The views expressed in this presentation are the views of the author and do not necessarily reflect the

views or policies of
the Asian Development Bank Institute (ADBI), the Asian Development Bank (ADB), its Board of Directors, or the
governments they represent. ADBI does not guarantee the accuracy of the data included in this paper and accepts no
responsibility for any consequences of their use. Terminology used may not necessarily be consistent with ADB official
terms.

Role of the Credit Risk Database in Japan.

Creation of 〔New Economic Activity Indicators〕

through dynamic monitoring of bank accounts How Digital Innovation Can
Help Small and Medium-sized
Enterprise (SME) Access to
Finance in Asia

29-30 October 2018

Radisson Blu. Plaza Bangkok
Thailand

jointly established by
volunteer private financial
institutions in 2000

President
Yutaka Ohkubo
 The Risk Data Bank of Japan, Limited (RDB)

RDB is a database company that was

jointly established by volunteer private Large quantity of high quality
financial institutions for Business big data
Process Re-engineering (BPR).
<database sizes (Unit: 1,000 records)>
The RDB members are varied,
including all mega banks, regional N ot Defa u lt Defau lt
484 510
538 566 589
4 5 4
financial institutions, major trading 3 1 8 347 372
400 426

263 290
companies, and leasing companies. 82
138
209 239
93 113 131 153
2 45 259 2 70 2 80 288 2 94
182 207 227
43 61 76
They currently include about 70% of all
24
'0 1 '0 2 '0 3 '0 4 '0 5 '0 6 '0 7 '0 8 '0 9 '1 0 '1 1 '1 2 '1 3 '1 4 '1 5 '1 6 '1 7

Japanese banks.

RDB provides solution services to support all aspects of business such as

provision of statistical models built using our databases, instruction and
inspection of models for financial institutions’ internal rating systems,
consulting, and development and maintenance of loan screening systems in
＜RDB企業デフォルト率の推移＞
addition to managing the joint databases.

1
 Succeeding in Making the Largest and Best Collection
“When You Can Collect Data” and “When You Must Collect Data”
4％ Completely collecting in time series “large-scale bankruptcies” in which Japan suffered from
Default the largest distress (including two financial crises in the 21st century)
rate
Mar.2009 301
294 (Thousand
3.34% 288 cases)
280
270
3％
Aug.2002 259 Number of companies
2.69% 245 going into default
RDB default 227
rate
207
589
615
566 (Thousand
182 cases)
Mar.2001 538
2％ 2.15% 153 510
Dec.2006 484
Number of companies
1.95% 454
not going into default
Aug.2004 131 426
113 400
1.67%
372
93 347 Jun.2018
76 318 0.89%
290
1% 61 263
Enforcement of Companies
Announcement of Program

Announcement of revised

Second monetary easing

239

Start of monetary easing

Raising consumption tax

Start of applying Basel II

Start of applying Basel II

Service Privatization Act

Finance Facilitation Act

Bankruptcy of Lehman
Dot-com bubble burst

Negative interest rate

Trump administration
Abe’s cabinet formed
Enforcement of SME

209
for Financial Revival

Enactment of Postal

Downgrading Greek
USA/ Simultaneous

Brothers Securities

43
Great East Japan
government bond
Basel Committee

(Lehman shock)
terrorist attacks

Earthquake

138
24

formed
(AIRB)
(FIRB)
frame

policy
rate
Act

82
2000 2001 2002 2004 2005 2006 2007 20082008 2009 2010 2011 2012 2013 2014 2014 2016 2017
May Sep Oct Jun Oct May Mar Mar Sep Nov Apr Mar Dec Apr Apr Oct Jan Jan

2
 Japan SME Credit Score by S&P Global Market Intelligence

Only for unlisted Japanese SMEs

with annual sales of 500 million yen or more.
Ratings are divided into 7 stages,
symbolized in lower case letters, “ aaa” to “ccc”.
Rating Certificate Icon

Verify SME’s Merits for SMEs

Financial institution Request to verify
existence
their existence
1 Develop Sales and Enhance client relations
application
S&P 2 Secure Qualified human resources
RDB SME
Certificates and
(Business center)
financial report,etc.
3 Receive an Objective financial analysis

Financial Report S&P Web Publishing

3
 RDB

RDB models for corporations/

Credit models for SMEs sole proprietors
<RDB’s scoring model> Model (with S&P GMI) (Evaluation models for credit risks)

RIBA
(Rating models for
Large
enterprises
large enterprises) Databases for
LGD estimation corporations
models
RIBA (models for
中堅企業
large enterprises)
Rent estimation Databases for
Credit models for SMEs
models sole proprietors Database
SMEs RDB default rate
Databases for
collecting bad loans
RDB models for corporations “Credit Research”
constructed only by RDB
“Operational Risk Research”
Individuals (Investigation & analysis in Japan
reports for members) Databases for
RDB models for sole proprietors
operational risks
Ratings for Japanese SMEs
(with S&P GMI)
(Credit ratings only for SMEs) RDB C-Voice <Structure of joint databases>
(Databases for
<Information bulletins local governments)
TORA (Technical Operational Risk Assessment)
for members> (Support for enhancing operational risk management)
Dynamic databases for
Credit Research Internal rating system commercial distribution Member
Advisory for construction RDB DynaMIC financial
and inspections
(Dynamic monitoring system) institutions
DynaMIC LOAN constructed only by RDB Submitted
Research (Screening engine module) in Japan
Returned (After data
data anonymization)
&
Solutions
Dynamic Monitoring
for Bank Accounts Databases

4
 Structural Challenges of Static Monitoring and
the Necessity of “Bank Accounts Dynamic Monitoring”

Financial statements are evidence presented by companies. Credit rating systems are built on evaluations
based on financial statements. However, the following challenges have been identified.

[Discrete fixed point observation]

 Internal rating systems based on settlement year information lack a method capable of evaluating
customers’ credit information as it changes from day to day.
 As a result, a credit investigation is necessary when financial needs occur, but it is unclear where such
Structural challenges

investigations should focus.

[Occurrence of constant time lag]
 There is a constant time lag of “9 to 21 months” in confirming customers’ business conditions because
financial statements are received three months after accounts settlement.
 Thus, growth capital and working capital cannot be lent in a timely manner.
[No difference in individual companies’ information]
 Financial statements received from customers are “passive” information that is “identical” to what other
financial institutions have, which does not lead to differentiation.
 Credit ratings are determined based on only an individual company’s information; they do not include its
business relationships with its suppliers and customers as a whole.

Digitizing all transactions Need to develop Drastic improvement in

instantly computer processing capabilities

One More FinTech

Dynamic monitoring for settlement bank accounts

RDB-DynaMIC
5
 One More FinTech Dynamic Monitoring for Bank Accounts

 Fund transfer information of bank accounts indicates current economic situations.

 RDB stores all information as data with the linkage structure among commercial distribution.

The only and first [Actual measurement] Commercial distribution

implementation in Japan Financial and economic network
activities Structural monitoring
Fund transfer Mathematical 財務 D yn aM IC 　R A N K

Credit risks
Analog to
ランク 1 2 3 4 5 6 7 8 9 10 総計
1 0 .0 0 % 0 .0 0 % 0 .0 0 % 0 .0 0 % 0 .0 0 % 0 .0 0 % 0 .0 0 % 0 .0 0 % 0 .0 0 % 0 .0 0 % 0 .0 0 %

information (AI)models
2 0 .0 0 % 0 .2 0 % 0 .3 4 % 0 .0 0 % 0 .0 0 % 0 .0 0 % 1 .6 7 % 0 .0 0 % 0 .0 0 % 0 .0 0 % 0 .1 4 %

digital Dynamic
3 0 .0 0 % 0 .0 0 % 0 .3 1 % 0 .0 0 % 0 .0 0 % 0 .0 0 % 0 .0 0 % 1 .4 3 % 0 .0 0 % 3 .8 5 % 0 .1 4 %
4 0 .0 0 % 0 .0 0 % 0 .0 0 % 0 .7 1 % 0 .4 2 % 0 .0 0 % 0 .6 5 % 2 .3 8 % 0 .0 0 % 4 .6 5 % 0 .3 6 %
5 0 .0 0 % 0 .0 0 % 0 .3 4 % 0 .3 4 % 0 .0 0 % 0 .7 9 % 1 .4 6 % 2 .3 8 % 1 .9 6 % 7 .1 4 % 0 .7 7 %
6 0 .0 0 % 0 .0 0 % 0 .8 1 % 0 .4 1 % 0 .0 0 % 1 .0 3 % 1 .1 8 % 1 .0 7 % 3 .6 0 % 9 .1 7 % 1 .2 3 %

converter
7 0 .0 0 % 0 .0 0 % 0 .5 3 % 0 .8 9 % 1 .6 3 % 0 .9 5 % 2 .2 6 % 1 .6 3 % 1 .0 7 % 6 .2 1 % 1 .5 1 %

monitoring
8 0 .0 0 % 1 .5 3 % 2 .0 3 % 1 .6 8 % 1 .9 3 % 2 .7 1 % 3 .7 9 % 1 .7 1 % 5 .2 2 % 8 .1 4 % 3 .1 9 %
9 0 .0 0 % 0 .0 0 % 1 .1 8 % 0 .7 4 % 3 .0 2 % 3 .9 4 % 3 .2 4 % 3 .3 2 % 5 .7 7 % 1 0 .0 5 % 4 .4 6 %
10 7 .4 1 % 0 .0 0 % 7 .3 2 % 3 .9 5 % 3 .9 4 % 5 .3 8 % 6 .7 4 % 8 .9 0 % 9 .8 2 % 1 8 .8 6 % 1 0 .8 8 %
総計 0 .0 6 % 0 .1 1 % 0 .6 0 % 0 .6 4 % 1 .0 6 % 1 .7 7 % 2 .7 9 % 3 .3 0 % 5 .1 2 % 1 2 .4 0 % 2 .2 8 %

Economic activity
× ＝ indicators
Dynamic monitoring
for PDCA for
restoration in case of
natural disasters
and other crises

Vividly describe the dynamic state of economic effects of “exchange rate appreciation/depreciation,
initiation of policies, bank loans, and macro-economic trends” on each economic entity.
Constantly monitoring from “compound-eye macro-viewpoints” as well as “freely drill-down micro-
viewpoints ” based on the records of one company and one person.

The Risk Data Bank of Japan. 6

Structuring many micro prudence into consistent macro prudence
<by company/person> ⇒ (by region)⇒ (by industry) ⇒ (by size) ⇒<national economy>

 Automatically continuously monitoring what kind of issues national/regional economy has

at present, which business type of companies and which size of companies are short of funds, and
which age group and which occupation have low income levels.
 Automatically monitoring which region/ business type/ size/ occupation are largely affected by
fluctuations in foreign exchange rates by visualizing clouds approaching, rain falling.
 Making after-the-fact verifiable policies by scientifically measuring the economic effect of
policy implementation.
 To realize a better, healthy society, finding out rationally and timely “vulnerable groups
in society” required for safety.
 Monetary and fiscal policies to inject necessary funds into necessary factors at a pin point like
“catheter operation”.

Confirmation of an intimacy
level between companies

1,000 kinds of dynamic news are continuously

transmitted automatically
by computer with AI.
Confirmation of fictitious
transaction relationships
Confirmation of effects
of chain bankruptcies

7
 Development of SMEs through Dynamic Monitoring of Bank Accounts
(Unification of MICROPRUDENCE and MACROPRUDENCE) Micro
prudence

Bidirectional consistent monitoring (drill down/ drill up)

Micro information

By company

Commercial network
information
Specifying
“core companies”
in regions

Macro information
Dynamic economic
activity indicators Macro
prudence

8
 Dynamic New Economic
Activity Indicators

 Making economics “Real Science”

 Scientific unification of micro prudence and macro
prudence Economic
Economic
temperature  Realizing after-the-fact verifiable PDCA policies weather
 Implementing policies like MRI & catheter
operations
 Comprehensive, continuous and scientific
measures taken for restoration in disasters
Economic
Economic forecast
humidity
PDCA of monetary &
Economic fiscal policies
atmospheric
pressure
9
 July 2018 West Japan
Torrential Rain
Dead: 223 people
Wounded: 442 people
House collapse/under water:
50,000 houses
July 2017 Northern Kyushu
Torrential Rain
Dead: 44 people
Wounded: 39 people
House collapse/under water:
3,800 houses
June 2018 Typhoon No.21
Disaster in Japan August 2018 Hokkaido Eastern
Iburi Earthquake
(2016–2018)
April 2016 Kumamoto Earthquake
Dead: 41 people
Wounded: 689 people
Dead: 271 people House collapse/under water:
Wounded: 2,808 people 4,700 houses
House collapse: 200,000 houses
August 2016 Typhoon No.10
(Tohoku)
Dead: 29 people
Wounded: 14 people
House collapse/under water:
6,000 houses
June 2018 Northern Osaka Earthquake
Dead: 13 people Dead: 5 people
Wounded: 912 people Wounded: 454 people
House collapse/under water: House collapse/under water:
22,000 houses 54,000 houses
10
“Dynamic Monitoring for Bank Accounts” Can Immediately Grasp
the Situation and Begin Restoration in Emergencies,
Including Natural Disasters.

Visualization of business continuity statuses and restoration per company and

region when economic crises and natural disasters occur.
Visualization of the effects of large companies’ business slumps, regional
withdrawals, and exchange rates crises, etc. on commercial distribution-
related companies.

(1) Recognition of
Physical limitation of targets
site investigation
 Visiting is very risky Indexation of
 Impossible to cover (4) Dynamic achievement (2) Recovery
everything monitoring for plan
levels of
restoration restoration

(3) Undertaking
restoration

Concrete beforehand simulation of financial training for earthquakes

and other disasters can be performed.
11
 Restoration from Kumamoto Earthquake
Restoration policies through (unification of micro prudence and macro prudence)
Macro
prudence
Changes in deposits of operating funds (by industry)
 The construction industry
obtained a chance for restoration
Construction needs.
industry  The “accommodation industry”
rapidly worsened. ⇒ Physical
bottleneck of restoration
Earthquake
occurrence Accommodation
industry Bidirectional consistent
monitoring
(drill down/drill up)
★Specification of companies to be supported Micro
prudence The red circle shows
Deposits of operating funds by company (compared to the previous
month/year) – Accommodation industry – accommodation suppliers that
decreased deposits.
➨Expected that the suppliers are not
in full operation and probably
Compared to the
previous month

require various types of support.

The blue circle shows

accommodation suppliers that
Bipolarization
increased deposits.
phenomenon ➨The suppliers probably require
operating funds etc.
Compared to the previous year
12
Changing wage payments into Dynamic Indicators (DI) Changing wage receipts into Dynamic Indicators (DI)
based on information of companies (payers) based on information of individuals (recipients)

Continuously monitoring rising (declining) wage trends Continuously monitoring wage rising (declining) trends
(by industry), (by company size), and (by region), etc. (by age), (by income level), and (by address), etc.

Wage Payment DI (by industry) Salary Receipt DI (by age)

Teens

Service industry
All generations
20s–30s

Manufacturing industry

40s–50s

 Automatically monitors continuously which business type of companies and which size of companies show poor
business performance, and which age group and which occupation have low income levels in the present
national/regional economy.
 To realize a better, healthy society, finding out “vulnerable groups in society” requiring safety nets in “various standards”
 Monetary and fiscal policies to inject necessary funds into necessary factors at a pin point like a “catheter operation”

13
 Relativity analysis
(mutual complementation)

[Monitoring financial results trends] [Dynamic monitoring of economy]

(Normalization/standardization as (Creating new economic indicators based
“economic indicators”) on deposit/withdrawal of accounts)

Relativity analysis
 Changes in renewed credit  Changes in deposit/withdrawal
rating of companies with the of accounts
fiscal year ended  Changes in dynamic relations/
 Changes in creditworthiness core commerce/ inter-economic
derived from RDB credit scores zone relations/ wage payment
 Changes in profits/sales etc. of  Changes in cash
companies with the fiscal year positions/risks
ended  Prediction of financial results
財務 D yn aM IC 　R A N K
ランク 1 2 3 4 5 6 7 8 9 10 総計
1 0 .0 0 % 0 .0 0 % 0 .0 0 % 0 .0 0 % 0 .0 0 % 0 .0 0 % 0 .0 0 % 0 .0 0 % 0 .0 0 % 0 .0 0 % 0 .0 0 %
2 0 .0 0 % 0 .2 0 % 0 .3 4 % 0 .0 0 % 0 .0 0 % 0 .0 0 % 1 .6 7 % 0 .0 0 % 0 .0 0 % 0 .0 0 % 0 .1 4 %
3 0 .0 0 % 0 .0 0 % 0 .3 1 % 0 .0 0 % 0 .0 0 % 0 .0 0 % 0 .0 0 % 1 .4 3 % 0 .0 0 % 3 .8 5 % 0 .1 4 %
4 0 .0 0 % 0 .0 0 % 0 .0 0 % 0 .7 1 % 0 .4 2 % 0 .0 0 % 0 .6 5 % 2 .3 8 % 0 .0 0 % 4 .6 5 % 0 .3 6 %
5 0 .0 0 % 0 .0 0 % 0 .3 4 % 0 .3 4 % 0 .0 0 % 0 .7 9 % 1 .4 6 % 2 .3 8 % 1 .9 6 % 7 .1 4 % 0 .7 7 %
6 0 .0 0 % 0 .0 0 % 0 .8 1 % 0 .4 1 % 0 .0 0 % 1 .0 3 % 1 .1 8 % 1 .0 7 % 3 .6 0 % 9 .1 7 % 1 .2 3 %
7 0 .0 0 % 0 .0 0 % 0 .5 3 % 0 .8 9 % 1 .6 3 % 0 .9 5 % 2 .2 6 % 1 .6 3 % 1 .0 7 % 6 .2 1 % 1 .5 1 %
8 0 .0 0 % 1 .5 3 % 2 .0 3 % 1 .6 8 % 1 .9 3 % 2 .7 1 % 3 .7 9 % 1 .7 1 % 5 .2 2 % 8 .1 4 % 3 .1 9 %
9 0 .0 0 % 0 .0 0 % 1 .1 8 % 0 .7 4 % 3 .0 2 % 3 .9 4 % 3 .2 4 % 3 .3 2 % 5 .7 7 % 1 0 .0 5 % 4 .4 6 %
10 7 .4 1 % 0 .0 0 % 7 .3 2 % 3 .9 5 % 3 .9 4 % 5 .3 8 % 6 .7 4 % 8 .9 0 % 9 .8 2 % 1 8 .8 6 % 1 0 .8 8 %
総計 0 .0 6 % 0 .1 1 % 0 .6 0 % 0 .6 4 % 1 .0 6 % 1 .7 7 % 2 .7 9 % 3 .3 0 % 5 .1 2 % 1 2 .4 0 % 2 .2 8 %

14