[AIG DRAFT OF AUGUST 8, 2008]

August 8, 2008

Mr. Jim B. Rosenberg
Senior Assistant Chief Accountant
United States
Securities and Exchange Commission
Division of Corporation Finance
100 F Street, NE
Mail Stop 6010
Washington, D. C. 20549

Re:		American International Group, Inc.
		Form 10-K for the Year Ended December 31, 2007
		Form 10-Q for the Quarter Ended March 31, 2008
		File No. 1-8787

Dear Mr. Rosenberg:

We are in receipt of your letter dated June 25, 2008 and thank you for your additional comments concerning American International Group, Inc.’s (AIG) captioned filings. We are pleased to respond to your questions and to have the opportunity to work with the Staff of the Commission to further enhance the overall presentation and disclosure in our filings.

As more fully described below, we have expanded certain of our disclosures in our Form 10-Q for the quarter ended June 30, 2008. Our responses and those disclosures are provided below. If further clarification or information is required, please let us know at your earliest convenience.

AIG acknowledges that the adequacy and accuracy of the disclosure in AIG’s filings is the responsibility of AIG, that Staff comments or changes to disclosure in response to Staff comments do not foreclose the Commission from taking any action with respect to the filings and that Staff comments may not be asserted by AIG as a defense in any proceeding initiated by the Commission or any person under the Federal securities laws of the United States.

We have repeated your questions below to facilitate your review.

Form 10-Q for the Fiscal Quarter Ended March 31, 2008

Note 4—Shareholders’ Equity and Earnings (Loss) Per Share, page 19

1.

You disclose that you shortened the vesting period of outstanding awards under your share-based employee compensation plans during the first quarter of fiscal 2008. Please provide us your evaluation of the modification

[AIG DRAFT OF AUGUST 8, 2008]

			of these awards under paragraph 51 of SFAS 123(R) demonstrating why the unamortized portion of the awards should be amortized over a shorter period. Explain why there was apparently no incremental compensation expense that was required to be recorded as a result of the modifications.
			AIG Response:

On March 11, 2008 AIG shortened the vesting schedules of certain outstanding non-vested awards of restricted stock units (RSUs) made under various share-based employee compensation plans. The vesting schedules of outstanding stock options were not modified.

AIG’s RSU awards to employees do not participate in dividends during the vesting periods and their original grant-date fair values were discounted to reflect this feature as required by paragraph B93 of SFAS 123(R). Because the vesting schedules—and consequently the discount periods—of the awards were reduced, the fair values of the awards were greater immediately after the modification compared with the fair values immediately before the modification, resulting in incremental compensation cost of $26 million, as required by paragraph 51a.

The vesting-schedule modification applied broadly to all holders of each class of award modified. Accordingly, AIG determined the vesting-schedule modification to be a “Type I: Probable-to-Probable” modification as discussed in paragraph B188a. As a result, AIG continues to amortize the original grant-date fair value of the awards (together with the incremental compensation cost discussed in the paragraph above) as required by paragraphs 51b and A160, but over the shorter requisite service (vesting) periods, as required by paragraph 39.

The cost of the increase in the number of awards expected to vest (i.e., the lower forfeiture rate) that resulted from the shorter vesting schedules was not significant.

AIG expanded its disclosures in the Second Quarter 2008 Form 10-Q (page 23) to include the following:

During the first quarter of 2008, AIG reviewed the vesting schedules of its share-based employee compensation plans, and on March 11, 2008, AIG’s management and the Compensation and Management Resources Committee of AIG’s Board of Directors determined that, to fulfill the objective of attracting and retaining high quality personnel, the vesting schedules of certain awards outstanding under these plans and all awards made in the future under these plans should be shortened.

For accounting purposes, a modification of the terms or conditions of an equity award is treated as an exchange of the original award for a new award. As a result of this modification, the incremental compensation cost related to the affected awards totaled $24 million and will, together with the unamortized originally-measured compensation cost, be amortized over shorter periods. AIG estimates the modifications will increase the amortization of this cost by $106 million and $46 million in 2008 and 2009, respectively, with a related reduction in amortization expense of $128 million in 2010 through 2013.

[AIG DRAFT OF AUGUST 8, 2008]

Financial Services, page 38

	2.		Please refer to prior comment three. You state that your credit-based analyses estimate potential realized credit pre-tax losses of approximately $1.2 billion to approximately $2.4 billion, as compared to the $20.6 billion of unrealized market losses recognized in the fourth quarter of 2007 and first quarter of 2008. Please expand your disclosure to explain the specific factors causing the divergence between your estimates of fair value and amounts to be ultimately realized upon settlement maturity.
			AIG Response:
			The unrealized market valuation losses of $20.6 billion recorded on AIGFP’s super senior credit default swap portfolio represent the cumulative change in fair value of these derivatives. Consistent with the definition of fair value in the Statement of Financial Accounting Standards No. 157, “Fair Value Measurements” (FAS 157), the unrealized market valuation loss of $20.6 billion represents AIG’s best estimate of the amount it would need to pay to a willing, able and knowledgeable third-party to assume the obligations under AIGFP’s super senior credit default swap portfolio as of March 31, 2008.
			On the other hand, AIG’s estimate of the potential pre-tax realized credit losses of approximately $1.2 billion to approximately $2.4 billion as of March 31, 2008 represents its then current estimate of the potential credit losses that it may incur if the multi-sector CDO portion of AIGFP’s super senior credit default swap portfolio and the referenced obligations acquired by AIGFP in extinguishing its obligations under the swaps are held to maturity. Credit losses represent a point-in-time estimate, using information available at that particular date, of the potential shortfall of principal and/or interest cash flows on the referenced obligations and credits underlying the portfolio that will not be recovered assuming the credit derivative portfolio and referenced obligations are held to maturity. At March 31, 2008, AIG derived its estimates of the potential pre-tax realized credit losses by applying two distinct methods, a ratings-based static stress test and a roll rate analysis.
			At March 31, 2008, the estimates for the range of the potential realized credit losses were lower than the fair value of AIGFP’s super senior multi-sector CDO credit default swap portfolio, a net loss of $19.3 billion at March 31, 2008. The fair value of AIGFP’s super senior multi-sector CDO credit default swap portfolio is based upon fair value accounting principles, which rely on third-party prices for both the underlying collateral securities and the CDOs that AIGFP’s super senior credit default swaps wrap. These prices currently incorporate liquidity premiums, risk aversion elements and credit risk modeling, which in some instances may use more conservative assumptions than those used by AIG in its roll rate stress testing. Due to the ongoing disruption in the U.S. residential mortgage market and credit markets and the downgrades of RMBS and CDOs by the rating agencies,

[AIG DRAFT OF AUGUST 8, 2008]

			the market continues to lack transparency around the pricing of these securities. These prices are not necessarily reflective of the ultimate potential realized credit losses AIGFP could incur in the future related to the AIGFP super senior multi-sector CDO credit default swap portfolio, and AIG believes they incorporate a significant amount of market-driven risk aversion.
			In conducting its risk analyses as of June 30, 2008, AIG discontinued use of the rating-based static stress test and used only the roll rate stress test because it believes that the roll rate stress test provides a more reasonable analysis methodology to illustrate potential realized credit losses than the rating-based static stress test used previously.
			AIG expanded its disclosures in the Second Quarter 2008 Form 10-Q (pages 121-123) to include the following:

The unrealized market valuation losses of $26.1 billion recorded on AIGFP’s super senior multi-sector credit default swap portfolio represents the cumulative change in fair value of these derivatives, which represents AIG’s best estimate of the amount it would need to pay to a willing, able and knowledgeable third party to assume the obligations under AIGFP’s super senior multi-sector CDO credit default swap portfolio as of June 30, 2008.

Stress Testing/Sensitivity Analysis

At June 30, 2008, AIG used a roll rate analysis to stress the AIGFP super senior multi-sector CDO credit default swap portfolio for potential pre-tax realized credit losses that it may incur if the multi-sector CDO super senior credit default swap portfolio and the referenced obligations acquired by AIGFP in extinguishing its obligations under the swaps are held to maturity. Credit losses represent an estimate of the potential shortfall of principal and/or interest cash flows on the referenced obligations and credits underlying the portfolio that will not be recovered assuming the portfolio and referenced obligations are held to maturity. Two scenarios illustrated in this process resulted in potential pre-tax realized credit losses of approximately $5.0 billion (Scenario A) and approximately $8.5 billion (Scenario B). Actual ultimate realized credit losses are likely to vary, perhaps materially, from these scenarios, and there can be no assurance that the ultimate realized credit losses related to the AIGFP super senior multi-sector CDO credit default swap portfolio will be consistent with either scenario or that such realized credit losses will not exceed the potential realized credit losses illustrated by Scenario B.

In prior quarters, AIG conducted risk analyses of the AIGFP super senior multi-sector CDO credit default swap portfolio using certain ratings-based static stress tests, which centered around scenarios of further stress on the

[AIG DRAFT OF AUGUST 8, 2008]

portfolio resulting from downgrades by the rating agencies from current levels on the underlying collateral in the CDO structures supported by AIGFP’s credit default swaps. During the first quarter of 2008, AIG developed an additional methodology to conduct stress tests for potential realized credit losses from AIGFP’s super senior multi-sector CDO credit default swap portfolio that combined a roll rate estimate of the losses emanating from the subprime and Alt-A RMBS collateral securities in the multi-sector CDOs, plus an estimate of losses arising from CDO securities (inner CDOs) and other ABS, such as CMBS, credit card and auto loan ABS, held by the CDOs. In conducting its risk analyses as of June 30, 2008, AIG discontinued use of the rating-based static stress test and used only the roll rate stress test because it believes that the roll rate stress test provides a more reasonable analysis methodology to illustrate potential realized credit losses than the rating-based static stress test used previously.

In the second quarter of 2008, AIG stressed the AIGFP super senior multi-sector CDO credit default swap portfolio using a roll rate analysis as applied to all RMBS collateral including subprime, Alt-A and prime residential mortgages that comprise the subprime, Alt-A and prime RMBS. This analysis assumed that certain percentages of actual delinquent mortgages will roll into default and foreclosure. It also assumed that certain percentages of non-delinquent mortgages will become delinquent and default over time, with those delinquency percentages depending on the age of the mortgage pool. To those assumed defaults AIG applied loss severities (one minus recovery) to derive estimated ultimate losses for each mortgage pool comprising a subprime, Alt-A and prime RMBS. Because subprime, Alt-A and prime RMBS have differing characteristics, the roll rates and loss severities differed. AIG then estimated tranche losses from these roll rate losses by applying the pool losses up through the capital structure of the RMBS. In this estimate of tranche losses, AIG introduced in the quarter an enhancement to the roll rate analysis to take into account the cash flow waterfall and to capture the potential effects, both positive and negative, of cash flow diversion within each CDO. To these estimated subprime, Alt-A and prime RMBS losses AIG added estimated credit losses on the inner CDOs and other ABS, such as CMBS, credit card and auto loan ABS, calculated by using rating-based static percentages, in the case of inner CDOs varying by vintage and type of CDO, and, in the case of other ABS, by rating. In addition to the foregoing, the analysis incorporates the effects of certain other factors such as mortgage prepayment rates, excess spread and delinquency triggers. The total of the roll rate losses and the losses on the inner CDOs and other ABS using two different scenarios of assumptions yielded estimated potential pre-tax realized credit losses of approximately $5.0 billion and approximately $8.5 billion.

At March 31, 2008, AIG’s credit-based analyses estimated potential pre-tax realized credit losses at approximately $1.2 billion to $2.4 billion. The

[AIG DRAFT OF AUGUST 8, 2008]

estimate of $2.4 billion was derived using the roll rate stress test described above. The increase in the estimated potential realized credit loss illustrated by Scenarios A and B was the result of both enhancements to the model and changes in the assumptions used. The model was enhanced by inclusion of prime RMBS into the portfolio of securities subjected to the roll rate analysis and the introduction of analytics to capture the potential effects of the cash flow waterfall. Changes in assumptions included revisions to the roll rate percentages and loss severities on subprime and Alt-A mortgages in view of deteriorating real estate market conditions, as well as a higher stress to other ABS collateral and the use of current inner CDO ratings in the rating-based static percentage. The potential realized credit loss illustrated by Scenario B is the result of applying different, more highly stressed assumptions to the roll rate analysis model than those used in Scenario A.

Due to the dislocation in the market for CDO and RMBS collateral, AIG does not use the market values of the underlying CDO collateral in estimating its potential realized credit losses. The use of factors derived from market-observable prices in models used to determine the estimates for future realized credit losses could result in materially higher estimates of potential realized credit losses.

Under the terms of most of these credit derivatives, credit losses to AIG would generally result from the credit impairment of the referenced obligations that AIG would acquire in extinguishing its swap obligations. Based upon its most current analyses, AIG believes that any credit losses which may emerge over time at AIGFP will not be material to AIG’s consolidated financial condition, but could be material to AIG’s liquidity. Other types of analyses or models could result in materially different estimates. AIG is aware that other market participants have used different assumptions and methodologies to estimate the potential realized credit losses on AIGFP’s super senior multi-sector CDO credit default swap portfolio, resulting in significantly higher estimates than those resulting from AIG’s roll rate stress testing scenarios. Actual ultimate realized credit losses are likely to vary, perhaps materially, from AIG’s roll rate stress testing scenarios, and there can be no assurance that the ultimate realized credit losses related to the AIGFP super senior multi-sector CDO credit default swap portfolio will be consistent with either scenario or that such realized credit losses will not exceed the potential realized credit losses illustrated by Scenario B.

The potential realized credit losses illustrated in Scenarios A and B are lower than the fair value of AIGFP’s super senior multi-sector CDO credit default swap portfolio, a net loss of $26.1 billion at June 30, 2008. The net loss represents AIG’s best estimate of the amount it would need to pay to a willing third party to assume the obligations under AIGFP’s super senior multi-sector CDO credit default swap portfolio. The fair value of AIGFP’s

[AIG DRAFT OF AUGUST 8, 2008]

super senior multi-sector CDO credit default swap portfolio is based upon fair value accounting principles, which rely on third-party prices for both the underlying collateral securities and the CDOs that AIGFP’s super senior credit default swaps wrap. These prices currently incorporate liquidity premiums, risk aversion elements and credit risk modeling, which in some instances may use more conservative assumptions than those used by AIG in its roll rate stress testing. Due to the ongoing disruption in the U.S. residential mortgage market and credit markets and the downgrades of RMBS and CDOs by the rating agencies, the market continues to lack transparency around the pricing of these securities. These prices are not necessarily reflective of the ultimate potential realized credit losses AIGFP could incur in the future related to the AIGFP super senior multi-sector CDO credit default swap portfolio, and AIG believes they incorporate a significant amount of market-driven risk aversion.

In the remainder of this response, we explain in further detail AIGFP’s fair valuation methodology that is used to estimate the fair value of AIGFP’s super senior multi-sector CDO credit default swap portfolio and the methods that AIG has applied to derive its estimates of the potential pre-tax realized credit losses.

FAS 157 COMPLIANT FAIR VALUATION

As disclosed in AIG’s First Quarter Form 10-Q, AIG determines the fair value of the AIGFP super senior multi-sector CDO credit default swap portfolio using a combination of internal valuation models and third party prices.

AIG utilizes a modified version of the binomial expansion technique (“BET”) model to value its credit default swaps written on the super senior securities of collateralized debt obligations (“CDO”). The BET model uses credit spreads to imply the evolution of the loss distribution of the underlying collateral pool through time and hence to value the CDO tranches. The model uses a Monte Carlo simulation to generate sample paths of the distribution of defaults in the reference collateral of the CDO through time, calculates the interest and principal cash flows arising from the reference collateral and when appropriate applies a cash flow waterfall algorithm to model the distribution of cash flows to the liabilities of the CDO conditional on the simulated defaults. The BET methodology relies on the use of a single diversification measure, the diversity score, which allows a mapping of the underlying collateral securities of a CDO structure into a hypothetical portfolio comprising a smaller number of idealized, independent, homogeneous securities. The number of such idealized securities is given by the diversity score. Diversity scores are supplied by the trustees for some of the multi-sector CDO deals. Where a diversity score is not available, a default value is applied. Due to the assumed independence of the defaults of the idealized securities, a binomial distribution can be used to calculate the loss distribution of the portfolio to value the various liabilities of the CDO.

[AIG DRAFT OF AUGUST 8, 2008]

AIG has adapted the BET model inputs so that the model accepts the price for the securities underlying the CDO as an input and converts the price to a credit spread using parameters consistent with those used for valuation. The calculations include a methodology through which prices are used to determine credit spread, which is used to determine implied probabilities of default. These probabilities of default are used to determine expected losses on the credit default swap:

The most significant assumption used in developing the estimate is the pricing of the securities within the CDO collateral pools. Prices for the individual securities held by a CDO are obtained in most cases from the CDO collateral managers, to the extent available. The CDO collateral managers obtain these prices from various sources, which include dealer quotations, third-party pricing services and in-house valuation models. To the extent there is a lag in the prices provided by the collateral managers, AIGFP rolls forward these prices to the end of the quarter using data provided by a third-party pricing service. Where a price for an individual security is not provided by the CDO collateral manager, AIGFP derives the price from a matrix that averages the prices of the various securities at the level of ABS category, vintage and the rating of the referenced security.

As mentioned in the previous paragraph, given the lack of price transparency surrounding ABS and CDO securities, AIG considers all available information to it in addition to the results of the BET model. A key set of data points that AIG collects are indicators from dealers and counterparties on the value of the super senior CDO securities referenced in AIGFP’s credit default swaps. Prices obtained from counterparties generally represent prices implied from collateral calls. Hence AIG compares those valuations to the valuations derived by the BET. Since December 31, 2007, AIG has applied a consistent methodology in determining when an adjustment to the BET-derived valuations is necessary. AIG compares the BET-derived valuations to the highest third party price it has obtained for each individual super senior bond. When more than one third party price is obtained for a particular transaction, AIG selects the highest third party price as AIG is expected to transact at the most advantageous price available to it. To the extent that the BET-derived valuations and the highest third party price approximate each other, AIG values its derivative using the average of both prices. To the extent that BET-derived valuations and the highest third party price do not approximate each other, AIG values its derivative using the lower of the two prices. In limited instances, AIG may make further adjustments to the third party prices if its analyses determine that the third party price is not a reasonable price.

Given the limited trading of CDO securities and inconsistency of pricing methodologies and assumptions among the key marketplace participants, the structured credit markets continue to lack transparency and convergence in price for this asset class. Although AIG attempts to consider all available information to derive its best estimate of the fair value of these derivatives, the quality and reliability of these valuations are dependent on the quality and reliability of the

[AIG DRAFT OF AUGUST 8, 2008]

prices it obtains from collateral managers, pricing vendors, its counterparties and dealers.

STATIC STRESS TEST

The objective of AIG’s ratings-based static stress test performed in the first quarter of 2008 and the year ended December 31, 2007 was to determine an estimate of the potential realized credit losses that AIG might experience over time against the credit default swap portfolio written on the super senior tranches of CDO securities. AIG applied a series of write-offs and discounts varying by rating to distinct types of collateral securities within the CDO. A static stress test is a point-in-time estimate.

AIG applied a total of six stresses to the sub-prime residential mortgage-backed securities (RMBS), Alt-A RMBS and inner CDO securities that collectively comprised a large percentage of the underlying securities to the multi-sector CDO portfolio. Under these stresses, AIG assumed that it would lose a specific percentage of the outstanding principal balance on certain securities held by a CDO with an assumption of no recoveries. These stresses were represented by the following write-off assumptions:

1.

100 percent of the outstanding principal balance of sub-prime RMBS rated BB+ and lower (BB+ through D) that were issued during the first six months of 2005;

2.

50 percent of the outstanding principal balance of all sub-prime RMBS rated BBB+ and lower (BBB+ through D) that were issued during the last six months of 2005.

3.

100 percent of the outstanding principal balance of all sub-prime RMBS rated A+ and lower (A+ through D) and 50% of the outstanding principal balance of all rated AAs (AA+ through AA-) that were issued during the first six months of 2006;

4.

100 percent of the outstanding principal balance of all sub-prime RMBS rated less than AAA that were issued during the last six months of 2006 and during 2007;

5.

100 percent of the outstanding principal balance of all Alt-A RMBS rated A+ or lower that were issued during 2006 and 2007;

6.

100 percent of the outstanding principal balance of all inner CDOs of high grade and mezzanine asset-backed securities (ABS) with sub-prime RMBS collateral, if these inner CDOs were rated A+ or lower.

ROLL RATE ANALYSIS AT MARCH 31, 2008

The second test AIG developed starting at March 31, 2008 was the AIG roll rate analysis. AIG developed this test because it believed that it may provide a more reasonable estimate of potential realized credit losses using highly stressed assumptions than the rating-based static stress test. This analysis starts with

[AIG DRAFT OF AUGUST 8, 2008]

estimating credit losses from actual performance data on the pool of U.S. residential mortgages that collateralize each sub-prime and Alt-A RMBS security within a CDO collateral securities pool. Furthermore, AIG added to this estimated loss from the roll rate analysis losses from application of rating-based haircuts to the inner CDOs of ABS and a 50 percent discount for single B or lower rated securities within the other asset classes in the CDO collateral pool. These three parts are further described as follows:

Part One: The sub-prime and Alt-A roll rate analysis begins by accessing three types of data to estimate losses in these RMBS securities. First, AIG collects the actual to-date losses which have been realized from U.S. residential mortgage defaults and recoveries for a given RMBS pool. Second, AIG collects actual delinquency data for the RMBS pool, starting with mortgages that are past due in payment for 30 days or more and including mortgages that are in foreclosure proceedings or that are in a real estate-owned status (REO). AIG then assumes that certain percentages of these mortgages, varying by how past due they may be, will transition from delinquent status into default. This part of the analysis is sometimes called “pipeline default” analysis. Third, AIG assumes that certain performing mortgages, mortgages that are not past due, will become past due over time and will also transition to default during the mortgage’s life. AIG estimates these future delinquencies by reviewing the age of the mortgage (how long ago was it made) and then estimating, based on loss timing curves drawn from the period 1998 to 2002, how many are likely to become delinquent and ultimately default. Loss timing curves are based on the empirical observation that mortgages become delinquent as they age, that an average of 75 percent of defaults occur within the first three years of a mortgage pool’s life and an average 95 to99 percent of a mortgage pool’s total defaults occur within the first five years. Finally, the roll rate analysis applies loss severities to the estimate of defaults from both the delinquent and non-delinquent mortgage pool. These loss severities incorporate the costs the servicer of the mortgage incurs to maintain the home during its time in default and the loss that is incurred in disposing or selling the home.

The roll rate assumptions and loss severity assumptions were based on several different sources, including the research of a well known investment bank. AIG also compared the assumptions to rating agency roll rate models and assumptions, with which they were comparable.

AIG used the following roll rate assumptions in applying this analysis, which varied depending on the year in which the mortgage was originated:

Vintage		Pre-2005					2005				2006				2007
U.S. sub-prime mortgages
30+ days delinquent			60	%			65	%			70	%			71	%
60+ days delinquent			70	%			75	%			80	%			83	%
90+ days delinquent + borrower bankruptcies			70	%			75	%			80	%			83	%
Foreclosed/REO mortgages			100	%			100	%			100	%			100	%

[AIG DRAFT OF AUGUST 8, 2008]

Vintage		Pre-2005				2005				2006					2007
U.S. Alt-A mortgages
30+ days delinquent			48	%			48	%			48	%			48	%
60+ days delinquent			60	%			70	%			77	%			78	%
90+ days delinquent + borrower bankruptcies			70	%			80	%			85	%			87	%
Foreclosed/REO mortgages			100	%			100	%			100	%			100	%

AIG also used the following loss severity assumptions in this roll rate analysis:

Vintage		Pre 2H 2004				2H 2004				1H 2005				2H 2005				2006/2007
Sub-Prime Loss Severity			40	%			45	%			50	%			55	%			60	%
Alt-A Loss Severity			35	%			35	%			40	%			45	%			45	%

A simplified illustration of how this roll rate analysis would proceed is as follows.

a.

Assume that all sub-prime mortgages in a RMBS pool were originated in the first half of 2005.

b. Assume that the mortgage pool has realized losses to date of 1 percent.

c.

Assume that after three years of seasoning 80 percent of the mortgages that will default have defaulted and that, after applying loss timing curves, 5 percent of the remaining performing (non-delinquent) pool mortgages will default in the future.

d. Assume the following actual pool delinquencies:

30+ delinquent: 5 percent
60+ delinquent: 4 percent
90+ delinquent: 4 percent
Bankrupt borrowers: 1 percent
Mortgages in foreclosure and REO: 15 percent

Thus, AIG computes the “pipeline” defaults from the actual delinquent mortgage behavior by selecting the applicable roll rate and loss severity percentages from the table above and making the following calculations:

30+ days: 5 percent times 65 percent equals 3.25 percent
60+ days: 4 percent times 75 percent equals 3 percent
90+ days: 4 percent times 75 percent equals 3 percent
Bankrupt: 1 percent times 75 percent equals 0.75 percent
Foreclosure/REO: 15 percent times 100 percent equals 15 percent
Total estimated defaults = 25 percent
Loss severity for IH 2005 = 50 percent

[AIG DRAFT OF AUGUST 8, 2008]

In this illustration estimated losses from actual pool delinquencies equal the sum of estimated defaults times the loss severity. Defaults are computed at 25 percent and assumed loss severity is 50 percent. Thus, estimated losses are 12.5 percent.

AIG assumes that applying loss timing curves to the underlying mortgage data results in an estimate of 5 percent defaults of the performing mortgages. After applying the 50 percent loss severity, AIG estimates the loss from these performing mortgages at 2.5 percent. Thus, total mortgage losses against this one RMBS pool equal actual losses to date of 1 percent; pipeline losses of 12.5 percent; and projected losses from the performing (non-delinquent) mortgage pool of 2.5 percent. The aggregate losses sum to 15.5 percent.

These mortgage losses are then applied against the RMBS tranches upwards in order of their subordination, starting with the equity tranche, then moving to the B and BB rated, BBB, A, AA, junior AAA and ultimately AAA classes. If the losses produced by the roll rate modeling exhaust the protection (subordinated tranches) in the CDO pool, then that amount of residual loss is applied first against the subordinated layers of the CDO and subsequently, only to the extent those subordinated tranches have been exhausted, against the most senior (super senior) tranche.

In addition to the foregoing, it is important to note that the roll rate analysis incorporates the effects of certain other factors, such as mortgage prepayment rates, excess spread and delinquency triggers that are modeled using Intex RMBS modeling software, the industry standard.

Part Two: As mentioned above, based on the research of the same investment bank, AIG also added to the estimated RMBS roll rate losses rating-based discounts or loss estimates for the inner CDOs of ABS with sub-prime RMBS collateral based on their type and vintage. Note that these discounts are applied against updated ratings data and, because the rating agencies have actively downgraded CDOs of ABS since mid-2007 and because the investment bank based its discounts on original ratings data, AIG believed they were conservatively stated. Nonetheless, because the rating agencies continued to downgrade CDOs of ABS as U.S. real estate market conditions were worsening AIG believed this conservatism was prudent. These discounts were as follows:

a.

High grade and mezzanine CDOs from 2006 and 2007: AAA: 50 percent; AA: 93 percent; A: 96 percent; BBB and below: 97 percent.

b.

High grade and mezzanine CDOs from 2005 and 2004: AAA: 8 percent; AA: 43 percent; A: 64 percent; BBB and below: 82 percent.

Part Three: Finally, AIG added a 50 percent discount or loss estimate to all B+ rated and lower collateral securities of all other types not covered by parts 1 and 2 above. This loss estimate attempts to account for some credit deterioration in

[AIG DRAFT OF AUGUST 8, 2008]

		highly subordinated tranches of other collateral securities not included in the roll rate and inner CDO analysis.
		ROLL RATE ANALYSIS AT JUNE 30, 2008
		At June 30, 2008 AIG used the roll rate analysis described above to stress the AIGFP super senior multi-sector credit default swap portfolio for potential realized credit losses under two scenarios, resulting in estimates of potential pre-tax realized credit losses of approximately $5 billion and approximately $8.5 billion. The increase in the estimates from the March estimate of approximately $2.4 billion was the result of both enhancements to the model and changes in the assumptions used. The model was enhanced to include prime RMBS into the portfolio of securities subjected to the roll rate analysis and the introduction of analytics to capture the potential effects of the CDO cash flow waterfall. Changes in assumptions included revisions to the roll rate percentages and loss severities on sub-prime and Alt-A mortgages in view of deteriorating U.S. real estate market conditions, as well as a higher stress to other ABS collateral and the effects of the use of current inner CDO ratings in the rating based static percentages applied to those securities.
		These changes to assumptions and enhancements to the methodology, which increased the credit loss estimate under the first scenario to approximately $5 billion from the approximately $2.4 billion estimate at March 31, 2008, are comprised of five elements totaling approximately $2.6 billion and are as follows:

	1.		Increases to roll rate assumptions to account for faster mortgage movement to default and of foreclosure, especially of Alt-A mortgages; and increases in loss severities to account for falling real estate values: approximately $700 million.
	2.		The modeling for the first time in AIG’s roll rate analysis of prime mortgages, which had not been explicitly modeled in the first quarter of 2008: approximately $200 million.
	3.		The effect of the updating of inner CDO ratings, such that a larger potential loss was computed because of further rating agency downgrades during the quarter: approximately $600 million.
	4.		An increase in the rating-based threshold and discount for other ABS securities, by which all such securities rated BB+ and lower are fully written off: approximately $100 million.
	5.		An enhancement of the roll rate analysis to incorporate CDO cash flow waterfall and to capture the potential effects, both positive and negative, of cash flow diversion: approximately $1.0 billion.
			The revised roll rate assumptions for the Q2 roll rate analysis are as follows:

[AIG DRAFT OF AUGUST 8, 2008]

Vintage		Pre-2005				2005				2006				2007
U.S. sub-prime mortgages
30+ days delinquent			60	%			70	%			80	%			80	%
60+ days delinquent			70	%			80	%			80	%			80	%
90+ days delinquent + borrower bankruptcies			70	%			80	%			90	%			90	%
Foreclosed/REO mortgages			100	%			100	%			100	%			100	%

Vintage		Pre-2005				2005				2006				2007
U.S. Alt-A mortgages
30+ days delinquent			50	%			70	%			80	%			80	%
60+ days delinquent			60	%			80	%			80	%			80	%
90+ days delinquent + borrower bankruptcies			70	%			80	%			90	%			90	%
Foreclosed/REO mortgages			100	%			100	%			100	%			100	%

          The revised loss severity assumptions for the Q2 roll rate analysis are as follows

Vintage		Pre 2H 2004				2H 2004				1H 2005				2H 2005				2006/2007
Sub-Prime Loss Severity			40	%			45	%			50	%			55	%			60	%
Alt-A Loss Severity			35	%			35	%			40	%			45	%			45	%

The prime mortgage roll rate assumptions for the roll rate analysis are as follows:

Vintage		Pre-2005				2005				2006				2007
U.S. prime mortgages
60+ days delinquent			60	%			60	%			60	%			60	%
90+ days delinquent + borrower bankruptcies			70	%			70	%			70	%			70	%
Foreclosed/REO mortgages			100	%			100	%			100	%			100	%

          The prime mortgage loss severity assumptions for the roll rate analysis are as follows:

Vintage		Pre 2H 2004				2H 2004				1H 2005				2H 2005				2006/2007
Prime Loss Severity			25	%			25	%			25	%			30	%			35	%

The estimate under the second scenario of approximately $8.5 billion pre-tax was based on a combined 20 percent increase in roll rates and loss severities beyond the figures above across all RMBS, irrespective of type and vintage, and is designed to be an even more conservative estimate of AIG’s potential credit losses from the multi-sector CDO credit default swap portfolio than the already highly stressed assumptions used in Scenario A.

[AIG DRAFT OF AUGUST 8, 2008]

		BREAKDOWN OF THE SUPER SENIOR MULTI-SECTOR CDO CREDIT DEFAULT SWAP PORTFOLIO

The super senior multi-sector CDO credit default swap portfolio can be disaggregated into the following sub-portfolios as of June 30, 2008:

(in billions)
		Notional				Fair Value
Multi-sector CDOs		Amount					June 30, 2008
CDOs of ABS with high grade collateral including subprime		$	42.0			$	14.1
CDOs of ABS with mezzanine collateral including subprime		$	15.8			$	6.9
Total CDOs of ABS including subprime		$	57.8			$	21.0
CDOs of ABS with high grade collateral without subprime		$	21.7			$	3.5
CDOs of ABS with mezzanine collateral without subprime		$	0.8			$	0.3
Total CDOs of ABS without subprime		$	22.5			$	3.8
Total CDOs of ABS		$	80.3			$	24.8

Valuation of Level 3 Assets and Liabilities, page 46

	3.		Please refer to prior comment two. You have attributed the additional $9.1 billion unrealized market valuation loss for your super senior credit default swap portfolio to ongoing disruption in the US residential mortgage and credit markets and downgrades of RMBS and CDO securities. Please explain and quantify the effects on the valuation of the specific changes made to your model in the first quarter of 2008, particularly how changes in key assumptions during the period affected the unrealized market valuation loss of $9.1 billion, as well as the impact of the roll rate analysis that was introduced in the first quarter of 2008.
			AIG Response:
			The principal cause of the unrealized market valuation loss of $9.1 billion in the first quarter of 2008 was the continued significant widening in spreads driven by the credit concerns resulting from U.S. residential mortgages, the severe liquidity crises affecting the markets and the effects of rating agency downgrades on structured securities. While AIG continued to refine its valuation methodologies during the first quarter of 2008, the refinements made had only a de minimus effect on the unrealized market valuation loss.
			The change in fair value of AIGFP’s credit default swaps was caused by the significant widening in spreads and the downgrades of RMBS and CDO securities by rating agencies in the six-month period ended June 30, 2008 driven by the credit concerns resulting from U.S. residential mortgages, the severe liquidity crisis affecting the markets and the effects of rating agency downgrades on structured securities.
			AIG’s estimate of the pre-tax unrealized market valuation loss for the three-month period ended June 30, 2008 is $5.6 billion. The primary drivers behind the

[AIG DRAFT OF AUGUST 8, 2008]

			incremental loss in the second quarter of 2008 continue to be similar to those experienced in the first three months of the fiscal year. Additionally, during the second quarter of 2008, AIGFP implemented further refinements to the cash flow waterfall used by the BET model and the assumptions used therein. These refinements reflected the ability of a CDO to use principal proceeds to cover interest payment obligations on lower-rated tranches, the ability of a CDO to use principal proceeds to cure a breach of an overcollateralization test, the ability of a CDO to amortize certain senior CDO tranches on a pro-rata or sequential basis and the preferential payment of management fees. To the extent there is a lag in the prices provided by the collateral managers, AIG refines those prices by rolling them forward to the end of the quarter using prices provided by a third-party pricing service. The net effect of these refinements was an increase in the unrealized market valuation loss of $342 million.
			AIG expanded its disclosures in the Second Quarter Form 10-Q (page 87) to include the following:

During the second quarter of 2008, AIGFP implemented further refinements to the cash flow waterfall used by the BET model and the assumptions used therein. These refinements reflected the ability of a CDO to use principal proceeds to cover interest payment obligations on lower-rated tranches, the ability of a CDO to use principal proceeds to cure a breach of an overcollateralization test, the ability of a CDO to amortize certain senior CDO tranches on a pro-rata or sequential basis and the preferential payment of management fees. To the extent there is a lag in the prices provided by the collateral managers, AIG refines those prices by rolling them forward to the end of the quarter using prices provided by a third-party pricing service. The net effect of these refinements was an increase in the unrealized market valuation loss of $342 million. Refinements made during the first quarter of 2008 had only a de minimus effect on the unrealized market valuation loss.

	4.		Given the additional unrealized market valuation loss of $9.1 billion in the first quarter of 2008, please tell us why you believe that a five point variation in the assumed price of securities within the CDO collateral pools is an appropriate measure of the reasonably likely variation in this model input. Revise your disclosures as appropriate.
			AIG Response:
			In the judgment of management, a five point variation in the assumed prices of securities within the CDO pools was deemed reasonable as of March 31, 2008 based on what was observed with respect to asset-backed security prices during late March into April. Generally, declines in price quotations for such assets experienced a significant slow down during this period. Approximately $8.0 billion of the approximately $9.1 billion in unrealized market valuation losses recognized in the first quarter of 2008 related to transactions where the referenced

[AIG DRAFT OF AUGUST 8, 2008]

		obligation is a super senior CDO security. A significant portion of the approximately $8.0 billion in unrealized market valuation losses was incurred as a result of the decline in asset-backed security prices during the first part of the first quarter. Due to the ongoing disruption in the U.S. residential mortgage market and credit markets and the downgrades of residential mortgage-backed securities and CDO securities by rating agencies, the market continues to lack transparency around the pricing of these securities. Given the current difficult market conditions, AIG cannot predict reasonably likely changes in the prices of the underlying collateral securities held within a CDO at this time. Actual results in any period are likely to vary, perhaps materially, from the modeled scenarios, and there can be no assurance that the unrealized market valuation loss related to the AIGFP super senior credit default swap portfolio will be consistent with any of the sensitivity analyses.
		The remaining $1.1 billion in unrealized market valuation losses related to credit default swaps written on corporate debt obligations or collateralized loan obligations. These derivatives are principally valued using the CDX and iTraxx indices. During the month of April 2008, these indices showed significant recovery, indicating that AIG would have recognized a gain for the month of April.
		AIG expanded its disclosures in the Second Quarter 2008 Form 10-Q (pages 52-53) to include the following:

			Valuation Sensitivity
			Set forth in the paragraphs below are sensitivity analyses that estimate the effects of using alternative pricing and other key inputs on AIG’s calculation of the unrealized market valuation loss related to the AIGFP super senior credit default swap portfolio. While AIG believes that the ranges used in these analyses are reasonable, given the current difficult market conditions, AIG is unable to predict which of the scenarios is most likely to occur. AIG is also unable to assess the effect, if any, that recent transactions involving sales of large portfolios of CDOs will have on the pricing of the AIGFP super senior credit default swap portfolio. Actual results in any period are likely to vary, perhaps materially, from the modeled scenarios, and there can be no assurance that the unrealized market valuation loss related to the AIGFP super senior credit default swap portfolio will be consistent with any of the sensitivity analyses.
			The most significant assumption used in developing the estimate is the pricing of the securities within the CDO collateral pools. These prices are used to derive default probabilities and expected losses that are used in the BET model. If the actual pricing of the securities within the collateral pools differs from the pricing used in estimating the fair value of the super senior credit default swap portfolio, there is potential for material variation in the

[AIG DRAFT OF AUGUST 8, 2008]

			fair value estimate. A decrease by five points (for example, from 87 cents per dollar to 82 cents per dollar) in the aggregate price of the underlying collateral securities would cause an additional unrealized market valuation loss of approximately $4.0 billion, while an increase in the aggregate price of the underlying collateral securities by five points (for example, from 90 cents per dollar to 95 cents per dollar) would reduce the unrealized market valuation loss by approximately $3.9 billion. Any further declines in the value of the underlying collateral securities held by a CDO will similarly affect the value of the super senior CDO securities given their significantly depressed valuations. Given the current difficult market conditions, AIG cannot predict reasonably likely changes in the prices of the underlying collateral securities held within a CDO at this time.
			The following table presents other key inputs used in the valuation of the credit default swap portfolio written on the super senior securities issued by multi-sector CDOs, and the potential increase (decrease) to the unrealized market valuation loss at June 30, 2008 calculated using the BET model for changes in these key inputs:

		Increase
		(Decrease) To
		Unrealized Market
(in millions)		Valuation Loss
Weighted average lives
Effect of an increase of 1 year		$	519
Effect of a decrease of 1 year			(905	)
Recovery rates
Effect of an increase of 10%			(18	)
Effect of a decrease of 10%			254
Diversity scores
Effect of an increase of 5			(84	)
Effect of a decrease of 5			261
Discount curve
Effect of an increase of 100 basis points			181

			These results are calculated by stressing a particular assumption independently of changes in any other assumption. No assurance can be given that the actual levels of the key inputs will not exceed, perhaps significantly, the ranges assumed by AIG for purposes of the above analysis. No assumption should be made that results calculated from the use of other changes in these key inputs can be interpolated or extrapolated from the results set forth above.
			In the case of credit default swaps written on investment grade corporate debt and CLOs, AIGFP estimates the value of its obligations by reference to the relevant market indices or third-party quotes on the underlying super senior tranches where available.

[AIG DRAFT OF AUGUST 8, 2008]

The following table represents the relevant market credit indices and index CDS maturity used in the valuation of the credit default swap portfolio written on investment-grade corporate debt and the increase (decrease) to the unrealized market valuation loss at June 30, 2008 corresponding to changes in these market credit indices and maturity:

		Increase (Decrease) To
(in millions)		Unrealized Market Valuation Loss
CDS maturity (in years)			5				7				10
CDX Index
Effect of an increase of 10 basis points		$	(23	)		$	(48	)		$	(10	)
Effect of a decrease of 10 basis points			23				49				10
iTraxx Index
Effect of an increase of 10 basis points			(11	)			(37	)			(8	)
Effect of a decrease of 10 basis points			11				37				8

These results are calculated by stressing a particular assumption independently of changes in any other assumption. No assurance can be given that the actual levels of the indices and maturity will not exceed, perhaps significantly, the ranges assumed by AIG for purposes of the above analysis. No assumption should be made that results calculated from the use of other changes in these indices and maturity can be interpolated or extrapolated from the results set forth above.

Contractual Obligations, page 86

6. Please refer to prior comment one. Please revise your disclosure in footnote (f) to include the related amounts of credit default swap liabilities that are recorded in the financial statements, along with the potential credit impairment pre-tax losses that you expect to realize.

AIG Response:

AIG revised its disclosure in footnote (f) in the Second Quarter 2008 Form 10-Q as follows (page 102):

(f)

The majority of AIGFP’s credit default swaps require AIGFP to provide credit protection on a designated portfolio of loans or debt securities. At June 30, 2008, AIG had recorded $26.1 billion of cumulative unrealized market valuation losses in its financial statements relating to AIGFP’s super senior credit default swap portfolio, net of amounts realized in extinguishing derivative obligations. However, AIG’s credit-based stress testing scenarios illustrate potential pre-tax realized credit losses from these contracts at approximately $5.0 billion and approximately $8.5 billion at that date. Due to the long-term maturities of these credit default swaps, AIG is unable to make reasonable estimates of the periods during which any payments would be made.

Very truly yours,

Kathleen E. Shannon
Senior Vice President, Secretary & Deputy General Counsel

cc: Frank Wyman, Staff Accountant

      Carl Tartar, Accounting Branch Chief

      (Securities and Exchange Commission)