Finance Receivables

The following summarizes the components of finance receivables:



	December 31,
	2012		2011
	($ in thousands)
Principal Balances	$	1,601,710	$	1,466,680
Accrued Interest	16,414		12,971
Loan Origination Costs	16,498		15,689
Finance Receivables	$	1,634,622	$	1,495,340

Our finance receivables are defined as one segment and class of loan, which is the sub-prime consumer auto loan. Therefore, the disaggregation of information into portfolio segment and classes for assets with different risk characteristics is limited, and the level of risks inherent in our financing receivables are managed as one homogeneous pool and further segmented with our proprietary credit scoring system as described below —Credit Quality Indicators. We have chosen our internal customer credit scoring model as our key credit quality indicator because it has a direct and prominent impact in managing our portfolio receivables and monitoring its performance.

Finance receivables pledged as collateral associated with liabilities in our warehouse facilities, asset backed securitizations, and bank term financings, are provided in Note 8—Debt Obligations. We do not place loans on nonaccrual status, nor do we classify loans as impaired, since accounts are charged-off when the loan becomes contractually past due under our charge-off policy. We do not have loans that meet the definition of troubled debt restructurings.

Credit quality information for our finance receivables portfolio is provided as of the dates indicated below:

Age Analysis of Past Due Finance Receivables



	December 31,
Days Delinquent:	2012				2011
	($ in thousands)
	Percent of Portfolio		Loan Principal		Percent of Portfolio		Loan Principal
Current	49.1	%	$	786,765	66.8	%	$	979,155
01-30 Days	33.0	%	528,300		22.0	%	322,670
31-60 Days	10.0	%	161,157		7.1	%	104,574
61-90 Days	5.1	%	81,378		3.8	%	55,881
91-120 Days	2.8	%	44,110		0.3	%	4,400
Total Past Due	50.9	%	$	814,945	33.2	%	$	487,525
Total Finance Receivables	100.0	%	$	1,601,710	100.0	%	$	1,466,680

An account is considered delinquent if a substantial portion of a scheduled payment has not been received by the date such payment was contractually due. Delinquencies may vary from period to period based upon the average age or seasoning of the portfolio, seasonality within the calendar year and economic factors. Delinquencies are presented on a Sunday-to-Sunday basis, which reflects delinquencies as of the nearest Sunday to period end. Sunday is used to eliminate any impact of the day of the week on delinquencies since delinquencies tend to be higher mid-week.

Credit Quality Indicators

We monitor our portfolio performance and the credit grade mix of originations. Our proprietary credit grading system segments our customers into eight credit grades. We control the grade mix of originations through the finance terms provided to our customers, which are established centrally by our risk management team, and applied consistently throughout our dealership network. Our loans have an average original term of approximately 57 months and an average life of 31 months, due to charge-offs and pay-offs. The average life of our loans enables us to closely monitor credit trends and make appropriate adjustments to both the grade mix and pricing of our originations.

Many companies use FICO scores as a standard metric to assess the credit risk of customers. Our internal scoring models include the use of alternative data sources along with traditional credit bureau data which allow us the ability to separate the credit risk levels of the subprime auto segment into different categories. Our centralized proprietary credit scoring models are currently used to classify customers into various risk grades that are linked to financing parameters. The scoring models are periodically updated to account for changes in loan performance, data sources, geographic presence, economic cycles, and business processes.

Prior to each sale, we require our customers to complete a credit application. Upon entering the customer information into our origination system, our proprietary credit scoring system determines the customer’s credit grade, which is used by the dealership to help select vehicles that fit the required finance terms and the customer’s needs. The customer’s credit grade and type of vehicle determine the term, maximum installment payment, and minimum down payment amounts. The annual percentage rate (APR) charged is a function of the customer’s credit grade, down payment, and model year of the vehicle. Our centralized risk management and pricing departments set these terms. The static-pool tracking of portfolio loss performance is also monitored by credit grade. Our scoring model is comprised of eight credit grades ranging from A+ to D-, with A+ being the lowest risk credit grade and a D- being the highest risk credit grade. Generally, the lower the risk grade, the lower the unit loss rate.

Concentration of Credit Risk

At December 31, 2012, a summary of our portfolio by our internally assigned credit risk ratings was as follows:



Grade	Average FICO Score (1)	Percentage of Portfolio Contracts		Total Contracts	Percentage of Portfolio Principal		Total Portfolio Principal
							($ in thousands)
A+	556	10.4	%	14,660	10.6	%	$	169,023
A	539	18.5	%	25,998	18.8	%	301,173
B	517	37.3	%	52,476	38.2	%	612,084
C	503	28.4	%	40,066	27.8	%	445,497
C-	488	3.9	%	5,447	3.3	%	53,512
D+/D/D-	478	1.5	%	2,101	1.3	%	20,421
		100.0	%	140,748	100.0	%	$	1,601,710

At December 31, 2011, a summary of our portfolio by our internally assigned credit risk ratings was as follows:



Grade	Average FICO Score (1)	Percentage of Portfolio Contracts		Total Contracts	Percentage of Portfolio Principal		Total Portfolio Principal
							($ in thousands)
A+	557	10.4	%	14,270	10.7	%	$	156,935
A	538	18.2	%	24,954	18.7	%	274,269
B	516	36.8	%	50,582	38.3	%	561,738
C	500	28.2	%	38,756	27.1	%	397,470
C-	486	4.7	%	6,400	3.8	%	55,734
D+/D/D-	475	1.7	%	2,331	1.4	%	20,534
		100.0	%	137,293	100.0	%	$	1,466,680


(1)	Average FICO score is provided as an external metric of credit quality, but is generally not utilized to determine internal credit grade. Our internal scoring model is the primary driver of our internal credit grade. Average FICO presented excludes originations with no FICO.

Our portfolio concentration by state was as follows:



December 31, 2012
State	Percent of Portfolio		Loan Principal
			($ in thousands)
Texas	23.0	%	$	369,021
Florida	15.4	%	247,281
North Carolina	9.9	%	157,670
Georgia	7.6	%	122,027
Arizona	6.8	%	108,792
Virginia	6.7	%	106,749
Tennessee	4.6	%	72,967
California	4.4	%	71,005
Nevada	4.0	%	63,346
South Carolina	3.6	%	58,163
New Mexico	3.0	%	48,421
Alabama	2.8	%	44,787
Oklahoma	2.3	%	36,109
Colorado	2.2	%	35,268
Indiana	1.3	%	21,603
Ohio	1.1	%	17,417
Mississippi	1.0	%	15,847
Arkansas	0.2	%	3,218
Missouri	0.1	%	2,019
	100.0	%	$	1,601,710



December 31, 2011
State	Percent of Portfolio		Loan Principal
			($ in thousands)
Texas	23.9	%	$	351,210
Florida	16.6	%	243,024
North Carolina	10.3	%	150,920
Arizona	7.6	%	112,097
Virginia	7.3	%	106,715
Georgia	7.3	%	106,415
California	4.9	%	72,137
Nevada	4.6	%	67,631
Tennessee	3.9	%	56,925
New Mexico	3.6	%	53,484
Colorado	2.8	%	40,558
South Carolina	2.6	%	37,987
Oklahoma	1.9	%	28,229
Alabama	1.7	%	25,455
Indiana	0.5	%	6,843
Mississippi	0.4	%	5,844
Ohio	0.1	%	1,206
Arkansas	—	%	—
Missouri	—	%	—
	100.0	%	$	1,466,680

Allowance for Credit Losses

We maintain an allowance for credit losses on an aggregate basis at a level we consider sufficient to cover probable credit losses inherent in our portfolio of receivables as of each reporting date. The allowance takes into account historical credit loss experience, including timing, frequency and severity of losses. This estimate of existing probable credit losses inherent in the portfolio is primarily based on static pool analyses by month of origination based on origination principal, credit grade mix and deal structure, including down payment and term. The evaluation of the adequacy of the allowance also considers factors and assumptions regarding the overall portfolio quality, delinquency status, the value of the underlying collateral, current economic conditions that may affect the borrowers’ ability to pay, and the overall effectiveness of collection efforts.

The static pool loss curves by grade are adjusted for actual performance to date and historical seasonality patterns. The forecasted periodic loss rates, which drive the forecast for estimated gross losses (before recoveries) are calculated by factoring amortization speed and origination terms. Recoveries are estimated using historical unit and dollar static pool recovery activity to forecast recoveries for estimated charge-offs at the balance sheet date. The forecasted recovery rates (on a per unit basis) are based on the historical unit recovery trend by recovery type as adjusted for estimated impacts of economic and market conditions.

The allowance model is sensitive to changes in assumptions such that an increase or decrease in our forecasted net charge-offs would increase or decrease the allowance as a percentage of principal outstanding required to be maintained. The amount of our allowance is sensitive to losses within credit grade, recovery values, deal structure, the loss emergence period and overall credit grade mix of the portfolio.

The following table sets forth the rollforward of the allowance for credit losses for the periods indicated:



	Years Ended December 31,
	2012			2011
	($ in thousands)
Allowance Activity:
Balance, beginning of period	$	221,533		$	208,000
Provision for credit losses	253,603			207,198
Net charge-offs	(222,546		)	(193,665		)
Balance, end of period	$	252,590		$	221,533
Allowance as a percent of ending principal	15.7		%	15.1		%
Charge-off Activity:
Principal balances	$	(372,493	)	$	(338,357	)
Recoveries, net	149,947			144,692
Net charge-offs	$	(222,546	)	$	(193,665	)