Usage of SARAS data in scoring models
Transcript of Usage of SARAS data in scoring models
Usage of SARAS data in scoring models
Dmitry Borodin
Dmitry Borodin
Head of Risk Analytics
@borodin.dm
Numerous scorecard development projects in countries including Estonia, Georgia, Iceland, Indonesia, Iran, Jamaica, Kenya, Latvia, Kazakhstan, Morocco, Palestine, Sri Lanka, Tanzania, Ukraine, etc.
MIDDLE EAST
HELPING CLIENTS IN 50+ COUNTRIES
Credit bureau services
Information solutions
Business informationDecision analytics
Psychometrics
EUROPE
LATIN AMERICA & CARIBBEAN
AFRICA
CENTRAL ASIA
Kazakhstan
KyrgyzstanGeorgia
Afghanistan
CENTRAL ASIA
ASIA
United Arab Emirates
IranIraq
Morocco
SenegalMali
Burkina Faso
Niger
Guinea BissauIvory Coast
Benin
TogoSudan
South Sudan
Zimbabwe
India
Indonesia
United Kingdom
GermanySpain
Poland
Czechia
SlovakiaRomania
Malta
MonacoIceland
Estonia
Latvia
LithuaniaUkraine
Turkey
RussiaGuyana
JamaicaBarbados
Mexico
Kenya
TanzaniaSouth Africa
EXPERTS IN CREDIT RISK MANAGEMENT
Credit Bureaus Decision Analytics
& Consultancy
Business Information
& Information solutions
Due to our lean structure, optimized processes and innovative
nature, we are the efficiency leaders in facilitating access to finance.
FinTech Innovation
Providing Business Scores
• Creditinfo has a strong focus on evaluating businesses
– Creditinfo Credit Bureau (10+)
– Creditsafe Credit Bureau (5)
– CIT Leasing (7 European Countries)
– Credit Agricole Bank
– Wells Fargo (UK)
• From both the perspective of
– Classical business financial scorecards
– Credit bureau behavioural data
• In some situations they have been first generation scores
– We understand and meet the need for training and
knowledge sharing
– We understand that the data has never have been truly
reviewed and we must anticipate data weaknesses
Creditinfo has the
knowledge and experience
in developing highly
predictive Business
Scorecards on differing
data sources
Business Data Sources
Data Sources for which
we have Identified Strong
Trends, some are relevant
only to certain markets or
company size. Most data
have incremental
improvement in
predictive power.
Application
data
Internal bank
transaction
data
Credit
bureau
reports
Financials
(Balance sheet,
P&L, ratios)
Ecommerce
platforms
(Amazon,
Alibaba)
Business Info
(Age of business,
employees, banks etc.)
Owners/ Directors
Related companies
Company
• Cash Ratio
• Debt ratio
• Debt to equity ratio
• Turnover growth rate
• ROI
• Sales to assets
• EBITDA to debt
Predictive Financial Ratios
The better lenders know
their borrowers, the
more they reduce credit
risks. This might increase
costs, but in the long run
it will ensure that loans
get repaid.
These ratios can be computed from SARAS data!
The wealth of Credit Bureau data will allow to build
strong predictive models!
Examples of Scoring Variables Based on Financial Statements
Note; Presented variables are computed as of disbursement dates of loans. Risk Measure depicts relative percentage change relative to average risk.
-80%
-60%
-40%
-20%
0%
20%
40%
60%
80%
100%
No recent
data
Low Medium High
Re
lativ
e R
isk M
ea
sure
Equity Capital (Greece)
Financial variables typically help to increase Gini by 30% leading to better and faster credit decisions!
-100%
-80%
-60%
-40%
-20%
0%
20%
40%
60%
80%
100%
Very low Low Medium High Very high
Re
lativ
e R
isk M
ea
sure
Total Assets (Iceland)
Holistic Score based on
Credit Bureau and SARAS data will improve quality
and increase speed of
credit decisions in
Georgia!
CRB and SARAS Data Enabling Better Decisions!
Financial
institution
(FI)
Holistic Score
built on CRB
and SARAS data
API callScore
returned
FI performs
informed
decision
Closing Remarks
• Financial data offers a “holistic”
view of a business and allows to
achieve a lift in predictive power
• Proved through statistical
evidence from Creditinfo
• Creditinfo will develop highly
predictive scoring models on Georgian financial data and
deliver increased value to its Clients!
Creditinfo · www.creditinfo.com · [email protected]