lender perspectives on credit reporting 1 2 3 · –india cibil started 2005, now experian, equifax...
TRANSCRIPT
Lender Perspectives on Credit
Reporting
Originations / Acquisitions2
Global Learning Opportunities1Originations / Acquisitions2
Account Management3
Negative,
Bureau Development Path
Demographics
Demographics
+ Criteria
Generic /
Custom
Scorecards
TimeNo bureau
Negative only
Negative + positive
Negative, positive, scorecards, non-bank data
3
Demographics
Emerging Markets Themes
• Governments interested in growing consumer economies
• Poor understanding of consumer behavior (repayment,
revenue, collections) restrains consumer opportunites &
bank growth
• Taking cues / approaches from advanced economies• Taking cues / approaches from advanced economies
• Key constraint: lack of formal economy
– Cash, barter and other styles of non-digital transaction
– Consumers avoiding tax & traceability via cash
• Countries have unique approaches, but can fall within
regional themes
4
Country Variety; Regional Themes
DevelopedPrivacy
Focused
Government /
Private
5
Private
Industry
Emerging
Private
Inside North America
• Private Bureau usage: competitive requirement– Prescreening / Marketing, Originations, Account Management
• Generic scores gamed by consumers– Consumers gaining bargaining power to lower interest rates & fees
– Increases need for custom models
• Current focus on collections, risk management• Current focus on collections, risk management– Validation discipline
– Generic odds charts vs. analysis for cutoff selection
• Longer term challenge: growing beyond the well
banked– New data vectors from alternative sources
– Cross border connections for immigrant populations
– How can we do verifications on consumer volunteered data?
6
Developing World Challenges
• Data sharing constraints by local champions (anti-competitive, anti-consumer)
• Data cost leads to lack of use
– Originations filtering
– Account management only in extreme cases
• Lack of infrastructure
– No address standardization– No address standardization
– Use of National ID (SSN, SIN) can be tricky
• Lack of Government support & understanding
– Aggregated vs. private credit information not recognized as distinct
– Need to create digital records, possibly via preferential tax treatments
• Name & account matching algorithms unique by country:
(India: V.S. Balakrishnasubramaniam; China: Lim, Im;
Latin America: Mother-Father vs. Father-Mother)
7
Emerging Market Bureaus
• All too often – negative only bureaus
– Low Hit rates, less predictive of loss, revenue, & other behaviors
– Partial, skewed, view consumer’s financial picture
– Risk management screen only, useless for finding & rewarding
responsible consumers
– Invasion of consumer privacy with no benefit to consumer
– Does not encourage consumer spending or economic growth
• Lack of understanding of bureau predictive power by banks
– Originations risk management focused on demographics
– Account management batch data: too expensive?
– Central bank regulators unaware / less aware of utility
• Too little consumer information, influence, or participation
8
Inside Asia
• India & China just getting started
–India CIBIL started 2005, now Experian, Equifax and Innovis
entering market
– Beijing and Shanghai regional bureaus, national to come?
• Japan and Korea fragmented, but fully functional
– Hard to gather complete picture of consumer’s financial picture– Hard to gather complete picture of consumer’s financial picture
• Name and trade line matching processes hard and
vary by country
• Lack of digital transactions from cash / barter / informal
economy restrains lender understanding of consumer
behavior
• Digital risk management culture emerging, but new in
all but most advanced economies
9
Developing World Originations
• Heavy focus on upper income customers
– Bank conservatism
– Frequently a high fraction of GDP informal – hence no data on
middle & lower income consumers
• Bureaus frequently ordered only on high / middle scoring applicants after they’ve passed an app. score, not generally on all applicants after they’ve passed an app. score, not generally on all applicants due to cost. (cents in USA, up to several Euros outside)
• Marketing often done to consumer’s employer or other group underwriting filter – self-employed ignored too often
• Country dependent, originations frequently focused on socio-demographic (application) data
• Information is applicant-volunteered, time & money spent confirming with onsite & phone verifications
10
Originations With Positive &
Negative Credit Bureau Criteria
<2 or >4 Revolving
>1 Revolving Accts >60 Days Past Due
Serious Delq / BK Last 18 Months
Business, Prison, P.O. Box Address
Yes
Yes
Yes
No
No
No
11
RejectRejectReject RejectAction Reject Acc’ptAcc’pt
Application Score
>5 Inquiries last 3 months
<2 or >4 Revolving Accounts
Reject Acc’ptReject
Yes
Low HighMedLow HighMed
No
Yes No
• Summarize value of bureau data into simple index
• Developing world: less useful as bureau is developing
– Big guys get in last
– Attribute counts & models less stable until they do…
Originations with Bureau Scores
High
LowCredit Bureau Score
Med High
12
Cre
dit B
ure
au S
core
Application Score
Reject Accept
Low High
High
Action
Application Score
Reject TestReject
Low HighMed
Reject Acc’ptAcc’pt
Low HighMed
Reject Acc’ptAcc’pt
Low HighMed
Emerging Account Management
• In developing countries more political will inside banks for customer level management as compared to US, - single cross-product country officer / head / leader. Not product-
focused
• Systems remain a key constraint, as in developed world
– Account focused– Account focused
– Little “talking” across products
• Strategies focus on internally available data, batch data costs often similar (or equal) to online: uneconomic to do quarterly or monthly refreshes
• Risk management culture looking at both internal & external data for account management limited
• Collections very culturally-specific on timing, scripting, contact methods, legal procedures, foreclosure/repo
13
Negative Only Credit Line Increase
Collections Acct Last 3 Months
BK / Writeoff Last 12 Months
>2 Revolving Late Payments Last 12 Mos.
Negative File Hit No Yes
Yes No
Yes No
Yes No
Action
Behavior Score
>4 Inquiries Last 60 Days
Months
Any Open Accts >60 DPD
No Action
No Action
No Action
BehScore
No Action
No Action
+10%+15%No Action
Low HighMed
Yes No
Yes No
Collections with Negative Data
External Delq
External Accts Currently Past Due
Negative File Hit
Cycles Delinquent 1
YesNo
0,12+
15
Decel
StdStdStd AccelAccelerated, Standard Decelerated
Accel LtrAccel
Behavior Score
Internal Delq Balance Amount
External Delq Balance Amount
Low HighMed
High LowHigh Low
NoneDecel
StdDecel
LowHigh
Low HighMed
High LowHigh Low
Ltr
LowHigh
Banned for Marketing?
• Frequently bureau association rules or national law prohibit
marketing uses for CB data
• How does your regulator and your lawyer feel about aggregated
CB data vs. personalized data?
• If permitted, typical uses: • If permitted, typical uses:
– Target audience selection beyond income
• The middle classes
• Self-employed, small company employees
• Underserved / recent immigrant populations
– Location of physical presence points: Sales offices,
branches, Kiosks
– Invitation to apply to a region / postcode
16
Negative Data Marketing / Promotions
Postcode High Hit Rate
%
Postcode High
Bankruptcy %
Prison, P.O. Box
Address
Yes
Yes
Yes
No
No
No
Kiosk, Branch
No Action
No Action
No ActionAction
No Action
Postcode High %
Revolving Delq
Postcode High % of 90
Days Past Due
Yes
No Action
No
Yes No
What Lenders Could Improve
• Encourage governments to formalize
economies – tax break for opening an
account?
• Use credit bureaus to move beyond
univariate approach on income + knockout univariate approach on income + knockout
rules – enfranchise consumers
• Focus on complete credit risk spectrum –
create benefits for low risk customers as
well as manage higher risk customers
• Use bureau data to make smarter line,
overlimit, collections & renewals
What Bureaus Could Improve
• Price offline data cheaper than online to
encourage analytics and account management
usage
• Give consumers a voice: actively create
opportunities to consumers to learn about, see opportunities to consumers to learn about, see
and modify their files
• Assemble data from non-bank industries that
show consumer monthly behaviors: insurance,
utility, telecomm, wireless, tax
Summing Up
• Bureaus are key piece of infrastructure for safety
& soundness & growth, but data must be
economic for lenders to use in all decision areas
• Developed markets need to move beyond
bureaus for growth – verifications, employment, bureaus for growth – verifications, employment,
customer level management
• Developing markets need more digital data on
consumers to move further into scientific risk
management