cpss_corr-102711.htm


 
19500 Jamboree Road
Irvine, California 92612
Facsimile  (949) 753-6897
Telephone (949) 753-6800

October 26, 2011
 
VIA EDGAR
 
Matt S. McNair, Attorney-Adviser
Securities and Exchange Commission
Division of Corporation Finance
100 F Street, N.E.
Washington, D.C. 20549

Re:
Consumer Portfolio Services, Inc.
 
Form 10-K for Fiscal Year Ended December 31, 2010
 
Filed March 31, 2011
 
Form 10-K/A for Fiscal Year Ended December 31, 2010
 
Filed May 2, 2011
 
Form 10-Q for Fiscal Quarter Ended June 30, 2011
 
Filed August 15, 2011
 
File No. 001-14116
 
Dear Mr. McNair:
 
We refer to your letter dated September 16, 2011, to our responsive letter filed Wednesday, October 5, and to the informal discussions we had with Messrs. Irving and Schroeder on Friday, October 14.  In this letter, we respond to the concerns raised in that discussion.
 
1.           Pro-Forma Delinquency Table Assuming No Extensions
 
In our discussion, the Staff expressed the desire to see what the performance of our portfolio would have been, had our accounts never been granted extensions.  As we attempted to prepare such a pro forma presentation, we realized that all we could do was to mechanically shifted accounts into later stages of delinquency. It became clear that such a presentation would not likely be helpful or meaningful: neither to us as managers of the company nor to readers of our financial statements.
 
Consider, for example, an account that may have been active and not delinquent at December 31, 2010.  Assume further that this account had twice been granted a one-month extension, with the later of the two occurring over 12 months prior to December 31, 2010.  In this example, a mechanical pro forma presentation of delinquency, without giving effect to extensions, would report this account as two payments past due at year-end 2010, even though the account was legitimately current and had made twelve consecutive monthly payments.  Moreover, such a presentation is potentially misleading in that we would only be speculating as to what might have occurred without the extensions that were granted on the account.
 

 
 

 

 
U.S. Securities and Exchange Commission
File No. 001-14116
October 26, 2011
Page 2


Without one or both of the extensions, the account might have remained past due. The obligor might have dug deeper into his pocket and found a way to catch up by one payment, or by both payments. Or the account might have deteriorated further, even to such an extent that we might have been forced to repossess and liquidate the vehicle, resulting in a charge off.  (Following charge off, the account would not have appeared on any subsequent delinquency schedule.)
 
In any case, we do not know now what would have happened in the absence of the extensions.  Because a pro forma presentation cannot present the unknowable effects on obligor behavior, we believe that it would be inherently misleading.
 
We understand, however, that it may be helpful to have more information regarding the extent to which we use extensions to maximize the performance of our portfolio.  We considered other metrics that would be meaningful, such as quantifying the number of extensions granted.  Our three-year practice is shown in the table below:
 
    Year Ended December 31,  
   
2010
   
2009
   
2008
 
                   
Average number of extensions granted per month
    1,883       2,665       2,892  
                         
Average number of outstanding accounts
    95,282       126,240       159,756  
                         
Average monthly extensions as % of average outstandings
    2.0%       2.1%       1.8%  
 
As the table shows, the number of monthly extensions granted is not significant relative to the numbers of accounts in our outstanding portfolio.  Moreover, the volume of extensions as a percentage of the overall portfolio has varied little from year to year, with no clear trend.  (By contrast, the cumulative number and cumulative percentage of extensions do vary with the age of our portfolio.)
 
2.           Extension Success Rates
 
In our discussion, the Staff asked that we provide more information regarding whether our extension policies have been successful.  Accordingly, we conducted a new analysis to measure the performance of accounts subsequent to their having been granted an extension.  To perform the analysis, we selected all of the accounts that have had an extension since January 1, 2008 and assessed their status as of December 31, 2010.  A summary of our findings is shown in the table below:
 

 
 

 

 
U.S. Securities and Exchange Commission
File No. 001-14116
October 26, 2011
Page 3
 
                                                       
Period of Extension
 
# Extensions Granted
   
Extended Accounts Paid Off at 12/31/10
   
Extended Accounts Active at 12/31/10
   
Pure Success Rate
   
Charged Off > 6 Months After Extension
   
Charge Off Success Rate
   
Charged Off <= 6 Months After Extension
   
Failure Rate
   
Avg Months to Charge Off Post Extension
 
                                                       
                                                       
2008
    34,706       2,670       12,425     43.5%       14,792     42.6%       4,819     13.9%       12  
                                                                         
2009
    31,980       2,005       15,854     55.8%       8,357     26.1%       5,764     18.0%       8  
 
We view these results as a confirmation of the success of our extension program.  For the accounts receiving extensions in 2008, 43.5% were either paid off in full, or active and performing, three years later at December 31, 2010.  Each of these successful accounts represent continued payments of interest and principal (through payment in full in many cases), where without the extension we likely would have incurred a substantial loss and no interest revenue subsequent to the extension.  In our view, the 2008 and 2009 extension accounts that were active or paid off at December 31, 2010 represent a “Pure Success” of the extension on those accounts.
 
For the 2008 and 2009 extension accounts that ultimately charged off, we consider any that charged off more than six months after the extension to be at least partially successful, or a “Charge Off Success”.  Moreover, for the 2008 extensions, of the accounts that charged off, the charge off was incurred, on average, 12 months after the extension, indicating that even in the cases of an ultimate loss, the obligor serviced the account with additional payments of principal and interest for an average of 12 months after the extension.  Although we ultimately incurred a loss, it was likely substantially mitigated as a result of the payments made on the account subsequent to the extension.  We would concede that any accounts that charged off in six or fewer months after the extension represent a failure.  Nevertheless, we believe that the table shows that the overall successes of the extensions significantly outweigh the failures.
 
The dropoff in average months to charge off for loans extended in 2009 does not reflect a deterioration in results so much as it reflects compression of time available: there are no accounts extended in 2009 that charged off many months later, as of December 31, 2010, just because there aren’t enough months in between.
 
Not surprisingly, the quantity of accounts charged off from the population of 2009 extensions is somewhat greater than the 2008 extensions.  We do not view this as a weakness of the extensions granted during that period, but rather as a result of the severe economic climate during 2009 and 2010.
 
The Staff also asked that we stratify our analysis of extension results.  To do that, we reached back three years prior to our most recent audit date, and reviewed the subsequent performance of the 25,932 accounts in our portfolio that had received one or more extensions as of December 31, 2007.  We stratified the pool between those accounts with one extension and those accounts with two or more as of that date.  We then assessed the performance of those accounts through December 31, 2010, and found that there was little variance in the performance characteristics between accounts with one extension compared to accounts that had two or more.  The results are summarized below: 
 

 
 

 

 
U.S. Securities and Exchange Commission
File No. 001-14116
October 26, 2011
Page 4
 
 
As of December 31, 2007
 
Total
   
Active or Paid Off at 12/31/10
   
Charged Off in 2009 or 2010
   
Charged Off in 2008
 
                                           
Accounts with one extension
    21,555       11,032       51%       5,049       23%       5,474       25%  
                                                         
Accounts with two or more extensions
    4,377       2,300       53%       852       19%       1,225       28%  
                                                         
      25,932       13,332               5,901               6,699          

Again we would view these results as a validation of our program, because (i) a majority of extension accounts in both groups were performing three years later and (ii) a significant portion of the accounts that ultimately charged off performed for a significant period subsequent to this measurement date of December 31, 2007.
 
3.           Expand on Accounts with Two or More Extensions
 
The Staff asked that we quantify our extensions past 30 days.  We have accordingly prepared the table below, which breaks out our existing disclosure regarding the numbers of accounts with two or more extensions:
 
   
December 31, 2010
   
December 31, 2009
   
December 31, 2008
 
   
Number of Contracts
   
Amount
   
Number of Contracts
   
Amount
   
Number of Contracts
   
Amount
 
               
(Dollars in thousands)
             
                                     
Contracts with one extension
    17,749       135,204       26,528       266,081       30,160       354,329  
Contracts with two extensions
    9,142       73,945       10,190       102,606       7,277       77,337  
Contracts with three extensions
    3,420       27,071       2,400       22,193       1,222       10,713  
Contracts with four extensions
    650       4,539       281       1,947       124       845  
Contracts with five extensions
    12       74       12       104       13       87  
Contracts with six extensions
    2       8       1       3       3       6  
      30,975       240,841       39,412       392,934       38,799       443,317  
                                                 
Gross servicing portfolio
    84,601       681,157       111,105       1,057,348       145,564       1,665,036  
 
As you can see, the incidence of accounts with 3 or 4 extensions is fairly small and those with 5 or 6 extensions are insignificant. We discuss below the effect of portfolio age on these categories.  At this point, we wish only to note that a portfolio with higher average age will tend to have more contracts in the multiple extension buckets: no contract moves from three extensions down to two, but some do move from two extensions to three.
 

 
 

 

 
U.S. Securities and Exchange Commission
File No. 001-14116
October 26, 2011
Page 5
 
 
4.           Expand on Discussion Regarding Portfolio Age
 
The Staff asked that we discuss the effect of increased age of our portfolio.   As we state in our filings, the amounts and percentage levels of our delinquencies and charge offs during 2008, 2009 and 2010 were significantly influenced not only by broad economic weakness and high unemployment during those periods, but also by the decreasing size and increasing age of our managed portfolio.
 
The receivables we purchase have remained nearly homogeneous over the past ten years, comprising retail installment motor vehicle purchase contracts with interest rates around 18% to 20%, requiring fixed monthly payments from the obligors, and average terms to maturity of 60 to 70 months.  Our experience throughout our history has been that losses are incurred at low rates in the first few months after an obligor purchases a vehicle, at increasing rates through approximately the 28th to 30th month, and that incremental losses thereafter decrease as the contract approaches its maturity, and as the amount owed drops down to the fair market value of the vehicle, and ultimately to zero.   Throughout this typical pattern, delinquencies tend to lead losses by a few months.
 
We use a fixed policy for recognizing losses, which is described in Note 1 to our financial statements: in most cases we charge off a receivable when we repossess and sell the related vehicle, but there are backstop requirements that require charge off and recognition of loss upon extended delinquency, even if we fail to obtain the vehicle or fail to sell it.
 
The average age of our portfolio of receivables increased rapidly over the two-year period reported on in our 10-K.  This was the inevitable effect of a combination of our holding a pool of self-amortizing assets and our decrease in contract purchases (as reported). After several years of moderate increases in contract purchases, we reduced our purchases with the onset of the financial crisis in late 2007, and took our purchases close to zero for the twelve months following September 30, 2008.  We thereafter began to purchase small quantities of new receivables in the fourth quarter of 2009, and slowly increased those quantities in 2010 and 2011.
 
Our quarterly purchases of receivables are set forth in the following table.  It should be plainly evident that, for the periods reported on in our 10-K, our portfolio has comprised largely receivables purchased prior to the first quarter of 2008.  That composition has just begun to change over the most recent few months.
 

 
 

 

 
U.S. Securities and Exchange Commission
File No. 001-14116
October 26, 2011
Page 6
 
 
 
Because we purchased many more contracts from 2005 - 2007 than we did in 2008-2010, our portfolio has been a snake with a lump in its belly, moving slowly to the rear. As the lump ages, our portfolio performance at any time resembles the performance of receivables with that age.
 
The weighted average age of our managed portfolio was 22 months, 33 months and 37 months, respectively as of December 31, 2008, 2009 and 2010.
 
As a portfolio of receivables ages, the denominator on which credit performance metrics are measured (the outstanding principal balance) continues to shrink relative to the number of delinquent accounts and the dollar amount of charge offs.
 
To illustrate further, during the 2009 and 2010, our portfolio decreased and aged primarily because we originated relatively few new finance receivables.  Typically new, unseasoned, receivables perform with relatively low levels of delinquency and few charge offs.  Such new receivables are represented in the denominator (but often not in the numerator) of our delinquency and charge off tables and tend to lead to lower delinquency and charge off percentage rates.  When new receivables are few relative to the size of the existing portfolio, the portfolio will amortize faster and become more seasoned until new receivables increase in volume.
 
5.           Use of a QSPE
 
The Staff asked that we clarify whether our treatment of any transaction was dependent on classification of an entity as a qualified special purpose entity (“QSPE”).  This is to confirm that none of the debt issued by any of our subsidiaries is excluded from our consolidated balance sheet on the basis that the issuer of the debt is a QSPE.
 
The only debt of any of our subsidiaries that we do not record on our consolidated balance sheet is the debt issued by CPS Auto Receivables Trust 2010-A (the “Trust”).  We determined that such debt is not properly consolidated with us because we have effectively surrendered control of the related assets.  We have no right or obligation to repurchase the related assets and our financial interest is limited: we receive a market rate base servicing fee to compensate us for our duties as servicer, and we have the prospect of a sharing in certain cash flows in the event that all of the Trust’s debt is ultimately repaid and that there are then some remaining residual cash proceeds.
 

 
 

 


 
U.S. Securities and Exchange Commission
File No. 001-14116
October 26, 2011
Page 7
 
 
6.           Relationship Between Allowance for Credit Losses and Age of Portfolio
 
The Staff asked that we provide further information regarding how we tie the amount of allowance for credit losses to the age/aging of the portfolio.    In general, it has been our policy to maintain an allowance for credit losses that is equal to our estimate of net losses likely to be incurred over the 12 months following the date of the estimate.  Through our long history, we have observed that net losses are small in the first few months following a customer’s purchase of a vehicle, but increase through approximately the 28th to 30th month, followed by reductions in incremental losses as installment contracts approach their maturity dates.  For example, when estimating net losses for 2011 at December 31, 2010, we reasoned (based on past performance data) that the receivables originated in the fourth quarter of 2010 would have fewer losses (as a percentage of the total pool) than receivables originated in the second quarter of 2010, since the second quarter 2010 receivables would have been more seasoned.  Conversely, at the same measurement date of December 31, 2010, we further reasoned (again, based on past performance data) that very old receivables, such as those originated in the fourth quarter of 2007 would have fewer net losses.  The table below illustrates relationships we have observed between portfolio age and expected losses.
 
 
Portfolio Age
Incremental % of Losses
Estimated for Next 12 Months (1)
 
Loss Window Months
     
New
2.5%
1 - 12
     
18 months
5.9%
19 - 30
     
24 months
6.0%
25 - 36
     
36 months
3.1%
37 - 48
     
48 months
-0.1%
49 - 60
     
(1) Percentage based on the original principal amount of the static pool
 
As discussed above, the weighted average age of our managed portfolio was 22 months, 33 months and 37 months, respectively as of December 31, 2008, 2009 and 2010.  Consequently, the estimate of future losses, and the related allowance for credit losses, tended to decrease during these periods in accordance with the net loss expectations as summarized in the above table.
 

 
*********
 

 
 

 

 
U.S. Securities and Exchange Commission
File No. 001-14116
October 26, 2011
Page 8
 

As part of our response, we acknowledge that the registrant is responsible for the adequacy and accuracy of the disclosures in its filings; that the staff’s comments or changes to disclosure in response to staff comments do not foreclose the Commission from taking any action with respect to the filing; and that we may not assert staff comments as a defense in any proceeding initiated by the Commission or any person under the federal securities laws of the United States. 
 
   
 
Sincerely,
   
 
/s/ Jeffrey P. Fritz
 
Jeffrey P. Fritz
 
Chief Financial Officer
 
 
cc (email)
David Lyon (lyond@sec.gov)
 
David Irving (irvingd@sec.gov)
 
William Schroeder (schroederw@sec.gov)