ALLOWANCE FOR CREDIT LOSSES, NONPERFORMING ASSETS, AND CONCENTRATIONS OF CREDIT RISK | NOTE 4 - ALLOWANCE FOR CREDIT LOSSES, NONACCRUING LOANS AND LEASES, AND CONCENTRATIONS OF CREDIT RISK Allowance for Credit Losses Management’s estimate of expected credit losses in the Company’s loan and lease portfolios is recorded in the ALLL and the reserve for unfunded lending commitments (collectively the ACL). See Note 5 in the Company’s 2019 Form 10-K for a detailed discussion of the ACL reserve methodology and estimation techniques as of December 31, 2019. Upon adoption of CECL effective January 1, 2020, the Company’s ACL reserve methodology changed to estimate expected credit losses over the contractual life of loans and leases. The ACL is maintained at a level the Company believes to be appropriate to absorb expected lifetime credit losses over the contractual life of the loan and lease portfolios and on the unfunded lending commitments. The determination of the ACL is based on periodic evaluation of the loan and lease portfolios and unfunded lending commitments that are not unconditionally cancelable considering a number of relevant underlying factors, including key assumptions and evaluation of quantitative and qualitative information. Key assumptions used in the ACL measurement process include the use of a two-year reasonable and supportable economic forecast period followed by a one-year period during which the expected credit losses revert to long-term historical macroeconomic inputs. The evaluation of quantitative and qualitative information is performed through assessments of groups of assets that share similar risk characteristics and certain individual loans and leases that do not share similar risk characteristics with the collective group. Loans are grouped generally by product type (e.g., commercial, commercial real estate, residential mortgage, etc.), and significant loan portfolios are assessed for credit losses using econometric models. The quantitative evaluation of the adequacy of the ACL utilizes a single economic forecast as its foundation, and is primarily based on econometric models that use known or estimated data as of the balance sheet date and forecasted data over the reasonable and supportable period. Known and estimated data include current PD, LGD and EAD (for commercial loans and leases), timing and amount of expected draws (for unfunded lending commitments), FICO, LTV, term and time on books (for retail loans), mix and level of loan balances, delinquency levels, assigned risk ratings, previous loss experience, current business conditions, amounts and timing of expected future cash flows, and factors particular to a specific commercial credit such as competition, business and management performance. Forward-looking economic assumptions include real gross domestic product, unemployment rate, interest rate curve, and changes in collateral values. This data is aggregated to estimate expected credit losses over the contractual life of the loans and leases, adjusted for expected prepayments. In highly volatile economic environments historical information, such as commercial customer financial statements or consumer credit ratings, may not be as important to estimating future expected losses as forecasted inputs to the models. The ACL may also be affected materially by a variety of qualitative factors that the Company considers to reflect current judgment of various events and risks that are not measured in the statistical procedures including uncertainty related to the economic forecasts used in the modeled credit loss estimates, loan growth, back testing results, credit underwriting policy exceptions, regulatory and audit findings, and peer comparisons. The qualitative allowance is further informed by multiple alternative scenarios to support the period-end ACL balance. The measurement process results in specific or pooled allowances for loans, leases and unfunded lending commitments, and qualitative allowances that are judgmentally determined and applied across the portfolio. There are certain loan portfolios that may not need an econometric model to enable the Company to calculate management’s best estimate of the expected credit losses. Less data intensive, non-modeled approaches to estimating losses are considered more efficient and practical for portfolios that have lower levels of outstanding balances (e.g., runoff or closed portfolios, new products or products that are not significant to the Company’s overall credit risk exposure). Loans and leases that do not share similar risk characteristics are individually assessed for expected credit losses. Nonaccruing commercial and commercial real estate loans with an outstanding balance of $5 million or greater and all commercial and commercial real estate TDRs (regardless of size) are assessed on an individual loan level basis. Generally, the measurement of ACL on individual loans and leases is the present value of its future cash flows or the fair value of its underlying collateral, if the loan or lease is collateral dependent. Loans that are deemed to be collateral dependent are written down to the fair value, less costs to sell, if sale of the collateral is expected as of the evaluation date and are reassessed each subsequent period to determine if a change to the ACL is required. Subsequent evaluations may result in an increase or decrease to the ACL, based on a corresponding change in the fair value of the collateral during the period. Any subsequent decrease to the ACL (because of an increase to the collateral-dependent loan’s fair value) is limited to the total amount previously written off for that loan. For retail TDRs that are not collateral dependent, the ACL is developed using the present value of expected future cash flows compared to the amortized cost basis in the loans. Expected re-default factors are considered in this analysis. Retail TDRs that are deemed collateral dependent are written down to fair market value less cost to sell. Expected recoveries are considered in management’s estimate of the ACL and may result in a negative adjustment (i.e., reduction) to the ACL balance. A loan is collateral dependent if repayment is expected to be provided substantially through the operation or sale of the collateral when the borrower is experiencing financial difficulty as of the evaluation date. Generally, repayment would be expected to be provided substantially by the sale or continued operation of the underlying collateral if cash flows to repay the loan from all other available sources (including guarantors) are expected to be no more than nominal. If repayment is dependent only on the operation of the collateral, the fair value of the collateral would not be adjusted for estimated costs to sell. If a loan is considered collateral dependent, the ACL is calculated as the difference between the fair value of collateral (adjusted for the costs to sell if the sale of the collateral is expected) and the amortized cost basis as of the evaluation date. It is possible to have a negative ACL for a collateral dependent loan if the fair value of the collateral increases in a subsequent reporting period. The negative ACL cannot exceed the total amount previously charged off. Accrued interest receivable on loans and leases is excluded from asset balances used to calculate the ACL. All accrued and uncollected interest is immediately reversed against interest income when a loan or lease is placed on nonaccrual status. Uncollectible interest is written off timely in accordance with regulatory guidelines. Generally, loans and leases are placed on nonaccrual status when contractually past due 90 days or more, or earlier if management believes that the probability of collection is insufficient to warrant further accrual. Residential mortgages are placed on nonaccrual status when contractually past due 120 days or more, or sooner if deemed collateral dependent, unless guaranteed by the Federal Housing Administration. Loans in COVID-19 pandemic-related forbearance programs continue to accrue interest during the forbearance period; a reserve is established for interest income expected to be uncollectible following forbearance. The amount of accrued interest receivable reversed against interest income for the three months ended September 30, 2020 was $4 million for total commercial and retail, respectively. Accrued interest reversed against interest income for the nine months ended September 30, 2020 was $7 million and $13 million for total commercial and retail, respectively. The Company estimates expected credit losses associated with off-balance sheet financial instruments such as standby letters of credit, financial guarantees and unfunded loan commitments that are not unconditionally cancelable. Off-balance sheet financial instruments are subject to individual reviews and are analyzed and segregated by risk according to the Company’s internal risk rating scale. These risk classifications, in conjunction with historical loss experience, current and future economic conditions, timing and amount of expected draws, and performance trends within specific portfolio segments, result in the estimate of the reserve for unfunded lending commitments. The Company does not recognize a reserve for future draws from credit lines that are unconditionally cancelable (e.g., credit cards). The ALLL and the reserve for unfunded lending commitments are reported on the Consolidated Balance Sheets in the allowance for loan and lease losses and in other liabilities, respectively. Provision for credit losses related to the loan and lease portfolios and the unfunded lending commitments are reported in the Consolidated Statements of Operations as provision for credit losses. The following table presents a summary of changes in the ALLL and the reserve for unfunded lending commitments for the three and nine months ended September 30, 2020 : Three Months Ended September 30, 2020 Nine Months Ended September 30, 2020 (in millions) Commercial Retail Total Commercial Retail Total Allowance for loan and lease losses, beginning of period $1,235 $1,213 $2,448 $674 $578 $1,252 Cumulative effect of change in accounting principle — — — (176 ) 629 453 Allowance for loan and lease losses, beginning of period, adjusted 1,235 1,213 2,448 498 1,207 1,705 Charge-offs (171 ) (86 ) (257 ) (292 ) (319 ) (611 ) Recoveries 1 37 38 7 101 108 Net charge-offs (170 ) (49 ) (219 ) (285 ) (218 ) (503 ) Provision charged to income 224 89 313 1,076 264 1,340 Allowance for loan and lease losses, end of period $1,289 $1,253 $2,542 $1,289 $1,253 $2,542 Reserve for unfunded lending commitments, beginning of period $69 $10 $79 $44 $— $44 Cumulative effective of change in accounting principle — — — (3 ) 1 (2 ) Reserve for unfunded lending commitments, beginning of period, adjusted 69 10 79 41 1 42 Provision for unfunded lending commitments 83 32 115 111 41 152 Reserve for unfunded lending commitments, end of period $152 $42 $194 $152 $42 $194 The following table provides additional detail on the cumulative effect of change in accounting principle on the ACL and related coverage ratios: December 31, 2019 January 1, 2020 September 30, 2020 (in millions) Amortized Cost Basis ACL Balance Coverage Impact of Adoption of CECL ACL Balance Coverage Amortized Cost Basis ACL Balance Coverage Commercial (1) $41,479 $575 1.4 % ($199 ) $376 0.9 % $45,185 $826 1.8 % Commercial real estate 13,522 124 0.9 (57 ) 67 0.5 14,889 548 3.7 Leases 2,537 19 0.7 77 96 3.8 2,288 67 2.9 Total commercial loans and leases 57,538 718 1.2 (179 ) 539 0.9 62,362 1,441 2.3 Residential 19,083 35 0.2 95 130 0.7 19,633 133 0.7 Home equity 13,154 83 0.6 73 156 1.2 12,322 156 1.3 Automobile 12,120 123 1.0 83 206 1.7 12,035 221 1.8 Education 10,347 116 1.1 298 414 4.0 11,631 386 3.3 Other retail 6,846 221 3.2 81 302 4.4 6,088 399 6.6 Total retail loans 61,550 578 0.9 630 1,208 2.0 61,709 1,295 2.1 Total loans and leases $119,088 $1,296 1.1 % $451 $1,747 1.5 % $124,071 $2,736 2.2 % (1) The commercial coverage ratio includes a 21 basis point reduction associated with PPP loans as of September 30, 2020 . The difference in ACL as of September 30, 2020 as compared to December 31, 2019 continues to be driven by the COVID-19 pandemic and associated lockdowns and the resulting economic impacts from March to September 2020, as well as the Company’s adoption of CECL on January 1, 2020. Citizens added $451 million in ACL upon adoption of CECL, and has since added an additional $989 million in the nine months ended September 30, 2020, resulting in an ending ACL balance of $2.7 billion . The increase in commercial net charge-offs in the nine months ended September 30, 2020 as compared to the nine months ended September 30, 2019 was driven by charge-offs in CRE, metals and mining, oil and gas, and casual dining industry sectors. Retail net charge-offs were flat in the nine months ended September 20, 2020 as compared to the nine months ended September 30, 2019. To determine the ACL as of September 30, 2020, Citizens utilized an economic scenario that generally reflects real GDP growth of 4.5% over 2021, returning to fourth quarter 2019 real GDP levels by the first quarter of 2022. The scenario also projects the fourth quarter 2020 unemployment rate to be in the range of 9% to 9.5%, and falling to 7% to 7.5% by the fourth quarter of 2021. While the macroeconomic forecast was slightly improved relative to the second quarter 2020 forecast, Citizens continued to apply management judgment to adjust the modeled reserves in the commercial industry sectors most impacted by the COVID-19-related lockdowns, including in retail and hospitality, casual dining, retail trade, price-sensitive energy and related, and educational services, as well as in certain retail products. The following table presents a summary of changes in the ALLL and the reserve for unfunded lending commitments for the three and nine months ended September 30, 2019 : Three Months Ended September 30, 2019 Nine Months Ended September 30, 2019 (in millions) Commercial Retail Total Commercial Retail Total Allowance for loan and lease losses, beginning of period $680 $547 $1,227 $690 $552 $1,242 Charge-offs (35 ) (124 ) (159 ) (106 ) (347 ) (453 ) Recoveries 3 43 46 17 128 145 Net charge-offs (32 ) (81 ) (113 ) (89 ) (219 ) (308 ) Provision charged to income 64 85 149 111 218 329 Allowance for loan and lease losses, end of period $712 $551 $1,263 $712 $551 $1,263 Reserve for unfunded lending commitments, beginning of period $93 $— $93 $91 $— $91 Provision for unfunded lending commitments (48 ) — (48 ) (46 ) — (46 ) Reserve for unfunded lending commitments, end of period $45 $— $45 $45 $— $45 Credit Quality Indicators Loan and lease portfolio segments and classes, excluding LHFS, are presented by credit quality indicator and vintage year. Citizens defines the vintage date for the purpose of this disclosure as the date of the most recent credit decision. In general, renewals are categorized as new credit decisions and reflect the renewal date as the vintage date. Loans modified in a TDR are considered to be a continuation of the original loan and vintage date corresponds with the initial loan origination date. For commercial loans and leases, Citizens utilizes regulatory classification ratings to monitor credit quality. Regulatory classification ratings are assigned at loan origination and are periodically re-evaluated by Citizens utilizing a risk-based approach, or at any time management becomes aware of information affecting the borrowers' ability to fulfill their obligations. Both quantitative and qualitative factors are considered in this review process. Loans with a “pass” rating are those that the Company believes will be fully repaid in accordance with the contractual loan terms. Commercial loans and leases that are “criticized” are those that have some weakness or potential weakness that indicate an increased probability of future loss. “Criticized” loans are grouped into three categories, “special mention,” “substandard” and “doubtful.” Special mention loans have potential weaknesses that, if left uncorrected, may result in deterioration of the Company’s credit position at some future date. Substandard loans are inadequately protected loans; these loans have well-defined weaknesses that could hinder normal repayment or collection of the debt. Doubtful loans have the same weaknesses as substandard, with the added characteristics that the possibility of loss is high and collection of the full amount of the loan is improbable. The following table presents the amortized cost basis of commercial loans and leases, by vintage date and regulatory classification rating, as of September 30, 2020 : Term Loans by Origination Year Revolving Loans (in millions) 2020 2019 2018 2017 2016 Prior to 2016 Within the Revolving Period Converted to Term Total Commercial Pass (1) $7,478 $6,823 $4,838 $2,747 $1,515 $2,453 $15,115 $182 $41,151 Special Mention 58 275 237 149 92 222 892 125 2,050 Substandard 60 220 326 100 85 140 630 19 1,580 Doubtful 52 22 35 43 20 42 186 4 404 Total commercial 7,648 7,340 5,436 3,039 1,712 2,857 16,823 330 45,185 Commercial real estate Pass 1,774 3,406 3,608 1,553 825 1,112 1,073 — 13,351 Special Mention 11 216 128 237 171 9 77 — 849 Substandard 68 1 170 50 53 66 60 — 468 Doubtful 20 38 53 — 36 3 24 47 221 Total commercial real estate 1,873 3,661 3,959 1,840 1,085 1,190 1,234 47 14,889 Leases Pass 332 327 265 169 200 917 — — 2,210 Special Mention — 2 3 6 5 25 — — 41 Substandard — 2 2 5 4 — — — 13 Doubtful — — 9 1 3 11 — — 24 Total leases 332 331 279 181 212 953 — — 2,288 Total commercial loans and leases Pass (1) 9,584 10,556 8,711 4,469 2,540 4,482 16,188 182 56,712 Special Mention 69 493 368 392 268 256 969 125 2,940 Substandard 128 223 498 155 142 206 690 19 2,061 Doubtful 72 60 97 44 59 56 210 51 649 Total commercial loans and leases $9,853 $11,332 $9,674 $5,060 $3,009 $5,000 $18,057 $377 $62,362 (1) Includes $4.7 billion of PPP loans designated as pass that are fully guaranteed by the SBA originating in 2020. For retail loans, Citizens utilizes credit scores provided by FICO which are generally refreshed on a quarterly basis and the loan’s payment and delinquency status to monitor credit quality. Management believes FICO credit scores are considered the strongest indicator of credit losses over the contractual life of the loan as the scores are based on current and historical national industry-wide consumer level credit performance data, and assist management in predicting the borrower’s future payment performance. The following table presents the amortized cost basis of retail loans, by vintage date and FICO scores, as of September 30, 2020 : Term Loans by Origination Year Revolving Loans (in millions) 2020 2019 2018 2017 2016 Prior to 2016 Within the Revolving Period Converted to Term Total Residential mortgages 800+ $1,971 $1,983 $751 $1,297 $1,833 $2,061 $— $— $9,896 740-799 2,351 1,341 463 612 776 977 — — 6,520 680-739 596 408 193 207 317 500 — — 2,221 620-679 101 97 44 51 69 235 — — 597 <620 17 23 34 56 54 197 — — 381 No FICO available (1) 4 1 — 1 — 12 — — 18 Total residential mortgages 5,040 3,853 1,485 2,224 3,049 3,982 — — 19,633 Home equity 800+ 3 9 11 7 5 245 4,319 360 4,959 740-799 1 7 7 7 4 204 3,224 340 3,794 680-739 — 4 9 14 8 197 1,671 293 2,196 620-679 — 8 16 19 12 146 420 201 822 <620 1 15 27 30 18 129 117 213 550 No FICO available (1) — — — — — — 1 — 1 Total home equity 5 43 70 77 47 921 9,752 1,407 12,322 Automobile 800+ 780 862 472 358 209 91 — — 2,772 740-799 1,130 1,127 606 406 214 87 — — 3,570 680-739 1,028 1,012 527 327 170 69 — — 3,133 620-679 523 557 300 185 102 46 — — 1,713 <620 90 245 202 157 96 48 — — 838 No FICO available (1) 1 1 — — — 7 — — 9 Total automobile 3,552 3,804 2,107 1,433 791 348 — — 12,035 Education 800+ 1,191 1,276 810 772 586 776 — — 5,411 740-799 1,307 1,151 621 449 292 443 — — 4,263 680-739 385 378 217 160 110 239 — — 1,489 620-679 27 52 42 36 31 108 — — 296 <620 2 7 13 14 12 56 — — 104 No FICO available (1) 6 — — — — 62 — — 68 Total education 2,918 2,864 1,703 1,431 1,031 1,684 — — 11,631 Other retail 800+ 286 474 174 79 17 49 313 — 1,392 740-799 419 611 217 97 23 33 621 2 2,023 680-739 356 409 141 60 13 17 555 6 1,557 620-679 195 156 51 19 3 6 181 7 618 <620 18 45 24 9 2 4 81 9 192 No FICO available (1) 36 1 — — — — 267 2 306 Total other retail 1,310 1,696 607 264 58 109 2,018 26 6,088 Retail 800+ 4,231 4,604 2,218 2,513 2,650 3,222 4,632 360 24,430 740-799 5,208 4,237 1,914 1,571 1,309 1,744 3,845 342 20,170 680-739 2,365 2,211 1,087 768 618 1,022 2,226 299 10,596 620-679 846 870 453 310 217 541 601 208 4,046 <620 128 335 300 266 182 434 198 222 2,065 No FICO available (1) 47 3 — 1 — 81 268 2 402 Total retail $12,825 $12,260 $5,972 $5,429 $4,976 $7,044 $11,770 $1,433 $61,709 (1) Represents loans for which an updated FICO score was unavailable (e.g., due to recent profile changes). Nonaccrual and Past Due Assets The following table presents nonaccrual loans and leases and loans accruing and 90 days or more past due: As of September 30, 2020 As of December 31, 2019 (in millions) Nonaccrual loans and leases 90+ days past due and accruing Nonaccrual with no related ACL Nonaccrual loans and leases Commercial $435 $3 $33 $240 Commercial real estate 323 — 3 2 Leases 2 — — 3 Total commercial loans and leases 760 3 36 245 Residential mortgages 131 17 101 93 Home equity 265 — 192 246 Automobile 80 — 18 67 Education 16 2 5 18 Other retail 25 6 1 34 Total retail 517 25 317 458 Total loans and leases $1,277 $28 $353 $703 Interest income is generally not recognized for loans and leases that are on nonaccrual status. The Company reverses accrued interest receivable with a charge to interest income upon classifying the loan or lease as nonaccrual. The following table presents an analysis of the age of both accruing and nonaccruing loan and lease past due amounts: September 30, 2020 December 31, 2019 Days Past Due Days Past Due (in millions) Current-29 30-59 60-89 90 or More Total Current-29 30-59 60-89 90 or More Total Commercial $44,845 $105 $129 $106 $45,185 $41,340 $45 $27 $67 $41,479 Commercial real estate 14,743 90 — 56 14,889 13,520 1 1 — 13,522 Leases 2,284 1 1 2 2,288 2,498 37 — 2 2,537 Total commercial loans and leases 61,872 196 130 164 62,362 57,358 83 28 69 57,538 Residential mortgages 19,430 64 16 123 19,633 18,947 35 17 84 19,083 Home equity 12,007 51 29 235 12,322 12,834 91 40 189 13,154 Automobile 11,825 147 52 11 12,035 11,788 227 81 24 12,120 Education 11,585 28 12 6 11,631 10,290 30 15 12 10,347 Other retail 6,005 33 22 28 6,088 6,729 45 31 41 6,846 Total retail loans 60,852 323 131 403 61,709 60,588 428 184 350 61,550 Total $122,724 $519 $261 $567 $124,071 $117,946 $511 $212 $419 $119,088 The Company estimates expected credit losses based on the fair value of collateral for collateralized loans that management believes will not be paid under the terms of the original loan contract. These loans are considered to be collateral dependent, and the estimated credit loss is calculated as the difference between the loan’s amortized cost basis and the fair value of the collateral as of each evaluation date. Collateral values for residential mortgage and home equity loans are based on refreshed valuations which are updated at least every 90 days less estimated costs to sell. At September 30, 2020 and December 31, 2019 , the Company had collateral-dependent residential mortgage and home equity loans totaling $489 million and $227 million , respectively. For collateral-dependent commercial loans, the ACL is individually assessed based on the fair value of the collateral. Various types of collateral are used, including real estate, inventory, equipment, accounts receivable, securities and cash, among others. For commercial real estate loans, collateral values are generally based on appraisals which are updated based on management judgment under the specific circumstances on a case-by-case basis. At September 30, 2020 and December 31, 2019 , the Company had collateral-dependent commercial loans totaling $144 million and $85 million , respectively. The amortized cost basis of mortgage loans collateralized by residential real estate property for which formal foreclosure proceedings were in process was $129 million and $152 million as of September 30, 2020 and December 31, 2019 , respectively. Troubled Debt Restructurings TDR is the classification given to a loan that has been restructured in a manner that grants a concession to a borrower experiencing financial hardship that the Company would not otherwise make. Citizens implemented various retail and commercial loan modification programs to provide borrowers relief from the economic impacts of COVID-19. The CARES Act and bank regulatory agencies provided guidance stating certain loan modifications to borrowers experiencing financial distress as a result of COVID-19 may not be accounted for as TDRs under U.S. GAAP. In accordance with the CARES Act, Citizens has elected to not apply TDR classification to any COVID-19 related loan modifications performed after March 1, 2020 for borrowers who were current as of December 31, 2019. In addition, for loans modified in response to the COVID-19 pandemic that are not eligible for relief from TDR classification under the CARES Act, the Company elected to apply the guidance issued by the bank regulatory agencies. Under this guidance, deferral of principal and interest for up to six months to borrowers who were current as of March 1, 2020 and impacted by COVID-19 are not classified as TDRs. For loan modifications that include a payment deferral and are not TDRs, the borrower’s past due and nonaccrual status will not be impacted during the deferral period. Interest income will continue to be recognized over the contractual life of the loan. The following table summarizes TDRs by class and total unfunded commitments: (in millions) September 30, 2020 December 31, 2019 Commercial $294 $297 Retail 715 667 Unfunded commitments related to TDRs 49 42 The following tables below summarize how loans were modified during the three and nine months ended September 30, 2020 and 2019 . The reported balances represent the post-modification outstanding amortized cost basis and can include loans that became TDRs during the period and were paid off in full, charged off, or sold prior to period end. Pre-modification balances for modified loans approximate the post-modification balances shown. Three Months Ended September 30, 2020 Primary Modification Types Interest Rate Reduction (1) Maturity Extension (2) Other (3) (dollars in millions) Number of Contracts Recorded Investment Number of Contracts Recorded Investment Number of Contracts Recorded Investment Commercial — $— 12 $103 2 $1 Commercial real estate — — — — — — Total commercial loans — — 12 103 2 1 Residential mortgages 47 9 41 6 19 4 Home equity 23 2 52 4 104 6 Automobile 25 1 47 — 1,119 18 Education — — — — 140 3 Other retail 410 1 — — 74 — Total retail loans 505 13 140 10 1,456 31 Total 505 $13 152 $113 1,458 $32 Three Months Ended September 30, 2019 Primary Modification Types Interest Rate Reduction (1) Maturity Extension (2) Other (3) (dollars in millions) Number of Contracts Amortized Cost Number of Contracts Amortized Cost Number of Contracts Amortized Cost Commercial 2 $— 6 $1 6 $15 Commercial real estate — — — — — — Total commercial loans 2 — 6 1 6 15 Residential mortgages 12 2 8 2 25 4 Home equity 63 6 16 1 120 6 Automobile 46 1 4 — 309 4 Education — — — — 131 2 Other retail 805 5 — — 218 — Total retail loans 926 14 28 3 803 16 Total 928 $14 34 $4 809 $31 Nine Months Ended September 30, 2020 Primary Modification Types Interest Rate Reduction (1) Maturity Extension (2) Other (3) (dollars in millions) Number of Contracts Amortized Cost Number of Contracts Amortized Cost Number of Contracts Amortized Cost Commercial — $— 18 $106 34 $95 Commercial real estate — — — — — — Total commercial loans — — 18 106 34 95 Residential mortgages 139 26 149 27 60 11 Home equity 96 8 107 8 365 21 Automobile 108 2 48 — 2,212 35 Education — — — — 373 9 Other retail 1,916 8 — — 251 2 Total retail loans 2,259 44 304 35 3,261 78 Total 2,259 $44 322 $141 3,295 $173 Nine Months Ended September 30, 2019 Primary Modification Types Interest Rate Reduction (1) Maturity Extension (2) Other (3) (dollars in millions) Number of Contracts Amortized Cost Number of Contracts Amortized Cost Number of Contracts Amortized Cost Commercial 3 $— 18 $2 24 $102 Commercial real estate — — 1 — — — Total commercial loans 3 — 19 2 24 102 Residential mortgages 25 6 29 5 87 13 Home equity 148 15 66 10 358 21 Automobile 111 2 16 — 933 13 Education — — — — 211 5 Other retail 2,362 14 — — 362 — Total retail loans 2,646 37 111 15 1,951 52 Total 2,649 $37 130 $17 1,975 $154 (1) Includes modifications that consist of multiple concessions, one of which is an interest rate reduction. (2) Includes modifications that consist of multiple concessions, one of which is a maturity extension (unless one of the other concessions was an interest rate reduction). (3) Includes modifications other than interest rate reductions or maturity extensions, such as lowering scheduled payments for a specified period of time, principal forgiveness, and capitalizing arrearages. Also included are the following: deferrals, trial modifications, certain bankruptcies, loans in forbearance and prepayment plans. Modifications can include the deferral of accrued interest resulting in post modification balances being higher than pre-modification. The net change to ALLL resulting from modification of loans for the three months ended September 30, 2020 and 2019 was ($30) million and $3 million , respectively. The net change to ALLL resulting from modifications of loans for the nine months ended September 30, 2020 and 2019 was ($21) million and $7 million , respectively. Charge-offs may also be recorded on TDRs. Citizens recorded $43 million and $1 million of charge-offs resulting from modification of loans in the three months ended September 30, 2020 and 2019 , respectively. Citizens recorded $49 million and $3 million for the nine months ended September 30, 2020 and 2019 , respectively. A payment default refers to a loan that becomes 90 days or more past due under the modified terms. Loan data includes loans meeting the criteria that were paid off in full, charged off, or sold prior to September 30, 2020 and 2019 . For commercial loans, recorded investment in TDRs that defaulted within 12 months of their modification date for the three months ended September 30, 2020 were $14 million and there were none for the three months ended September 30, 2019. There were $53 million and $1 million in the nine months ended September 30, 2020 and 2019 . For retail loans, there were $22 million and $9 million of loans which defaulted within their restructuring date for the three months ended September 30, 2020 and 2019 , respectively, and $47 million and $28 million of loans which defaulted within 12 months of their restructuring date for the nine months ended September 30, 2020 and 2019 , respectively. Concentrations of Credit Risk As of September 30, 2020 , under the Company’s COVID-19-related forbearance and other customer accommodation programs that are guided by the CARES Act as well as banking regulator interagency guidance, Citizens deferred payments on approximately $2.4 billion , or 3.8% , of the retail portfolio, approximately $795 million , or 1.4% , of the commercial portfolio, and approximately $464 million , or 8.1% , of the Business Banking portfolio. The vast majority of these deferrals are not classified as TDRs. Most of the Company’s lending activity is with customers located in the New England, Mid-Atlantic and Midwest regions. Generally, loans are collateralized by assets including real estate, inventory, accounts receivable, other personal property and investment securities. As of September 30, 2020 and December 31, 2019 , Citizens had a significant amount of loans collateralized by residential and commercial real estate. There were no significant concentration risks within the commercial loan or retail loan portfolios. Exposure to credit losses arising from lending transactions may fluctuate with fair values of collateral supporting loans, which may not perform according to contractual agreements. The Company’s policy is to collateralize loans to the extent necessary; however, unsecured loans are also granted on the basis of the financial strength of the applicant and the facts surrounding the transaction. Certain loan products, including residential mortgages, home equity loans and lines of credit, and c |