Whats the Score or Do We Want to Know _____

1
Whats the Score or Do We Want to Know?_____

• Amy Kerwin, GLHEGC
• Mike Chamberlain, OSI

2
Great Lakes Use of Scoring as a Default Aversion
Tool _____

• Amy Kerwin,
• Great Lakes Higher Education Guaranty Corporation

3
Default Aversion Scoring Objectives

• Increase our cure rate
• Maximize the use of limited resources
• spend less effort on borrowers that are likely to
  cure without significant assistance
• segment borrower populations that may benefit
  from targeted default aversion strategies

4
Scoring Challenges

• Scoring for default aversion purposes is
  different than for credit granting purposes
• Our experimentation with new default aversion
  strategies requires a dynamic modeling process
• Student loans have a significant number of
  variables that can be used in modeling

5
Great Lakes Approach to Scoring

• Partner with Fair Isaac Corporation (FIC) to
  develop default aversion scoring models
• FIC created 3 custom scoring models based on 50
  student loan data elements
• 60-120 days delinquent model
• 121-210 days delinquent model
• 211-330 days delinquent model

6
Great Lakes Approach to Scoring

• We provide FIC with daily data updates on each
  account with an outstanding default aversion
  request
• FIC provides daily scores that indicate the
  borrowers likelihood to cure or default
• We use the scores to generate our daily
  autodialer campaigns

7
Implementation Issues

• GL determines our daily/monthly dialer capacity
  FIC assigns scores within these stated capacities
• GL notifies FIC of changes in default aversion
  strategies
• GL and FIC establishes monthly autodialer
  penetration strategies based on scores

8
Critical Success Factors

• Continuous model feedback loop that considers the
  following
• changes in incoming default aversion requests
• changes in our default aversion strategies
• effectiveness of our Loan Counselors
• effectiveness of scores generated by the models
• changes in terms of which accounts cure, when
  they cure and which activities we performed

9
Critical Success Factors

• Staffing
• Creation of autodialer management team
• Focus on minimizing turnover
• Training and Development
• Effective use of the GL call model
• Use of individual development plans and
  calibration sessions

10
Critical Success Factors

• Discipline
• Utilization rate and talk/type time goals
• Minimize unscheduled time off
• Technology
• Timely transfer of data to/from FIC
• Reliance on vendors for autodialer support, phone
  line availability

11
Early Results

• Too soon to draw conclusions!
• Scores first used in August 2003 need 320 days
  before new default aversion requests scored in
  August work their way through the loan life cycle

12
Next Steps

• Develop new strategies targeted at specific
  populations of borrowers
• borrowers who break multiple promises to cure
• borrowers for whom a deferment is rejected
• Integrate scores into our skiptracing strategies

13
The Evolution of Traditional Scoring Practices

• Mike Chamberlain
• OSI Education Services, Inc.

14
Why Score?

• Lending
• Identifies future customer risk
• Determine appropriate interest rate
• Collections
• Improves liquidation
• Expense reduction
• Generate additional revenue

15
Traditional Credit Scores

• Information derived solely from credit report
• The foundation of FICO as a credit granting model
• Other examples include Experian, Beacon, and
  Empirica
• Driven by five data categories

16
FICO Scoring Variables
17
FICO Score Comparison
18
The Next Generation

• Recovery score
• A numerical representation of a borrowers
  likelihood to pay based on the integration of
  credit data and other relevant factors

19
The Recovery Score Recipe

• Credit score plus/minus other attributes
• POE phone number
• Date of last payment
• Length of employment
• Income
• Mortgage ownership
• Credit inquiries

20
Recovery Score - Now What?

• Segmentation
• The identification, separation and prioritization
  of scored accounts into groups for the purpose of
  applying different recovery treatment strategies

21
Benefits of Scoring and Segmentation

• Competitive advantages
• - Prioritizes accounts with highest return and
  greatest netback to GAs
• - Optimizes alignment between collector skill
  set and target segment
• - Maximizes use of best collection talent
• - Improves overall portfolio liquidation

22
Benefits Continued.

• Margin implications
• Reduces the need for additional headcount
• Requires lower operating expense
• Creates an opportunity to increase overall
  portfolio revenue without increasing expense

23
Segmentation Process Flow
Basic Segmentation Process
24
Collection Yield Ranges
25
Collected vs. Accounts Worked
26
Not a Perfect Science

• The human element
• Can become over-reliant on the plan
• May require incentive compensation adjustments
• Simpler is better
• Fewer segments enable improved execution
• Err on the side of liquidation performance over
  margin emphasis

27
(No Transcript)
28
Questions?
29
Thank You! _____