Small Business Administration.
This final notice implements the Small Business Administration's (SBA's) risk rating system (Risk Rating System) as an internal tool to assist SBA in assessing the risk of each active 7(a) Lender's and Certified Development Company's (CDC's) SBA loan operations and loan portfolio. The Risk Rating System will enable SBA to monitor 7(a) Lenders and CDCs (collectively, “SBA Lenders”) on a uniform basis and identify those institutions whose SBA loan operations and portfolio require additional SBA monitoring or other action. It is also a vehicle for assessing the aggregate strength of SBA's 7(a) and 504 portfolios. Under the Risk Rating System, SBA will assign each SBA Lender a composite rating based on certain portfolio performance factors, which may be overridden in some cases due to SBA Lender specific factors that may be indicative of a higher or lower level of risk. SBA Lenders will have access to their own ratings through SBA's Lender Portal (Portal).
This notice is effective June 15, 2007.Start Further Info
FOR FURTHER INFORMATION CONTACT:
Bryan Hooper, Director, Office of Lender Oversight, U.S. Small Business Administration, 409 Third Street, SW., Washington, DC 20416, (202) 205-3049.End Further Info End Preamble Start Supplemental Information
On May 1, 2006, SBA published a notice and request for comment in the Federal Register seeking comments on a proposed SBA internal Risk Rating System for assessing an SBA Lender's SBA loan portfolio (i.e., loan portfolio performance). 71 FR 25624 Notice. SBA published a subsequent notice extending the comment period for the proposed Risk Rating System to July 15, 2006. 71 FR 34674. The Risk Rating System is an internal tool that uses data in SBA's Loan and Lender Monitoring System (L/LMS) to assist SBA in assessing the risk of an SBA Lender's SBA loan performance on a uniform basis and identify those SBA Lenders whose portfolio performance demonstrate the need for additional SBA monitoring or other action. The Risk Rating System will also serve as a vehicle to measure the aggregate strength of SBA's overall 7(a) and 504 loan portfolios and to assist SBA in managing the related risk. In addition, SBA will use risk ratings to make more Start Printed Page 27612effective use of its on-site and off-site lender review and assessment resources.
As discussed in greater detail in the Notice under the Risk Rating System, SBA will assign each SBA Lender a composite rating. The composite rating reflects SBA's assessment of the potential risk to the government of that SBA Lender's SBA portfolio performance. A rating of 1 will indicate strong portfolio performance, least risk, and that the least degree of SBA management oversight is needed (relative to other SBA Lenders in the peer group), while a 5 rating will indicate weak portfolio performance, highest risk and therefore, the highest degree of SBA management oversight.
For 7(a) Lenders, SBA will base the composite rating on four common components or factors. The common factors for 7(a) Lenders will be as follows: (i) 12 month actual purchase rate; (ii) problem loan rate; (iii) three month change in the small business predictive score (SBPS), which is a small business credit score on loans in the 7(a) Lender's portfolio; and (iv) projected purchase rate derived from the SBPS. On a lender-specific basis, the existence of additional factors may cause SBA to override the composite rating and either increase or decrease the composite rating.
For CDCs, SBA will base the composite rating on three common components or factors. The common factors for CDCs will be as follows: (i) 12 month actual purchase rate; (ii) problem loan rate; and (iii) average SBPS on loans in the CDC's portfolio. The third factor replaces the third and fourth factors used for 7(a) Lenders because it was found, during the testing process, to be more predictive of SBA purchases for CDCs. On a CDC-specific basis, the existence of additional factors may cause SBA to override the composite rating and either increase or decrease the composite rating.
In general, the factors described above reflect both historical SBA Lender performance and projected future performance. SBA will perform quarterly calculations on the common factors for each SBA Lender, so that SBA Lenders' composite risk ratings will be updated on a quarterly basis.
The composite risk rating is a measure of how each SBA Lender's portfolio performance compares to the portfolio performance of its peers. Thus, an individual SBA Lender's overall portfolio performance (using all common factors) will be compared to its peers to derive that SBA Lender's composite risk rating. SBA Lenders whose overall portfolio performance (using all of the common factors) is worse than their peers will receive a worse, or higher score, while SBA Lenders whose overall portfolio performance is better than their peers will receive a better, or lower, score. In order to prevent the inequitable comparison of differently-sized SBA Lenders, which may be affected differently by similar changes in their portfolio performance, SBA has separated both 7(a) Lenders and CDCs into different peer groups based upon their SBA loan portfolio size.
All SBA Lenders will be given access to their composite risk rating and component results through SBA's Lender Portal, which is available on line. The proposed notice described the Portal information that SBA will provide and how SBA lenders can access this information.
Comments Received and Changes Made
SBA received 51 comments on the proposed Risk Rating System. Twenty-three of the comments were from CDCs. Thirteen of the comments were from 7(a) Lenders other than Small Business Lending Companies (SBLCs). Six comments were from trade organizations. Five of the comments were from SBLCs. Finally, four comments were from individuals. Twenty-three of the commenters were generally supportive of an SBA Lender rating system. Comments generally covered the following areas: (i) The Portal; (ii) the rating components; (iii) use of the override; (iv) peer groupings; (v) the comparative nature of the system; (vi) static pool analysis; and (vii) other comments.
The purpose of the Portal is to communicate SBA Lender performance to SBA Lenders. The Portal will allow SBA Lenders to view their own quarterly performance data, including their most current composite risk rating. The Portal will also allow SBA Lenders to access data on peer group and portfolio averages. Consequently, an SBA Lender will be able to gauge its performance relative to its peer group and the portfolio norm, although SBA Lenders will not be able to view the individual ratings and performance indicators of other SBA Lenders. The quarterly performance data is updated approximately six to eight weeks after a calendar quarter ends.
Several commenters requested that SBA provide additional detail to facilitate reconciliation of the Portal performance results with performance results from other SBA and SBA Lender accounting systems. They also requested that SBA provide a process for correcting errors uncovered in the reconciliation process. SBA has provided that information on its Web site at http://www.sba.gov/olo/outstanding.pdf. As indicated on the website, L/LMS incorporates data from many different sources in order to calculate the common factors that are used to develop each SBA Lender's composite rating. As a result, some portfolio performance data in the Portal may not appear to be the same as that provided to SBA Lenders from other official sources (e.g. 504 LAMP and its Management Reports; Sacramento Loan Processing Center's ratios, Risk database reports.). An explanation of the potential differences between data in the Portal and data provided by other sources may also be found on SBA's Web site at http://www.sba.gov/idc/groups/public/documents/sba_program_office/olo_portal_data.pdf.
A few commenters requested that SBA Lenders be able to access previous quarters' data. The commenters explained that access to previous data would facilitate trend analysis. SBA has considered this comment and has added all previous quarters' data to the portal.
A few commenters suggested that SBA provide more than one user account per SBA Lender. Multi-bank holding companies, and SBA Lenders with centralized SBA loan processing or servicing, stated that it would be helpful to have additional user accounts for managers with various SBA lending responsibilities. SBA is working with its contractor on the possibility of allowing SBA Lenders more than one user account.
A few commenters suggested that it would be helpful if users had access to peer group performance statistics for all peer groups in the user's lending program [7(a) or 504], rather than the performance of only the user's peer group. SBA believes that providing portfolio performance information on all peer groups may be informative for SBA Lenders, and is therefore making that information available through the Portal.
Several commenters discussed SBA's proposed component factors and suggested that SBA consider other components for the Risk Rating System. Commenters suggested that SBA consider the following as additional or alternative components: (i) Historical loss rate; (ii) a longer term purchase rate; (iii) value of pledged collateral; (iv) credit scores for all principals and guarantors; (v) consideration of SBA's Start Printed Page 27613social mission; and (vi) removal of the problem loan rate.
(i) Historical Loss Rate
Several commenters suggested that incorporation of actual historical losses as a component would increase model accuracy. Ten commenters suggested substituting actual historical loss rate for the 12-month purchase rate component. In developing the risk rating model, SBA considered the use of historical loss rate as a component. It was found that while historical loss rates are somewhat predictive of future purchases, their use in combination with the other component factors provided little additional predictiveness. In addition, loss is a lagging indicator. Actual losses are not recorded until all collateral has been liquidated and normal collection efforts have been exhausted, sometimes years after the default and purchase. This may have negative implications for the calculation of losses and the SBA Lender's historical loss rate. Specifically, negative events such as loan origination fraud or poor underwriting decision-making under previous management may adversely impact an SBA Lender's risk rating for several years; conversely, improved origination or underwriting practices will only slowly be reflected in that SBA Lender's risk rating. On the other hand, the 12-month purchase component factor, where both positive and negative events will be reflected in the SBA Lender's risk rating more quickly than they would with a historical loss rate factor. In addition, the time lag inherent in a historical loss rate factor may result in the rate not reflecting the SBA Lender's current portfolio. For example, if a 7(a) Lender had originated most of its loans under the former Low-Doc program, its historical loss rates would continue to reflect losses from that program for several years, even if the 7(a) Lender's current portfolio were predominantly comprised of EXPRESS loans. Finally, SBA believes that use of historical loss rates may not reflect some of the costs borne by SBA and the Federal Government, such as the cost of funds used for loan purchases and the administrative costs borne by SBA in its liquidation oversight and charge-off activities.
A few commenters that sell their SBA loans in the secondary market believed that the use of purchase rates in the component factors and composite ratings, rather than recovery or loss rates, was a disadvantage to them given that SBA purchases all defaulted loans from the secondary market. These commenters also stated that their recovery rates should be higher than other 7(a) Lenders, since loans are purchased by SBA out of the secondary market earlier in the default curve. SBA agrees that loss rates may provide some evidence of SBA Lender risk, since the rates may be an indicator of poor origination, servicing, or liquidation on the part of the SBA Lender. In addition, the rates—over time—do show SBA's actual losses from an SBA Lender's portfolio. Therefore, SBA is reviewing its data to determine how to incorporate some measure of losses into SBA Lenders' composite risk ratings. At this time, we cannot identify the form such a measure would take, or how the measure would be considered within the Risk Rating System. For example, SBA may use net loss or recovery rates, or we may use a calculation of net cash flows to account for the revenues provided to SBA from guaranty fees and other fees. Once SBA has developed its data measurements and determined what it believes to be the best measure of losses, it will submit the proposal in the form of a notice in the Federal Register. At least until then, SBA will use the purchase rate as a key component because it is a more leading indicator, it indicates purchase, liquidation, and charge-off costs, and has tested as a better predictor of future purchases.
(ii) Longer Term Purchase Rate
A few commenters recommended that SBA continue to use purchase rates as a rating component, but proposed a longer term purchase rate of 36 months, rather than the 12 month purchase rate. During the Risk Rating System development process, SBA considered using both 24 and 36 month historical purchase rates; however, the 12 month historical purchase rate was selected because it proved to be more predictive of future purchases than either of the other two terms.
(iii) Value of Pledged Collateral
A few commenters recommended that the value of pledged collateral should be considered as a component factor. SBA considered the use of value of pledged collateral in its Risk Rating System. However, SBA believes that the use of pledged collateral should not be considered a possible component factor for several reasons. First, SBA does not regularly collect information on the value of pledged collateral on all of its loans. Second, each SBA Lender has its own individual policy regarding how it values pledged collateral; for example, different SBA Lenders will assign different market value rates to the same form of collateral. Finally, even where SBA collects data on pledged collateral, it only does so for one tax identification number, which may understate the amount of collateral actually pledged. For these reasons, SBA has determined not to use pledged collateral as part of its composite risk ratings.
(iv) Credit Scores for All Principals/Guarantors
A few commenters requested that SBA include credit information on all principals and guarantors associated with a particular loan, rather than the business and the principal owner. These commenters surmised that without credit information on all of the principals of the business, SBA might understate the loan's credit strength. Currently, SBA can only collect information on one additional principal or guarantor. SBA is in the process of increasing the number of principals and guarantors whose credit information will be used, when available.
(v) Consideration of Economic Development Goals
Several commenters stated that the ratings failed to take into consideration the economic development goals of SBA's lending programs as may be evidenced through SBA Lenders' historical loan volume. SBA appreciates the critical role that SBA Lenders play in helping to achieve SBA's economic development goals. However, the Risk Rating System is intended as a means to help SBA measure SBA Lender risk and program risk. Thus, incorporating a factor that measures SBA Lenders' success in helping SBA achieve its mission is not appropriate within the Risk Rating System.
(vi) Problem Loan Rate
Seven commenters expressed concern that including the problem loan rate as a component will be a disincentive to working with borrowers to save a business or maximize recovery on the loan during the liquidation process. SBA believes that this should not be a concern, because it is in an SBA Lender's interest as holder of a remaining percentage in the loan (generally 15% to 50%) to maximize recovery and minimize losses. Further, under SBA Standard Operating Procedure (SOP) 50-50-4 (Loan Servicing), Chpt. 7, para 1(c) and SOP 50-51-2A (Loan Liquidation and Acquired Property), Chpt. 8, para. 1(a)(4), an SBA Lender should work with borrowers to either allow the borrower to retain their business or, failing that objective, to reduce both the SBA Lender's and SBA's losses to the Start Printed Page 27614greatest extent possible. Therefore, application of the Problem Loan Rate as a component factor for all SBA Lenders should not serve as a disincentive to working with borrowers and maximizing recoveries.
Use of the Override Component
The May 1, 2006 notice proposed that the occurrence of certain factors may lead SBA to conclude that an individual SBA Lender's composite rating is not fully reflective of the SBA Lender's true risk. Therefore, the proposal provided for consideration of overriding factors. The use of the overriding factors will enable SBA to include key risk factors that are not necessarily applicable to all SBA Lenders, but which indicate a greater or lower level of risk from a particular SBA Lender than the calculated score will provide. Use of overriding factors will occur on a case-by-case basis in SBA's discretion. One of the most important overriding factors may be an SBA Lender's on-site risk-based reviews/assessments. Another important overriding factor may be the institution of enforcement actions by a regulator or other authority. Examples of other overriding factors that may be considered are: Early loan default trends; purchase rate or projected purchase rate trends; abnormally high default, purchase or liquidation rates; denial of liability occurrences; lending concentrations; rapid growth of SBA lending; inadequate, incomplete, or untimely reporting to SBA; inaccurate submission of required fees to SBA; and audits or investigations conducted by the SBA Office of Inspector General.
Commenters were generally supportive of the concept of allowing SBA to override an SBA Lender's risk rating should circumstances indicate that the SBA Lender's rating may not truly reflect SBA's risk. One commenter suggested that SBA should provide additional information on the override process. As stated in the proposal, SBA will notify an SBA Lender in the event SBA plans to override that SBA Lender's risk rating, and provide the SBA Lender with an explanation of the reason(s) for the override. If the SBA Lender disagrees with the override, it may ask SBA to reconsider the override, and provide to SBA all supporting information.
The Notice proposed the separation of SBA Lenders into peer groups based on SBA loan portfolio size, as determined by outstanding SBA guaranteed dollars. SBA based the peer groups on portfolio size for several reasons. First, it allows the peer groups to reflect each peer group's relative risk to SBA—SBA Lenders in large peer groups will generally represent a greater risk to SBA, in terms of potential dollars of loans that SBA may be required to purchase, than SBA Lenders in smaller sized peer groups. Second, basing peer groups by portfolio sizes will significantly reduce the possibility of the same event having a different impact on SBA Lenders in the same peer group. For example, the effect of the purchase of one loan by SBA will have a minimal impact on the purchase rates of SBA Lenders in a large peer group; the purchase of one loan would have a similar impact for any SBA Lender in a small peer group. Third, the size groups selected allowed SBA to split both 7(a) Lenders and CDCs into peer groups that were large enough to maintain a statistically valid number of SBA Lenders within each peer group. Finally, splitting SBA Lenders into peer groups based on the size of SBA-guaranteed loan dollars enables SBA to better monitor those SBA Lenders in the largest peer groups that represent the overwhelming majority of guaranteed dollars at risk, and allows SBA to make the best use of its oversight resources.
SBA received several comments suggesting that SBA use alternative or additional characteristics to set the peer groups. Most suggested using geographic or regional characteristics. Others suggested establishing peer groups based on loan originations, use of loan proceeds, local economic events and conditions, portfolio industry segment concentration, SBA delivery method, average loan term (months), SBA Lenders' past contribution to SBA's success in meeting its public objectives, SBA Lenders' underwriting quality, SBA Lenders' workout standards and experience, new vs. experienced SBA Lenders, average SBA loan size, SBA Lenders' business model, and organizational structure.
A number of commenters suggested that there may be a number of alternative peer groups that might be established. However, portfolio size is the only necessary alternative. This is due to the large variance in performance measures of smaller sized portfolios. Since Lenders with few loans are more likely to have extremely high or low performance measures, all lenders in the largest two peer group would only receive average ratings—none would receive above average or below average ratings. Further, as additional factors are added to further segment the peer groups, the reduced peer group size would reduce the statistical validity of the peer groups (particularly for CDCs). As the number of SBA Lenders in each peer group declines, the performance of individual SBA Lenders within each peer group will become more evident to its peers, and may affect competitive advantages or disadvantages held by each SBA Lender. Also, most of the suggested peer group factors do not provide additional measures of risk, or correlate to increased purchases on the part of SBA. We, therefore, believe basing the peer groups at this time on one metric, portfolio size, is the best measure of potential purchase risk.
SBA agrees that one or more of the alternative peer grouping categories that were suggested may be useful in understanding the problems that have resulted in an SBA Lender having a poor risk rating. However, the reasons for those risk ratings will vary from SBA Lender to SBA Lender; therefore, it is difficult to isolate one particular category among those suggested that may impact most SBA Lenders' peer ratings, and that thus would be useful in the peer groupings. As noted above, trying to implement peer groupings based upon several factors, in order to explain all possible reasons for an SBA Lender's poor risk rating, could destroy the statistical validity of the model. Therefore, SBA feels that the types of factors mentioned by commenters would be more useful in discussions between SBA and the SBA Lender as an explanation of the reasons for the SBA Lender's specific portfolio performance issues. Consequently, SBA will take such factors into account during the corrective action process, to determine the causes and remedies for the weaknesses resulting in the poor risk rating, as well as when determining whether to take any enforcement action against an SBA Lender.
Several commenters, accepting of SBA's use of portfolio size as the basis for determining peer groupings, suggested increasing the number of groups. Many of these commenters were concerned that the dollar size range of certain peer groups was broad enough to include SBA Lenders with different types and scales of operation, and thus could yield an inaccurate comparison of SBA Lenders within the peer group. SBA understands the concern; however, further segmentation of the size-based peer groups will result in many of the same problems as those noted in the preceding discussion regarding alternative or additional peer group segmentation. As SBA was developing its Risk Rating System, it was clear that each peer group would have to contain a statistically significant number of SBA Lenders to ensure the validity of the statistical model and methodologies used to risk rate SBA Lenders. Further Start Printed Page 27615splitting of the current peer groups would jeopardize the model's validity at either one or several of the peer group levels. For example, as of June 30, 2006, there were a total of eight 7(a) Lenders with portfolios of more than $500 million in SBA guaranteed dollars. In order to maintain the statistical validity of the largest dollar peer group, it was necessary to set that peer group size at $100 million or more, rather than $500 million or more.
Some commenters noted that rating peers on a curve causes some SBA Lenders in each group to have risk ratings that indicate relatively weak portfolio performance. Commenters stated that an SBA Lender with a certain risk rating in one peer group will not be comparable to another SBA Lender with the same risk rating in a different peer group. This is generally true. The nature of the Risk Rating System does not lend itself to direct comparisons between SBA Lenders in different peer groups. The Risk Rating System uses step-wise regression analysis to determine the relative weighting of each of the component factors that optimizes the system's predictiveness of future loan purchases. For each peer group, the weighting of each component factor in predicting future purchases will vary according to the relative weights that yield the greatest level of predictiveness for that specific peer group. Thus, the relative weightings of each of the component factors will change from peer group to peer group, making a direct comparison of SBA Lenders across peer groups less useful. SBA does not intend to evaluate or compare SBA Lenders across different peer groups, or against the overall portfolio. Rather, SBA will evaluate each SBA Lender according to its performance as measured against those in its peer group.
Some of these commenters suggested that SBA consider establishing benchmarks, either in lieu of, or in conjunction with, the comparative ratings. Commenters expressed that SBA Lenders should not have a poor risk rating if their portfolio performance was only slightly worse than their peers, but still within an acceptable range. For example, one commenter noted that by using the comparative analysis, some SBA Lenders could be rated relatively poorly even if they were in compliance with SBA's program. The commenter was concerned that SBA would unnecessarily spend time and resources monitoring the risk of “compliant” SBA Lenders when overall program performance was acceptable. Conversely, the concern was that there would not be enough oversight when overall program performance became unacceptable.
The comment appears to suggest that SBA should not dedicate resources to program and SBA Lender monitoring while the program is performing well. However, there is no definition of acceptable program performance; SBA would first have to develop subjective measures of program performance in order to determine whether the program meets the definition of “acceptable performance.” These measures would have to be continually monitored and replaced, as program and economic conditions change. Given the process required for implementation of new measurements and standards, the measures might easily become outdated by the time they are implemented. The comparative analysis in the current Risk Rating System adjusts to changes in program and economic conditions, so there is little possibility that the risk ratings will be based on outdated performance measures.
Second, if program performance (and the performance of the participating Lenders) is deemed “acceptable”, it is implied that SBA will reduce its monitoring of its Lenders. However, this reduction in monitoring could result in SBA failing to detect negative performance trends that could point to unacceptable performance in the future. Without ongoing monitoring, SBA may be forced to react too late to negative performance and then have to devote even greater resources to resolve entrenched SBA Lender deficiencies. Using a relative performance rating recognizes that there are always SBA Lenders that present relatively higher risk, and that SBA Lender oversight is an ongoing process to help ensure that SBA Lenders with poorly performing portfolios (relative to the peer group) improve—which will help ensure that the entire portfolio continues to perform well. By taking preventative measures to monitor lower-rated SBA Lenders when portfolio performance is relatively strong, SBA can reduce the likelihood of overall portfolio deterioration, help keep SBA losses down, and reduce SBA lending program costs.
Finally, it would be premature to develop the Risk Rating System with benchmarks at this time. This is because the System has not been available throughout an entire economic cycle. Benchmarks will be more meaningful and equitable if developed based upon long-term portfolio performance that reflects all stages of an economic cycle. We do not believe the Risk Rating System has enough historical performance information to establish meaningful benchmarks for the components. Once that data is developed, SBA may consider incorporating benchmarks. SBA will publish a notice for comments should SBA decide to propose benchmarks.
Static Pool Measurements
Some commenters suggested that SBA include all originated loans in its component factor measures, even those loans that have prepaid or been liquidated and charged-off by SBA. These commenters believe that measuring historical loan purchases as a percentage of all loans, for example, would present a more accurate picture of the quality of loans originated by SBA Lenders, because it would include good loans that had improved their credit quality so much that the loan had become eligible for conventional financing and had paid-off.
It is SBA's opinion that using only those loans still in the SBA Lender's portfolio is a better indicator of an SBA Lender's risk for the simple reason that, once a loan is paid-off, SBA no longer retains any risk of purchase. In addition, SBA believes that such an approach would be unfair to new SBA Lenders that do not have historical prepayment history to offset high purchase rates. Finally, SBA believes that prepayments affect all SBA Lenders, so the impact of one SBA Lender's prepayment history should have a minimal effect on that SBA Lender's risk rating relative to its peers.
Several respondents asked for more information on how the model weighs factors so they could better understand and evaluate L/LMS. As described above, in order to maximize the predictiveness of the Risk Rating System within each peer group, each of the component factors has a different weighting from peer group to peer group, and the weighting can vary from quarter to quarter. Commenters were also unfamiliar with the SBPS that is a key part of the model, and wanted to learn how it works in credit evaluation. The SBPS is a proprietary portfolio management (not origination) credit score based upon a borrower's business credit report and principal's consumer credit report. It is compatible with Fair, Isaac & Co.'s “Liquid Credit” origination score, which is a commercially available, off-the-shelf product used by many small business lenders.
Several commenters requested an appeals process of the rating generated by the Risk Rating System. An appeals process presumes that enforcement Start Printed Page 27616actions will be automatically generated as a direct result of an SBA Lender's risk rating. However, SBA generally does not intend to use the Risk Rating System as the sole basis for taking enforcement actions against SBA Lenders. The primary purpose of the system is to focus SBA's oversight resources on those SBA Lenders whose portfolio performance (as shown by the Risk Rating System) demonstrate a need for further review and evaluation by SBA. SBA expects that enforcement actions would typically be taken only after SBA has engaged the SBA Lender, and generally will not be taken until after the SBA Lender has had an opportunity to eliminate the problem through a corrective action process.
Text of the SBA Lender Risk Rating System
Under SBA's Risk Rating System, SBA assigns all SBA Lenders a composite rating. The composite rating reflects SBA's assessment of the potential risk to the government of that SBA Lender's SBA portfolio performance.
For 7(a) Lenders, the SBA composite rating is based on four common components or factors. The common factors for 7(a) Lenders are as follows: (i) 12 month actual purchase rate; (ii) problem loan rate; (iii) three month change in the small business predictive score (SBPS), which is a small business credit score on loans in the 7(a) Lender's portfolio; and (iv) projected purchase rate derived from the SBPS.
For CDCs, the SBA composite rating is based on three common components or factors. The common factors for CDCs are as follows: (i) 12 month actual purchase rate; (ii) problem loan rate; and (iii) average SBPS on loans in the CDC's portfolio. The third factor replaces the third and fourth factors used for 7(a) Lenders because it was found, during the testing process, to be more predictive of SBA purchases for CDCs. These factors for 7(a) Lenders and CDCs are discussed in more detail in the section entitled “Rating Components” below.
In general, these factors reflect both historical SBA Lender performance and projected future performance. The factors are derived through formulas developed using regression analysis validated and tested by industry experts. SBA performs quarterly calculations on the common factors for each SBA Lender, so SBA Lenders' composite risk ratings are updated on a quarterly basis. Each of the factors is described in more detail in the Rating Components section below.
The composite risk rating is a measure of how each SBA Lender's loan performance compares to the loan performance of its peers. Thus, an individual SBA Lender's overall loan performance (using all common factors) is compared to its peers to derive that SBA Lender's composite risk rating. SBA Lenders whose overall portfolio performance (using all of the common factors) is worse than their peers will receive a worse, or higher score, while SBA Lenders whose overall portfolio performance is better than their peers will receive a better, or lower, score.
SBA recognizes that it may be inequitable to compare all SBA Lenders in a risk rating system, without separating them into peer groups, because changes in loan performance would have dramatically different impacts on the portfolio performance of SBA Lenders of different sizes. For example, the purchase of one loan from an SBA Lender will have a much higher impact on the actual purchase rate component of an SBA Lender with a small portfolio than it will on the actual purchase rate of an SBA Lender with a large portfolio. Therefore, SBA has established peer groups to minimize the differences that could result from changes in loan performance for portfolios of different sizes. The peer groups are as follows (based on outstanding SBA guaranteed dollars):
|7(a) Lender peer groups||CDC peer groups|
|$100,000,000 or more||$100,000,000 or more.|
|$0-$999,999 [7(a) Lenders disbursed at least one loan in past 12 months]||Less than $5,000,000.|
|$0-$999,999 [7(a) Lenders did not disburse at least one loan in past 12 months]|
As noted above, the common components are used to derive a composite risk rating for each 7(a) Lender and CDC. No single component factor normally decides an SBA Lender's composite rating. However, depending upon the size of the peer group, and the variation between an SBA Lender's performance and that of its peers, a single factor can carry a disproportionate weight among the three or four components.
SBA assigns a composite rating of 1 to 5 to each SBA Lender based upon its portfolio performance. A rating of 1 indicates strong portfolio performance, least risk, and that the least degree of SBA management oversight is needed (relative to other SBA Lenders in their peer group), while a 5 rating indicates weak portfolio performance, highest risk, and therefore, the highest degree of SBA management oversight. SBA provides the following definitions for the composite ratings.
Composite 1—The SBA operations of an SBA Lender rated 1 are considered strong in every respect, and typically score well above average than their peer group averages in all or nearly all of the rating components described in this Notice. An SBA Lender rated 1 generally has relatively stable component factors and overall composite rating from one quarter to the next. Since the component factors measure previous performance, and also attempt to predict future performance, an SBA Lender rated 1 is more likely to have well below average historical purchase rates (as compared to its peers), as well as well below average current problem loan rates that predict lower than average future purchase rates. Overall, loans in the portfolio of an SBA Lender rated 1 demonstrate highly acceptable credit quality and/or credit trends as measured by credit scores and portfolio performance. An SBA Lender rated 1 typically also has a well managed SBA loan program as demonstrated through on-site or off-site reviews and assessments (of mid-size and large SBA Lenders). Based on the strengths outlined in this composite rating, SBA Lenders rated a 1 present SBA with the least amount of risk, and thus are subject to the lowest level of SBA oversight compared to other SBA Lenders in the same peer group.
Composite 2—The SBA operations of an SBA Lender rated 2 are considered good, and typically are above average in all or nearly all of the rating components described in this Notice. An SBA Lender rated a 2 has component factors and a composite Start Printed Page 27617rating that typically are relatively stable from one quarter to the next. An SBA Lender rated 2 is more likely to have below average previous (12 months) purchase rates (as compared to its peers), as well as below average current problem loan rates that predict lower than average future purchase rates. Generally, loans in the portfolio of an SBA Lender rated 2 demonstrate better-than-acceptable credit quality and/or credit trends as measured by credit scores and portfolio performance. An SBA Lender rated 2 has a generally well managed (i.e., a few minor exceptions or findings) SBA loan program as demonstrated through on-site or off-site reviews and assessments (of mid-size and large SBA Lenders). Based on the strengths outlined in this composite rating, SBA Lenders rated a 2 present SBA with a lower level of risk, and thus are subject to a lower level of SBA oversight compared to other SBA Lenders in the same peer group.
Composite 3—The SBA operations of an SBA Lender rated 3 are considered about average in all or nearly all of the rating components described in this Notice. An SBA Lender rated a 3 has, on average, component factors and an overall composite rating that generally are relatively stable from one quarter to the next. An SBA Lender rated 3 likely has average previous (12 months) purchase rates (as compared to its peers), as well as average current problem loan rates that predict future purchase rates in line with SBA peer averages. Generally, loans in the portfolio of an SBA Lender rated 3 demonstrate acceptable credit quality and/or credit trends as measured by credit scores and peer performance. An SBA Lender rated 3 has an adequate (i.e., some minor exceptions or findings, but few if any major exceptions or findings, which can be corrected in the normal course of business) SBA loan program as demonstrated through on-site or off-site reviews and assessments (of mid-size and large SBA Lenders). However, SBA Lenders rated a 3 have room for improvement, should monitor their portfolios closely, and consider methods to improve loan performance. Based on the strengths and weaknesses outlined in this composite rating, SBA Lenders rated a 3 present SBA with an acceptable level of risk, and are thus subject to standard SBA oversight compared to other SBA Lenders in the same peer group. Oversight may include requests for corrective action plans.
Composite 4—The SBA operations of an SBA Lender rated 4 are considered below average in all or nearly all of the rating components described in this Notice. An SBA Lender rated a 4 may have several changes in any of its component factor rates; the component factors and overall composite rating may demonstrate instability or negative performance from one quarter to the next. An SBA Lender rated 4 is likely to have above average previous (12 months) purchase rates (as compared to its peers), as well as above average current problem loan rates that predict future purchase rates above SBA portfolio averages. Generally, loans in the portfolio of an SBA Lender rated 4 demonstrate somewhat less-than-acceptable credit quality and/or credit trends as measured by credit scores and portfolio performance. An SBA Lender rated 4 likely has a poorly managed (i.e., both minor exceptions or findings, and major exceptions or findings) SBA loan program as demonstrated through on-site or off-site reviews and assessments (of mid-size and large SBA Lenders). Based on the weaknesses outlined in this composite rating, SBA Lenders rated a 4 present SBA with a less-than-acceptable level of risk, and are thus subject to greater than normal SBA oversight compared to other SBA Lenders in the same peer group. Oversight measures can include (but are not limited to) additional reviews or assessments, requests for corrective action plans, and/or removal from delegated loan programs, depending upon the level of activity and peer group.
Composite 5—The SBA operations of an SBA Lender rated 5 are considered well below average in all or nearly all of the rating components described in this Notice. An SBA Lender rated a 5 is most likely to have changes in any of its component factor rates, and have the greatest likelihood to have its component factors and overall composite rating demonstrate instability or negative performance from one quarter to the next. An SBA Lender rated 5 probably has well above average previous (12 months) purchase rates, and well above average current problem loan rates that predict future purchase rates above its peer group. Generally, loans in the portfolio of an SBA Lender rated 5 demonstrate less-than-acceptable credit quality and/or credit trends as measured by credit scores and portfolio performance. An SBA Lender rated 5 likely has a record of significant SBA program compliance issues as demonstrated through on-site or off-site reviews and assessments (of mid-size and large SBA Lenders). Based on the substantial weaknesses outlined in this composite rating, SBA Lenders rated a 5 present SBA with the highest level of risk, and are thus subject to extensive SBA oversight compared to other SBA Lenders in the same peer group. Oversight measures can include (but are not limited to) additional reviews or assessments, requests for corrective action plans, and/or removal from delegated loan programs, depending upon the level of activity and peer group.
The descriptions within each composite rating are not meant as definitions of the ratings, but are given to provide, in general, the characteristics an SBA Lender receiving a particular rating may exhibit. Consequently, an SBA Lender assigned a particular composite rating may not exhibit every characteristic described for that rating, nor is SBA's action limited to those stated in the descriptions.
In some cases, SBA may have reason to believe that an SBA Lender's calculated composite rating may not fully reflect the level of risk that an individual SBA Lender presents. In those cases, SBA may override the composite risk rating (either positively or negatively) and assign a different composite score. Should a decision be made to override the composite score, SBA will provide the SBA Lender with an explanation of the reason(s) for the override. More information on overrides of composite ratings is provided in the overriding factors section of this Notice.
SBA's composite ratings system utilizes a numeric scale similar to rating systems used by bank regulators and other federal loan guarantors. For example, SBA's composite rating of 1 is similar to that of a bank regulator in that it is indicative of an institution with strong performance and requiring limited regulatory oversight. SBA's rating system is similar to those of other federal loan guarantors because it measures risk and portfolio performance of loan portfolios guaranteed by SBA, rather than measuring the quality of the entire institution.
The 4 Common Components for 7(a) Lenders
SBA's Risk Rating System for 7(a) Lenders features four common component factors. The four common rating components are defined below.
(i) Past 12 Months Actual Purchase Rate—The Past 12 Months Actual Purchase Rate is an historical measure of SBA purchases from the 7(a) Lender in the preceding 12 months. Thus, this component provides a measure of 7(a) Lender performance and risk as indicated by actual SBA purchases. SBA calculates this ratio by dividing the sum Start Printed Page 27618of total gross dollars of the 7(a) Lender's loans purchased during the past 12 months (numerator) by the sum of total gross outstanding dollars of their SBA loans outstanding at the end of the 12-month period, plus gross dollars purchased during the past 12 months (denominator).
(ii) Problem Loan Rate—The Problem Loan Rate provides an indication of current 7(a) Lender risk. This problem loan indicator helps measure 7(a) Lender performance and risk by showing current delinquencies and liquidations, as well as predicting potential future purchases by SBA. Calculated using a numerator of total gross dollars of loans 90 days or more delinquent plus gross dollars in liquidation. The denominator is total gross dollars outstanding. Active purchases, dollars that are purchased but not yet charged off, are excluded from this figure.
(iii) 3 Months Change in Small Business Predictive Scores (SBPS)—The SBPS is a portfolio management (not origination) credit score based upon a borrower's business credit report and principal's consumer credit report. SBPS is a proprietary calculation provided by Dun & Bradstreet, under contract with SBA, and is compatible with Fair, Isaac & Co.'s “Liquid Credit” origination score. This component signals increasing or declining purchase risk by measuring changes in borrower credit trends, and acts as a predictor of possible future loan delinquencies, liquidations, and SBA purchases. The 3 months change in SBPS is calculated by measuring the percentage change, on a dollar-weighted average basis, of the SBPS on all outstanding SBA loans held by the 7(a) Lender, from the previous quarter to the current quarter.
(iv) Projected Purchase Rate—The Projected Purchase Rate is a predictive measure of the probability of the amount of SBA guaranteed dollars in a 7(a) Lender's portfolio that are likely to be purchased by SBA. This factor uses credit bureau data on a 7(a) Lender's individual SBA loans to project the purchase rate of a 7(a) Lender's SBA portfolio. It is a 12-month projection of future performance based on the most current credit data on a borrower's payment history. For each of a 7(a) Lender's SBA loans outstanding, SBA multiplies the amount of guaranteed loan dollars outstanding by the probability of its purchase (as determined by the SBPS of the individual loan) and totals the sum of each individual loan outstanding. This total (numerator) is then divided by the 7(a) Lender's total SBA-guaranteed dollars outstanding (denominator).
The 3 Common Components for CDCs
SBA's quantitative Risk Rating System for CDCs features three common component factors. The three common rating components are defined below.
(i) Past 12 Months Actual Purchase Rate—The Past 12 Months Actual Purchase Rate is an historical measure of SBA purchases from the CDC in the preceding 12 months. Thus, this component provides a measure of CDC performance and risk as indicated by actual SBA purchases. SBA calculates this ratio by dividing the sum of total SBA gross dollars of the CDC's loans purchased during the past 12 months (numerator) by the sum of total SBA gross dollars of their SBA loans outstanding at the end of the 12-month period, plus total SBA gross dollars purchased during the past 12 months (denominator).
(ii) Problem Loan Rate—The Problem Loan Rate provides an indication of current CDC risk. This problem loan indicator helps measure CDC performance and risk by showing current delinquencies and liquidations, as well as predicting potential future purchases by SBA. Calculated using a numerator of total gross dollars of loans 90 days or more delinquent plus gross dollars in liquidation. The denominator is total gross dollars outstanding. Note that for 504 only, active purchases, dollars that are purchased but not yet charged off, that are in liquidation (loan status of Liquidation or Purchase Pending) must be added back into the denominator, as they are not included in the outstanding figure. (This is because as a normal function of 504, nearly all loans in Liquidation are active purchases.)
(iii) Average Small Business Predictive Scores (SBPS)—The SBPS is a portfolio management (not origination) credit score based upon a borrower's business credit report and principal's consumer credit report. SBPS is a proprietary calculation provided by Dun & Bradstreet, under contract with SBA, and is compatible with Fair, Isaac & Co.'s “Liquid Credit” origination score. This component provides an indication of the relative credit quality of the loans in a CDC's SBA portfolio. The score is calculated from the average SBPS score of the loans in a CDC's portfolio, weighted by each loan's guaranteed loan dollars outstanding.
Each of the common components described above reflects a different means of measuring an SBA Lender's risk to SBA in terms of loan purchase data. Loan purchase metrics provide a core gauge of SBA lending success and program risk. SBA believes a Risk Rating System emphasizing purchase indicators provides a good measure of SBA lending risk because purchases are a strong indicator of the cost to SBA, and when tested correlated with net losses (purchase less recoveries). In addition, loan purchases are resource intensive and an administrative expense to SBA that may affect SBA's ability to provide further assistance to small businesses. Finally, SBA is a “gap” lender, and purchases can be a prime indicator of the failure of the financing to assist in the growth and development of small businesses.
In addition to the common components calculated through the use of loan performance factors, the Risk Rating System allows for consideration of additional factors. The occurrence of these factors may lead SBA to conclude that an individual SBA Lender's composite rating is not fully reflective of its true risk. Therefore, the Risk Rating System provides for the consideration of overriding factors, which may only apply to a particular SBA Lender or group of SBA Lenders, and permit SBA to adjust an SBA Lender's overall composite rating. The allowance of overriding factors in helping determine an SBA Lender's risk rating enables SBA to use key risk factors that are not necessarily applicable to all SBA Lenders, but indicate a greater or lower level of risk from a particular SBA Lender than that which the calculated score provides.
One of the most important overriding factors is an SBA Lender's on-site risk-based reviews/assessments usually performed on SBA's relatively large SBA Lenders, or that may (under extraordinary circumstances) be performed on other SBA Lenders whose performance demonstrates a highly unusual deviation from their peer group. SBA conducts on-site reviews of large SBA Lenders, performs safety and soundness examinations of SBA Supervised Lenders (SBLCs and Non-Federally Regulated Lenders), and uses certain off-site evaluation measures for less active SBA Lenders. Consequently, these assessments, as a factor, may only be available for a fraction of SBA's approximately 5,101 SBA Lenders (as of 12/31/2006). Examples of other overriding factors that may be considered are: Early loan default trends; purchase rate or projected purchase rate trends; abnormally high default, purchase or liquidation rates; denial of liability occurrences; lending concentrations; rapid growth of SBA lending; inadequate, incomplete, or untimely reporting to SBA or inaccurate Start Printed Page 27619submission of required fees to SBA; and enforcement actions of regulators or other authority. This list is not all inclusive; however, SBA does not expect any of the overriding factors to affect a significant number of composite scores.
SBA has and will continue to perform annual validation testing on the Risk Rating System, and will further refine the system as necessary to improve the predictability of its risk scoring.
SBA communicates SBA Lender performance to SBA Lenders through the use of SBA's Lender Portal (Portal). The Portal allows SBA Lenders to view their own quarterly performance data, including their current historical composite risk rating. SBA Lenders can also access data on peer group and portfolio averages. Consequently, an SBA Lender is able to gauge its performance relative to its peer group and the portfolio norm. While SBA Lenders may view their ratings, their performance indicators, and peer and portfolio averages, they are not able to view the individual ratings and performance indicators of other SBA Lenders. SBA has added all previous quarters' data to the portal.
SBA updates the Portal data each quarter approximately six to eight weeks after a calendar quarter ends. SBA Lenders can now access up to eight quarters of data on SBA Lender performance.
Correcting Portal Data
Portal data includes both summary performance and credit quality data. Because summary performance data is largely derived from data that SBA Lenders provide to SBA through 1502 and 172 Reports, SBA Lenders bear much of the responsibility for ensuring data accuracy. If an SBA Lender reviews its performance components and they do not comport with its own data records, the SBA Lender should confirm the accuracy of the underlying data. If the SBA Lender determines that the data is inaccurate, it should seek to amend any incorrect data through SBA's normal processing channels (for example—for loan performance data, SBA Lender should contact SBA's fiscal and transfer agent).
Credit quality data used to help establish certain component scores is derived from credit bureau reports of the borrower business and its principals or guarantors. To the extent that credit quality data relies on information that an SBA Lender provides on the business, its principals, or guarantors contained in the loan application and as required to be updated by the SBA Lender, the SBA Lender must take responsibility for ensuring this information is correct, complete, and updated. SBA recognizes that underlying borrower credit data cannot be changed by SBA or an SBA Lender. Therefore, any changes to data provided to credit bureaus must be reported directly to Dun & Bradstreet or Trans Union, as appropriate, by the borrower. All corrections to the Portal data (both summary performance and credit quality data) will be reflected in the quarterly update following the quarter in which the correction is entered.
SBA Lenders with at least one outstanding SBA loan may apply for the Portal access. Currently, SBA issues only one Portal user account per SBA Lender; however, we are working with our contractors on the possibility of increasing the number of Portal user accounts per SBA Lender. SBA will provide a notice to SBA Lenders if we are able to provide multiple user accounts. SBA Lenders must submit initial requests for a Portal user account (or requests to switch or terminate a user) by regular or overnight mail to SBA at the following address: Office of Lender Oversight—Capital Access, Suite 8200; Mail Code 7011, ATTN: Lender Portal, U.S. Small Business Administration, 409 Third Street, SW., Washington, D.C. 20416.
SBA Lenders must take the following steps in requesting Portal access:
1. Request must be made by a senior officer of the SBA Lender (Senior VP or above).
2. Request must be sent via regular or overnight mail to the address provided above.
3. Request must be made using the SBA Lender's stationery.
4. Request must include the user's business card.
5. The stationery and business card should include the SBA Lender's name and address.
6. The request should include the following data:
(a) SBA FIRS ID Number(s).
(b) Account user's name.
(c) Account user's title.
(d) Account user's mailing address at the SBA Lender.
(e) Account user's telephone number at the SBA Lender.
(f) Account user's e-mail address at the SBA Lender.
(g) Requesting officer's name.
(h) Requesting officer's title.
(i) Requesting officer's mailing address at the SBA Lender.
(j) Requesting officer's telephone number at the SBA Lender.
(k) Requesting officer's e-mail address at the SBA Lender.
Once SBA receives and approves the user request, the Agency will forward the approval to SBA's Portal contractor for issuance of a user account name and password. The Portal contractor will e-mail the user his or her user name and password within approximately two weeks of account approval. The user can then access its data by logging into the SBA Lender Portal web page at https://pdp.dnb.com/pdpsba/pdplogin.asp.
SBA Lender Portal Responsibilities
SBA Lenders are responsible for complying with SBA's requirements in obtaining and maintaining the Portal user accounts and passwords as set forth below and as published from time to time. SBA Lenders are also responsible for timely informing SBA to terminate or switch an account if the person to whom it was issued no longer holds that responsibility for the SBA Lender. Upon accessing the SBA Lender Portal, SBA Lenders must take full responsibility for protecting the confidentiality of the user password and SBA Lender risk rating information and for ensuring the security of the data.
By clicking on the Portal log-in button to access the Portal, SBA Lender agrees to use the Confidential Information (defined in the Portal) contained in the Portal only for confidential use within its own immediate corporate organization, and to hold and maintain the Confidential Information in confidence in accordance with the terms of the Agreement. SBA Lender agrees to restrict access to the Confidential Information to those of its officers and employees who have a legitimate need to know such information for the purpose of assisting the SBA Lender in improving the SBA Lender's 7(a) or 504 program operations in conjunction with SBA's Lender Oversight Program and SBA's portfolio management (each referred to as a “permitted party”), and to those for whom SBA has approved access by prior written consent and for whom access is required by applicable law or legal process. If such law or process requires SBA Lender to disclose the Confidential Information to any person other than a permitted party, SBA Lender agrees to promptly notify SBA and SBA's Information Provider (defined below) in writing so that SBA Start Printed Page 27620and the Information Provider have, within their sole discretion, the opportunity to seek appropriate relief such as an injunction or protective order prior to SBA Lender's disclosure. In addition, SBA Lender agrees to ensure that each permitted party is aware of the requirements of the Agreement and to ensure that each such permitted party agrees to the terms and conditions. SBA Lender agrees not to disclose, and agrees to protect from disclosure, SBA Lender's password to enter the Portal. Further, any disclosure of Confidential Information other than as permitted by the Agreement may result in appropriate action as authorized by law. The Confidentiality Agreement also provides that SBA Lender agrees to indemnify and hold harmless each of SBA and any provider of the Confidential Information from and against any and all claims, demands, suits, actions, and liabilities to any degree based upon or resulting from the unauthorized use or disclosure of the Confidential Information. “Information Provider” means Dun & Bradstreet. (Mail Provider Information notice to Dun & Bradstreet, Legal Department, 103 JFK Parkway, Short Hills, NJ 07078.)
No information contained in the Portal shall be relied upon for any purpose other than SBA's lender oversight and SBA's portfolio management purposes. In addition, SBA Lender acknowledges and agrees that the Confidentiality Agreement is for the benefit not only of the SBA but also of any party providing the Confidential Information. Any such party shall have the right and standing to pursue all legal and equitable remedies against the SBA Lender in the event of unauthorized use or disclosure.
For general inquiries, an SBA Lender may submit its inquiry by e-mail to firstname.lastname@example.org. If an SBA Lender needs to speak to an individual on a non-technical matter, it may contact Paul Bishop, Institutional Financial Analyst at 202-205-7516. SBA advises an SBA Lender to state upfront its SBA Lender name, address, FIRS number, and user name to expedite processing of all inquiries.Start Signature
Dated: May 8, 2007.
Steven C. Preston,
[FR Doc. E7-9442 Filed 5-15-07; 8:45 am]
BILLING CODE 8025-01-P