Note 7.

Derivatives and Hedging Activities

Derivative Activities

Derivatives are instruments that derive their value from underlying asset prices, indices, reference rates and other inputs, or a combination of these factors. Derivatives may be traded on an exchange (exchange-traded) or they may be privately negotiated contracts, which are usually referred to as OTC derivatives. Certain of the firm’s OTC derivatives are cleared and settled through central clearing counterparties (OTC-cleared), while others are bilateral contracts between two counterparties (bilateral OTC).

Market-Making. As a market maker, the firm enters into derivative transactions to provide liquidity to clients and to facilitate the transfer and hedging of their risks. In this role, the firm typically acts as principal and is required to commit capital to provide execution, and maintains inventory in response to, or in anticipation of, client demand.

Risk Management. The firm also enters into derivatives to actively manage risk exposures that arise from its market-making and investing and lending activities in derivative and cash instruments. The firm’s holdings and exposures are hedged, in many cases, on either a portfolio or risk-specific basis, as opposed to an instrument-by-instrument basis. The offsetting impact of this economic hedging is reflected in the same business segment as the related revenues. In addition, the firm may enter into derivatives designated as hedges under U.S. GAAP. These derivatives are used to manage interest rate exposure in certain fixed-rate unsecured long-term and short-term borrowings, and deposits, and to manage foreign currency exposure on the net investment in certain non-U.S. operations.

The firm enters into various types of derivatives, including:

•

Futures and Forwards. Contracts that commit counterparties to purchase or sell financial instruments, commodities or currencies in the future.

•

Swaps. Contracts that require counterparties to exchange cash flows such as currency or interest payment streams. The amounts exchanged are based on the specific terms of the contract with reference to specified rates, financial instruments, commodities, currencies or indices.

•

Options. Contracts in which the option purchaser has the right, but not the obligation, to purchase from or sell to the option writer financial instruments, commodities or currencies within a defined time period for a specified price.

Derivatives are reported on a net-by-counterparty basis (i.e., the net payable or receivable for derivative assets and liabilities for a given counterparty) when a legal right of setoff exists under an enforceable netting agreement (counterparty netting). Derivatives are accounted for at fair value, net of cash collateral received or posted under enforceable credit support agreements (cash collateral netting). Derivative assets and liabilities are included in “Financial instruments owned, at fair value” and “Financial instruments sold, but not yet purchased, at fair value,” respectively. Realized and unrealized gains and losses on derivatives not designated as hedges under ASC 815 are included in “Market making” and “Other principal transactions” in Note 4.

The tables below present the gross fair value and the notional amounts of derivative contracts by major product type, the amounts of counterparty and cash collateral netting in the condensed consolidated statements of financial condition, as well as cash and securities collateral posted and received under enforceable credit support agreements that do not meet the criteria for netting under U.S. GAAP.


	As of March 2017					As of December 2016
$ in millions		Derivative Assets		Derivative Liabilities			Derivative Assets		Derivative Liabilities
Not accounted for as hedges
Exchange-traded		$ 461		$ 495			$ 443		$ 382

OTC-cleared		145,594		122,472			189,471		168,946

Bilateral OTC		292,542		271,992			309,037		289,491
Total interest rates		438,597		394,959			498,951		458,819
OTC-cleared		4,901		4,984			4,837		4,811

Bilateral OTC		19,850		16,746			21,530		18,770
Total credit		24,751		21,730			26,367		23,581
Exchange-traded		10		16			36		176

OTC-cleared		736		729			796		798

Bilateral OTC		80,074		81,025			111,032		106,318
Total currencies		80,820		81,770			111,864		107,292
Exchange-traded		3,018		3,100			3,219		3,187

OTC-cleared		189		236			189		197

Bilateral OTC		7,670		8,984			8,945		10,487
Total commodities		10,877		12,320			12,353		13,871
Exchange-traded		8,953		8,634			8,576		8,064

Bilateral OTC		37,790		43,259			39,516		45,826
Total equities		46,743		51,893			48,092		53,890
Subtotal		601,788		562,672			697,627		657,453
Accounted for as hedges
OTC-cleared		4,216		185			4,347		156

Bilateral OTC		3,858		10			4,180		10
Total interest rates		8,074		195			8,527		166
OTC-cleared		22		31			30		40

Bilateral OTC		22		83			55		64
Total currencies		44		114			85		104
Subtotal		8,118		309			8,612		270
Total gross fair value		$ 609,906		$ 562,981			$ 706,239		$ 657,723
Offset in condensed consolidated statements of financial condition
Exchange-traded		$ (10,406	)	$ (10,406	)		$ (9,727	)	$ (9,727	)

OTC-cleared		(128,189	)	(128,189	)		(171,864	)	(171,864	)

Bilateral OTC		(349,176	)	(349,176	)		(385,647	)	(385,647	)
Total counterparty netting		(487,771	)	(487,771	)		(567,238	)	(567,238	)
OTC-cleared		(26,963	)	(233	)		(27,560	)	(2,940	)

Bilateral OTC		(49,978	)	(35,970	)		(57,769	)	(40,046	)
Total cash collateral netting		(76,941	)	(36,203	)		(85,329	)	(42,986	)
Total amounts offset		$(564,712	)	$(523,974	)		$(652,567	)	$(610,224	)
Included in condensed consolidated statements of financial condition
Exchange-traded		$ 2,036		$ 1,839			$ 2,547		$ 2,082

OTC-cleared		506		215			246		144

Bilateral OTC		42,652		36,953			50,879		45,273
Total		$ 45,194		$ 39,007			$ 53,672		$ 47,499
Not offset in condensed consolidated statements of financial condition
Cash collateral		$ (338	)	$ (1,306	)		$ (535	)	$ (2,085	)

Securities collateral		(13,403	)	(9,032	)		(15,518	)	(10,224	)
Total		$ 31,453		$ 28,669			$ 37,619		$ 35,190


	Notional Amounts as of
$ in millions		March 2017	December 2016
Not accounted for as hedges
Exchange-traded		$ 7,103,874	$ 4,425,532

OTC-cleared		16,790,485	16,646,145

Bilateral OTC		11,633,305	11,131,442
Total interest rates		35,527,664	32,203,119
OTC-cleared		382,658	378,432

Bilateral OTC		1,009,141	1,045,913
Total credit		1,391,799	1,424,345
Exchange-traded		16,912	13,800

OTC-cleared		77,380	62,799

Bilateral OTC		6,552,532	5,576,748
Total currencies		6,646,824	5,653,347
Exchange-traded		291,840	227,707

OTC-cleared		3,961	3,506

Bilateral OTC		206,679	196,899
Total commodities		502,480	428,112
Exchange-traded		701,664	605,335

Bilateral OTC		1,024,526	959,112
Total equities		1,726,190	1,564,447
Subtotal		45,794,957	41,273,370
Accounted for as hedges
OTC-cleared		55,009	55,328

Bilateral OTC		32,754	36,607
Total interest rates		87,763	91,935
OTC-cleared		2,266	1,703

Bilateral OTC		9,674	8,544
Total currencies		11,940	10,247
Subtotal		99,703	102,182
Total notional amounts		$45,894,660	$41,375,552

In the tables above:

•

Gross fair values exclude the effects of both counterparty netting and collateral, and therefore are not representative of the firm’s exposure.

•

Where the firm has received or posted collateral under credit support agreements, but has not yet determined such agreements are enforceable, the related collateral has not been netted.

•

Notional amounts, which represent the sum of gross long and short derivative contracts, provide an indication of the volume of the firm’s derivative activity and do not represent anticipated losses.

•

Total gross fair value of derivatives includes derivative assets and derivative liabilities of $11.91 billion and $15.46 billion, respectively, as of March 2017, and derivative assets and derivative liabilities of $19.92 billion and $20.79 billion, respectively, as of December 2016, which are not subject to an enforceable netting agreement or are subject to a netting agreement that the firm has not yet determined to be enforceable.

Pursuant to a rule change at a clearing organization in the first quarter of 2017, transactions with this clearing organization are considered settled each day. The impact of reflecting transactions with this clearing organization as settled would have been a reduction in gross interest rate and credit derivative assets and liabilities as of December 2016 of $24.58 billion and $27.36 billion, respectively, and a corresponding decrease in counterparty and cash collateral netting, with no impact to the condensed consolidated statements of financial condition.

Valuation Techniques for Derivatives

The firm’s level 2 and level 3 derivatives are valued using derivative pricing models (e.g., discounted cash flow models, correlation models, and models that incorporate option pricing methodologies, such as Monte Carlo simulations). Price transparency of derivatives can generally be characterized by product type, as described below.

•

Interest Rate. In general, the key inputs used to value interest rate derivatives are transparent, even for most long-dated contracts. Interest rate swaps and options denominated in the currencies of leading industrialized nations are characterized by high trading volumes and tight bid/offer spreads. Interest rate derivatives that reference indices, such as an inflation index, or the shape of the yield curve (e.g., 10-year swap rate vs. 2-year swap rate) are more complex, but the key inputs are generally observable.

•

Credit. Price transparency for credit default swaps, including both single names and baskets of credits, varies by market and underlying reference entity or obligation. Credit default swaps that reference indices, large corporates and major sovereigns generally exhibit the most price transparency. For credit default swaps with other underliers, price transparency varies based on credit rating, the cost of borrowing the underlying reference obligations, and the availability of the underlying reference obligations for delivery upon the default of the issuer. Credit default swaps that reference loans, asset-backed securities and emerging market debt instruments tend to have less price transparency than those that reference corporate bonds. In addition, more complex credit derivatives, such as those sensitive to the correlation between two or more underlying reference obligations, generally have less price transparency.

•

Currency. Prices for currency derivatives based on the exchange rates of leading industrialized nations, including those with longer tenors, are generally transparent. The primary difference between the price transparency of developed and emerging market currency derivatives is that emerging markets tend to be observable for contracts with shorter tenors.

•

Commodity. Commodity derivatives include transactions referenced to energy (e.g., oil and natural gas), metals (e.g., precious and base) and soft commodities (e.g., agricultural). Price transparency varies based on the underlying commodity, delivery location, tenor and product quality (e.g., diesel fuel compared to unleaded gasoline). In general, price transparency for commodity derivatives is greater for contracts with shorter tenors and contracts that are more closely aligned with major and/or benchmark commodity indices.

•

Equity. Price transparency for equity derivatives varies by market and underlier. Options on indices and the common stock of corporates included in major equity indices exhibit the most price transparency. Equity derivatives generally have observable market prices, except for contracts with long tenors or reference prices that differ significantly from current market prices. More complex equity derivatives, such as those sensitive to the correlation between two or more individual stocks, generally have less price transparency.

Liquidity is essential to observability of all product types. If transaction volumes decline, previously transparent prices and other inputs may become unobservable. Conversely, even highly structured products may at times have trading volumes large enough to provide observability of prices and other inputs. See Note 5 for an overview of the firm’s fair value measurement policies.

Level 1 Derivatives

Level 1 derivatives include short-term contracts for future delivery of securities when the underlying security is a level 1 instrument, and exchange-traded derivatives if they are actively traded and are valued at their quoted market price.

Level 2 Derivatives

Level 2 derivatives include OTC derivatives for which all significant valuation inputs are corroborated by market evidence and exchange-traded derivatives that are not actively traded and/or that are valued using models that calibrate to market-clearing levels of OTC derivatives.

The selection of a particular model to value a derivative depends on the contractual terms of and specific risks inherent in the instrument, as well as the availability of pricing information in the market. For derivatives that trade in liquid markets, model selection does not involve significant management judgment because outputs of models can be calibrated to market-clearing levels.

Valuation models require a variety of inputs, such as contractual terms, market prices, yield curves, discount rates (including those derived from interest rates on collateral received and posted as specified in credit support agreements for collateralized derivatives), credit curves, measures of volatility, prepayment rates, loss severity rates and correlations of such inputs. Significant inputs to the valuations of level 2 derivatives can be verified to market transactions, broker or dealer quotations or other alternative pricing sources with reasonable levels of price transparency. Consideration is given to the nature of the quotations (e.g., indicative or firm) and the relationship of recent market activity to the prices provided from alternative pricing sources.

Level 3 Derivatives

Level 3 derivatives are valued using models which utilize observable level 1 and/or level 2 inputs, as well as unobservable level 3 inputs. The significant unobservable inputs used to value the firm’s level 3 derivatives are described below.

•

For the majority of the firm’s interest rate and currency derivatives classified in level 3, significant unobservable inputs include correlations of certain currencies and interest rates (e.g., the correlation between Euro inflation and Euro interest rates) and specific interest rate volatilities.

•

For level 3 credit derivatives, significant unobservable inputs include illiquid credit spreads and upfront credit points, which are unique to specific reference obligations and reference entities, recovery rates and certain correlations required to value credit derivatives (e.g., the likelihood of default of the underlying reference obligation relative to one another).

•

For level 3 commodity derivatives, significant unobservable inputs include volatilities for options with strike prices that differ significantly from current market prices and prices or spreads for certain products for which the product quality or physical location of the commodity is not aligned with benchmark indices.

•

For level 3 equity derivatives, significant unobservable inputs generally include equity volatility inputs for options that are long-dated and/or have strike prices that differ significantly from current market prices. In addition, the valuation of certain structured trades requires the use of level 3 correlation inputs, such as the correlation of the price performance of two or more individual stocks or the correlation of the price performance for a basket of stocks to another asset class such as commodities.

Subsequent to the initial valuation of a level 3 derivative, the firm updates the level 1 and level 2 inputs to reflect observable market changes and any resulting gains and losses are classified in level 3. Level 3 inputs are changed when corroborated by evidence such as similar market transactions, third-party pricing services and/or broker or dealer quotations or other empirical market data. In circumstances where the firm cannot verify the model value by reference to market transactions, it is possible that a different valuation model could produce a materially different estimate of fair value. See below for further information about significant unobservable inputs used in the valuation of level 3 derivatives.

Valuation Adjustments

Valuation adjustments are integral to determining the fair value of derivative portfolios and are used to adjust the mid-market valuations produced by derivative pricing models to the appropriate exit price valuation. These adjustments incorporate bid/offer spreads, the cost of liquidity, credit valuation adjustments and funding valuation adjustments, which account for the credit and funding risk inherent in the uncollateralized portion of derivative portfolios. The firm also makes funding valuation adjustments to collateralized derivatives where the terms of the agreement do not permit the firm to deliver or repledge collateral received. Market-based inputs are generally used when calibrating valuation adjustments to market-clearing levels.

In addition, for derivatives that include significant unobservable inputs, the firm makes model or exit price adjustments to account for the valuation uncertainty present in the transaction.

Fair Value of Derivatives by Level

The tables below present the fair value of derivatives on a gross basis by level and major product type as well as the impact of netting, included in the condensed consolidated statements of financial condition.

As of March 2017

$ in millions

Level 1

Level 2

Level 3

Total

Assets

Interest rates

$ 65

$ 446,015

$ 591

$ 446,671

Credit

—

20,152

4,599

24,751

Currencies

—

80,673

191

80,864

Commodities

—

10,531

346

10,877

Equities

—

46,321

422

46,743

Gross fair value

603,692

6,149

609,906

Counterparty netting in levels

)

(484,967

)

(1,199

)

(486,172

)

Subtotal

$ 59

$ 118,725

$ 4,950

$ 123,734

Cross-level counterparty netting

(1,599

)

Cash collateral netting

(76,941

)

Net fair value

$ 45,194

Liabilities

Interest rates

$ (52

)

$(394,229

)

$ (873

)

$(395,154

)

Credit

—

(19,370

)

(2,360

)

(21,730

)

Currencies

—

(81,717

)

(167

)

(81,884

)

Commodities

—

(12,052

)

(268

)

(12,320

)

Equities

(419

)

(49,090

)

(2,384

)

(51,893

)

Gross fair value

(471

)

(556,458

)

(6,052

)

(562,981

)

Counterparty netting in levels

484,967

1,199

486,172

Subtotal

$(465

)

$ (71,491

)

$(4,853

)

$ (76,809

)

Cross-level counterparty netting

1,599

Cash collateral netting

36,203

Net fair value

$ (39,007

)


	As of December 2016
$ in millions		Level 1		Level 2		Level 3		Total
Assets
Interest rates		$ 46		$ 506,818		$ 614		$ 507,478

Credit		—		21,388		4,979		26,367

Currencies		—		111,762		187		111,949

Commodities		—		11,950		403		12,353

Equities		1		47,667		424		48,092
Gross fair value		47		699,585		6,607		706,239

Counterparty netting in levels		(12	)	(564,100	)	(1,417	)	(565,529	)
Subtotal		$ 35		$ 135,485		$ 5,190		$ 140,710

Cross-level counterparty netting								(1,709	)

Cash collateral netting								(85,329	)
Net fair value								$ 53,672
Liabilities
Interest rates		$ (27	)	$(457,963	)	$ (995	)	$(458,985	)

Credit		—		(21,106	)	(2,475	)	(23,581	)

Currencies		—		(107,212	)	(184	)	(107,396	)

Commodities		—		(13,541	)	(330	)	(13,871	)

Equities		(967	)	(49,083	)	(3,840	)	(53,890	)
Gross fair value		(994	)	(648,905	)	(7,824	)	(657,723	)

Counterparty netting in levels		12		564,100		1,417		565,529
Subtotal		$(982	)	$ (84,805	)	$(6,407	)	$ (92,194	)

Cross-level counterparty netting								1,709

Cash collateral netting								42,986
Net fair value								$ (47,499	)

In the tables above:

•

The gross fair values exclude the effects of both counterparty netting and collateral netting, and therefore are not representative of the firm’s exposure.

•

Counterparty netting is reflected in each level to the extent that receivable and payable balances are netted within the same level and is included in counterparty netting in levels. Where the counterparty netting is across levels, the netting is reflected in cross-level counterparty netting.

•

Derivative assets are shown as positive amounts and derivative liabilities are shown as negative amounts.

Significant Unobservable Inputs

The table below presents the amount of level 3 assets (liabilities), and ranges, averages and medians of significant unobservable inputs used to value the firm’s level 3 derivatives.


	Level 3 Assets (Liabilities) and Range of Significant Unobservable Inputs (Average/Median) as of
$ in millions		March 2017	December 2016
Interest rates, net		$(282)	$(381)

Correlation		(10)% to 86% (56%/60%)	(10)% to 86% (56%/60%)

Volatility (bps)		31 to 151 (84/57)	31 to 151 (84/57)
Credit, net		$2,239	$2,504

Correlation		36% to 90% (67%/70%)	35% to 91% (65%/68%)

Credit spreads (bps)		1 to 962 (88/49)	1 to 993 (122/73)

Upfront credit points		0 to 99 (41/35)	0 to 100 (43/35)

Recovery rates		20% to 97% (59%/65%)	1% to 97% (58%/70%)
Currencies, net		$24	$3

Correlation		25% to 70% (50%/55%)	25% to 70% (50%/55%)
Commodities, net		$78	$73

Volatility		10% to 59% (29%/28%)	13% to 68% (33%/33%)

Natural gas spread		$(1.68) to $3.47 ($(0.22)/$(0.13))	$(1.81) to $4.33 ($(0.14)/$(0.05))

Oil spread		$(9.31) to $63.63 ($7.62/$(0.41))	$(19.72) to $64.92 ($25.30/$16.43)
Equities, net		$(1,962)	$(3,416)

Correlation		(30)% to 89% (42%/41%)	(39)% to 88% (41%/41%)

Volatility		5% to 80% (23%/22%)	5% to 72% (24%/23%)

In the table above:

•

Derivative assets are shown as positive amounts and derivative liabilities are shown as negative amounts.

•

Ranges represent the significant unobservable inputs that were used in the valuation of each type of derivative.

•

Averages represent the arithmetic average of the inputs and are not weighted by the relative fair value or notional of the respective financial instruments. An average greater than the median indicates that the majority of inputs are below the average. For example, the difference between the average and the median for credit spreads and oil spread inputs indicates that the majority of the inputs fall in the lower end of the range.

•

The ranges, averages and medians of these inputs are not representative of the appropriate inputs to use when calculating the fair value of any one derivative. For example, the highest correlation for interest rate derivatives is appropriate for valuing a specific interest rate derivative but may not be appropriate for valuing any other interest rate derivative. Accordingly, the ranges of inputs do not represent uncertainty in, or possible ranges of, fair value measurements of the firm’s level 3 derivatives.

•

Interest rates, currencies and equities derivatives are valued using option pricing models, credit derivatives are valued using option pricing, correlation and discounted cash flow models, and commodities derivatives are valued using option pricing and discounted cash flow models.

•

The fair value of any one instrument may be determined using multiple valuation techniques. For example, option pricing models and discounted cash flows models are typically used together to determine fair value. Therefore, the level 3 balance encompasses both of these techniques.

•

Correlation within currencies and equities includes cross-product correlation.

•

Natural gas spread represents the spread per million British thermal units of natural gas.

•

Oil spread represents the spread per barrel of oil and refined products.

Range of Significant Unobservable Inputs

The following is information about the ranges of significant unobservable inputs used to value the firm’s level 3 derivative instruments:

•

Correlation. Ranges for correlation cover a variety of underliers both within one market (e.g., equity index and equity single stock names) and across markets (e.g., correlation of an interest rate and a foreign exchange rate), as well as across regions. Generally, cross-product correlation inputs are used to value more complex instruments and are lower than correlation inputs on assets within the same derivative product type.

•

Volatility. Ranges for volatility cover numerous underliers across a variety of markets, maturities and strike prices. For example, volatility of equity indices is generally lower than volatility of single stocks.

•

Credit spreads, upfront credit points and recovery rates. The ranges for credit spreads, upfront credit points and recovery rates cover a variety of underliers (index and single names), regions, sectors, maturities and credit qualities (high-yield and investment-grade). The broad range of this population gives rise to the width of the ranges of significant unobservable inputs.

•

Commodity prices and spreads. The ranges for commodity prices and spreads cover variability in products, maturities and delivery locations.

Sensitivity of Fair Value Measurement to Changes in Significant Unobservable Inputs

The following is a description of the directional sensitivity of the firm’s level 3 fair value measurements to changes in significant unobservable inputs, in isolation:

•

Correlation. In general, for contracts where the holder benefits from the convergence of the underlying asset or index prices (e.g., interest rates, credit spreads, foreign exchange rates, inflation rates and equity prices), an increase in correlation results in a higher fair value measurement.

•

Volatility. In general, for purchased options, an increase in volatility results in a higher fair value measurement.

•

Credit spreads, upfront credit points and recovery rates. In general, the fair value of purchased credit protection increases as credit spreads or upfront credit points increase or recovery rates decrease. Credit spreads, upfront credit points and recovery rates are strongly related to distinctive risk factors of the underlying reference obligations, which include reference entity-specific factors such as leverage, volatility and industry, market-based risk factors, such as borrowing costs or liquidity of the underlying reference obligation, and macroeconomic conditions.

•

Commodity prices and spreads. In general, for contracts where the holder is receiving a commodity, an increase in the spread (price difference from a benchmark index due to differences in quality or delivery location) or price results in a higher fair value measurement.

Due to the distinctive nature of each of the firm’s level 3 derivatives, the interrelationship of inputs is not necessarily uniform within each product type.

Level 3 Rollforward

The table below presents a summary of the changes in fair value for all level 3 derivatives.


	Three Months Ended March
$ in millions		2017		2016
Total level 3 derivatives
Beginning balance		$(1,217	)	$ 495

Net realized gains/(losses)		(15	)	(79	)

Net unrealized gains/(losses)		769		461

Purchases		79		115

Sales		(458	)	(1,825	)

Settlements		871		106

Transfers into level 3		(10	)	(16	)

Transfers out of level 3		78		798
Ending balance		$ 97		$ 55

In the table above:

•

Changes in fair value are presented for all derivative assets and liabilities that are classified in level 3 as of the end of the period.

•

Net unrealized gains/(losses) relate to instruments that were still held at period-end.

•

If a derivative was transferred into level 3 during a reporting period, its entire gain or loss for the period is classified in level 3. Transfers between levels are reported at the beginning of the reporting period in which they occur.

•

Positive amounts for transfers into level 3 and negative amounts for transfers out of level 3 represent net transfers of derivative assets. Negative amounts for transfers into level 3 and positive amounts for transfers out of level 3 represent net transfers of derivative liabilities.

•

A derivative with level 1 and/or level 2 inputs is classified in level 3 in its entirety if it has at least one significant level 3 input.

•

If there is one significant level 3 input, the entire gain or loss from adjusting only observable inputs (i.e., level 1 and level 2 inputs) is classified in level 3.

•

Gains or losses that have been classified in level 3 resulting from changes in level 1 or level 2 inputs are frequently offset by gains or losses attributable to level 1 or level 2 derivatives and/or level 1, level 2 and level 3 cash instruments. As a result, gains/(losses) included in the level 3 rollforward below do not necessarily represent the overall impact on the firm’s results of operations, liquidity or capital resources.

The table below disaggregates, by major product type, the information for level 3 derivatives included in the summary table above.


	Three Months Ended March
$ in millions		2017		2016
Interest rates, net
Beginning balance		$ (381	)	$ (398	)

Net realized gains/(losses)		(22	)	(11	)

Net unrealized gains/(losses)		103		28

Purchases		4		3

Sales		(9	)	(10	)

Settlements		46		17

Transfers into level 3		(10	)	—

Transfers out of level 3		(13	)	(12	)
Ending balance		$ (282	)	$ (383	)
Credit, net
Beginning balance		$ 2,504		$ 2,793

Net realized gains/(losses)		43		(26	)

Net unrealized gains/(losses)		(174	)	210

Purchases		16		33

Sales		(20	)	(57	)

Settlements		(135	)	(75	)

Transfers into level 3		13		8

Transfers out of level 3		(8	)	(65	)
Ending balance		$ 2,239		$ 2,821
Currencies, net
Beginning balance		$ 3		$ (34	)

Net realized gains/(losses)		(22	)	(21	)

Net unrealized gains/(losses)		(13	)	(5	)

Purchases		2		6

Sales		—		(1	)

Settlements		51		61

Transfers into level 3		(2	)	—

Transfers out of level 3		5		3
Ending balance		$ 24		$ 9
Commodities, net
Beginning balance		$ 73		$ (262	)

Net realized gains/(losses)		—		(5	)

Net unrealized gains/(losses)		20		41

Purchases		13		47

Sales		(13	)	(18	)

Settlements		(21	)	(37	)

Transfers into level 3		(9	)	(26	)

Transfers out of level 3		15		(31	)
Ending balance		$ 78		$ (291	)
Equities, net
Beginning balance		$(3,416	)	$(1,604	)

Net realized gains/(losses)		(14	)	(16	)

Net unrealized gains/(losses)		833		187

Purchases		44		26

Sales		(416	)	(1,739	)

Settlements		930		140

Transfers into level 3		(2	)	2

Transfers out of level 3		79		903
Ending balance		$(1,962	)	$(2,101	)

Level 3 Rollforward Commentary

Three Months Ended March 2017. The net realized and unrealized gains on level 3 derivatives of $754 million (reflecting $15 million of net realized losses and $769 million of net unrealized gains) include gains/(losses) of $848 million and $(94) million reported in “Market making” and “Other principal transactions” respectively.

The net unrealized gain on level 3 derivatives for the three months ended March 2017 was primarily attributable to gains on certain equity derivatives, reflecting the impact of an increase in equity prices.

Transfers into level 3 derivatives during the three months ended March 2017 were not material.

Transfers out of level 3 derivatives during the three months ended March 2017 primarily reflected transfers of certain equity derivative liabilities to level 2, principally due to certain unobservable volatility inputs not being significant to the valuation of these derivatives.

Three Months Ended March 2016. The net realized and unrealized gains on level 3 derivatives of $382 million (reflecting $79 million of realized losses and $461 million of unrealized gains) include gains/(losses) of $393 million and $(11) million reported in “Market making” and “Other principal transactions” respectively.

The net unrealized gain on level 3 derivatives for the three months ended March 2016 was primarily attributable to gains on certain credit derivatives, reflecting the impact of changes in interest rates and widening of certain credit spreads, and gains on certain equity derivatives, reflecting the impact of changes in equity prices.

Transfers into level 3 derivatives during the three months ended March 2016 were not material.

Transfers out of level 3 derivatives during the three months ended March 2016 primarily reflected transfers of certain equity derivative liabilities to level 2, principally due to certain unobservable inputs no longer being significant to the valuation of these derivatives.

OTC Derivatives

The table below presents the fair values of OTC derivative assets and liabilities by tenor and major product type.


$ in millions	Less than 1 Year		1 - 5 Years		Greater than 5 Years		Total
As of March 2017
Assets
Interest rates	$ 5,419		$17,823		$77,258		$100,500

Credit	1,295		3,476		4,161		8,932

Currencies	9,715		6,246		8,154		24,115

Commodities	3,103		1,252		175		4,530

Equities	3,299		6,955		1,294		11,548

Counterparty netting in tenors	(2,965	)	(5,174	)	(3,966	)	(12,105	)
Subtotal	$19,866		$30,578		$87,076		$137,520

Cross-tenor counterparty netting							(17,421	)

Cash collateral netting							(76,941	)
Total							$ 43,158
Liabilities
Interest rates	$ 5,086		$ 9,749		$34,116		$ 48,951

Credit	1,298		3,082		1,531		5,911

Currencies	11,591		8,502		5,035		25,128

Commodities	2,387		1,067		2,437		5,891

Equities	7,859		6,298		2,859		17,016

Counterparty netting in tenors	(2,965	)	(5,174	)	(3,966	)	(12,105	)
Subtotal	$25,256		$23,524		$42,012		$ 90,792

Cross-tenor counterparty netting							(17,421	)

Cash collateral netting							(36,203	)
Total							$ 37,168
As of December 2016
Assets
Interest rates	$ 5,845		$18,376		$79,507		$103,728

Credit	1,763		2,695		4,889		9,347

Currencies	18,344		8,292		8,428		35,064

Commodities	3,273		1,415		179		4,867

Equities	3,141		9,249		1,341		13,731

Counterparty netting in tenors	(3,543	)	(5,550	)	(3,794	)	(12,887	)
Subtotal	$28,823		$34,477		$90,550		$153,850

Cross-tenor counterparty netting							(17,396	)

Cash collateral netting							(85,329	)
Total							$ 51,125
Liabilities
Interest rates	$ 5,679		$10,814		$38,812		$ 55,305

Credit	2,060		3,328		1,167		6,555

Currencies	14,720		9,771		5,879		30,370

Commodities	2,546		1,555		2,315		6,416

Equities	7,000		10,426		2,614		20,040

Counterparty netting in tenors	(3,543	)	(5,550	)	(3,794	)	(12,887	)
Subtotal	$28,462		$30,344		$46,993		$105,799

Cross-tenor counterparty netting							(17,396	)

Cash collateral netting							(42,986	)
Total							$ 45,417

In the table above:

•

Tenor is based on expected duration for mortgage-related credit derivatives and generally on remaining contractual maturity for other derivatives.

•

Counterparty netting within the same product type and tenor category is included within such product type and tenor category.

•

Counterparty netting across product types within the same tenor category is included in counterparty netting in tenors. Where the counterparty netting is across tenor categories, the netting is reflected in cross-tenor counterparty netting.

Credit Derivatives

The firm enters into a broad array of credit derivatives in locations around the world to facilitate client transactions and to manage the credit risk associated with market-making and investing and lending activities. Credit derivatives are actively managed based on the firm’s net risk position.

Credit derivatives are generally individually negotiated contracts and can have various settlement and payment conventions. Credit events include failure to pay, bankruptcy, acceleration of indebtedness, restructuring, repudiation and dissolution of the reference entity.

The firm enters into the following types of credit derivatives:

•

Credit Default Swaps. Single-name credit default swaps protect the buyer against the loss of principal on one or more bonds, loans or mortgages (reference obligations) in the event the issuer (reference entity) of the reference obligations suffers a credit event. The buyer of protection pays an initial or periodic premium to the seller and receives protection for the period of the contract. If there is no credit event, as defined in the contract, the seller of protection makes no payments to the buyer of protection. However, if a credit event occurs, the seller of protection is required to make a payment to the buyer of protection, which is calculated in accordance with the terms of the contract.

•

Credit Options. In a credit option, the option writer assumes the obligation to purchase or sell a reference obligation at a specified price or credit spread. The option purchaser buys the right, but does not assume the obligation, to sell the reference obligation to, or purchase it from, the option writer. The payments on credit options depend either on a particular credit spread or the price of the reference obligation.

•

Credit Indices, Baskets and Tranches. Credit derivatives may reference a basket of single-name credit default swaps or a broad-based index. If a credit event occurs in one of the underlying reference obligations, the protection seller pays the protection buyer. The payment is typically a pro-rata portion of the transaction’s total notional amount based on the underlying defaulted reference obligation. In certain transactions, the credit risk of a basket or index is separated into various portions (tranches), each having different levels of subordination. The most junior tranches cover initial defaults and once losses exceed the notional amount of these junior tranches, any excess loss is covered by the next most senior tranche in the capital structure.

•

Total Return Swaps. A total return swap transfers the risks relating to economic performance of a reference obligation from the protection buyer to the protection seller. Typically, the protection buyer receives from the protection seller a floating rate of interest and protection against any reduction in fair value of the reference obligation, and in return the protection seller receives the cash flows associated with the reference obligation, plus any increase in the fair value of the reference obligation.

The firm economically hedges its exposure to written credit derivatives primarily by entering into offsetting purchased credit derivatives with identical underliers. Substantially all of the firm’s purchased credit derivative transactions are with financial institutions and are subject to stringent collateral thresholds. In addition, upon the occurrence of a specified trigger event, the firm may take possession of the reference obligations underlying a particular written credit derivative, and consequently may, upon liquidation of the reference obligations, recover amounts on the underlying reference obligations in the event of default.

As of March 2017, written and purchased credit derivatives had total gross notional amounts of $669.66 billion and $722.22 billion, respectively, for total net notional purchased protection of $52.56 billion. As of December 2016, written and purchased credit derivatives had total gross notional amounts of $690.47 billion and $733.98 billion, respectively, for total net notional purchased protection of $43.51 billion. Substantially all of the firm’s written and purchased credit derivatives are credit default swaps.

The table below presents certain information about credit derivatives.


	Credit Spread on Underlier (basis points)
$ in millions		0 - 250	251 - 500		501 - 1,000		Greater than 1,000		Total
As of March 2017
Maximum Payout/Notional Amount of Written Credit Derivatives by Tenor
Less than 1 year		$207,898	$ 8,016		$ 1,629		$ 5,065		$222,608

1 - 5 years		337,949	10,995		8,711		7,428		365,083

Greater than 5 years		73,908	5,972		1,564		528		81,972
Total		$619,755	$24,983		$11,904		$13,021		$669,663
Maximum Payout/Notional Amount of Purchased Credit Derivatives
Offsetting		$531,632	$16,116		$10,521		$10,873		$569,142

Other		138,756	9,863		2,082		2,380		153,081
Fair Value of Written Credit Derivatives
Asset		$ 14,252	$ 650		$ 192		$ 59		$ 15,153

Liability		1,844	538		904		4,271		7,557
Net asset/(liability)		$ 12,408	$ 112		$ (712	)	$ (4,212	)	$ 7,596
As of December 2016
Maximum Payout/Notional Amount of Written Credit Derivatives by Tenor
Less than 1 year		$207,727	$ 5,819		$ 1,016		$ 8,629		$223,191

1 - 5 years		375,208	17,255		8,643		7,986		409,092

Greater than 5 years		52,977	3,928		1,045		233		58,183
Total		$635,912	$27,002		$10,704		$16,848		$690,466
Maximum Payout/Notional Amount of Purchased Credit Derivatives
Offsetting		$558,305	$20,588		$10,133		$15,186		$604,212

Other		119,509	7,712		1,098		1,446		129,765
Fair Value of Written Credit Derivatives
Asset		$ 13,919	$ 606		$ 187		$ 45		$ 14,757

Liability		2,436	902		809		5,686		9,833
Net asset/(liability)		$ 11,483	$ (296	)	$ (622	)	$ (5,641	)	$ 4,924

In the table above:

•

Fair values exclude the effects of both netting of receivable balances with payable balances under enforceable netting agreements, and netting of cash received or posted under enforceable credit support agreements, and therefore are not representative of the firm’s credit exposure.

•

Tenor is based on expected duration for mortgage-related credit derivatives and on remaining contractual maturity for other credit derivatives.

•

The credit spread on the underlier, together with the tenor of the contract, are indicators of payment/performance risk. The firm is less likely to pay or otherwise be required to perform where the credit spread and the tenor are lower.

•

Offsetting purchased credit derivatives represent the notional amount of purchased credit derivatives that economically hedge written credit derivatives with identical underliers and are included in offsetting.

•

Other purchased credit derivatives represent the notional amount of all other purchased credit derivatives not included in offsetting.

Impact of Credit Spreads on Derivatives

On an ongoing basis, the firm realizes gains or losses relating to changes in credit risk through the unwind of derivative contracts and changes in credit mitigants.

The net gain, including hedges, attributable to the impact of changes in credit exposure and credit spreads (counterparty and the firm’s) on derivatives was $11 million and $132 million for the three months ended March 2017 and March 2016, respectively.

Bifurcated Embedded Derivatives

The table below presents the fair value and the notional amount of derivatives that have been bifurcated from their related borrowings.


	As of
$ in millions		March 2017	December 2016
Fair value of assets		$ 734	$ 676

Fair value of liabilities		981	864
Net liability		$ 247	$ 188
Notional amount		$8,495	$8,726

In the table above, these derivatives, which are recorded at fair value, primarily consist of interest rate, equity and commodity products and are included in “Unsecured short-term borrowings” and “Unsecured long-term borrowings” with the related borrowings. See Note 8 for further information.

Derivatives with Credit-Related Contingent Features

Certain of the firm’s derivatives have been transacted under bilateral agreements with counterparties who may require the firm to post collateral or terminate the transactions based on changes in the firm’s credit ratings. The firm assesses the impact of these bilateral agreements by determining the collateral or termination payments that would occur assuming a downgrade by all rating agencies. A downgrade by any one rating agency, depending on the agency’s relative ratings of the firm at the time of the downgrade, may have an impact which is comparable to the impact of a downgrade by all rating agencies.

The table below presents the aggregate fair value of net derivative liabilities under such agreements (excluding application of collateral posted to reduce these liabilities), the related aggregate fair value of the assets posted as collateral and the additional collateral or termination payments that could have been called at the reporting date by counterparties in the event of a one-notch and two-notch downgrade in the firm’s credit ratings.


	As of
$ in millions		March 2017	December 2016
Net derivative liabilities under bilateral agreements		$28,517	$32,927

Collateral posted		$24,530	$27,840

Additional collateral or termination payments:
One-notch downgrade		$ 367	$ 677

Two-notch downgrade		$ 1,880	$ 2,216

Hedge Accounting

The firm applies hedge accounting for (i) certain interest rate swaps used to manage the interest rate exposure of certain fixed-rate unsecured long-term and short-term borrowings and certain fixed-rate certificates of deposit and (ii) certain foreign currency forward contracts and foreign currency-denominated debt used to manage foreign currency exposures on the firm’s net investment in certain non-U.S. operations.

To qualify for hedge accounting, the hedging instrument must be highly effective at reducing the risk from the exposure being hedged. Additionally, the firm must formally document the hedging relationship at inception and test the hedging relationship at least on a quarterly basis to ensure the hedging instrument continues to be highly effective over the life of the hedging relationship.

Fair Value Hedges

The firm designates certain interest rate swaps as fair value hedges. These interest rate swaps hedge changes in fair value attributable to the designated benchmark interest rate (e.g., London Interbank Offered Rate (LIBOR) or Overnight Index Swap Rate), effectively converting a substantial portion of fixed-rate obligations into floating-rate obligations.

The firm applies a statistical method that utilizes regression analysis when assessing the effectiveness of its fair value hedging relationships in achieving offsetting changes in the fair values of the hedging instrument and the risk being hedged (i.e., interest rate risk). An interest rate swap is considered highly effective in offsetting changes in fair value attributable to changes in the hedged risk when the regression analysis results in a coefficient of determination of 80% or greater and a slope between 80% and 125%.

For qualifying fair value hedges, gains or losses on derivatives are included in “Interest expense.” The change in fair value of the hedged item attributable to the risk being hedged is reported as an adjustment to its carrying value and is subsequently amortized into interest expense over its remaining life. Gains or losses resulting from hedge ineffectiveness are included in “Interest expense.” When a derivative is no longer designated as a hedge, any remaining difference between the carrying value and par value of the hedged item is amortized to interest expense over the remaining life of the hedged item using the effective interest method. See Note 23 for further information about interest income and interest expense.

The table below presents the gains/(losses) from interest rate derivatives accounted for as hedges, the related hedged borrowings and deposits, and the hedge ineffectiveness on these derivatives, which primarily consists of amortization of prepaid credit spreads resulting from the passage of time.


	Three Months Ended March
$ in millions		2017		2016
Interest rate hedges		$(754	)	$ 1,990

Hedged borrowings and deposits		554		(2,028	)
Hedge ineffectiveness		$(200	)	$ (38	)

Net Investment Hedges

The firm seeks to reduce the impact of fluctuations in foreign exchange rates on its net investments in certain non-U.S. operations through the use of foreign currency forward contracts and foreign currency-denominated debt. For foreign currency forward contracts designated as hedges, the effectiveness of the hedge is assessed based on the overall changes in the fair value of the forward contracts (i.e., based on changes in forward rates). For foreign currency-denominated debt designated as a hedge, the effectiveness of the hedge is assessed based on changes in spot rates.

For qualifying net investment hedges, the gains or losses on the hedging instruments, to the extent effective, are included in “Currency translation” in the condensed consolidated statements of comprehensive income.

The table below presents the gains/(losses) from net investment hedging.


	Three Months Ended March
$ in millions		2017		2016
Hedges:
Foreign currency forward contract		$(349	)	$(356	)

Foreign currency-denominated debt		$ (82	)	$(150	)

The gain/(loss) related to ineffectiveness and the gain/(loss) reclassified to earnings from accumulated other comprehensive loss were not material for the three months ended March 2017 or March 2016.

As of March 2017 and December 2016, the firm had designated $1.77 billion and $1.69 billion, respectively, of foreign currency-denominated debt, included in “Unsecured long-term borrowings” and “Unsecured short-term borrowings,” as hedges of net investments in non-U.S. subsidiaries.