The Company accounts for credit losses on loans in accordance with ASC 326 - Financial Instruments - Credit Losses, to determine the ACL. ASC 326 requires the Company to recognize estimates for lifetime losses on loans and unfunded loan commitments at the time of origination or acquisition. The recognition of losses at origination or acquisition represents the Company’s best estimate of the lifetime expected credit loss associated with a loan given the facts and circumstances associated with the particular loan, and involves the use of significant management judgement and estimates. The Company uses a discounted cash flow model when determining estimates for the ACL for commercial real estate loans and commercial loans, which are the majority of the loan portfolio, and uses a historical loss rate model for retail loans. The Company also utilizes proxy loan data in its ACL model where the Company’s own historical data is not available.

The discounted cash flow model is applied on an instrument-by-instrument basis, and for loans with similar risk characteristics, to derive estimates for the lifetime ACL for each loan. The discounted cash flow methodology relies on several significant components essential to the development of estimates for future cash flows on loans and unfunded commitments. These components consist of: (i) the estimated probability of default, (ii) the estimated loss given default, which represents the estimated severity of the loss when a loan is in default, (iii) estimates for prepayment activity on loans and (iv) the estimated exposure to the Company at default (“EAD”). These components are also heavily influenced by changes economic forecasts employed in the model over a reasonable and supportable period. The Company’s ACL methodology for unfunded loan commitments also includes assumptions concerning the probability the unfunded commitment will be drawn upon by the borrower. These assumptions are based on the Company’s historical experience.

The Company’s discounted cash flow ACL model for commercial real estate and commercial loans uses internally derived estimates for prepayments in determining the amount and timing of future contractual cash flows to be collected. The estimate of future cash flows also incorporates estimates for contractual amounts the Company believes may not be collected, which are based on assumptions for PD, LGD and EAD. EAD is the estimated outstanding balance of the loan at the time of default. It is determined by the contractual payment schedule and expected payment profile of the loan, incorporating estimates for expected prepayments and future draws on revolving credit facilities. The Company discounts cash flows using the effective interest rate on the loan. The effective interest rate represents the contractual rate on the loan; adjusted for any purchase premiums, purchase discounts, and deferred fees and costs associated with the origination of the loan. The Company has made an accounting policy election to adjust the effective interest rate to take into consideration the effects of estimated prepayments. The ACL for loans is determined by measuring the amount by which a loan’s amortized cost exceeds its discounted cash flows.

The PD for commercial real estate loans is based largely on a model provided by a third party, using proxy loan information. The PDs generated by this model are reflective of current and expected changes in economic conditions and conditions in the commercial real estate market, and how they are expected to impact loan level and property level attributes, and ultimately the likelihood of a default event occurring. Significant loan and property level attributes include: loan to value ratios, debt service coverage, loan size, loan vintage and property types.

The PD for commercial loans is based on an internally developed PD rating scale that assigns PDs based on the Company’s internal risk grades for loans. This internally developed PD rating scale is based on a combination of the Company’s own historical data and observed historical data from the Company’s peers, which consist of banks that management believes align with our business profile. As credit risk grades change for loans in the commercial segment, the PD assigned to them also changes. As with commercial real estate loans, the PD for commercial loans is also impacted by current and expected economic conditions.

LGDs for commercial real estate loans are derived from a third party, using proxy loan information, and are based on loan and property level characteristics in the Company’s loan portfolio, such as: loan to values, estimated time to resolution, property size and current and estimated future market price changes for underlying collateral. The LGD is highly dependent upon loan to value ratios, and incorporates estimates for the expense associated with managing the loan through to resolution. LGDs also incorporate an estimate for the loss severity associated with loans where the borrower fails to meet their debt obligation at maturity, such as through a balloon payment or the refinancing of the loan through another lender. External factors that have an impact on LGDs include: changes in the CRE Price Index, GDP growth rate, unemployment rates and the Moody’s Baa rating corporate debt interest rate spread. LGDs are applied to each loan in the commercial real estate portfolio, and in conjunction with the PD, produce estimates for net cash flows not expected to be collected over the estimated term of the loan.

LGDs for commercial loans are also derived from a third party that has a considerable database of credit related information specific to the financial services industry and the type of loans within this segment, and is used to generate annual default information for commercial loans. These proxy LGDs are dependent upon data inputs such as: credit quality, borrower industry, region, borrower size and debt seniority. LGDs are then applied to each loan in the commercial portfolio, and in conjunction with the PD, produce estimates for net cash flows not expected to be collected over the estimated term of the loan.

U.S. GAAP requires the Company to develop reasonable and supportable forecasts of future conditions, and estimate how those forecasts are expected to impact a borrower’s ability to satisfy their obligation to the Bank and the ultimate collectability of future cash flows over the life of the loan. The Company uses economic scenarios from Moody’s Analytics in its estimation of a borrower’s ability to repay a loan in future periods. These scenarios are based on past events, current conditions and the likelihood of future events occurring. These scenarios typically are comprised of: (1) a base-case scenario, (2) an upside scenario, representing slightly better economic conditions than currently experienced and (3) a downside scenario, representing recessionary conditions. Management periodically evaluates economic scenarios and may decide that a particular economic scenario or a combination of probability-weighted economic scenarios should be used in the Company’s ACL model. The economic scenario or scenarios chosen for the model and, to the extent more than one scenario is used, the weights that are assigned to them, are based on the Company’s estimate of the probability of each scenario occurring, which is based in part on analysis performed by an independent third-party. Economic scenarios, as well as assumptions within those scenarios, whether to use more than one scenario and the relative weighting or multiple scenarios, if used, can vary based on changes in current and expected economic conditions and due to the occurrence of specific events such as the COVID-19 pandemic. The Company recognizes the non-linearity of credit losses relative to economic performance and thus the Company believes the consideration and, if appropriate under the circumstances, use of multiple probability-weighted economic scenarios is appropriate in estimating credit losses over the forecast period. The approach of considering and, if appropriate, using multiple probability-weighted scenarios is based on certain assumptions. The first assumption is that no single forecast of the economy, however detailed or complex, is completely accurate over a reasonable forecast time-frame, and is subject to revisions over time. By considering multiple scenario outcomes and, when utilizing multiple scenario outcomes is appropriate under the circumstances, assigning reasonable probability weightings to them, some of the uncertainty associated with a single scenario approach, the Company believes, is mitigated.

As of January 1, 2020, upon the adoption of ASC 326, the Company’s ACL model used three probability-weighted scenarios representing a base-case scenario, an upside scenario and a downside scenario. The weightings assigned to each scenario were as follows: the base-case scenario, or most likely scenario, was assigned a weighting of 40%, while the upside and downside scenarios were each assigned weightings of 30%. As of March 31, 2020, and in response to the COVID-19 pandemic, economic forecasts used in the Company’s ACL model incorporate assumptions associated with the estimated economic effects of the pandemic. The Company, with the assistance of an independent third party, determined it appropriate to include an economic scenario that is reflective of the estimated economic effects of the pandemic, including the responses to contain the pandemic. This scenario is referred to as the critical pandemic scenario. Additionally, the Company evaluated the weightings of each economic scenario in the current period with the assistance of an independent third party, and revised those weightings to levels it believes appropriately reflect the likelihood of outcomes for each scenario given the change in the current economic environment. As such, for the three months ended March 31, 2020, The Company’s ACL model incorporated three economic scenarios comprised of: the critical pandemic scenario weighted 30%, the more severe (as compared to the critical pandemic scenario) downside scenario weighted 32.5% and the base-case scenario weighted 37.5%. The weightings of economic scenarios in the current period are reflective of rapidly changing economic conditions, economic uncertainty and volatility in financial markets brought on by the COVID-19 pandemic, and the estimated likelihood that each scenario may occur as of March 31, 2020.

The Company currently forecasts economic conditions over a two-year period, which we believe is a reasonable and supportable period. Beyond the point which the Company can provide for a reasonable and supportable forecast, economic variables revert to their long-term averages. The Company has reflected this reversion over a period of three years in each of its economic scenarios used to generate the overall probability-weighted forecast. Changes in economic forecasts impact the PD, LGD and EAD for each loan, and therefore influence the amount of future cash flows for each loan the Company does not expect to collect.

The Company derives the economic forecasts it uses in its ACL model from an independent third party that has a large team of economists, data-base managers and operational engineers with a history of producing monthly economic forecasts for over 25 years. The forecasts produced by this third party have been widely used by banks, credit unions, government agencies and real estate developers. These economic forecasts cover all states and metropolitan areas in the Unites States, and reflect changes in economic variables such as: GDP growth, interest rates, employment rates, changes in wages, retail sales, industrial production, metrics associated with the single-family and multifamily housing markets, vacancy rates, changes in equity market prices and energy markets.

It is important to note that the Company’s ACL model relies on multiple economic variables, which are used under several economic scenarios. Although no one economic variable can fully demonstrate the sensitivity of the ACL calculation to changes in the economic variables used in the model, the Company has identified certain economic variables that have significant influence in the Company’s model for determining the ACL. As of March 31, 2020, the Company’s ACL model incorporated the following assumptions for key economic variables in the base-case and downside scenarios:

The Company recognizes that historical information used as the basis for determining future expected credit losses may not always, by themselves, provide a sufficient basis for determining future expected credit losses. The Company, therefore, periodically considers the need for qualitative adjustments to the ACL. Qualitative adjustments may be related to and include, but not be limited to, factors such as: (i) management’s assessment of economic forecasts used in the model and how those forecasts align with management’s overall evaluation of current and expected economic conditions, (ii) organization specific risks such as credit concentrations, collateral specific risks, regulatory risks and external factors that may ultimately impact credit quality, (iii) potential model limitations such as limitations identified through back-testing, and other limitations associated with factors such as underwriting changes, acquisition of new portfolios and changes in portfolio segmentation and (iv) management’s overall assessment of the adequacy of the ACL, including an assessment of model data inputs used to determine the ACL. As of March 31, 2020, qualitative adjustments included in the ACL totaled $6.0 million. These adjustments relate to management’s assessment of the economic forecasts used in the model. Management determined through additional review of economic indicators and forecasts of economic conditions that there is potential for economic conditions under the base-case scenario to worsen. Management observed through this additional review that certain key economic indicators may experience additional declines over the intermediate term. Real GDP may experience a decrease of approximately 18%, the U.S. unemployment rate may reach 9%, and the CRE price index may also experience further weakness. Management concluded the potential for additional weakness in these economic indicators under a base-case scenario may contribute to higher lifetime losses in CRE non-owner-occupied, construction and land and franchise loans as of March 31, 2020. Management reviews the need for and appropriate level of qualitative adjustments on a quarterly basis, and as such, the amount and allocation of qualitative adjustments may change in future periods.



	For the Three Months Ended March 31, 2020
	Beginning ACL Balance ⁽⁶⁾		Adoption of ASC 326			Charge-offs			Recoveries		Provision for Loan Losses			Ending ACL Balance
	(Dollars in thousands)
Investor loans secured by real estate
CRE non-owner-occupied	$	1,899	$	8,423		$	(387	)	$	—	$	5,961		$	15,896
Multifamily	729		9,174			—			—		4,819			14,722
Construction and land	4,484		(124		)	—			—		4,862			9,222
SBA secured by real estate ⁽¹⁾	1,915		(1,401		)	—			—		421			935
Business loans secured by real estate ⁽²⁾
CRE owner-occupied	2,781		20,166			—			12		3,834			26,793
Franchise real estate secured	592		5,199			—			—		1,712			7,503
SBA secured by real estate ⁽³⁾	2,119		2,207			(315		)	71		(38		)	4,044
Commercial loans ⁽⁴⁾
Commercial and industrial	13,857		87			(490		)	5		2,283			15,742
Franchise non-real estate secured	5,816		9,214			—			—		1,586			16,616
SBA non-real estate secured	445		218			(236		)	4		85			516
Retail loans
Single family residential ⁽⁵⁾	655		541			—			—		(59		)	1,137
Consumer loans	406		1,982			(8		)	—		(84		)	2,296
Totals	$	35,698	$	55,686		$	(1,436	)	$	92	$	25,382		$	115,422



	For the Three Months Ended March 31, 2019
	Beginning ALLL Balance		Charge-offs			Recoveries		Provision for Credit Losses			Ending ALLL Balance
	(Dollars in thousands)
Investor loans secured by real estate
CRE non-owner-occupied	$	1,624	$	—		$	—	$	44		$	1,668
Multifamily	740		—			—		(71		)	669
Construction and land	5,964		—			—		(4		)	5,960
SBA secured by real estate ⁽¹⁾	1,827		—			—		877			2,704
Business loans secured by real estate ⁽²⁾
CRE owner-occupied	1,908		—			8		53			1,969
Franchise real estate secured	743		—			—		1,430			2,173
SBA secured by real estate ⁽³⁾	1,824		—			—		142			1,966
Commercial loans ⁽⁴⁾
Commercial and industrial	13,695		(302		)	67		127			13,587
Franchise non-real estate secured	6,066		—			—		(368		)	5,698
SBA non-real estate secured	654		—			3		(154		)	503
Retail loans
Single family residential ⁽⁵⁾	808		—			—		(50		)	758
Consumer loans	219		(5		)	1		(14		)	201
Totals	$	36,072	$	(307	)	$	79	$	2,012		$	37,856

The change in the ACL during the three months ended March 31, 2020 of $79.7 million is reflective of a $55.7 million adjustment associated with the Company’s adoption of ASC 326 on January 1, 2020, which was recorded through a cumulative effect adjustment to retained earnings, as well as $25.4 million in provisions for credit losses on loans, and net charge-offs of $1.3 million. The provision for loan losses during the three months ended March 31, 2020 is reflective of unfavorable changes in economic forecasts used in the Company’s ACL model related to the COVID-19 pandemic.



	December 31, 2019
	Loans Evaluated Individually for Impairment		ALLL Attributed to Individually Evaluated Loans		Loans Evaluated Collectively for Impairment		ALLL Attributed to Collectively Evaluated Loans
	(Dollars in thousands)
Investor loans secured by real estate
CRE non-owner-occupied	$	1,088	$	—	$	2,069,053	$	1,899
Multifamily	—		—		1,575,726		729
Construction and land	—		—		438,786		4,484
SBA secured by real estate ⁽¹⁾	390		—		68,041		1,915
Business loans secured by real estate ⁽²⁾
CRE owner-occupied	—		—		1,846,554		2,781
Franchise real estate secured	—		—		353,240		592
SBA secured by real estate ⁽³⁾	1,517		—		86,864		2,119
Commercial loans ⁽⁴⁾
Commercial and industrial	7,529		—		1,385,741		13,857
Franchise non-real estate secured	10,834		—		553,523		5,816
SBA non-real estate secured	1,118		—		16,308		445
Retail loans
Single family residential ⁽⁵⁾	366		—		254,658		655
Consumer loans	—		—		50,975		406
Totals	$	22,842	$	—	$	8,699,469	$	35,698



	Commercial Real Estate Term Loans by Vintage
	2020		2019		2018		2017		2016		Prior		Revolving		Revolving Converted to Term During the Period		Total
March 31, 2020	(Dollars in thousands)
Investor loans secured by real estate
CRE non-owner-occupied
0% - 5.00%	$	104,083	$	373,078	$	365,350	$	282,641	$	200,867	$	629,141	$	11,644	$	—	$	1,966,804
>5.00% - 10.00%	—		8,928		16,401		19,982		6,318		5,364		—		—		56,993
Greater than 10%	—		—		318		356		—		15,727		—		—		16,401
Multifamily
0% - 5.00%	99,379		268,267		308,116		241,432		291,126		356,870		987		—		1,566,177
>5.00% - 10.00%	5,829		40,102		3,522		506		—		4,263		—		—		54,222
Greater than 10%	—		—		3,824		—		—		1,459		—		—		5,283
Construction and Land
0% - 5.00%	2,250		62,553		14,444		12,823		—		7,408		391		—		99,869
>5.00% - 10.00%	—		26,376		21,927		40,704		—		298		—		—		89,305
Greater than 10%	—		29,621		102,969		54,505		—		1,256		—		—		188,351
SBA secured by real estate ⁽¹⁾
0% - 5.00%	494		10,726		12,158		16,888		7,160		11,664		—		—		59,090
>5.00% - 10.00%	—		—		1,660		—		—		523		—		—		2,183
Greater than 10%	—		—		—		—		392		—		—		—		392
Total investor loans secured by real estate	$	212,035	$	819,651	$	850,689	$	669,837	$	505,863	$	1,033,973	$	13,022	$	—	$	4,105,070
Business loans secured by real estate ⁽²⁾
CRE owner-occupied
0% - 5.00%	$	114,657	$	296,223	$	274,727	$	303,164	$	244,460	$	515,875	$	5,573	$	—	$	1,754,679
>5.00% - 10.00%	—		18,905		29,663		16,248		23,617		31,941		247		—		120,621
Greater than 10%	—		—		3,635		965		2,180		5,302		250		—		12,332
Franchise real estate secured
0% - 5.00%	19,326		78,086		74,790		105,669		30,643		38,631		—		—		347,145
>5.00% - 10.00%	—		1,575		1,069		3,534		1,009		7,553		—		—		14,740
Greater than 10%	—		7,913		728		—		—		902		—		—		9,543
SBA secured by real estate ⁽³⁾
0% - 5.00%	2,109		7,723		13,559		16,143		8,948		23,564		364		—		72,410
>5.00% - 10.00%	—		—		694		2,260		2,430		2,780		—		—		8,164
Greater than 10%	—		—		—		1,033		413		1,620		—		—		3,066
Total business loans secured by real estate	$	136,092	$	410,425	$	398,865	$	449,016	$	313,700	$	628,168	$	6,434	$	—	$	2,342,700



	Commercial Real Estate Term Loans by Vintage
	2020		2019		2018		2017		2016		Prior		Revolving		Revolving Converted to Term During the Period		Total
March 31, 2020	(Dollars in thousands)
Commercial Loans ⁽⁴⁾
Commercial and industrial
0% - 5.00%	$	31,625	$	114,904	$	94,899	$	67,070	$	34,540	$	110,090	$	881,734	$	1,204	$	1,336,066
>5.00% - 10.00%	—		10,607		7,471		38,578		2,603		6,021		27,259		370		92,909
Greater than 10%	—		524		2,769		869		1,915		3,027		19,640		1,250		29,994
Franchise non-real estate secured
0% - 5.00%	9,938		210,415		121,918		77,413		48,552		44,294		2,062		—		514,592
>5.00% - 10.00%	323		2,370		4,036		1,977		4,174		3,430		—		—		16,310
Greater than 10%	—		—		—		13,056		—		3,835		—		—		16,891
SBA not secured by real estate
0% - 5.00%	265		2,383		1,326		2,258		601		3,503		1,572		285		12,193
>5.00% - 10.00%	—		—		469		—		387		230		—		—		1,086
Greater than 10%	—		88		141		261		—		1,709		787		—		2,986
Total commercial loans	$	42,151	$	341,291	$	233,029	$	201,482	$	92,772	$	176,139	$	933,054	$	3,109	$	2,023,027



	Term Loans by Vintage
	2020		2019		2018		2017		2016		Prior		Revolving		Revolving Converted to Term During the Period		Total
March 31, 2020	(Dollars in thousands)
Investor loans secured by real estate
CRE non-owner-occupied
55% and below	$	56,085	$	174,693	$	137,386	$	179,286	$	144,084	$	515,589	$	11,085	—		$	1,218,208
>55-65%	27,173		114,711		104,810		97,883		52,649		104,802		559		—		502,587
>65-75%	20,825		92,602		136,934		23,455		10,238		29,568		—		—		313,622
Greater than 75%	—		—		2,939		2,355		214		273		—		—		5,781
Multifamily
55% and below	26,605		108,109		133,038		93,938		72,385		159,449		511		—		594,035
>55-65%	32,070		140,824		121,292		83,532		85,911		160,662		—		—		624,291
>65-75%	46,533		49,354		40,497		62,560		132,830		37,319		476		—		369,569
Greater than 75%	—		10,082		20,635		1,908		—		5,162		—		—		37,787
Construction and land
55% and below	2,250		114,230		116,970		32,535		—		7,581		391		—		273,957
>55-65%	—		4,320		19,439		57,988		—		—		—		—		81,747
>65-75%	—		—		1,887		17,509		—		—		—		—		19,396
Greater than 75%	—		—		1,044		—		—		1,381		—		—		2,425
SBA secured by real estate ⁽¹⁾
55% and below	—		1,150		658		844		725		439		—		—		3,816
>55-65%	—		3,194		710		3,869		980		4,079		—		—		12,832
>65-75%	494		3,730		8,821		6,523		4,371		3,601		—		—		27,540
Greater than 75%	—		2,652		3,629		5,652		1,476		4,068		—		—		17,477
Total investor loans secured by real estate	$	212,035	$	819,651	$	850,689	$	669,837	$	505,863	$	1,033,973	$	13,022	$	—	$	4,105,070
Business loan secured by real estate ⁽²⁾
CRE owner-occupied
55% and below	$	44,737	$	113,567	$	154,056	$	194,126	$	143,274	$	361,762	$	4,951	—		$	1,016,473
>55-65%	14,963		91,000		85,230		61,426		75,933		87,639		—		—		416,191
>65-75%	15,560		89,702		44,579		59,813		47,717		60,839		1,119		—		319,329
Greater than 75%	39,397		20,859		24,160		5,012		3,333		42,878		—		—		135,639
Franchise real estate secured
55% and below	9,342		18,866		14,141		16,765		11,413		21,859		—		—		92,386
>55-65%	—		11,447		11,696		24,812		7,010		6,891		—		—		61,856
>65-75%	—		44,892		27,393		10,261		12,102		17,098		—		—		111,746
Greater than 75%	9,984		12,369		23,357		57,365		1,127		1,238		—		—		105,440
SBA secured by real estate ⁽³⁾
55% and below	2,109		7,723		14,253		19,436		11,791		27,964		364		—		83,640
>55-65%	—		—		—		—		—		—		—		—		—
>65-75%	—		—		—		—		—		—		—		—		—
Greater than 75%	—		—		—		—		—		—		—		—		—
Total business loans secured by real estate	$	136,092	$	410,425	$	398,865	$	449,016	$	313,700	$	628,168	$	6,434	$	—	$	2,342,700



	Term Loans by Vintage
	2020		2019		2018		2017		2016		Prior		Revolving		Revolving Converted to Term During the Period		Total
March 31, 2020	(Dollars in thousands)
Retail Loans
Single family residential⁽¹⁾
Greater than 740	$	3,466	$	9,868	$	13,973	$	11,586	$	32,709	$	90,860	$	29,710	—		$	192,172
>680 - 740	—		1,196		3,786		4,824		2,678		10,893		9,273		—		32,650
>580 - 680	—		—		—		521		3,169		6,549		1,319		—		11,558
Less than 580	—		—		—		—		—		763		37		—		800
Consumer loans
Greater than 740	89		126		866		55		25		2,682		2,366		—		6,209
>680 - 740	—		64		8		37,990		—		501		1,765		—		40,328
>580 - 680	—		24		—		—		4		155		110		—		293
Less than 580	—		—		—		—		—		21		41		—		62
Total retail loans	$	3,555	$	11,278	$	18,633	$	54,976	$	38,585	$	112,424	$	44,621	$	—	$	284,072