2. Select PERC Supporters Include…
Foundations
& Nonprofits
Government &
Multilaterals
Trade
Associations
Private
Organizations
2
3. Our Footprint
Africa
Cameroon
Kenya
South Africa
Tanzania
North America/
Caribbean
Canada
Mexico
Trinidad & Tobago
United States of America
Asia
Brunei
China
Hong Kong
India
Indonesia
Japan
Malaysia
Philippines
Singapore
Sri Lanka
Thailand
Australia/Oceani
a
Australia
New Zealand
Europe
France
Central/South
America
Bolivia
Brazil
Chile
Colombia
Guatemala
Honduras
3
4. PERC’s
Alternative
Data
Initiative
(ADI)
PERC advocates the inclusion of alternative data for use in
credit granting
alternative = regular bill payment data from telecoms, energy utilities, rental
payments and other such non-financial services that are valuable inputs for credit
decisions
5. Q: Who benefits from ADI?
A: The credit-underserved population
The credit-underserved population is estimated to include the
estimated 54 to 70 million Credit Invisible:
Immigrants
Students and young adults
Elderly Americans
Consumers operating on a cash basis
Minorities
Consumers trying to establish a good credit rating without new debt
5
6. PERC’s ADI Research
Select ADI Publications
2004 Giving Underserved Consumers Better Access to Credit Systems
2006 Give Credit where Credit is Due (w/Brookings Institution)
2008 You Score You Win
2009 New to Credit from Alternative Data
2009 Credit Reporting Customer Payment Data
2012 A New Pathway to Financial Inclusion
2012 The Credit Impacts on Low-Income Americans from
Reporting Moderately Late Payment Data
6
7. Research has shown that using alternative data for
credit granting results in:
Increased, Safer, Sounder,
Fairer and Broader
Lending
What have we found?
7
8. A New Pathway to Financial Inclusion:
8
ALTERNATIVE DATA,
CREDIT BUILDING, AND
RESPONSIBLE LENDING
IN THE WAKE OF THE
GREAT RECESSION
June 2012
11. 11
VantageScore Tier Change with Alt Data
Uses the ‘ABC’ Tiers:
900-990 is an A
800-899 is a B
700-799 is a C
600-699 is a D
501-599 is an F
Unscoreable defined
as lowest tier
More tier rises than falls
15. 15
VantageScore Score Change with Alt Data,
Helps those with damaged credit (PR & 90+ dpd)
55.8% see score increases, 30.2% see decreases
16. Research Consensus Confirms
Benefits of Alternative Data
16
March 2015!
Research(Consensus(Confirms(
Benefits(of(Alternative(Data(
!
March(2015(
Authors:(
Michael(A.(Turner,(Ph.D.(
Robin(Varghese,(Ph.D.(
Patrick(Walker,(M.A.(
17. Many Organizations Examined Alternative Data
Types of Data Examined: Utility payments, Rent
Payments, Telecom Payments, Pay TV, Cable, and
Underutilized Public Records
18. Broad Findings…A Consensus
How Big of an Issue is Credit Invisibility?
Who are the Credit Invisible?
At least tens of millions
Disproportionately low income, young, elderly, ethnic minority
What is the Risk Profile of the Credit Invisible?
Somewhat riskier than average, has a smaller superprime group, but
contains a large number of moderate to low risk consumers. The
group is NOT monolithically high risk.
How Can Alternative Data Help Eliminate Credit Invisibility?
Alternative data is found to be predictive of future performance of
financial accounts…alternative data can be used to underwrite credit…
majority of Credit Invisible can become scoreable with alternative data
20. Alt Data is Predictive of Financial Accounts
30+ DPD Delinquency Rate or Public Record
(July 2009- July 2010)
On time and severely delinquent Alt Data Payers
(Utility + Telecom) measured prior to July 2009
21. 30+ DPD Delinquency Rate on Mortgage Accounts
(July 2009- July 2010)*
Alt Data is Predictive of Mortgages
*Only includes those with an active mortgage
22. 30+ DPD Delinquency Rate on a previously Clean Mortgage
Accounts (July 2009-July 2010)*
Alt Data is Predictive of Clean Mortgages
*Only includes those with an active mortgage, Clean Mortgage defined as no delinquencies reported for
mortgages for the 24 months prior to July 2009
23. 30+ DPD Delinquency Rate on previously Clean Mortgage Accounts
(July 2009- July 2010) by VantageScore Credit Score*
*Only includes those with an active mortgage, Clean Mortgage defined as no delinquencies reported for
mortgages for the 24 months prior to July 2009, VantageScore used here only includes Traditional Data
Alt Data is Predictive of Clean Mortgages after
Accounting for Traditional Data
24. Shares of Previously Clean Mortgage Sample with / without Previous
90+ DPDs
Previously Clean Mortgage Delinquency Rates with / without
Previous 90+ DPDs
Alt Data Contains New, Useful Information
That may not be found in Traditional Accounts
Consumers with Past Alt Data Delinquencies but no Past
Financial Acct Delinquencies are not seen by lenders
but are higher risk…
25. 25
‘Consumer Friendly’ Reporting
For instance:
• Use restriction (not for employment
screening or insurance underwriting)
• Exclude all negatives less than 90 days
• Report assistance as “paid as agreed” or
exclude (e.g. LIHEAP)
• Exclude unpaid balances on closed
accounts (e.g. <$100)
26. 26
Other Alternative Data
Being Used
Rental data
United States (certain locations)
Colombia (in Bogota area)
South Africa (Johannesburg area)
Trade supply (not trade credit) for FMCG
Agricultural supply data (for rural lending)
Some fit into credit bureau model, others do not
27. 27
Digital Data Being Tested/Used
Promise of improving credit access for urban and rural
poor in emerging economies:
Mobile microfinance
Development of mobile based interface for financial services offers
new opportunities for risk assessment
Unified platform for application and distribution
Data
o Payment and prepayment patterns
o Social collateral from call log data
Smart (Philippines), M-Shwari (Kenya), Cignifi (Brazil)
Mobile data in bank lending First Access (Tanzania)
28. Hurdles to Reporting (US)
28
Technological barriers to reporting:
Complex billing cycles (footprint dependent)
Legacy IT systems
Regulatory barriers:
Some states have statutory prohibitions
Regulatory uncertainty
Jurisdictional issues—FCC, state PUCs/PSCs, CFPB
Economic barriers:
Compliance costs—FCRA data furnisher obligations
Customer service costs from lenders scaring
customers substantial
Incentives, what do you get for sharing data?
29. 29
How Should We Approach Alt Data
For traditional providers, Incentives are different.
Banks are users of the data, so they get something
for what they give.
Confidentiality concerns are different—banks are backed by
regulation, by safety and soundness concerns, and by a post-
paid relationship. Not so with alt data furnishers.
Fairness: why should these sources give a bureau data for free,
so that a bureau can make money off of it?
Here’s where regulators can help, in pushing financial inclusion
mission, and in helping the system develop trust.
30. 30
Big Data and Data Fiefdoms
Some observations from the field:
McKinsey effect
› Growing belief that every firm is sitting on a gold mine.
› Seeking to monetize data assets.
Data Fiefdoms
› Data becoming more fragmented (MNOs, banks on SME credit, banks)
› All want to be CRA/info service provider
Muddy Waters
› “Traditional” alternative data vs. “Fringe” alternative data (Robinson+Yu)
› Sensing increased uncertainty among regulators/policymakers
Here’s where regulators can help—in pushing financial inclusion
mission, and in helping the system develop trust.
31. 302 East Pettigrew Street
Suite 130
Durham, NC 27701
www.perc.net
(919) 338-2798 x803