Commonwealth Bank of Australia (CMWAY) CORRESP Correspondence with SEC

28 September 2006

Mr Donald Walker
Senior Assistant Chief Accountant
United States Securities and Exchange Commission
Division of Corporation Finance
100F Street NE
Washington DC 20549

Re: Commonwealth Bank of Australia
Form 20-F for the Fiscal Year Ended June 30, 2005
Filed December 15, 2005
File No: 001-02419

Dear Sirs

We are in receipt of the letter from the staff of the Securities and Exchange Commission, dated September 14, 2006, regarding our annual report on Form 20-F for the financial year ended June 30, 2005. For your convenience, we have included the staff’s comment directly below and have presented our response below the comment.

“Note 48 – Differences between Australian and United States Accounting PrinciplesLoan Impairment Provision, page 220

We read your response to prior Comment 4 in our letter dated June 30, 2006. We infer from your response that you believe, with regard to accounting for your loan impairment provision, that Australian accounting principles and US GAAP are equivalent and that US GAAP and AIFRS are equivalent, but that Australian accounting principles and AIFRS are not equivalent. Please tell us how you determined that US GAAP is consistent with both Australian accounting principles and AIFRS, but that these two bases are not equivalent.

As a related matter, your response indicates that in order to comply with AIFRS, you have made major enhancements and improvements to add more precision to your estimation process. You indicate that the resulting adjustment is a change in accounting principle for AIFRS, but a change in accounting estimate for US GAAP. However, it appears that if you are recording such an adjustment upon the implementation of AIFRS due to this change in accounting principle, and if AIFRS and US GAAP are considered to be equivalent, then recording a similar adjustment for US GAAP purposes should also result from a change in accounting principle. SAB 102, issued July 2001, has very clear and explicit guidance about the development of a systematic methodology for determining the loan loss allowance. Other US GAAP guidance for determining the allowance for loan losses, including the AICPA Audit and Accounting Guide – Depository and Lending Institutions and EITF D-80, has not recently changed. Therefore, the implementation of such an adjustment for purposes of US GAAP reporting suggests that those principles may not have been properly applied in prior years. Please tell us:

• how you concluded that the additional enhancements you made to your process, as outlined in your response to prior comment 4 of our letter dated June 30, 2006, were not required to have been part of your historical US GAAP methodology for the loan loss allowance;

• in a comparative tabular format, how you determined each portion of the US GAAP allowance before and after your 2005 changes; and

• how you determined that this adjustment does not indicate that there was an error in the prior application of US GAAP.”

2

Introduction

The Bank’s approach to loan impairment provisioning has always been to assess the “probable losses” consistent with “management’s best estimate”, based on the best information and methodology available at that time. This is consistent with the US GAAP approach to assessment of the amount of FAS 5 and FAS 114 “probable losses” using “management’s best estimate” under EITF D-80. Similar to other banks, our loan provisioning methodology is constantly being refined. During transition to AIFRS, further enhancements were made to our provisioning process that resulted in refinements to our estimate of loan loss allowance for the financial year ended 30 June 2005.

The Bank takes its US reporting requirements very seriously. The financial statements within the Form 20-F for the financial year ended June 30, 2005 were audited by Ernst & Young, who concurred with our US GAAP treatment of the loan impairment provision. We also engage US legal advisors (Sullivan & Cromwell) and have obtained specialist accounting advice when necessary.

Consistency of previous Australian GAAP, AIFRS and US GAAP

In determining that the loan impairment provisioning accounting principles under US GAAP are consistent with both previous Australian GAAP and AIFRS, the following considerations were taken into account:

Previous Australian GAAP consistency with US GAAP

The Australian GAAP loan impairment provision included a provision for bad debts covering non-identifiable probable losses and latent risks inherent in the overall portfolio of loans, advances and other credit transactions.

The Bank calculated the loan loss provision under Australian GAAP based on loss estimates calculated using historical loss experience adjusted for the possible impact of general economic factors on the lending portfolio.

For reporting periods prior to 30 June 2005, the Australian GAAP loan impairment provision was consistent with the US GAAP allowance for loan losses for the following reasons:

• The US GAAP loan impairment provision is an inherently complex calculation involving a significant degree of management judgment in the use of assumptions and requires management to determine a best estimate. The Bank’s Australian GAAP provisioning approach incorporated the key elements of US GAAP to ensure it was consistent with US GAAP loan impairment provisioning.

• The relevant Australian Accounting Standard (AASB 10 ‘Recoverable Amount of Non-Current Assets’) was very broad in nature requiring only that assets be written down “when its carrying amount is greater than recoverable amount” (AASB 10.5.1). This was broad enough to allow the principles set out within US GAAP to be applied in calculating the previous Australian GAAP provision.

• Specifically, the Australian GAAP model met the primary FAS 5 “probability of loss” test, and no part of the provision related to “general or unspecified business risks” specifically prohibited by that Standard. The model also complied with the documentation, calculation methodology and internal control requirements set out within SAB 102.

AIFRS consistency with US GAAP

Under AIFRS (specifically, AASB 139 ‘Financial Instruments: Recognition and Measurement’, sections 58 to 70 and AG84 to AG93) more specific prescriptive requirements for accounting for loan loss provisioning were incorporated into the accounting standards. One of the requirements was that the Bank was specifically required to show objective evidence of the

3

existence of impairment before any collective provisioning for bad debts could be raised. This requirement led to an improvement in the granularity of our source data (for example, arrears data is now gathered at individual client level rather than portfolio level) and, in turn, a better estimate of our loan loss allowance.

The notion of US GAAP is consistent with AIFRS accounting for loan impairment provisioning. However, US GAAP loan impairment provisions represent best estimates, within a range of probable outcomes, due in part to the high degree of complexity and management judgment involved in the loan loss provisioning process. Supervisory letter SR99-13, issued by the US Federal Reserve, and EITF D-80 both state this recognition.

Given the complexity and subjectivity inherent in arriving at an estimate, it is reasonable that significant differences may arise when re-calculating loan impairment provisions using re-evaluated underlying assumptions and additional data. In developing the Bank’s AIFRS-compliant loan impairment provision, these underlying aspects of the calculation have been enhanced in order to provide the Bank with a loan impairment provision which is a better estimate of an incurred loss model. Key elements of the provision calculation which have improved in this respect were set out in our previous response letter dated 12 July 2006.

In summary, given the literal consistency of US GAAP and AIFRS, and given that the AIFRS loan impairment calculation is a better estimate of US GAAP loan impairment provisioning than Australian GAAP, the Bank has used AIFRS as the basis for the US GAAP calculation for the 30 June 2005 Financial Year and for future periods.

Enhancements made to our loan impairment calculation

The enhancements to our process set out within our response letter dated 12 July 2006 represent refinements to our previous methodology rather than any fundamental change in calculation approach. As explained above, the loan impairment calculation under both Australian GAAP and AIFRS resulted in provision calculations which were in compliance with US GAAP estimates.

Aside from the specific improvements to the calculation made as a result of AIFRS transition, as part of our normal internal processes the Bank regularly reviews the credit risk factors used in the assessment of the loan impairment provision. This can involve extraction of additional data, data cleansing, additional segmentation and a range of other activities, with the aim of continuously improving our estimate of loan impairment provisions.

How we determined each portion of the US GAAP allowance

A table setting out in comparative form the details of the provision/loan loss allowance is attached as Appendix One. The table highlights the specific areas where the Bank has added more precision to its estimation process in order to comply with AIFRS. The overall methodology for assessment of the provision balances, being the application of probabilities of loss and other relevant loss factors to our credit exposures, remains the same among previous Australian GAAP, US GAAP and AIFRS.

Error versus change in estimate

Due to the inherent complexity involved in the calculation of the loan impairment provision, a range of subjective outcomes exists from which management need to determine a best estimate based on the information available at that time. This estimate is by nature subject to continual re-evaluation and re-estimation.

The relevant US GAAP pronouncement (APB Opinion 20) states that:

“Errors in financial statements result from mathematical mistakes, mistakes in the application of accounting principles, or oversight or misuse of facts that existed at the time the financial statements were prepared. In contrast, a change in accounting estimate

4

results from new information or subsequent developments and accordingly from better insight or improved judgment. Thus an error is distinguishable from a change in estimate.”

We believe that the use of AIFRS represents a refinement of the previous Australian GAAP calculation used to support the US GAAP loan provisioning estimate. This refinement was not in response to any mathematical mistakes, mistakes in the application of accounting principles, or oversight or misuse of facts that existed at the time the financial statements were prepared. The refinement did, however, arise due to new information being available that resulted in an improved estimate. As such we believe that the change in accounting for loan impairment provisioning should not be characterised as an error, but as a change in accounting estimate.

Yours faithfully

/s/ Ralph J. Norris
Chief Executive Officer
Commonwealth Bank of Australia

cc.	Amanda Roberts
	Staff Accountant
	United States Securities and Exchange Commission Division of Corporation Finance

APPENDIX ONE

PORTFOLIO SEGMENT	AGAAP/USGAAP	AIFRS/USGAAP
RETAIL SEGMENT
(Principally comprising: Housing Loans; Personal Loans; Credit Cards; and Leasing)	Estimate was based on a five year rolling average (the previous financial year and preceding four years) of portfolio write-offs and arrears	Enhancements to the approach based on segmenting the portfolio into arrears segments — arrears representing a loss trigger. Nil arrears through 30 day intervals to the 180 +days in arrears segment. Arrears segmentation enhances the provisioning outcome as it is more reflective of current portfolio quality as opposed to 5 year average losses driven by increased exposure levels
	The probability of loss (“PoL”) was determined by the addition of the annual losses and arrears position as at year end expressed as a percentage of the portfolio exposure and averaged for the relevant five year period	A probability of loss factor derived from individual account level data for the preceding average term of the portfolio and PoL factors derived for each arrears segment — nil arrears, up to 30 days, up to 60 days etc
	The loss given default (“LGD”) was derived from the annual loss, expressed as a percentage of the average balance of the portfolio, expressed as a percentage relative to the loss probability percentage, averaged for the relevant five year period	The loss given loss factor derived from the same data set after taking account the cost of recovery over time
	The resultant outcome as at the date of assessment of exposure X PoL X LGD was then multiplied by the estimated duration of each portfolio and allowing for the prevailing economic conditions over that period of time.	The resultant outcome of the exposure in each arrears category X PoL X LGL is aggregated together with the application of management judgement, using a statistical confidence interval adjustment, to the relevant portfolio segment after taking into account the relevant portfolio circumstances eg model risk, economic factors, arrears and write-off trends etc.

2

PORTFOLIO SEGMENT	AGAAP/USGAAP	AIFRS/USGAAP
CREDIT RISK RATED SEGMENT (Subject to individual risk rating, principally comprising: Institutional Lending; and Corporate & Business Lending)	Credit exposure for all of a client’s facilities except for the Retail segments mentioned above	The requirements of IFRS are more prescriptive than the generally accepted industry practice that constituted AGAAP.
	Probability of Loss (“PoL”) factors assessed on a regular basis. The PoL factors were derived for a one year probability of loss. The PoL factors used in the June 2005 assessment were derived from 10 years of client default and loss data over the period from December 1994 through to December 2004.	Additional data, cleansing of data, more granularity of data and greater segmentation has been provided via the Basle ll project by enhancing the data sets used to form the basis of new Probability of Loss (“PoL”) factors. The Basle ll data set, from December 1996 to December 2004, has been used for the updated assessment.
	Loss Given Default factors were derived from client level write-offs and recoveries data over the same period.	The same data has been utilised
	The resultant outcome as at the date of assessment of exposure X PoL X LGD was then multiplied by the estimated duration of each portfolio segment — a disaggregation into 10 major industry segments — and allowing for the prevailing economic conditions over that period of time.	Rather than a disaggregation into high level segments and the use of dated estimated durations, a specific ‘trigger’ or loss event has been used at the client level. The ‘trigger’ is a downgrade in credit risk rating of an individual client in the portfolio and a discounted cash flow assessment via Term Probability of Defaults applied to the client facilities. Enhanced data has been made available through the Basel ll development work where more robust PoL factors have been derived. In addition to the identification of client downgrades an assessment of an incurred but not reported, or emergence period, set of loss factors have been derived and applied to the remainder of the portfolio. In both instances allowance has been made for the prevailing economic conditions and other portfolio factors.