Asset Quality

12 Months Ended

Dec. 31, 2017

ASSET QUALITY

We closely monitor economic conditions and loan performance trends to manage and evaluate our exposure to credit risk. Trends in delinquency rates may be a key indicator, among other considerations, of credit risk within the loan portfolios. The measurement of delinquency status is based on the contractual terms of each loan. Loans that are 30 days or more past due in terms of payment are considered delinquent. Loan delinquencies exclude loans held for sale, purchased impaired loans, nonperforming loans and loans accounted for under the fair value option which are on nonaccrual status, but include government insured or guaranteed loans and accruing loans accounted for under the fair value option.

Nonperforming assets include nonperforming loans and leases, OREO, foreclosed and other assets. Nonperforming loans are those loans accounted for at amortized cost whose credit quality has deteriorated to the extent that full collection of contractual principal and interest is not probable. Interest income is not recognized on these loans. Loans accounted for under the fair value option are reported as performing loans as these loans are accounted for at fair value. However, when nonaccrual criteria is met, interest income is not recognized on these loans. Additionally, certain government insured or guaranteed loans for which we expect to collect substantially all principal and interest are not reported as nonperforming loans and continue to accrue interest. Purchased impaired loans are excluded from nonperforming loans as we are currently accreting interest income over the expected life of the loans.

See Note 1 Accounting Policies for additional information on our loan related policies.

The following tables display the delinquency status of our loans and our nonperforming assets at December 31, 2017 and December 31, 2016, respectively.

Table 39: Analysis of Loan Portfolio (a)



	Accruing
Dollars in millions	Current or Less Than 30 Days Past Due			30-59 Days Past Due			60-89 Days Past Due			90 Days Or More Past Due			Total Past Due (b)				Nonperforming Loans			Fair Value Option Nonaccrual Loans (c)			Purchased Impaired Loans			Total Loans (d)
December 31, 2017
Commercial Lending
Commercial	$	109,989		$	45		$	25		$	39		$	109			$	429								$	110,527
Commercial real estate	28,826			27			2						29				123									28,978
Equipment lease financing	7,914			17			1						18				2									7,934
Total commercial lending	146,729			89			28			39			156				554									147,439
Consumer Lending
Home equity	26,561			78			26						$	104			818						$	881		28,364
Residential real estate	14,389			151			74			486			711			(b)	400			$	197		1,515			17,212
Credit card	5,579			43			26			45			114				6									5,699
Other consumer
Automobile	12,697			79			20			8			107				76									12,880
Education and other	8,525			105			64			159			328			(b)	11									8,864
Total consumer lending	67,751			456			210			698			1,364				1,311			197			2,396			73,019
Total	$	214,480		$	545		$	238		$	737		$	1,520			$	1,865		$	197		$	2,396		$	220,458
Percentage of total loans	97.29		%	.25		%	.11		%	.33		%	.69		%		.85		%	.09		%	1.08		%	100.00		%
December 31, 2016
Commercial Lending
Commercial	$	100,710		$	81		$	20		$	39		$	140			$	496					$	18		$	101,364
Commercial real estate	28,769			5			2						7				143						91			29,010
Equipment lease financing	7,535			29			1						30				16									7,581
Total commercial lending	137,014			115			23			39			177				655						109			137,955
Consumer Lending
Home equity	27,820			64			30						94				914						1,121			29,949
Residential real estate	12,425			159			68			500			727			(b)	501			$	219		1,726			15,598
Credit card	5,187			33			21			37			91				4									5,282
Other consumer
Automobile	12,257			51			12			5			68				55									12,380
Education and other	9,235			140			78			201			419			(b)	15									9,669
Total consumer lending	66,924			447			209			743			1,399				1,489			219			2,847			72,878
Total	$	203,938		$	562		$	232		$	782		$	1,576			$	2,144		$	219		$	2,956		$	210,833
Percentage of total loans	96.73		%	.27		%	.11		%	.37		%	.75		%		1.02		%	.10		%	1.40		%	100.00		%


(a)	Amounts in table represent recorded investment and exclude loans held for sale. Recorded investment in a loan includes the unpaid principal balance plus net accounting adjustments, less any charge-offs. Recorded investment does not include any associated valuation allowance.

(b)

Past due loan amounts exclude purchased impaired loans, even if contractually past due (or if we do not expect to receive payment in full based on the original contractual terms), as we are currently accreting interest income over the expected life of the loans. Past due loan amounts include government insured or guaranteed Residential real estate mortgages totaling $.6 billion at both December 31, 2017 and December 31, 2016, and Education and other consumer loans totaling $.3 billion and $.4 billion at December 31, 2017 and December 31, 2016, respectively.


(c)	Consumer loans accounted for under the fair value option for which we do not expect to collect substantially all principal and interest are subject to nonaccrual accounting and classification upon meeting any of our nonaccrual policies. Given that these loans are not accounted for at amortized cost, these loans have been excluded from the nonperforming loan population.


(d)	Net of unearned income, net deferred loan fees, unamortized discounts and premiums, and purchase discounts and premiums totaling $1.2 billion and $1.3 billion at December 31, 2017 and December 31, 2016, respectively.

In the normal course of business, we originate or purchase loan products with contractual characteristics that, when concentrated, may increase our exposure as a holder of those loan products. Possible product features that may create a concentration of credit risk would include a high original or updated LTV ratio, terms that may expose the borrower to future increases in repayments above increases in market interest rates, and interest-only loans, among others.

We originate interest-only loans to commercial borrowers. Such credit arrangements are usually designed to match borrower cash flow expectations (e.g., working capital lines, revolvers). These products are standard in the financial services industry and product features are considered during the underwriting process to mitigate the increased risk that the interest-only feature may result in borrowers not being able to make interest and principal payments when due. We do not believe that these product features create a concentration of credit risk.

At December 31, 2017, we pledged $18.7 billion of commercial loans to the Federal Reserve Bank (FRB) and $62.8 billion of residential real estate and other loans to the Federal Home Loan Bank (FHLB) as collateral for the ability to borrow, if necessary. The comparable amounts at December 31, 2016 were $22.0 billion and $60.8 billion, respectively.

Table 40: Nonperforming Assets



Dollars in millions	December 31 2017			December 31 2016
Nonperforming loans
Total commercial lending	$	554		$	655
Total consumer lending (a)	1,311			1,489
Total nonperforming loans	1,865			2,144
OREO, foreclosed and other assets	170			230
Total nonperforming assets	$	2,035		$	2,374
Nonperforming loans to total loans	.85		%	1.02		%
Nonperforming assets to total loans, OREO, foreclosed and other assets	.92		%	1.12		%
Nonperforming assets to total assets	.53		%	.65		%
Interest on nonperforming loans (b)
Computed on original terms	$	114		$	111
Recognized prior to nonperforming status	$	19		$	21


(a)	Excludes most consumer loans and lines of credit not secured by residential real estate, which are charged off after 120 to 180 days past due and are not placed on nonperforming status.


(b)	Amounts are for the year ended.

Nonperforming loans also include certain loans whose terms have been restructured in a manner that grants a concession to a borrower experiencing financial difficulties. In accordance with applicable accounting guidance, these loans are considered TDRs. See Note 1 Accounting Policies and the TDR section of this Note 3.

Total nonperforming loans in Table 40 include TDRs of $1.0 billion at December 31, 2017 and $1.1 billion at December 31, 2016. TDRs that are performing, including consumer credit card TDR loans, totaled $1.1 billion at both December 31, 2017 and December 31, 2016, and are excluded from nonperforming loans. Nonperforming TDRs are returned to accrual status and classified as performing after demonstrating a period of at least six months of consecutive performance under the restructured terms. Loans where borrowers have been discharged from personal liability through Chapter 7 bankruptcy and have not formally reaffirmed their loan obligations to us and loans to borrowers not currently obligated to make both principal and interest payments under the restructured terms are not returned to accrual status. See the TDRs section of this Note 3 for more information on TDRs.

Additional Asset Quality Indicators

We have two overall portfolio segments – Commercial Lending and Consumer Lending. Each of these two segments comprises multiple loan classes. Classes are characterized by similarities in initial measurement, risk attributes and the manner in which we monitor and assess credit risk. The Commercial Lending segment is composed of the commercial, commercial real estate and equipment lease financing loan classes. The Consumer Lending segment is composed of the home equity, residential real estate, credit card and other consumer loan classes.

Commercial Lending Asset Classes

Commercial Loan Class

For commercial loans, we monitor the performance of the borrower in a disciplined and regular manner based upon the level of credit risk inherent in the loan. To evaluate the level of credit risk, we assign an internal risk rating reflecting the borrower’s PD and LGD. This two-dimensional credit risk rating methodology provides granularity in the risk monitoring process on an ongoing basis. These ratings are reviewed and updated, generally at least once per year. Additionally, no less frequently than on an annual basis, we review PD rates related to each rating grade based upon internal historical data. These rates are updated as needed and augmented by market data as deemed necessary. For small balance homogeneous pools of commercial loans, mortgages and leases, we apply statistical modeling to assist in determining the probability of default within these pools. Further, on a periodic basis, we update our LGD estimates associated with each rating grade based upon historical data. The combination of the PD and LGD ratings assigned to commercial loans, capturing both the combination of expectations of default and loss severity in event of default, reflects the relative estimated likelihood of loss at the reporting date. In general, loans with better PD and LGD tend to have a lower likelihood of loss compared to loans with worse PD and LGD. The loss amount also considers an estimate of exposure at date of default, which we also periodically update based upon historical data.

Based upon the amount of the lending arrangement and our risk rating assessment, we follow a formal schedule of written periodic review. Quarterly, we conduct formal reviews of a market’s or business unit’s entire loan portfolio, focusing on those loans which we perceive to be of higher risk, based upon PDs and LGDs, or loans for which credit quality is weakening. If circumstances warrant, it is our practice to review any customer obligation and its level of credit risk more frequently. We attempt to proactively manage our loans by using various procedures that are customized to the risk of a given loan, including ongoing outreach, contact, and assessment of obligor financial conditions, collateral inspection and appraisal.

Commercial Real Estate Loan Class

We manage credit risk associated with our commercial real estate projects and commercial mortgages similar to commercial loans by analyzing PD and LGD. Additionally, risks associated with commercial real estate projects and commercial mortgage activities tend to be correlated to the loan structure and collateral location, project progress and business environment. As a result, these attributes are also monitored and utilized in assessing credit risk.

As with the commercial class, a formal schedule of periodic review is also performed to assess market/geographic risk and business unit/industry risk. Often as a result of these overviews, more in-depth reviews and increased scrutiny are placed on areas of higher risk, including adverse changes in risk ratings, deteriorating operating trends, and/or areas that concern management. These reviews are designed to assess risk and take actions to mitigate our exposure to such risks.

Equipment Lease Financing Loan Class

We manage credit risk associated with our equipment lease financing loan class similar to commercial loans by analyzing PD and LGD.

Based upon the dollar amount of the lease and the level of credit risk, we follow a formal schedule of periodic review. Generally, this occurs quarterly, although we have established practices to review such credit risk more frequently if circumstances warrant. Our review process entails analysis of the following factors: equipment value/residual value, exposure levels, jurisdiction risk, industry risk, guarantor requirements, and regulatory compliance as applicable.

Table 41: Commercial Lending Asset Quality Indicators (a)



			Criticized Commercial Loans
In millions	Pass Rated		Special Mention (b)		Substandard (c)		Doubtful (d)		Total Loans
December 31, 2017
Commercial	$	105,280	$	1,858	$	3,331	$	58	$	110,527
Commercial real estate	28,380		148		435		15		28,978
Equipment lease financing	7,754		77		102		1		7,934
Total commercial lending	$	141,414	$	2,083	$	3,868	$	74	$	147,439
December 31, 2016
Commercial	$	96,231	$	1,612	$	3,449	$	72	$	101,364
Commercial real estate	28,561		98		327		24		29,010
Equipment lease financing	7,395		89		91		6		7,581
Total commercial lending	$	132,187	$	1,799	$	3,867	$	102	$	137,955


(a)	Loans are classified as “Pass”, “Special Mention”, “Substandard” and “Doubtful” based on the Regulatory Classification definitions. We use PDs and LGDs to rate commercial loans.


(b)	Special Mention rated loans have a potential weakness that deserves management’s close attention. If left uncorrected, these potential weaknesses may result in deterioration of repayment prospects at some future date. These loans do not expose us to sufficient risk to warrant a more adverse classification at the reporting date.


(c)	Substandard rated loans have a well-defined weakness or weaknesses that jeopardize the collection or liquidation of debt. They are characterized by the distinct possibility that we will sustain some loss if the deficiencies are not corrected.


(d)	Doubtful rated loans possess all the inherent weaknesses of a Substandard loan with the additional characteristics that the weakness makes collection or liquidation in full improbable due to existing facts, conditions, and values.

Consumer Lending Asset Classes

Home Equity and Residential Real Estate Loan Classes

We use several credit quality indicators, including delinquency information, nonperforming loan information, updated credit scores, originated and updated loan-to-value (LTV) ratios, and geography, to monitor and manage credit risk within the home equity and residential real estate loan classes. We evaluate mortgage loan performance by source originators and loan servicers. A summary of asset quality indicators follows:

Delinquency/Delinquency Rates: We monitor trending of delinquency/delinquency rates for home equity and residential real estate loans. See Table 39 for additional information.

Nonperforming Loans: We monitor trending of nonperforming loans for home equity and residential real estate loans. See Table 39 for additional information.

Credit Scores: We use a national third-party provider to update FICO credit scores for home equity loans and lines of credit and residential real estate loans at least quarterly. The updated scores are incorporated into a series of credit management reports, which are utilized to monitor the risk in the loan classes.

LTV (inclusive of combined loan-to-value (CLTV) for first and subordinate lien positions): At least annually, we update the property values of real estate collateral and calculate an updated LTV ratio. For open-end credit lines secured by real estate in regions experiencing significant declines in property values, more frequent valuations may occur. We examine LTV migration and stratify LTV into categories to monitor the risk in the loan classes.

Historically, we used, and we continue to use, a combination of original LTV and updated LTV for internal risk management and reporting purposes (e.g., line management, loss mitigation strategies). In addition to the fact that estimated property values by their nature are estimates, given certain data limitations it is important to note that updated LTVs may be based upon management’s assumptions (e.g., if an updated LTV is not provided by the third-party service provider, home price index (HPI) changes will be incorporated in arriving at management’s estimate of updated LTV).

Updated LTV is estimated using modeled property values. The related estimates and inputs are based upon an approach that uses a combination of third-party automated valuation models (AVMs), broker price opinions (BPOs), HPI indices, property location, internal and external balance information, origination data and management assumptions. We generally utilize origination lien balances provided by a third-party, where applicable, which do not include an amortization assumption when calculating updated LTV. Accordingly, the results of the calculations do not represent actual appraised loan level collateral or updated LTV based upon lien balances held by others, and as such, are necessarily imprecise and subject to change as we refine our methodology.

Geography: Geographic concentrations are monitored to evaluate and manage exposures. Loan purchase programs are sensitive to, and focused within, certain regions to manage geographic exposures and associated risks.

A combination of updated FICO scores, originated and updated LTV ratios and geographic location assigned to home equity loans and lines of credit and residential real estate loans is used to monitor the risk in the loan classes. Loans with higher FICO scores and lower LTVs tend to have a lower level of risk. Conversely, loans with lower FICO scores, higher LTVs, and in certain geographic locations tend to have a higher level of risk.

The following table presents asset quality indicators for home equity and residential real estate balances, excluding consumer purchased impaired loans of $2.4 billion and $2.8 billion at December 31, 2017 and 2016, respectively, and government insured or guaranteed residential real estate mortgages of $.8 billion at both December 31, 2017 and 2016, respectively.

Table 42: Asset Quality Indicators for Home Equity and Residential Real Estate Loans – Excluding Purchased Impaired and Government Insured or Guaranteed Loans (a)



	Home Equity				Residential Real Estate		Total
December 31, 2017 – in millions	1st Liens		2nd Liens		Residential Real Estate		Total
Current estimated LTV ratios
Greater than or equal to 125% and updated FICO scores:
Greater than 660	$	108	$	385	$	126	$	619
Less than or equal to 660 (b)	21		64		23		108
Missing FICO	1		5		1		7
Greater than or equal to 100% to less than 125% and updated FICO scores:
Greater than 660	300		842		253		1,395
Less than or equal to 660 (b)	46		143		45		234
Missing FICO	2		9		5		16
Greater than or equal to 90% to less than 100% and updated FICO scores:
Greater than 660	331		890		324		1,545
Less than or equal to 660	55		134		55		244
Missing FICO	2		9		4		15
Less than 90% and updated FICO scores:
Greater than 660	13,954		8,066		13,445		35,465
Less than or equal to 660	1,214		774		507		2,495
Missing FICO	42		57		95		194
Total home equity and residential real estate loans	$	16,076	$	11,378	$	14,883	$	42,337



	Home Equity				Residential Real Estate		Total
December 31, 2016 – in millions	1st Liens		2nd Liens		Residential Real Estate		Total
Current estimated LTV ratios
Greater than or equal to 125% and updated FICO scores:
Greater than 660	$	161	$	629	$	174	$	964
Less than or equal to 660 (b)	32		110		35		177
Missing FICO	1		9		2		12
Greater than or equal to 100% to less than 125% and updated FICO scores:
Greater than 660	394		1,190		345		1,929
Less than or equal to 660 (b)	66		211		76		353
Missing FICO	3		10		7		20
Greater than or equal to 90% to less than 100% and updated FICO scores:
Greater than 660	453		1,100		463		2,016
Less than or equal to 660	77		171		78		326
Missing FICO	1		8		6		15
Less than 90% and updated FICO scores:
Greater than 660	14,047		7,913		11,153		33,113
Less than or equal to 660	1,323		822		586		2,731
Missing FICO	42		55		102		199
Missing LTV and updated FICO scores:
Greater than 600					1		1
Total home equity and residential real estate loans	$	16,600	$	12,228	$	13,028	$	41,856


(a)	Amounts shown represent recorded investment.

(b)

Higher risk loans are defined as loans with both an updated FICO score of less than or equal to 660 and an updated LTV greater than or equal to 100%. The following states had the highest percentage of higher risk loans at December 31, 2017: New Jersey 17%, Pennsylvania 13%, Illinois 13%, Ohio 9%, Maryland 8%, Florida 6%, North Carolina 5% and Michigan 4%. The remainder of the states had lower than 4% of the higher risk loans individually, and collectively they represent approximately 25% of the higher risk loans. The following states had the highest percentage of higher risk loans at December 31, 2016: New Jersey 16%, Pennsylvania 14%, Illinois 12%, Ohio 10%, Florida 7%, Maryland 6%, Michigan 4% and North Carolina 4%. The remainder of the states had lower than 4% of the high risk loans individually, and collectively they represent approximately 27% of the higher risk loans.

Credit Card and Other Consumer Loan Classes

We monitor a variety of asset quality information in the management of the credit card and other consumer loan classes. Other consumer loan classes include education, automobile, and other secured and unsecured lines and loans. Along with the trending of delinquencies and losses for each class, FICO credit score updates are generally obtained monthly, as well as a variety of credit bureau attributes. Loans with high FICO scores tend to have a lower likelihood of loss. Conversely, loans with low FICO scores tend to have a higher likelihood of loss.

Table 43: Credit Card and Other Consumer Loan Classes Asset Quality Indicators



	Credit Card				Other Consumer (a)
Dollars in millions	Amount		% of Total Loans Using FICO Credit Metric		Amount		% of Total Loans Using FICO Credit Metric
December 31, 2017
FICO score greater than 719	$	3,457	61	%	$	10,366	63	%
650 to 719	1,596		28	%	4,352		27	%
620 to 649	250		4	%	659		4	%
Less than 620	272		5	%	715		4	%
No FICO score available or required (b)	124		2	%	314		2	%
Total loans using FICO credit metric	5,699		100	%	16,406		100	%
Consumer loans using other internal credit metrics (a)					5,338
Total loan balance	$	5,699			$	21,744
Weighted-average updated FICO score (b)			735				741
December 31, 2016
FICO score greater than 719	$	3,244	61	%	$	10,247	65	%
650 to 719	1,466		28	%	3,873		25	%
620 to 649	215		4	%	552		3	%
Less than 620	229		4	%	632		4	%
No FICO score available or required (b)	128		3	%	489		3	%
Total loans using FICO credit metric	5,282		100	%	15,793		100	%
Consumer loans using other internal credit metrics (a)					6,256
Total loan balance	$	5,282			$	22,049
Weighted-average updated FICO score (b)			736				744

(a)

We use updated FICO scores as an asset quality indicator for non-government guaranteed or insured education loans, automobile loans and other secured and unsecured lines and loans. We use internal credit metrics, such as delinquency status, geography or other factors, as an asset quality indicator for government guaranteed or insured education loans and consumer loans to high net worth individuals, as internal credit metrics are more relevant than FICO scores for these types of loans.

(b)

Credit card loans and other consumer loans with no FICO score available or required generally refers to new accounts issued to borrowers with limited credit history, accounts for which we cannot obtain an updated FICO score (e.g., recent profile changes), cards issued with a business name and/or cards secured by collateral. Management proactively assesses the risk and size of this loan portfolio and, when necessary, takes actions to mitigate the credit risk. Weighted-average updated FICO score excludes accounts with no FICO score available or required.

Troubled Debt Restructurings (TDRs)

A TDR is a loan whose terms have been restructured in a manner that grants a concession to a borrower experiencing financial difficulty. TDRs result from our loss mitigation activities, and include rate reductions, principal forgiveness, postponement/reduction of scheduled amortization, and extensions, which are intended to minimize economic loss and to avoid foreclosure or repossession of collateral. Additionally, TDRs also result from borrowers that have been discharged from personal liability through Chapter 7 bankruptcy and have not formally reaffirmed their loan obligations to us. In those situations where principal is forgiven, the amount of such principal forgiveness is immediately charged off.

Some TDRs may not ultimately result in the full collection of principal and interest, as restructured, and result in potential incremental losses. These potential incremental losses have been factored into our overall ALLL estimate. The level of any subsequent defaults will likely be affected by future economic conditions. Once a loan becomes a TDR, it will continue to be reported as a TDR until it is ultimately repaid in full, the collateral is foreclosed upon, or it is fully charged off. We held specific reserves in the ALLL of $.2 billion and $.3 billion at December 31, 2017 and December 31, 2016, respectively, for the total TDR portfolio.

Table 44 quantifies the number of loans that were classified as TDRs as well as the change in the loans’ recorded investment as a result of becoming a TDR during the years 2017, 2016 and 2015. Additionally, the table provides information about the types of TDR concessions. The Principal Forgiveness TDR category includes principal forgiveness and accrued interest forgiveness. The Rate Reduction TDR category includes reduced interest rate and interest deferral. The Other TDR category primarily includes consumer borrowers that have been discharged from personal liability through Chapter 7 bankruptcy and have not formally reaffirmed their loan obligations to us, as well as postponement/reduction of scheduled amortization and contractual extensions for both consumer and commercial borrowers.

In some cases, there have been multiple concessions granted on one loan. This is most common within the commercial loan portfolio. When there have been multiple concessions granted in the commercial loan portfolio, the principal forgiveness concession was prioritized for purposes of determining the inclusion in Table 44. Second in priority would be rate reduction. In the event that multiple concessions are granted on a consumer loan, concessions resulting from discharge from personal liability through Chapter 7 bankruptcy without formal affirmation of the loan obligations to us would be prioritized and included in the Other type of concession in Table 44. After that, consumer loan concessions would follow the previously discussed priority of concessions for the commercial loan portfolio.

Table 44: Financial Impact and TDRs by Concession Type (a)



	Number of Loans	Pre-TDR Recorded Investment (b)		Post-TDR Recorded Investment (c)
During the year ended December 31, 2017 Dollars in millions	Number of Loans	Pre-TDR Recorded Investment (b)		Principal Forgiveness		Rate Reduction		Other		Total
Total commercial lending	120	$	293	$	18	$	7	$	227	$	252
Total consumer lending	11,993	248				146		97		243
Total TDRs	12,113	$	541	$	18	$	153	$	324	$	495
During the year ended December 31, 2016 Dollars in millions
Total commercial lending	143	$	524			$	57	$	413	$	470
Total consumer lending	11,262	245				157		76		233
Total TDRs	11,405	$	769			$	214	$	489	$	703
During the year ended December 31, 2015 Dollars in millions
Total commercial lending	158	$	284	$	22	$	4	$	198	$	224
Total consumer lending	10,962	311				190		106		296
Total TDRs	11,120	$	595	$	22	$	194	$	304	$	520


(a)	Impact of partial charge-offs at TDR date are included in this table.


(b)	Represents the recorded investment of the loans as of the quarter end prior to TDR designation, and excludes immaterial amounts of accrued interest receivable.


(c)	Represents the recorded investment of the TDRs as of the end of the quarter in which the TDR occurs, and excludes immaterial amounts of accrued interest receivable.

After a loan is determined to be a TDR, we continue to track its performance under its most recent restructured terms. We consider a TDR to have subsequently defaulted when it becomes 60 days past due after the most recent date the loan was restructured. The recorded investment of loans that were both (i) classified as TDRs or were subsequently modified during each 12-month period preceding January 1, 2017, 2016 and 2015, and (ii) subsequently defaulted during the 12-month period following each of January 1, 2017, 2016 and 2015, totaled $.1 billion, $.2 billion and $.1 billion, respectively.

Impaired Loans

Impaired loans include commercial and consumer nonperforming loans and TDRs, regardless of nonperforming status. TDRs that were previously recorded at amortized cost and are now classified and accounted for as held for sale are also included. Excluded from impaired loans are nonperforming leases, loans accounted for as held for sale other than the TDRs described in the preceding sentence, loans accounted for under the fair value option, smaller balance homogeneous type loans and purchased impaired loans. We did not recognize any interest income on impaired loans that have not returned to performing status, while they were impaired during the year ended December 31, 2017 and December 31, 2016. The following table provides further detail on impaired loans individually evaluated for impairment and the associated ALLL. Certain commercial and consumer impaired loans do not have a related ALLL as the valuation of these impaired loans exceeded the recorded investment.

Table 45: Impaired Loans



In millions	Unpaid Principal Balance		Recorded Investment		Associated Allowance		Average Recorded Investment (a)
December 31, 2017
Impaired loans with an associated allowance
Total commercial lending	$	580	$	353	$	76	$	419
Total consumer lending	1,061		1,014		195		1,072
Total impaired loans with an associated allowance	1,641		1,367		271		1,491
Impaired loans without an associated allowance
Total commercial lending	494		366				330
Total consumer lending	1,019		638				648
Total impaired loans without an associated allowance	1,513		1,004				978
Total impaired loans	$	3,154	$	2,371	$	271	$	2,469
December 31, 2016
Impaired loans with an associated allowance
Total commercial lending	$	742	$	477	$	105	$	497
Total consumer lending	1,237		1,185		226		1,255
Total impaired loans with an associated allowance	1,979		1,662		331		1,752
Impaired loans without an associated allowance
Total commercial lending	447		322				365
Total consumer lending	982		608				604
Total impaired loans without an associated allowance	1,429		930				969
Total impaired loans	$	3,408	$	2,592	$	331	$	2,721


(a)	Average recorded investment is for the years ended December 31, 2017 and 2016.