This document discusses how predictive analytics can help banks improve risk management. It begins by outlining the major risks banks face and the regulatory requirements around risk management. It then discusses how predictive analytics can enhance various aspects of enterprise risk management, including improving credit decisioning, enhancing credit quality, and enabling a 360-degree view of customers. The document provides examples of how social network analysis and big data can generate insights to better identify fraud and risk. Overall, the document argues that predictive analytics, when embedded into risk management frameworks, can help banks more efficiently identify and respond to risks.
2. With increasing levels of
economic volatility and global
interconnectivity, a “good”
economy can turn “bad” much
more quickly today than 25
years ago. Globalised
economies enable companies
in one country to tap into
markets in other countries,
thereby reducing their
exposure to local factors and
reducing point failures.
However the very same
interconnections also create
issues where a strong local
economy does not guarantee
the strength of the companies
based in that economy. As per
data from S&P Dow Jones
Indices, Foreign sales account
for more than 40% of the total
S&P 500 turnover, with 261
companies in the index
tallying more than 15 per cent
of their revenues outside of
the United States. As a healthy
financial services sector is vital
to a functioning economy it is
little wonder that banks are
mandated to comply with
such a wide range of global
standards and frameworks,
including Basel III, which
focuses on market, credit and
operational risks.
01 | Predictive analytics - The silver bullet in efficient risk management for banks?
Is Risk Management Today Only About Managing Risk?
Financial
Internal
External
Non-Financial
Credit
Liquidity
Market Compliance
Operational
Strategic
Reputational
Systemic
Exhibit 1: Major risks in banking sector
Foreign sales account for
more than 40% of the total
S&P 500 turnover, with 261
companies in the index
tallying more than 15 per
cent of their revenues outside
of the United States.
“
”
Globalised economies enable
companies in one country to
tap into markets in other
countries, thereby reducing
their exposure to local factors
and reducing point failures.
“
”
3. The cost of compliance
continues to rise. In 2013
HSBC reported that it was
going to more than double
the number of people in
compliance to 5,000 – a figure
which has now increased to
7,000. In 2014 Deutsche Bank
reported EUR1.3b in extra
regulatory-related spending
of which 400m was related to
additional staff. In 2015
Citigroup reported that about
half of the bank’s $3.4b
efficiency savings were being
‘consumed by additional
investments’ in regulatory
and compliance activities. So
what are the banks getting for
all this additional investment?
Exhibit 2 highlights some of
the insights from a survey
conducted by Ernst & Young
in 2014 that helps banks see
where they stand in
managing their risks.
With increased focus on
risk-return trade-offs, risk
management in banks has
changed from a compliance
driven role to a business
strategy defining function. In
a recent study grading
companies on efficient risk
management, the top 20
percent organizations were
found to perform three times
better on earnings before
interest, taxes, depreciation
and amortization (EBITDA)
than the bottom 20%. So how
can financial institutions
make their risk management
practice more efficient? The
whitepaper aims to highlight
the key aspects of traditional
enterprise risk management
and how the use of analytics,
can improve the effectiveness
of any risk management
program by enhancing credit
quality, improving credit
decisioning and enabling a
360 degree view of customer.
Exhibit 2: Insights from risk survey conducted by EY in 2014
57% of banks
have increased
the size of the
risk function.
Expected
increase is 53% in
the next year
58% of the banks
say that progress
is made on firm
level in setting up
of risk appetite
More than half of
banks' board
focusses on risks
such as credit,
stress testing,
ALM risk and
operational
risk.Risk appetite
is rated the
highest (80%)
Only 31% report
that risk appetite
is fully imbibed
into the business
25% have
reported losses
more than $M
500 due to
operational risks
75% have
increased their IT
spend
71% have
introduced new
stress tesing
methods
35% have
incorporated the
stress testing
results in their
business decision
making
There is an
increased focus
on risk manage-
ment post crisis.
56% of survey
respondents
indicated
heightened focus
Credit risk is the
top focus for the
CROs closely
followed by
regulatory and
operational risks
Board's View Risk Integration Focus on Risk Risk-borne costs Stress Testing
02 | Predictive analytics - The silver bullet in efficient risk management for banks?
4. Defining a multi-dimensional
Enterprise Risk Management
(ERM) framework is the
cornerstone for effective risk
management. The Committee
of Sponsoring Organizations
of the Treadway Commission
(COSO) established an
integrated framework (Exhibit
3) to help banks derive
business value while meeting
compliance requirements.
In alignment with the
framework, it is imperative
that banks focus on the key
issues that form the crux for
ERM. (Exhibit 4)
The Cornerstone for Effective Risk Management
ERM Risk
Culture
Risk
Appetite
Stress testing &
Capital Management
Assessment
& Reporting
Exhibit 4: Key issues of ERM
Internal Environment
Objective Setting
Event Identification
Risk Assessment
Risk Responses
Control Activities
Information &
Communication
Monitoring
Inter-related components
of Framework
Organizational Level
Objectives
Entry Level
Division
Business Unit
Subsidiary
Compliance
Reporting
Operations
Strategic
Exhibit 3: ERM Framework
03 | Predictive analytics - The silver bullet in efficient risk management for banks?
5. is the set of norms and
traditions that govern the
behaviour of the individuals
and groups of an entity to
determine how risks are
identified, understood and
responded to. It is about being
aware of ethics, best practices
and the risk appetite of the
organization. In the EY report,
“Shifting focus: Risk Culture at
the forefront of the banking”,
61% of the banks have aligned
their risk appetite by changing
their risk culture while 74% of
them termed enhanced
communication of risk values
to be one of the top initiatives
to strengthen the risk culture.
Risk Culture
is the amount of risk that the
firm is willing to accept in
pursuit of its goals and
objectives. It is determined by
the kind of risks the bank will
take or accept in differing
contexts. Further, risk appetite
statements with top-down or
bottom-up collaboration and
defined metrics are crucial for
embedding the risk appetite
throughout the organization.
They would help in monitoring
the performance of the
business groups or portfolios.
Although stress testing is a
regulatory mandate for capital
planning, it can also assist the
bank’s top management in
assessing the business
model‘s sustainability towards
market volatility and as a tool
for the strategic decision
making. There is a growing
necessity to refine stress
testing to improve balance
sheet and P&L forecasting
under different scenarios.
entralized testing models are
the need of the hour with the
integration of bank’s risk and
finance functions.
Risk appetite
Stress Testing and
Capital Management
This lies at the heart of the
risk management framework
that helps banks align their
business objectives with the
risk appetite or what experts
term as “embedding the risk”.
There are
business-intelligence tools
that provide insights into the
risk profile of the banks
Regulatory mandates like
Basel ensure that banks are
aware of and deal with the
conventional risks. However,
in order to have a holistic
view of the bank’s risk, some
of the non-financial risks like
reputational risk should also
be considered. In addition, the
methodologies and the
approaches adopted by the
banks should neither succumb
to the regulatory pressure nor
should they overly rely on
backward looking models.
Forward looking approaches
by considering varied
scenarios help banks in being
prepared for contingencies. It
provides a total understanding
of the top risk drivers and
throws light at the root causes
and early warning signals.
Risk Assessment &
Reporting
04 | Predictive analytics - The silver bullet in efficient risk management for banks?
Regulatory mandates like
Basel ensure that banks are
aware of and deal with the
conventional risks. However,
in order to have a holistic view
of the bank’s risk, some of the
non-financial risks like
reputational risk should also
be considered.
“
”
6. With the exponential growth
and availability of data, both
structured and unstructured,
big data comes into the
picture and can be combined
with historical transactional
data to uncover new
opportunities. CROs across
the globe are looking to use
structured and unstructured
data to make accurate risk
predictions along with
understanding the potential
impact of a range of risks.
They are also looking at linking
them better to the
organization’s strategy.
Currently, there are several
challenges impeding the banks
from applying ERM effectively.
For instance, extracting and
aggregating data continues to
be the top challenge in
improving stress testing.
Credible risks quite often go
unnoticed.
The intrinsic challenges in risk
management necessitate a
more cohesive ERM solution-
something can be made
possible with the usage of risk
analytics. While analytics
previously was synonymous
with business intelligence,
today the level of
sophistication has increased
with more focus on data
exploration, segmentation,
statistical clustering,
predictive modeling and event
simulation & scenario analysis
leading to better insights.
By embedding predictive
analytics into the ERM delivery
approach, organizations can
monitor performance through
risk sensitivity analysis, model
key risk events scenarios, and
become more risk intelligent in
developing intervention and
mitigation strategies. It helps
the bank chart the best course
of action for the future. Pricing
decisions can be made by the
use of analytics thereby giving
a deeper understanding of
risks. The bank can also use
analytics to fight against credit
risk and manage their
portfolios optimally. Exhibit 5
explains how analytics can add
value to the bank’s ERM
practices.
Evolution of Predictive Analytics
Build financial metrics
to examine the likely
risk scenarios
Measure risks by
establishing baseline
of data
Quantifying
Risks
Data exploration,
segmentation,
statistical clustering,
predictive modeling
and scenario analysis
Whats new?
Monitor performance
through risk sensitivity
analysis
Enable Risk
Intelligence to develop
mitigation strategies
Embedded to ERM
Enhances the accuracy
and quality of
forecasts
Improving reporting
mechanisms
Reporting
Using analytics to
gauge risk-adjusted
performance in line
with regulatory norms
Regulatory
Compliance
Exhibit 5: Value Delivered by Predictive Analytics to ERM
05 | Predictive analytics - The silver bullet in efficient risk management for banks?
7. With deteriorating credit
quality, addressing credit risk -
primarily due to default - has
become the top most priority
for the banks. This has
resulted in an increased focus
on internal stress testing over
the past 12 months.
Traditionally, banks rely
heavily on the credit bureau’s
score for making a loan
decision or, in the absence of a
credit bureau, on internal
scoring models. However,
scoring models from credit
bureaus and internal scoring
models are based on the
historical credit profile of the
borrower which may not
accurately reflect the current
situation and therefore might
not help the underwriter make
an informed decision. This
may lead to turning down
potential clients which reduces
profits and may damage the
bank’s reputation. On the
other hand, accepting
non-worthy businesses will
make matters worse by
creating future Non
Performing Loans.
In credit risk modelling,
scoring models are developed
using state-of-the-art
statistical techniques and data
aggregation from the bank’s
archives. Predictive
Analytics-based scorecards
allow the bank to rapidly
identify which loans should to
be approved, which loans
should be rejected and which
loans should be subject to
further investigation.
In Exhibit 6, the model score is
the internal score computed
by the bank using predictive
analytics. The decision
process for loan approval or
rejection becomes more
robust by devising a decision
map using both the model
score and the score from the
credit bureau.
Consider a customer who has
a medium Credit Bureau score
as well as a medium risk
model score. His case, by
default, falls into the Refer/on
hold bucket of the business
risk strategy map created
using statistical scoring
models. In such a case, the
underwriter usually sends the
application for further field
investigation leading to
increase in time and costs. In
the meantime the customer
may decide to take loan from
some other bank and thereby
the first bank loses a potential
good customer.
Improving Credit
Decisioning
Enabling a 360o
View
of Customer
Enhancing Credit
Quality
Driving Effective Risk Management in Financial
Organizations
Decision
Map
Bureau Score
Model
Scrore
Low Score Medium Score High Score
Very High Risk
High Risk
Medium Risk
Low
Reject
Reject
Reject
Reject
Refer
Refer
Refer
Approve Approve
ApproveReject
Reject
Exhibit 6: Decision map for auto-approval of loans
through predictive analytics
06 | Predictive analytics - The silver bullet in efficient risk management for banks?
8. By combining big data and
high-powered analytics, it is
possible to:
By leveraging big data in the
underwriter decision making
stage, the decisions for
refer/on hold applications can
be made after analysing the
current behavioural and risk
patterns of the customer. The
amount of investigations for
on hold applications is
reduced thus bringing down
the time and costs involved
and freeing up people to
focus on more important
activities. In addition,
fraudulent customers can be
detected easily as well.
Social media has changed the
way people interact and firms
across the globe are trying to
leverage social data in their
efforts to stay ahead of
competition. Social Network
Analysis (SNA) (Exhibit 7)
includes pattern analysis and
network linkage analysis to
uncover the large amount of
data that can be linked to
show relationships. To gain
customer insights, one looks
for clusters and how those
clusters are linked with the
other clusters. Public records
such as social media
behaviour, address change
Create a unified view of the
customer covering all
his/her touch points
including web crawling data,
call centre interactions,
social media activities,
branch interactions etc.
(Exhibit 7)
Recalculate entire risk
portfolios in minutes
Quickly identify valuable
customers
Detect fraudulent behaviour
using clickstream analysis
and text mining
Social
Network
Analysis
Facebook
Twitter
Linkedin
Email
Customer
Feedback
Transactional
Systems
Call Records
Exhibit 7: Extracting data from various sources
for network analysis modelling
07 | Predictive analytics - The silver bullet in efficient risk management for banks?
The Rise of Social: More
Data More Insights
Predictive Analytics-based
scorecards allow the bank to
rapidly identify which loans
should to be approved, which
loans should be rejected and
which loans should be
subject to further
investigation.
“
”
9. frequency, criminal records
and foreclosures are all data
sources that can be integrated
into the model.
This will generate many
insights at the time of
underwriting and therefore
the credit decision process
can be enhanced
substantially. By integrating
this with transactional
systems, even fraud risks can
be mitigated in real time.
Exhibit 8 shows the
mechanism of risk modelling
with SNA.
While some banks have begun
to see real benefits of these
enormous data sources, many
are still working in isolated
silos. Others, while having a
multidimensional and
ntegrated ERM framework,
are still not utilizing predictive
analytics at the optimal level.
With the exponential growth
and availability of data, banks
can gain a strategic advantage
by using predictive analytics
to make improved risk
predictions that are better
aligned to current and future
economic conditions, and
hence quickly adjust to
dynamic market conditions
and steer their portfolios
through uncertain times.
Nucleus Lending Analytics is
designed to provide
comprehensive business
insight into credit
risk-management of banks
and other financial
institutions. The solution uses
sophisticated credit scoring
models to allow credit risk
managers and credit analysts
create predictive scorecards. It
also incorporates defined
metrics that provide a unified
view of customers across lines
of businesses and channels.
The solution focuses on the
three key tenets of efficient
risk management in lending:
Informed Decisioning,
Enhanced Portfolio
Management and Fraud
Prevention.
08 | Predictive analytics - The silver bullet in efficient risk management for banks?
How can Nucleus Help?
Data
Aggregation
Data
Pre-
processing
Social
Network
Analysis
Pattern
Analysis
Exploratory
data
Analysis
Business
Rules & Alert
Generation
Exhibit 8: Risk Modelling using SNA
With the exponential growth
and availability of data,
banks can gain a strategic
advantage by using
predictive analytics to make
improved risk predictions
that are better aligned to
current and future economic
conditions
“
”
10. 09 | Predictive analytics - The silver bullet in efficient risk management for banks?
About the Author
Arup Das
Business Unit and Global Product Head (P&L Management) for Lending,
Nucleus Software
Arup, a Wharton MBA, is the Business Unit Head and Global Product Head
(P&L Management) for Lending at Nucleus Software where he is responsible
to lead the flagship product to the next level of global leadership. Before
joining Nucleus, he has played various roles in strategy and product
management with leading companies like CISCO, IPValue and Mphasis.
Author e-mail id: arup.das@nucleussoftware.com
Shivendu Shekhar Mishra
Lending Product Manager, Nucleus Software
As Product Manager, Shivendu’s major focus is to understand latest business
needs and opportunities in Lending market. Before joining Nucleus, Shivendu
worked in Product Development & Management, Business Consulting roles
with companies like Infosys, CSC and ZS Associates.
Author e-mail id: shivendu.mishra@nucleussoftware.com
Pavan Elchuri
Lending Product Analyst, Nucleus Software
As a lending product analyst at Nucleus Software, Pavan’s focus is on the
flagship lending product FinnOne Neo and Analytics. He has done his Masters
from IIM Lucknow and Bachelors in Engineering Physics from IIT Delhi.
Author e-mail id: pavan.elchuri@nucleussoftware.com
Atish Jain
Business Analyst, Nucleus Software
Atish is responsible for the implementation of advanced machine learning
algorithms in the Nucleus Lending Analytics suite . He has done his Bachelors
in Electronics and Communication Engineering from NIT Silchar.
Author e-mail id: atish.jain@nucleussoftware.com