and variant descriptions
- Scenario labels:
  - "2.6 SD from baseline (also consider prolonged slow growh over five years)"
  - "Two SD from baseline"
  - "Two SD from baseline (also consider prolonged slow growh over five years)"
- Outcome / path descriptions included in source:
  - "-1 percent in year 1 then V-shaped recovery"
  - "-3 percent in year 1 then V-shaped recovery"
  - "- 1 percent in year 2 then 2.7 percent in year 3"
  - "-5 percent cumulatively over 3 years"
  - "-1.6 percent in year 1 then a 0.3 percent increase"
  - "-2.3 percent in year 1 then       V - shapped recovery"
  - "negative growth (yoy) for nine quarters"

### Numeric shock magnitudes and SD measures
- Standalone numeric entries from scenario table:
  - "-4.8-1.6"
  - "-2.3"
  - "(number of SD from beginning year's outturn) 2/ -2.3-2.4-3.3-3.3-0.2-0.1-2.6"
  - "(number of SD from historical mean) 2/ -2.4-3.5 -2.3 -3.0-2.0-2.3-3.5"
  - "Unemployment 3/910.8"
  - "2.6 SD from baseline 9.715.8"
  - "Two SD from baseline 9.8 12"
  - "(number of SD from beginning year's outturn) 2/ 2.43.21.10.41.81.7"
  - "(number of SD from historical mean) 2/ 0.81.9-0.36/0.91.71.6"

### House price inflation and funding risk indicators
- House price inflation (cumulative) entries:
  - "House price inflation 4/-30-2505 /-20-33-25-14"
- Funding risk flags:
  - "Funding riskYesNoYesYesYesYesNoYes"
- Country labels appearing in table context:
  - "Australia"
  - "New ZealandIreland"

### Notes and footnotes extracted exactly as in source
- "Sources: Various stress test reports and IMF staff calculations."
- "1/ The lowest growth rate assumed."
- "2/ Based on the data from 1981-2005."
- "3/ The highest umemployment rate assumed."
- "4/ Cumulative."
- "5/ House prices in Germany have been flat for more than a decade."
- "6/ Owing due to double digit unemployment rates from 1982-1997. The average umployment rate for 2000-05 was 4.3 percent."

*Source: _wp1225 - 2.6 SD from baseline*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp1225.pdf_