Credit Underwriting and Default Management in Today

1
Credit Underwriting and Default Management in
Todays Private Student Loan EnvironmentPresented
by Michial Thompson Managing Director, Credit
Risk Management First Marblehead
2
How to Avoid Student Loan Defaults

• To determine how to prevent defaults, lets look
  at what the main drivers of default are
• Credit Policy Lenders make loans they expect to
  be paid back
• Collection Agency Management Ensure maximum
  performance when DQ loans are placed for
  collections
• Data Analytics Performance projections,
  reporting and collections placement streams
  driven by data analytics
• Student Loan Idiosyncrasies Deferment, youth,
  cosigners

3
PSL Credit, Data and Analytics
Historical First Marblehead
Underwriting Loans to (almost) anyone at (almost) any school with (almost) any cosigner. Student and cosigner both evaluated, and much more rigorously Quality of school considered
Credit scores Primarily cosigner FICO FICO of both student and cosigner used, and much higher values required. Many other credit attributes reviewed Custom scorecards
Collections Due diligence check the box style, modeled after federal program. Agencies compensated for carrying out tasks, not for performance. Driven by data and analytics Custom treatment streams driven by credit risk Similar to other asset classescredit cards Rigorous (micro)management of agencies and performance
Student loan specific Not much customization of credit policy or collections Especially complicated asset class to understand Deferment, forbearance, young borrowers. Large unsecured personal loan. Products custom designed with credit and portfolio management in mind Data, analytics, reporting and collections are custom designed to deal with student loan idiosyncrasies Analytical techniques specifically tailored for PSLs
Data Very few have data needed to understand credit and performance 17B in originations and performance data over 20 years Comprehensive and frequent credit bureau refreshes Robust data set of loans across multiple lenders, marketers , school lists, economic periods and credit policies
4
Credit Policy
5
Credit Policy

• Appropriate Assessment of Risk at Time of
  Application
• Beyond just FICO
• More granular credit bureau attributes
• Evaluate both student and cosigner
• Over-borrowing/loan amounts
• School types/programs
• Ability to repay

6
Credit Policy Skeletons in the Closet

• All of these are cosigned loans with cosigner
  FICO gt 750. The bars show what happens to
  defaults when we further segment these by student
  FICO.
• The student (skeleton in the closet) weighs
  heavily on the performance of the loan.
• Overall cosigned loans with cosigner FICO gt 750
  default at a higher rate than non-cosigned loans
  with student FICO gt 750.

Cosigner vs. CWS
Student FICO on Cosigned (gt750) Loans
7
Credit Policy Lend to Quality Schools

• Dropouts are twice as likely to default as
  graduates
• School, school type, and program of study are
  strong predictors of graduation rates
• Clearly graduates are more likely to get a higher
  paying job that will allow them to pay back the
  loan

8
Credit Policy Lend to Quality Schools
9
Credit Policy Control Over-Borrowing

• School certification greatly reduces
  over-borrowing compared to DTC
• Loan amount requested should be considered in
  credit decision
• Capacity metrics (such as DTI) further assess
  ability to repay and prevent excessive loan
  amounts

10
FMC Private Student Loan Scorecard Updated
Score Further Separates Risk
11
Agency Management
12
Aggressive Agency Management Approach
Define Strategy

• Define the agency type (experience, client base,
  management, etc)
• Performance drives future volume placements
• Incentive plan must be meaningful to agency to
  align performance

Develop Network

• Optimizing number of agencies per segment to
  foster competition
• Continuous refresh of agencies based on results
• Robust bullpen for quick change-out for
  performance or client need

• Goals and volume forecasts clearly communicated
• Monitoring in place for outcomes activity
  monitoring progressing
• Mutual transparency into operations
• Deep dives on root causes of performance gaps
• Volume shift algorithms for Recovery agencies
• Agencies now know they are being watched

Manage
13
Data Analytics
14
Data and Analytics

• NOT one-size-fits-all
• Collectability scorecard
• Origination, monthly performance, refreshed
  credit bureau data
• Probability of a delinquent loan curing
• Strategies driven by data
• When to place a file vs. leaving it with servicer
• Which collection agency to place with
• How long to leave loan at a given collection
  agency
• Which strategies (FB, MGRS, etc) available per
  customer
• Test-and-learn approach

15
Data and Analytics

• Agency level
• Daily, weekly, monthly
• Performance by batch, by risk segment, by
  placement stream/strategy
• Transparent view of competition
• Agent level
• Daily, weekly, monthly
• Keep track of what happens to top performers

16
Data and Analytics

• Data
• Dialer data
• Daily details of every call
• Skip-tracing
• Refreshed credit bureau data
• Phone, cell phone data
• USPS (and others) data to track relocations

17
Data and Analytics Example Agent level reporting

• Prevent best performer migration
• Plans for lower performers
• Resulted in 3 better supervisors transferred in
• They know we are watching

18
Data and Analytics Example Test-and-Learn

• Mailing Strategy Test
• Timing of communications strategy
• Borrower vs. Cosigner
• Delivery / Channel options
• Agency integration / talking points

No Cosigner
With Cosigner
19
Student Loan Idiosyncrasies
20
Student Loan Idiosyncrasies

• Deferment does not build a habit of making
  payments
• Credit policy should encourage cash-flowing loans
• Early Awareness Program
• Reach out to both student and cosigner before
  repayment
• Email, phone, mail package
• Most loans need a cosignerutilize this early and
  often
• Contact cosigner at any sign of trouble
• Include cosigner in all communications
• Require cosigner participation in FB or similar
  decisions

21
Student Loan Idiosyncrasies Example Deferment
22
Results A Case Study
23
Case Study FMD reduced delinquencies and
defaults for one major banks PSL portfolio by 50

• After taking over, delinquencies immediately
  improved. Within 6 months, annualized monthly
  charge-off rates were cut in half.