Data and Analytics

Innovations in
Credit Portfolio
Management
Presentation document
December 2022

CONFIDENTIAL AND PROPRIETARY

Any use of this material without specific permission of McKinsey & Company
is strictly prohibited
McKinsey and Scan range of industry practices on:
IACPM together  Emergence of alternative data sources for credit risk
completed a survey identification, assessment, and monitoring
on new  Related evolution of analytics tools
developments in
data and analytics Understand degree to which different data types and
for credit portfolio analytical approaches are in use/under consideration
management

The survey had 3 Develop insights on current state and path forward
main objectives for participants to incorporate next generation data
and analytics

McKinsey & Company 2

CONFIDENTIAL AND PROPRIETARY
Any use of this material without specific permission of McKinsey & Company is strictly prohibited
44 financial institutions participated in the survey from across
Americas, APAC, and EMEA

Americas Asia Pacific

Europe Middle East and Africa

Source: McKinsey/IACPM Survey on data and analytics innovations in Credit Portfolio Management – October 2021 McKinsey & Company 3
CONFIDENTIAL AND PROPRIETARY
Any use of this material without specific permission of McKinsey & Company is strictly prohibited
Agenda

Summary of survey results

Perspectives on selected topics

McKinsey & Company 4

CONFIDENTIAL AND PROPRIETARY
Any use of this material without specific permission of McKinsey & Company is strictly prohibited
Summary of results

A Trends Majority of participants expect significant increase in use of internally developed advanced techniques and new
types of data

B Challenges Data quality and talent management are top challenges for both advanced analytics and data solutions

C Climate Majority of participants believe that Climate and ESG are next big challenge for credit assessment

D Investment and
Strategic Goals
Top investment areas have been data tech and data acquisition. Participants expect a greater role of innovative
data and advanced analytics in improving credit strategy and customer experience

E Use cases Machine learning models are primarily gaining traction for risk scoring of SMEs and early warning across
the board
Innovative external data sources are more used for corporate segment while SME segment uses more
innovative internal data sources

F Impact Use of innovative data and/or advanced analytics improves model accuracy , turn-around-time, automated
decisioning and time spent on analysis, with higher benefit observed in SME segment

Source: McKinsey/IACPM Survey on data and analytics innovations in Credit Portfolio Management – October 2021 McKinsey & Company
CONFIDENTIAL AND PROPRIETARY
Any use of this material without specific permission of McKinsey & Company is strictly prohibited
A: Over next 2 years, majority of participants
expect significant increase in use of new types
of data

How has your firm’s data/analytics for credit decisions changed in the past 2 years and how In the past 2 years, over 60% of
do you expect it to change in the coming 2 years? the participants have seen an
% participants see increase in trends increase in the:

In the past 2 years Expectation for the next 2 years  Use of new types of internal
and external data
Increased use of new types
of internal data
74% 91%  Use of internally developed
advanced techniques
Increased use of internally developed
machine learning and other 72% 81%  Size of data and analytics
advanced analytics techniques team
Increased use of new types
66% 79% Over the next 2 years, even
external data
larger % of participants expect
Change in the size of the this trend to continue
64% 81%
data & analytics team

Change in share of automated

49% 66%
credit approvals
Increased use of vendor-developed
machine learning and other 32% 59%
advanced analytics techniques

Source: McKinsey/IACPM Survey on data and analytics innovations in Credit Portfolio Management – October 2021 McKinsey & Company 6
CONFIDENTIAL AND PROPRIETARY
Any use of this material without specific permission of McKinsey & Company is strictly prohibited
B: Data quality and talent management are top
challenges for use of both advanced analytics
and innovative data solutions

Currently, what are the major challenges faced by your firm that constrain the use of Major challenges for use
innovative data or advanced analytics (e.g., machine learning and AI)? of advanced analytics
solutions are:
Top 3 challenges for use of Top 3 challenges for use of
Percentage innovative data solutions advanced analytics solutions  Attract, retain and develop
resources
Data Quality 63 42
 Ability to explain
Resources 42 49
Cost of data 30 14  Data quality
Skepticism/Probability 28 26  Validation
Regulatory 28 26
While for using innovative data,
Validation difficulty 19 40 key challenge in data quality
assessment and talent
Unmet expectations 14 9
Ability to explain 14 42
Risk concerns 12 7
Other 2 2
Fragmentation 2 5

Source: McKinsey/IACPM Survey on data and analytics innovations in Credit Portfolio Management – October 2021 McKinsey & Company 7
CONFIDENTIAL AND PROPRIETARY
Any use of this material without specific permission of McKinsey & Company is strictly prohibited
C: 86% of participants believe Climate and
ESG are next big challenges for credit
portfolio management

What are the biggest challenges facing credit risk For incorporating the impact of climate risk ,are you One third of the participants
and credit portfolio management analytics in the using existing loss models with climate shock plan to use existing credit
next 2-3 years? applied to input variables? Or are you developing
models to translate the climate
new loss models to assess it?
impact to credit risk and
Percentage Percentage
another one third of the
participants plan to develop
Climate and ESG 86 new loss models for climate
assessment
Capital, Provisioning, or regulatory Not yet
58 Using existing
stress testing model requirements exploring
33 models
35
Post COVID-19 model adjustments
51
and uncertainties

Incorporating Machine Learning

models within regulatory and 42
risk constraints

Effectively competing with attackers 9

33
Developing new
loss models
Other – please specify 7

Source: McKinsey/IACPM Survey on data and analytics innovations in Credit Portfolio Management – October 2021 McKinsey & Company 8
CONFIDENTIAL AND PROPRIETARY
Any use of this material without specific permission of McKinsey & Company is strictly prohibited
C: > 50% of participants have implemented/or are planning to
implement climate stress loss analyses in non-SME segments

Have you implemented or are planning to implement in the next 12 months any changes to the credit assessment/
adjudication and monitoring models to capture the impact of climate change? (transition and physical risks)
Percentage of participants
SME Mid-Market Corporate CRE
N=33 N=29 N=41 N=35

No 36% 34% 20% 27%

Yes - Climate stress loss scenarios 39% 59% 54% 52%

Yes - Adjustments to obligor

12% 21% 22% 21%
rating methodology

Yes - Obligor climate risk scorecard 21% 28% 39% 39%

Yes - Other (1) 18% 21% 17% 15%

1: E.g., Adjustments to obligor rating methodology and climate stress loss scenarios, but beyond 12 months. Bucketing of risks (geography, industry, property
type segments).

Source: McKinsey/IACPM Survey on data and analytics innovations in Credit Portfolio Management – October 2021 McKinsey & Company 9
CONFIDENTIAL AND PROPRIETARY
Any use of this material without specific permission of McKinsey & Company is strictly prohibited
D: In last 2 years, top investment areas have
been data tech and data acquisition – this trend
is expected to continue
Largest 2nd largest 3rd largest

Where have you made the most investments in the past 2 years and where do you expect to In the past 2 years, the top
invest the most in the coming 2 years? investment areas for
participants were data tech and
data acquisition
Number of votes In the past 2 years Expectation for the next 2 years
And this trend in expected to
Data tech continue over the next 2 years
13 6 6 25 15 9 5 29
(e.g. cloud, visualization)
with higher expected
investment
Data acquisitions 5 10 7 22 7 6 7 20
Other top investment areas
Data scientists and engineers include talent for both
8 4 10 22 4 6 11 21
for development and validation development/ validation and
data processing
Data scientists and engineers
2 9 9 20 2 8 10 20
for data quality processes

3rd Party/Vendor models 6 5 0 11 5 5 1 11

Other, please specify 1 1

1 1
Source: McKinsey/IACPM Survey on data and analytics innovations in Credit Portfolio Management – October 2021 McKinsey & Company 10
CONFIDENTIAL AND PROPRIETARY
Any use of this material without specific permission of McKinsey & Company is strictly prohibited
E: Machine learning models are primarily
gaining traction for Risk Scoring of SMEs and
Early Warning across the board
Innovative data sources

What methodologies are being “used in production”, “validated” or “in pilot” for each of the listed use cases Expert based and statistical
for the Corporate and SME portfolio models are most widely applied
Other machine
Statistical Simulation-based Machine learning learning approaches across the
approaches (e.g., models (e.g., tree algorithms models (e.g. spectrum of use cases
Expert based linear or logistic monte-carlo (e.g. random support vector,
qualitative regression or approaches to forest, gradient NLP, neural
Percentage models CHAID) economic capital) boosting) network) Simulation based models
Corpo- Risk score/ Rating/ are more widely used for
44% 68% 10% 2% 0%
rate PD model stress testing
Early warning indicators 39% 37% 5% 12% 5%

Credit pricing 44% 24% 5% 0% 0% Machine learning models are

getting traction for Risk Scoring,
IFRS 9/ CECL 39% 56% 7% 0% 0%
Early Warning, and Pricing
Stress Testing 32% 46% 29% 0% 0%

SME Risk score/ Rating/

41% 72% 6% 13% 3%
PD model
Early warning indicators 44% 56% 9% 16% 3%

Credit pricing 47% 38% 3% 6% 0%

IFRS 9/ CECL 44% 66% 6% 3% 0%

Stress Testing 38% 53% 28% 0% 0%

Source: McKinsey/IACPM Survey on data and analytics innovations in Credit Portfolio Management – October 2021 McKinsey & Company 11
CONFIDENTIAL AND PROPRIETARY
Any use of this material without specific permission of McKinsey & Company is strictly prohibited
E: Innovative external data sources are more
used for Corporate segment while SME segment
uses more innovative internal data sources
Innovative data sources

Which of the following categories of data are being used in production, under pilot or under External data sources:
consideration for credit risk management use cases within the Corporate portfolio? For Corporates, over 50% of the
participants are using, piloting or
Percentage Corporate SME considering New media or social
Credit bureau and rating agency data 97% 91% media and 3rd Party account data, a
External
higher proportion than for SMEs
data
Economic and market forecasts 88% 63% Both segments use E-commerce
sources
Financial market data on issuers 76% 35% data at similar rates

News media or social media 59% 32%

Internal data sources:
3rd Party account data 53% 30%
For Corporates, over 70% of the
E-commerce data 33% 29% participants are using, piloting or
considering Automated client/ issuer
Other high frequency data 32% 19%
financials and internal credit behavior
Internal Automated client / issuer financials 88% 83% data, while for the SME segment in
data addition to above 2, internal cross
sources Internal credit behavior data 76% 91% product data also has large share
Internal cross product behavior 62% 76%
Internal client interaction data 56% 45%
Internal consumer-wholesale
44% 38%
crossover data

Source: McKinsey/IACPM Survey on data and analytics innovations in Credit Portfolio Management – October 2021 McKinsey & Company 12
CONFIDENTIAL AND PROPRIETARY
Any use of this material without specific permission of McKinsey & Company is strictly prohibited
F: Automated decisions are largely a feature of
SME portfolios
None and Not implemented Less than 10% 11-30% 31-50% Above 50%

In the past 3-5 years, what was the typical percentage of automated decisions Key insights
based on models in your portfolio?
Fully automated decisions for a
Percentage of participants, where applicable material portion of the portfolio
is almost exclusively a feature
of SME portfolios
18%

However, there are pockets of

34% 72% portfolio with full automation,
87% even for mid-market and others
89%

18%

19% 16%

11% 8% 13% 11%

4%
SME Mid-market Corporate CRE
N= 27 25 31 27

Source: McKinsey/IACPM Survey on data and analytics innovations in Credit Portfolio Management – October 2021 McKinsey & Company 13
CONFIDENTIAL AND PROPRIETARY
Any use of this material without specific permission of McKinsey & Company is strictly prohibited
F: Where implemented, use of innovative data
and/or advanced analytics typically achieved
increased automation
% of participants Implemented but no uplift in automation Increased by 11-50%
XX%
implemented Increased by up to 10% Increased by more than 50%

In terms of increasing automation, what benefit have you seen in the past 3-5 63% participants reported an
years from the use of innovative data and/or advanced analytics? increase in automation for the
SME segment, followed by Mid-
Percentage of participants that implemented
Market (27%), Corporate (19%)
70% 36% 26% 27% and CRE segment (14%)
80
Where implemented, use of
70
70 innovative data and/or
7 advanced analytics typically
60 improved automation by up
to 10%
50
37
40 36 Higher improvement (11-50%)
in automation is typically
30 9 27 observed in SME segment,
26
7 which involves the highest level
20 14 of direct automated decisioning
22 23
11
10 5 9 5
4 0 7 0 0
0
SME Mid-market Corporate CRE

Source: McKinsey/IACPM Survey on data and analytics innovations in Credit Portfolio Management – October 2021 McKinsey & Company 14
CONFIDENTIAL AND PROPRIETARY
Any use of this material without specific permission of McKinsey & Company is strictly prohibited
F: Where implemented, use of innovative data
and/or advanced analytics significantly
improved turn-around-time “TAT” for SMEs
% of participants Implemented but no reduction in approval time Decreased by 11-50%
XX%
implemented Decreased by upto 10% Decreased by more than 50%

In terms of accelerating TAT to decision, what benefit have you seen in the past 3- 56% participants reported
5 years from the use of innovative data and/or advanced analytics? decrease in TAT for the SME
segment, followed by Mid-
Percentage of participants that implemented
Market (43%), CRE (32%), and
63% 52% 41% 44% Corporate (31%)
70
63 Where implemented, use of
60 innovative data and/or
7
52 advanced analytics typically
50 improved turn-around-time
19 9 44 “TAT” by up to 10%
41
40 12
10 Higher improvement (11-50%)
30 in TAT is typically observed in
30
SME segment but also to some
30 20 extent in Mid-Market and
20
28 CRE segment
10 3
13 12
7 0 0 0
0
SME Mid-market Corporate CRE

Source: McKinsey/IACPM Survey on data and analytics innovations in Credit Portfolio Management – October 2021 McKinsey & Company 15
CONFIDENTIAL AND PROPRIETARY
Any use of this material without specific permission of McKinsey & Company is strictly prohibited
Agenda

Summary of survey results

Perspectives on selected topics

McKinsey & Company 16

CONFIDENTIAL AND PROPRIETARY
Any use of this material without specific permission of McKinsey & Company is strictly prohibited
Perspectives on specific topics : Climate Risk, Next generation data
and analytics, and Scenario planning and forecasting

Next generation data and Scenario Planning and

Climate Risk analytics Forecasting

Setting up a modular, portfolio specific Using cross-product data with help Building a flexible infrastructure to
methodology for scenario analyses is from AI/ML can drive both revenue forecast and optimize portfolio is more
critical, prioritization will depend on growth and automated credit decision critical than ever
exposure to high-risk sectors
For larger obligors, availability of Rather than waiting for a full-scale
Climate risk impact on portfolio analytics and accessible data is key solution, banks would want to
requires inter-disciplinary skills and for turnaround time reduction, establish analytics and organizational
mobilization across credit, front-line, however, doing so requires treating capabilities that enable rapid ‘what-if’
and model risk management data quality as more than a ‘regulator- analyses
required’ effort

McKinsey & Company 17

CONFIDENTIAL AND PROPRIETARY
Any use of this material without specific permission of McKinsey & Company is strictly prohibited
 Climate Risk

Key learnings on the materiality of Climate Risk on Credit,

based on our extensive work with banks

Risk is concentrated in  Both physical and transition risk lie in very targeted areas of the portfolio: for
“pockets” across the example, for a large global bank we identified that approximately ~15% of their loan book
was materially exposed to climate risk
portfolio; banks need to take
a targeted approach  Banks need to perform heatmapping to focus their efforts on the high risk portfolios
and risks: even within a CRE portfolio for a large US bank, we found that majority of the
credit impact came from 10% of the portfolio

The “average” impact is  We found that even for high risk industries, the average impact is moderate: For
moderate in the near-term, example, in a portfolio of upstream O&G companies, the impact by 2025 under below 2C
scenario was ~7% median impact on EBITDA
but there is high degree of
counterparty-level variability  However, the difference between winners and losers is stark: in the upstream O&G
example, we saw several counterparties with up to ~40% impact on EBITDA, while there
were other companies that saw a positive EBITDA impact

Most of the risk is in knock-  Direct damages are immaterial on credit; Knock-on effects can dwarf direct
on impacts that most banks impacts, e.g., 4.5x the impact of direct 1st order impact for a Muni flood example
do not model  Material risk drivers include community deterioration, geographic transition risk, and
broader ecosystem impacts (e.g., insurance cost and availability)
 Real estate losses are driven by asset pricing (property values and cap rates), not
physical damage
​Impact on Economic profit 1, CAD mn

Climate is not a “capital”  A ‘CCAR mindset’ of focusing on capital risks will underestimate the business ​~1100 ​-35%

problem; however, it can value risk and miss the opportunity to steer the business ​~700

have real impact on returns /  For a North American bank, we identified that 35% of economic profits could erode

​Manufacturing

​Transportation
​Current SVA

​Oil and Gas

​Commercial
Real Estate
(estimated)

​Residential
​Agriculture
by 2030 without taking action on key pockets of climate risk exposures

2030 SVA
​Expected
economic profit

​Utilities
​Mining
1. Calculated as the return on invested equity above the return on equity (estimated at 10.2% based on McKinsey CPAnalytics) times invested equity
(allocation based on exposure). Assumes that percentage NII decline = percentage net income decline

McKinsey & Company 18

CONFIDENTIAL AND PROPRIETARY
Any use of this material without specific permission of McKinsey & Company is strictly prohibited
 Climate Risk

Scenario analyses for corporate and mid-market obligors

needs to be sector-specific: Oil and gas example

Methodology isolates individual aspects of scenario analysis through modules that calculate Oil and Gas obligor impacts from:
 Oil and Gas demand changes across Upstream and Downstream Operations
 Clean Technology demand changes
 Carbon Costs
 Acute Physical Hazard damage costs

Scenario selection (e.g. NGFS Scenarios)

Scenario expansion and country downscaling (e.g., damage curve, transition pathways)

Module Specific Data

i
Upstream Oil &
Obligor Financials Gas
ii
Obligor Emissions Supply Chain Oil
Transition risk & Gas vii
impact Climate-stress
iii Integration
Oil and Gas Asset- Cleantech obligor
Level Data module
financials
iv
Clean Tech Market Carbon cost vi
Shares Competition Stage 2 profit
Physical risk v module revenue, cost
Abatement Cost Curves impact Acute impact

McKinsey & Company 19

CONFIDENTIAL AND PROPRIETARY
Any use of this material without specific permission of McKinsey & Company is strictly prohibited
Sanitized example: Demand destruction,
Climate Risk

direct carbon cost, and market impacts for a

specific obligor
Assuming no management/bank action

Delayed transition scenario – 2050

Valuation impact waterfall – average impacts weighted by exposure Takeaway
100% Demand destruction, direct
7% carbon cost and market
3% impacts are the top 3 drivers
40% of the climate impact to the
O&G portfolio

0%  Demand destruction is
41% related to the business
50%
composition and
30% breakeven cost
5%
 Direct carbon cost is related
to the carbon price and the
Current Physical Adaptation Demand Demand Direct Abatement Market Full impact amount of output that can
valuation impact destruction creation carbon costs impacts be sustained
Physical risk impact Revenue impact Carbon cost impact  Market impact is related to
Driven by current Driven by carbon emissions which the ability to pass cost to
business composition, drives carbon costs, abatement and customers and gain market
breakeven cost, and oil ability to pass cost to customers/ share from other players
and gas price market impacts

McKinsey & Company 20

CONFIDENTIAL AND PROPRIETARY
Any use of this material without specific permission of McKinsey & Company is strictly prohibited
 Next generation data analytics

Deposits data can be used to create an up-to-the-minute

estimate of SMB’s financials

Context Risk signals extracted from transactions Strategic implications

Deposits transactions can be SME financial Estimated revenues and profits Competition to become
analyzed using a transaction position the primary bank
Business seasonality
classifier to derive an estimate of will intensify
 SMB’s financials (e.g., revenues, Sectoral dependencies
In turn, banks that host
revenue growth, and profits) the deposit account
Using a reinforcement learning and Competitive Multi-banking clients where the payroll is
natural language processing, information deposited have a
Loans with other lenders competitive advantage
deposits transactions are analyzed to
provide structure to unstructured data Interest rate changes by other Deposit accounts are
lenders already increasingly
In a client application, this method
yield 95% transaction classification Fintech relationships being bundled with other
accuracy and over 300 risk signals (e.g. Paypal) financial products
and competitive indicators, resulting Fees for deposit
in substantial improvement in accounts are expected to
predictive power of credit risk models reduce further

McKinsey & Company 21

CONFIDENTIAL AND PROPRIETARY
Any use of this material without specific permission of McKinsey & Company is strictly prohibited
 Next generation data analytics

Case example: Leveraging transaction data to transform

credit decisioning
Illustrative; Client example

Traditional lending process Reimagined credit decisioning More predictive power

40%
Customer completes credit Customer completes online application additional predictive
application form (pre-populated, leveraging API enabled power in some
data sources) Up to subsegments

Customer provides bank account and ETB: Transactional data from own More customer insights
provides simple historic financial system, used to build synthetic
statements financial statements

95%
Accuracy of financial
inflows and outflows
NTB: Customer provides permission1
from classifier
to access bank transactions data to
build synthetic financial statements

RM manually reviews data and Algorithm integrates data and performs

More sophisticated understanding
performs credit assessment (scoring credit assessment (e.g., leveraging of risks
model, uses credit bureau, and transaction classier to build financial
historic/ narrow financial data) statements, link to risk signals)

>300
Predictive risk signals
– linked to
Manual decision communicated to Decision communicated to customer transaction
customer (instant time yes) classifications

McKinsey & Company 22

CONFIDENTIAL AND PROPRIETARY
Any use of this material without specific permission of McKinsey & Company is strictly prohibited
An AI-driven asset deployment driven approach
AI driven data quality

can accelerate data quality improvement

significantly

1 2 3
AI toolkit for data Deployment accelerators Training modules
Detection: Ready to implement Deployable as pipelines that can be A new way of working, including
package to assess data quality stand-alone for immediate results roles, talent, an a fast-paced Agile
Correction: Relationship discovery and integrated into Data Platforms operating model When implemented
and anomaly detection to find errors to continuously monitor and improve Co-development of solutions
data quality (e.g., Apache Airflow through build-operate-transfer to
together these 3
Repair: AI driven correction through
an open-loop process integration with Collibra), platform sustain the impact components significantly
agnostic deployment
accelerate data quality
capabilities

AI4D
Q
McKinsey & Company 23
CONFIDENTIAL AND PROPRIETARY
Any use of this material without specific permission of McKinsey & Company is strictly prohibited
 AI driven data quality

Data Quality can be addressed through modularized tools in

three core areas

Detection Correction Repair

Objectives Measure the quality of each record on a Recommend corrections to data quality Validate recommendations with experts
scale of Low, Medium, High data quality to errors prioritized by business impact and and incorporate changes into underlying data
help prioritize remediation measure the confidence (e.g., 95%) systems feeding reports

How machine Automatic generation of reports on the profile Rule-mining and clustering algorithms Validate corrections with experts (open-loop)
learning helps of data, inferred relationships between tables recommend corrections to data quality errors, confirming only lower confidence
and anomalies, root cause identification to quantify confidence, and help estimate recommendations manually and automated
prioritize upstream interventions business impact approval of validated changes

Illustration Detect top 1th percentile interest rate, Correct loan type with attribution, country of Flag and automatically fix reporting dates that
unusually late maturity date as a origination, and address with 95% confidence were corrected by experts repeated
potential error

Ready-to- Data relationship discovery Automated DQ rule generation Collaborative workflows

deploy tools
Entity disconnect identification Free-form text corrections Programmatic data validation

Root cause identification Comparison to 3rd party data

Corrections that integrate attribute,

Anomality detection
record, and database signals

McKinsey & Company 24

CONFIDENTIAL AND PROPRIETARY
Any use of this material without specific permission of McKinsey & Company is strictly prohibited
 AI driven data quality

High impact examples AI being used to improve data quality

Root cause detection Accelerating an Automated identification

and correction to errors enterprise data of errors in CRE loan
in 5M customer accounts transformation with AI data
Corrected 5M free-form Automated data quality detection Automated detection of data
occupations to accelerate AML and correction with AI (human-in- quality errors in over 100,000
Customer Risk Rating using a the-loop) to reduce data loans using time series anomaly
neural language model and transformation timeline by detection to detect issues in real
traditional fuzzy matching 30-40 percent time and prioritize remediation

McKinsey & Company 25

CONFIDENTIAL AND PROPRIETARY
Any use of this material without specific permission of McKinsey & Company is strictly prohibited
Scenario planning and forecasting approach needs to be
Scenario Planning and Forecasting

tailored to bank’s footprint flexible to incorporate different

driving factors
Full estimates from business drivers
Scenarios from underlying reasons Business drivers from scenarios and optimize across BUs

Interest Income
B1 A3 A4

B2 A1 A2

B3 B4 B5

Credit Loss
US US

UK UK

Canada Canada

… …

Develop inflation scenarios based on set of Identify drivers that are likely to be inflation- Develop analytics to project the underlying
underlying reasons (e.g., supply chain sensitive for business portfolios / geographies, drivers and the business portfolio financials
bottleneck leading to ‘cost-push’ inflation) and and direction of impact – e.g., inflation may drive Based on projection results, synthesize
consider incorporation of important macro- up transaction volume in the short term; implications for strategic decision-making
linkages (e.g., currency fluctuation risk) however, inflation may also reduce demand in
the long term

McKinsey & Company 26

CONFIDENTIAL AND PROPRIETARY
Any use of this material without specific permission of McKinsey & Company is strictly prohibited
Uncertainty in economic path require a forecasting
Scenario Planning and Forecasting

approach that is nimble, flexible, and responsive to

emerging risks like inflation

Versatility  Flexibility to expand current internal scenarios

Scenario Generation and Refinement
with new variables
Customized scenarios for portfolio/segments  Ability to add macro and event driven overlays
 Can be macro-economic or event driven  Ability to refresh scenarios weekly vs. monthly
 Recent macro-trends (e.g., EPOP ratio) core to design

Coherence  Model and analytics developed for relevant

Business driver forecast through models drivers and separately for relevant geography
Volume, Revenue, and Expense Forecast Models tied to business  Models capture time trend, macroeconomic
drivers (e.g., line utilization) and scenario conditioned factors, seasonality and latent portfolio factors
in a transparent manner

Flexibility  Designed to capture, test and visualize

business actions under different macro-
Implementation Engine conditions
One-click solution for aggregation, reporting and visualization  Structured to generate report on portfolio and
sub-segment level with different ‘what-if’

McKinsey & Company 27

CONFIDENTIAL AND PROPRIETARY
Any use of this material without specific permission of McKinsey & Company is strictly prohibited
In addition to standard capital or liquidity constraints,
Scenario Planning and Forecasting

exploring emissions related constraints can help in

portfolio alignment

Preliminary list of constraints incorporated in the Approach to incorporate emissions related constraints
Carbon emission
approach (to be refined based on observations
during design phase)  Include carbon limit based on benchmark scenario
Capital
at North America with relevant downscaling
Product Industry BU limit
 Add constraint for net zero target of total portfolio
emissions
Capital
 Develop functionality to add sector specific targets
and connect with potential sector-specific carbon
Carbon intensity metric
Risk Weighted Assets limit

Capital consumption
Expected Loss
Example output by industry at a loan level Loans out Loans in

Energy Materials Telecom

Origination/balance growth Carbon consumption

Liquidity coverage ratio

Emissions
Economic return

McKinsey & Company 28

CONFIDENTIAL AND PROPRIETARY
Any use of this material without specific permission of McKinsey & Company is strictly prohibited