Via Edgar*
October 3, 2012
United States Securities and Exchange Commission
Division of Corporation Finance
100 F Street, NE
Washington, D.C. 20549
| Attention: | Suzanne Hayes |
| Assistant Director |
| Division of Corporation Finance |
| Re: | The Goldman Sachs Group, Inc. |
| Form 10-K for the Fiscal Year Ended December 31, 2011 |
| Filed February 28, 2012 |
| Form 10-Q for the Quarterly Period Ended June 30, 2012 |
| Filed August 9, 2012 (“June 2012 Form 10-Q”) |
| File No. 001-14965 |
Dear Ms. Hayes:
We are in receipt of the letter, dated September 5, 2012, to Lloyd C. Blankfein, Chairman and Chief Executive Officer of The Goldman Sachs Group, Inc. (the “Company”), from the staff (the “Staff”) of the Securities and Exchange Commission (the “Commission”), regarding the above-referenced filings. We appreciate the Staff’s review of our filings and look forward to working with the Staff to resolve the Staff’s comments. For your convenience, we have included the Staff’s comments below, followed by our responses.
| * | Certain confidential portions of this letter were omitted by means of redacting a portion of the text. The word “[Redacted]” has been inserted in place of the portions so omitted. Copies of the letter containing the redacted portions have been filed separately with the Commission subject to a request for confidential treatment pursuant to Rule 83 of the Commission’s Rules on Information and Requests. |
Form 10-Q for the Quarterly Period Ended June 30, 2012
Notes to Condensed Consolidated Financial Statements
Note 5 – Fair Value Measurements, page 13
| 1. | In the tabular presentation disclosed on page 14, you present total financial liabilities at fair value on a net basis. To enhance the transparency of your disclosure, please present total financial liabilities at fair value similar to your presentation of total financial assets at fair value by disclosing total financial liabilities at fair value by gross basis then a separate line item showing the effect of any counterparty netting and/or cash collateral. Refer to ASC 820-10-50-3. |
| Response: |
| In accordance with ASC 820-10-50-3, we disclosed in the table on page 30 of our June 2012 Form 10-Q gross derivative liabilities and the effects of any counterparty and cash collateral netting. In future filings, we will also expand the table referenced in your comment to present total financial liabilities at fair value on a similar basis to our presentation of financial assets at fair value, with a separate line item showing the effect of counterparty netting and cash collateral. |
Note 6 – Cash Instruments, page 15
Level 3 Cash Instruments, page 15
| 2. | You disclose on pages 17, 29, and 41–42 the range of significant unobservable inputs for each major category of assets and liabilities measured at fair value on a recurring basis with a significant Level 3 balance, as well as qualitative information on the sensitivity of the fair value measurements to changes in the significant unobservable inputs. Given the wide range of assumptions for several of the categories, please consider revising your disclosure here and the section on page 29 in future filings to also provide a weighted average of the significant unobservable inputs reported, similar to the illustration provided in ASC 820-10-55-103, and state your basis for calculating the weighted average (e.g., weighted average by notional, principal, etc.). |
Response: |
| In future filings, for level 3 cash instruments we will disclose a weighted average of the significant unobservable inputs reported (consistent with the illustration included in ASC 820-10-55-103) and state our basis for calculating the weighted average. The basis for calculating the weighted average will be the fair value of the instruments, with each input equally weighted based on a financial instrument’s total fair value. |
| We do not believe disclosing weighted averages of significant unobservable inputs to level 3 derivative financial instruments is meaningful for the following reasons: |
| · | Weighting by notional value or quantity will distort the average because notional values are not directly linked to fair value (e.g., a transaction with a large notional value might have a fair value of $0). The resulting averages will be skewed by these large notional values and would not accurately represent the range of inputs used to value the underlying derivative portfolio. |
| · | Weighting by fair value creates variability in the average from period to period, despite the fact that level 3 inputs may not have changed, because of the significant impact changes in level 1 and level 2 inputs have on derivative valuations. For example, the valuation of derivatives is dependant on observable interest rates (which are level 2 inputs) used for discounting cash flows. |
| · | We also note that the illustration included in ASC 820-10-55-103 did not provide weighted averages for inputs to derivatives. |
2
| 3. | Further, in your future filings, please consider revising your fair value methodology tables, here and elsewhere in your filing as appropriate, quantify the dollar amount of financial instruments for which fair value is determined for each methodology listed. We believe such disclosure is helpful in allowing the reader and improves transparency. |
Response: |
| In future filings, we will add the level 3 balance to our fair value methodology tables for cash instruments and derivatives. |
Transfers Between Levels of the Fair Value Hierarchy, page 19
| 4. | We note during the three and six months ended June 30, 2012 transfers of cash instruments between Level 1 and 2 primarily reflected transfers to Level 2 of certain non-U.S. government obligations and certain public equity investments. Please revise to more clearly provide the disclosures required by ASC 820-10-50-2bb as it relates to a separate discussion of the transfers into and out of each level. In addition, to enhance the transparency of your disclosure, please quantify the amount when discussing the transfers in and out of Levels 1 and 2. For example, we believe it would be beneficial to investors to know the amount of non-U.S. government obligations and public entity investments that were transferred to Level 2 for the period ending June 30, 2012. |
Response: |
| In future filings, we will separately disclose total transfers into and out of levels 1 and 2 and will describe and quantify any significant transfers. The following is an example of what this disclosure would have been in our June 2012 Form 10-Q: |
| During the three months ended June 2012, transfers into level 2 from level 1 of cash instruments were $1.87 billion, primarily including transfers of non-U.S. government obligations of $1.81 billion, reflecting the level of market activity in these instruments. Transfers into level 1 from level 2 of cash instruments were $386 million, primarily including transfers of equity securities of $169 million, where the firm was able to obtain quoted prices for identical instruments, and transfers of non-U.S. government obligations of $164 million, reflecting the level of market activity in these instruments. |
| During the six months ended June 2012, transfers into level 2 from level 1 of cash instruments were $2.16 billion, primarily including transfers of non-U.S. government obligations of $1.37 billion, reflecting the level of market activity in these instruments, and transfers of public equity investments of $792 million, reflecting the impact of transfer restrictions. Transfers into level 1 from level 2 of cash instruments were $382 million, primarily including transfers of non-U.S. government obligations of $205 million, reflecting the level of market activity in these instruments, and transfers of equity securities of $156 million, where the firm was able to obtain quoted prices for identical instruments. |
3
Note 7 – Derivatives and Hedging Activities, page 25
Valuation Techniques for Derivatives, page 27
| 5. | We note that you value your Level 2 exchange-traded derivatives using models that calibrate to market-clearing levels of OTC derivatives. Furthermore, you discuss on page 28 that Level 3 OTC derivatives are valued using models which utilize observable Level 1 and/or Level 2 inputs, as well as unobservable inputs. In future filings, please expand your disclosure to specifically discuss the valuation methods used in valuing these instruments. Refer to ASC 820-10-50-2bbb. |
Response: |
| In future filings, we will expand our disclosure to specifically describe the valuation methods used in valuing our derivative instruments. These valuation methods include derivative pricing models (e.g., models that incorporate option pricing methodologies, Monte Carlo simulations and discounted cash flows). |
Note 8 – Fair Value Option, page 40
Fair Value of Other Financial Assets and Financial Liabilities by Level, page 42
| 6. | Please revise your future filings to more clearly provide the disclosures required by ASC 820-10-50-2bb as it relates to transfers between Level 1 and Level 2 of the fair value hierarchy. To the extent no such transfers occurred at the end of the reporting period, please disclose that fact in future filings. |
Response: |
| No transfers of other financial assets and financial liabilities between level 1 and level 2 occurred during the three or six month periods ended June 30, 2012. In future filings, we will disclose the amount of transfers of other financial assets and financial liabilities between level 1 and level 2 of the fair value hierarchy or we will disclose that no such transfers occurred. |
4
Note 13 – Goodwill and Identifiable Intangible Assets, page 62
| 7. | You disclose at the bottom of page 62 that, for purposes of your goodwill impairment test, the net book value of each reporting unit reflects the estimated amount of shareholders’ equity required to support the activities of the reporting unit. Please address the following regarding this disclosure in your response as well as your future filings: |
| · | To the extent the total capital allocated to your reporting units is more or less than the total shareholders’ equity of the company, please explain how you account for any difference and quantify such amount. In this regard, clarify whether the carrying value assigned to your reporting units is always equal to the risk capital determined for each reporting unit, or whether any excess or deficit of total shareholders’ equity of the Company over the total calculated risk capital amounts for each of the reporting units is allocated to the reporting units during certain periods, and if so, how. |
Response: |
| We are in the process of performing the goodwill impairment test for 2012, using a quantitative test. For purposes of this test, we derive the net book value of each reporting unit by estimating the amount of shareholders’ equity required to support the activities of each reporting unit, which includes a full allocation of shareholders’ equity. To the extent that the amount of shareholders’ equity required to support the activities of each reporting unit (risk capital) is not equivalent to the shareholders’ equity of the Company, the excess or deficit is allocated to all reporting units on a pro-rata basis. In prior years, our goodwill was not impaired even with a full allocation of shareholders’ equity. |
| · | Expand your disclosure in future filings to more clearly discuss how the carrying values of your reporting units are determined. Additionally, to the extent that you do not determine the individual assets and liabilities of your reporting units for purposes of determining the carrying value of the reporting unit, please expand your discussion of how you perform Step 2 of the goodwill impairment test under your methodology when you are required to do so. |
Response: |
| We will clarify in future filings that the carrying values of our reporting units are determined by estimating the amount of shareholders’ equity required to support the activities of each reporting unit under recent guidelines issued by the Basel Committee. If, in the future, the estimated fair value of a reporting unit is less than its estimated net book value, we will perform Step 2 of the goodwill impairment test to measure the amount of impairment, if any. In order to test for impairment, we will determine the individual assets and liabilities of the relevant reporting units and we will make the disclosures required by ASC 350 regarding such determination. |
5
Note 18 – Commitments, Contingencies and Guarantees, page 71
Commercial Lending, page 71
| 8. | You disclose that Sumitomo Mitsui Financial Group, Inc. (SMFG) provides the company with credit loss protection on certain approved loan commitments, primarily investment-grade commercial lending commitments. You also disclose that you use other financial instruments to mitigate credit risks related to certain commitments not covered by SMFG. Please revise your future filings to clarify the type of financial instruments used to mitigate credit risk on commitments not covered by SMFG. |
Response: |
| In future filings, we will disclose the type of other financial instruments used to mitigate credit risks related to certain commitments not covered by SMFG. These instruments primarily include credit default swaps that reference the same or similar underlying instrument or entity or credit default swaps that reference a market index. |
Management’s Discussion and Analysis, page 112
Critical Accounting Policies, page 117
Fair Value Hierarchy, page 117
| 9. | You disclose that appropriate valuation adjustments are made to the fair value for certain Level 2 and Level 3 financial assets and financial liabilities for a variety of factors including liquidity. Tell us the instruments for which you have made liquidity adjustments, the amount of such adjustments, your basis for the adjustment, and how you considered guidance in ASC 820-10-35-36B with respect to these adjustments. |
Response: |
[Redacted] |
| Approximately 25% of these LVAs are to account for certain characteristics of individual instruments as required by ASC 820-10-35-36B, specifically: i) a legally enforceable restriction on sale or transfer of the instrument and/or ii) a lack of marketability of the instrument as compared to more observable liquid market benchmarks. |
| The remaining LVAs relate to financial assets and financial liabilities that are managed as groups or portfolios as permitted by ASC 820-10-35-18D. This includes groups of derivatives and structured notes with embedded derivatives and their related hedges. |
6
Controls Over Valuation of Financial Instruments, page 117
| 10. | We note your policy of determining fair value, and the disclosures here regarding your control infrastructure that is “independent of the revenue producing units and is fundamental to ensuring that all of (y)our financial instruments are appropriately valued at market clearing levels.” Please tell us, and disclose in future filings, the percentage of prices that are verified as well as whether the independent verification process is only for Level 3 instruments or for all fair value measurements. |
| Response: |
| We will clarify in future filings that all of the firm’s financial instruments at fair value are subject to the price verification process, regardless of level in the fair value hierarchy. Level 1 financial instruments (approximately 17% of total assets as of June 30, 2012) are valued using quoted market prices for identical unrestricted instruments in active markets. Level 2 financial instruments (approximately 58% of total assets as of June 30, 2012) are verified to quoted prices, recent trading activity for identical or similar instruments, broker or dealer quotations or alternative pricing sources with reasonable levels of price transparency. Level 3 financial instruments (approximately 5% of total assets as of June 30, 2012) have one or more significant valuation inputs that are not observable and therefore do not meet the criteria of level 2. For substantially all of our level 3 financial instruments, we typically use one of the following price verification strategies: trade comparison, external price comparison, calibration to market comparables, relative value analyses, collateral analyses and backtesting. In future filings, we will add the following disclosure: |
“All financial instruments at fair value in levels 1, 2 and 3 of the fair value hierarchy are subject to our independent price verification process.” |
Value-at-Risk, page 164
| 11. | You disclose here that your Value-at-Risk (VaR) model is regularly reviewed and enhanced in order to incorporate changes in the composition of inventory positions, as well as variations in market conditions. Please respond to the following and expand your disclosure in future filings as appropriate: |
| · | Tell us the number of different VaR models that are used to determine your total trading VaR, and discuss the drivers regarding the need to use multiple different models. |
| · | Tell us how all of the different VaR models used are aggregated to arrive at your total trading VaR as well as the individual market risk categories disclosed. For example, clarify whether you simply aggregate the outputs from the different models or whether adjustments are made, and if so, how the adjustments are determined. |
| · | Tell us whether the VaR models used for regulatory capital purposes are the same as the VaR models used for your market risk disclosures. To the extent that certain of the models used for both purposes differ, please tell us the drivers behind those differences. |
| · | Discuss the process and validation procedures in place prior to implementing significant model and assumption changes. For example, discuss the approval process required, back-testing procedures performed, and periods of parallel model runs before implementation. |
| · | To the extent that all or some of your VaR models used for your market risk disclosures are different than those used to calculate regulatory capital, please tell us whether the model review process and model oversight processes are the same for both. As part of your response, please clarify when approval is required from any of your regulators regarding VaR model changes. |
| · | Given that your VaR models are evolving over time, tell us how you consider when disclosure is required under Item 305(a)(1)(iii)(4) of Regulation S-K regarding model, assumptions and parameter changes. |
7
Response: |
| Our responses to your comments are as follows: |
| · | We use a single VaR model. |
| · | Since we use a single VaR model, no aggregation process is required. |
| · | The VaR model used for regulatory capital purposes is the same as the VaR model used for our market risk disclosures but reflects appropriate differences (e.g., differences in parameters, as VaR used for regulatory capital purposes is calculated at a 99% confidence level over a ten-day time horizon while we disclose our firmwide VaR at a 95% confidence level and a 1-day time horizon). |
| · | Prior to implementing significant model or assumption changes that impact VaR, we perform model validation processes and test runs. Significant changes to our VaR model are also reviewed with the Chief Risk Officer and Chief Financial Officer and approved by the Firmwide Risk Committee. |
| · | The VaR model used for regulatory capital purposes is the same as the VaR model used for market risk disclosures. We notify our regulators (e.g., the Federal Reserve Board and the U.K.’s Financial Services Authority) of model changes in accordance with specific guidelines provided by these regulators. We also inform the Broker Dealer Risk Office within the Trading and Markets Division of the SEC of such changes. |
| · | We evaluate the impact of changes to our VaR model, including assumptions and parameters. If such changes are material to our firmwide VaR, we would disclose the changes in accordance with Item 305(a)(4) of Regulation S-K. |
| In future filings, we will add the following disclosure: |
| “We use a single VaR model. Prior to implementing significant changes to our VaR assumptions and/or model, we perform model validation and test runs. Significant changes to our VaR model are reviewed with the Chief Risk Officer and Chief Financial Officer and approved by the Firmwide Risk Committee.” |
Stress Testing, page 164
| 12. | We note your response to prior comment 2 from our letter dated April 27, 2012. In response to this comment, you have revised your disclosures on pages 164 and 169 to include specific examples of your indirect risk identified and considered in your stress scenarios that model a sovereign debt default. While you disclose your use of stress tests in various pages throughout your filing, we believe that a more transparent disclosure of your portfolio reviews and stress testing related to such indirect exposure would be useful to readers. Therefore, please revise your future filings to address the following: |
| · | More clearly disclose the nature of the portfolio reviews and stress testing you conducted related to your indirect exposures to these countries. |
| · | More clearly disclose the most significant parameters of your stress tests in order to give the reader a better understanding of your focus in this area. Identify the key risks that your reviews and stress tests were designed to evaluate. For example, discuss the extent to which you evaluated the likelihood and implications of one or more countries exiting the Eurozone, including redenomination and revaluation risk. |
| · | To the extent that redenomination and revaluation risk were considered in your reviews and stress tests, please clearly disclose how you considered the implications of such risks. Consider providing quantification of the assets and liabilities in these countries that are subject to redenomination risk. |
| · | Discuss the most significant implications of your indirect exposure as identified by your reviews and stress tests, and discuss how you are managing these risks. |
8
Response: |
| In future filings, we will add the following disclosure: |
| “We conduct stress tests intended to estimate the direct and indirect impact that might result from a variety of possible events involving the above countries, including sovereign defaults and the exit of one or more countries from the Euro area. In the stress tests, described in “Market Risk Management — Stress Testing” and “Credit Risk Management — Stress Tests/Scenario Analysis,” we estimate the direct impact of the event on our credit and market exposures resulting from shocks to risk factors including, but not limited to, currency rates, interest rates, and equity prices. The parameters of these shocks vary based on the scenario reflected in each stress test. We also estimate the indirect impact on our exposures arising from potential market moves in response to the event, such as the impact of credit market deterioration on corporate borrowers and counterparties along with the shocks to the risk factors described above. We review estimated losses produced by the stress tests in order to understand their magnitude, highlight potential loss concentrations, and assess and mitigate our exposures where necessary. |
| Euro area exit scenarios include analysis of the impacts on exposure that might result from the redenomination of assets in the exiting country or countries. Constructing stress tests for these scenarios requires many assumptions about how exposures might be directly impacted and how resulting secondary market moves would indirectly impact such exposures. Given the multiple parameters involved in such scenarios, losses from such events are inherently difficult to quantify and may materially differ from our estimates. In order to prepare for any market disruption that might result from a Euro area exit, we test our operational and risk management readiness and capability to respond to a redenomination event.” |
| We also note the following with respect to your comments: |
| We run numerous scenarios involving Europe and each scenario includes shocks to a large number of risk factors. Given the number of possible permutations it is not practical to disclose specific parameters of the stress tests. Discussion of the likelihood of any of these scenarios occurring would not be meaningful as there is generally no implied probability that any of our stress test scenarios, including those that we have conducted on possible exits from the Euro area, will occur. |
| Our existing disclosure of credit and market exposure on page 173 of our June 2012 Form 10-Q includes detailed information on our assets in European countries which might be considered to have heightened redenomination and revaluation risk. Further quantitative disclosure based on our stress tests would not be practical given the number of scenarios considered and the range of approaches and tools. |
9
The Company hereby acknowledges that: the Company is responsible for the adequacy and accuracy of the disclosure in the above-referenced filings; Staff comments or changes to disclosure in response to Staff comments do not foreclose the Commission from taking any action with respect to the above-referenced filings; and the Company may not assert Staff comments as a defense in any proceeding initiated by the Commission or any person under the federal securities laws of the United States.
* * *
Please feel free to call me (212-902-5675) if you have any questions about the foregoing, or if you would like to further discuss any of the matters raised in this response letter.
Sincerely,
| /s/ Sarah E. Smith | |
| Sarah E. Smith | |
| Principal Accounting Officer |
| cc: | Lloyd C. Blankfein, Chairman and Chief Executive Officer |
| David A. Viniar, Chief Financial Officer |
| (The Goldman Sachs Group, Inc.) |
10