Portfolio Management Package Insights A quarterly briefing with best practices and thought leadership concepts from your Portfolio Management Package (PMP) team
Contents 1. New Special Handling Code (First Time 31 Days Past Due) 3 2. Getting the most from your PMP investment...5 3. Product Spotlight Risk Triggers.7 4. Industry Comparisons..9 PMP Insights July 2010 Page 2
1. New Special Handling Code THE CHALLENGE We ve observed an interesting trend over the past two years and it involves customers who are traditionally the best scoring customers. A phenomenon occurs when these customers hit 31 days (2 cycles) past due for the first time. Many of our clients have seen their best scoring customers become their worst performing customers in terms of losses. When this occurs, the population of customers is often in a sink or swim position: they pay fairly quickly after collection efforts begin or they roll straight to loss. We call these customers straight rollers. While it seems somewhat counter-intuitive, straight-rollers are often facing some sort of major hardship (loss of job, medical issues, etc) and they simply do not know how to respond to their situation. The real dilemma for our clients is identifying the straight rollers in the portfolio before they become losses. In recent analysis we conducted for a regional banking client, we discovered an excellent separation in performance among a subset of accounts that rolled to 31 days past due and beyond. The chart below presents our findings. The analysis compares performance of accounts that are 31 days past due at an observation point and how they performed over the following 3-month period: Total 31+ Population Risk Loss % Bad % (91+, no loss) Bad % (61-90) Indet % (31-60) Cured (< 31 DPD) % of Accts Low 3.58% 3.05% 14.45% 44.10% 34.82% 29.26% Medium 8.79% 4.33% 23.53% 39.12% 24.24% 38.92% High 11.83% 5.02% 23.97% 34.40% 24.78% 31.82% First Time 31 Days Past Due Risk Loss % Bad % (91+, no loss) Bad % (61-90) Indet % (31-60) Cured (< 31 DPD) % of Accts Low 5.94% 3.96% 13.86% 11.88% 64.36% 26.37% Medium 11.26% 4.64% 11.92% 18.54% 53.64% 39.43% High 25.95% 6.11% 6.11% 17.56% 44.27% 34.20% Looking at the low risk group as an example, you will notice that the loss rate is 66% higher in the First Time 31 Days Past Due population compared to the Total 31+ Population (5.94% versus 3.58%). Loss is defined as accounts that charged off during the three months following the observation period. In addition, the cured rate for this same period is 85% higher for the bottom group compared to the total population (64.36% versus 34.82%). Our analysis supports the sink or swim phenomenon. PMP Insights July 2010 Page 3
SOLUTION Earlier this year, we were on a mission to isolate accounts that have a) rolled 31 days past due for the first time and b) have been delinquent (1 day past due) two times or less. Our work is now done! A new special handling code has been created within PMP to identify accounts in a client s specific portfolio that have rolled to 31 days past due for the first time. The new code (ASHP and BSHP) can be utilized by installment lending clients. Implementing the new codes is easy. Once the codes are turned on, they ll identify target accounts and rank the special handling code appropriately (i.e. should this new code take priority over existing special handling codes). The new code will come over on feedback files in the 4-byte strategy code field. New strategies can then be applied to these accounts. Strategies may include: Designating specialized collectors to proactively manage these accounts, conduct deep-dives into what is going on with the consumer and offer applicable hardship programs Offering temporary extensions Conducting field visits Implementing specialized letter campaigns If you are interested in deploying the new special handling code, please contact your PMP consultant. PMP Insights July 2010 Page 4
2. Getting the most from your PMP investment There are many ways you can use PMP to manage accounts through the collections process. We often find clients are interested in trying new strategies or segmenting accounts differently, but may not be aware of everything PMP has to offer. In future editions of PMP Insights, we ll take a look at different functionality and share examples of how Experian clients are using these capabilities to improve processes and decrease losses. Risk Grade Breaks OVERVIEW PMP is an efficiency tool designed to increase lift in the collections process by scoring accounts using a risk-based scorecard. PMP has other functional specifications designed to enhance collection efficiency gains including the use of risk grade breaks. The ability to differ risk grades among delinquency levels and portfolios is a capability that allows our clients to expand their strategies. During initial set up meetings, the PMP consultants work with the client to establish unique portfolios among the total account population. Most clients use one set of risk grade breaks in each of the portfolios for all delinquency levels during early strategy reviews. This approach is recommended to ensure that analysis performed after PMP has gone live includes quantifiable variables on which to measure a true return on investment. After several iterations of strategy review sessions, more sophisticated strategies are recommended. PMP was created to allow clients to set up differing risk grade breaks at each delinquency level within each portfolio. The number of risk grades also varies by client. A client may have up to five risk grades per portfolio. We typically see clients using three to five risk grades. We have found that different types of portfolios often behave very differently. Accounts that may score out as an A account in an auto portfolio may be more similar in performance to that of a B account in a mortgage portfolio. Further, it is advantageous for the client to differ risk grade breaks by delinquency level. Accounts are rescored at each delinquency level and separate score breaks could be warranted. Accounts that score as an A when they are 1 day or one cycle past due will often generate a very different score when they hit the next delinquency level. Varying score breaks can group accounts differently, and thus new strategies can be applied. CASE STUDY Recently, a captive automotive client experienced great success by expanding from four risk grades to five risk grades for all of their portfolios. In the four risk grade scheme, the client experienced good segmentation between each risk grade. Treatment for the top two risk grades was very lenient, while treatment for the bottom two risk grades was quite aggressive. A large concentration of accounts in the top two risk grades led to the test for five risk grades to identify any sub-population within the current scheme that could benefit from different treatment. Success was found in creating five risk grades that primarily segmented the bottom performing risk grade B accounts and top PMP Insights July 2010 Page 5
performing risk grade C accounts and applying a middle ground of treatment from the aggressive and lenient pathways. The following chart shows results of this test: The chart reflects 1 to 31 roll rates (cycle 1 to cycle 2) between champion and challenger. The champion test ran with four risk grades, while the challenger test ran with five risk grades. The circled area above reflects the performance of the newly segmented risk grades. For the period tested, the champion accounts fell into the bottom layer of the B risk grade, while the challenger accounts fell into the newly created C risk grade. The champion accounts were treated leniently, while the challenger accounts were given a moderate level of collection treatment. As you can see, the challenger accounts experienced a significantly lower roll rate with the more aggressive treatment. Further, the performance of the champion accounts within this top B segment performed more similarly to the champion top grade C accounts. This substantially higher roll rate showed that the performance of these accounts was far different than that of the top B champion accounts and that a more aggressive treatment was necessary. PMP Insights July 2010 Page 6
3. Product Spotlight Risk Triggers CHALLENGE Clients utilize industry-specific or custom scorecards to segment accounts by risk. Scorecard performance is monitored on a monthly basis to ensure the best possible solution is available and utilized, but there are always segmentation opportunities outside of a scorecard. Finding the best segmentation tools is critical in targeting accounts with the appropriate treatment. We have often found that there will be subsets of populations within segmentation schemes that perform quite differently than the population as a whole. There are various tools from Experian that can help identify these pockets of accounts. SOLUTION Experian has two types of triggers risk triggers and collection triggers. For this issue, we would like to highlight what risk triggers can potentially do for your portfolio. Risk triggers provide you with recent credit information about your consumer, such as new credit inquiries, negative reportings with other lenders, new public records, etc. When utilized, risk triggers will monitor accounts for thirty days. Any new information that appears on the consumer s credit file during that time will hit as a risk trigger. In recent analysis conducted for a regional banking client, we found that certain types of risk triggers can add significant segmentation value. By combining scorecard segmentation with risk triggers, pockets of populations within the portfolio can be identified and then treated separately. CASE STUDY REGIONAL BANK For our case study, we analyzed the population of accounts that were 31 days (2 cycles) past due for a particular month and monitored performance over the following three months. We combined trigger data for the same time period to analyze which triggers provided value. For this particular client, we identified several triggers that provided significant segmentation value. These triggers included a recent home equity inquiry (IQHEQ) and recent negative reportings at sixty days past due and beyond (PD060+) with another lender. PMP Insights July 2010 Page 7
Grade Flag % in Population Current-1 31-60 61-90 91+ Charge-off BK A No Special Condition 13.24% 42.42% 36.39% 11.31% 8.61% 0.57% 0.69% FPD3 2.06% 48.34% 14.76% 14.02% 20.66% 0.74% 1.48% IQHEQ 0.87% 30.43% 36.52% 6.09% 20.00% 2.61% 4.35% PD060+ 2.39% 27.39% 32.48% 21.66% 12.10% 1.91% 4.46% B No Special Condition 20.36% 29.01% 46.00% 18.26% 5.41% 0.19% 1.12% FPD3 1.92% 31.62% 31.62% 18.58% 14.23% 0.79% 3.16% IQHEQ 1.89% 21.37% 40.32% 18.15% 13.71% 1.61% 4.84% PD060+ 10.68% 22.92% 45.48% 20.43% 7.97% 0.57% 2.63% C No Special Condition 16.59% 24.20% 44.78% 23.10% 7.29% 0.18% 0.46% FPD3 1.17% 37.01% 27.27% 14.29% 17.53% 1.30% 2.60% IQHEQ 1.32% 19.08% 43.35% 22.54% 12.14% 0.00% 2.89% PD060+ 7.96% 15.95% 43.65% 25.12% 12.61% 0.19% 2.48% D No Special Condition 10.26% 13.19% 47.70% 29.85% 8.30% 0.22% 0.74% FPD3 0.21% 21.43% 42.86% 14.29% 10.71% 3.57% 7.14% IQHEQ 0.98% 15.50% 41.86% 31.78% 9.30% 0.00% 1.55% PD060+ 8.09% 11.94% 45.96% 29.70% 9.12% 0.28% 3.01% Total Population 25.39% 42.72% 21.06% 8.80% 0.42% 1.62% The table above reflects account performance for each risk grade with no special condition, meaning these accounts did not contain a trigger hit. This line is your base line for performance. Also listed are accounts that hit thirty-one days past due for the first time with this lender (FPD3), accounts where the recent home equity inquiry trigger hit (IQHEQ), and accounts that have recently been reported negatively for going sixty days past due and beyond with another lender (PD060+). If you look at the risk grade A accounts, you will notice that the IQHEQ accounts had a charge-off rate of 2.61%, versus accounts that had no special condition at 0.57%. Also, there is a much higher incidence of bankruptcy (BK) for the IQHEQ accounts at 4.35% compared to 0.69% for the no special condition accounts. Further, the percentage of accounts that cured (Current never more than 1 day past due) is much lower for the IQHEQ and PD060+ accounts. Knowing that these accounts show such differences in performance, the next questions are how do you treat these accounts differently and do you have the capacity to do so? Using the same example of accounts, you will see that a very small percentage of accounts hit these particular triggers. For IQHEQ in risk grade A, only 2.06% of the total population hit the trigger. For PD060+, less than 1% of the total population hit the trigger. Providing special treatment to this volume of accounts is very manageable. One option could be to use your most experienced collectors to perform deep dives on these accounts and potentially offer hardship programs. A second option could be to simply accelerate collection activity after the evaluation. By managing these small pockets of accounts differently you can generate very different levels of performance in your collections process. PMP Insights July 2010 Page 8
4. Industry Trends These charts show a number of delinquency metrics we have gathered using data generated by our monthly PMP reporting. PMP has been deployed in three vertical markets: mid-tier banks and credit unions, captive auto finance, and utility companies. The metrics we have gathered include historical roll rates for clients in each of the three vertical markets. PMP Insights July 2010 Page 9
Bank 1 to 31 Roll Rate 6.00% 5.50% 5.00% 4.50% 4.00% 3.50% 3.00% Average Comp 1 Comp 2 Comp 3 Comp 4 Comp 5 2.50% 2.00% Jul-07 Sep-07 Nov-07 Jan-08 Mar-08 May-08 Jul-08 Sep-08 Nov-08 Jan-09 Mar-09 May-09 Jul-09 Sep-09 Nov-09 Jan-10 Mar-10 May-10 Bank 31 to 61 Roll Rate 40.00% 35.00% 30.00% 25.00% 20.00% Average Comp 1 Comp 2 Comp 3 Comp 4 Comp 5 15.00% 10.00% Jul-07 Sep-07 Nov-07 Jan-08 Mar-08 May-08 Jul-08 Sep-08 Nov-08 Jan-09 Mar-09 May-09 Jul-09 Sep-09 Nov-09 Jan-10 Mar-10 May-10 PMP Insights July 2010 Page 10
Captive Auto 1 to 31 Roll Rate 8.50% 8.00% 7.50% 7.00% 6.50% 6.00% 5.50% 5.00% Average Comp 1 Comp 2 Comp 3 4.50% 4.00% 3.50% July-07 September-07 November-07 January-08 March-08 May-08 July-08 September-08 November-08 January-09 March-09 May-09 July-09 September-09 November-09 January-10 March-10 May-10 Captive Auto 31 to 61 Roll Rate 24.00% 22.00% 20.00% 18.00% 16.00% 14.00% 12.00% Average Comp 1 Comp 2 Comp 3 10.00% 8.00% July-07 September-07 November-07 January-08 March-08 May-08 July-08 September-08 November-08 January-09 March-09 May-09 July-09 September-09 November-09 January-10 March-10 May-10 PMP Insights July 2010 Page 11
Utilities Cycle 1 to Cycle 2 Roll Rate 38.00% 33.00% 28.00% 23.00% 18.00% Average Comp 1 Comp 2 Comp 3 13.00% 8.00% October-08 December-08 February-09 April-09 June-09 August-09 October-09 December-09 February-10 April-10 Utilities Cycle 2 to Cycle 3 50.00% 45.00% 40.00% 35.00% 30.00% Average Comp 1 Comp 2 Comp 3 25.00% 20.00% 15.00% Oct-08 Dec-08 Feb-09 Apr-09 Jun-09 Aug-09 Oct-09 Dec-09 Feb-10 Apr-10 PMP Insights July 2010 Page 12