Best Practices in Loan Portfolio Analysis

Through a fun and engaging presentation attendees will learn best practices in Loan Portfolio Analytics, including a discussion of real life applications, current examiner points of emphasis, and ‘how to guides’ on performing credit risk, concentration risk, static pool, loan quality ratings, and more.

Optimization Most Valuable:

-When interest rates are on the move

-When considering a new product

-To promote strategic thinking

Best Practice #3-A commitment to solve the big questions with data

Best Practice #4-Create and report against a "Model Portfolio"

Best Practice #6

Use the right metric, presented in the right way

Best Practice #7

Use basic analytics for consistent portfolio reporting and monitoring

Best Practice #8

-Use advanced analytics for periodic reporting and strategic planning

Analytics does not have to be intimidating. Start small and work towards continual improvement. The right data and a little know-how will help you remain competitive and relevant so that you can meet your overall purpose of helping members fulfill their financial dreams.

How many pieces of candy were in the jar?

Results: How do you currently use Analytics?

Don't Use..........................................................10.0%

To Meet Regulatory Requirements.....30.0%

To Make Decisions.......................................55.0%

Other...................................................................5.0%

Results: What is the most difficult part of analyzing a portfolio?

Getting the right data...........................................37.5%

Knowing the right analysis to perform.......40.0%

Communicating the results...............................6.3%

Knowing what to do with the results..........15.0%

Other...............................................................................1.3%

What do you expect will be the focus of your next exam? (Shown in Order of Most Common Response Grouping)

-Ability to Manage Risk

-Indirect Lending

-Concentration Risk

-Fair Lending

-Credit/ Risk Management Processes

-Business Lending

-Interest Rate Risk

-Information Technology/Systems

What are some questions in your organization you wish data/analytics could help you answer?

(Shown in Order of Most Common Response Grouping)

-Marketing/Opportunities

-Loan Pricing

-Risk Indentification

-Product Mix

-Merger/Acquisition Activity

Best Practices

in loan portfolio analytics

Introduction

What? Why? How? Q&A!

The "Right" Data

The Wisdom of Crowds-

Guess how many pieces of candy are in the jar?

Top 5 Mistakes

What is analytics?

1. Granularity of Loan Type Codes

2. Not recording Sr. Lien Info

3. Not recording Original Value

4. Only recording "Current" Credit Score

5. Changing Loan Code to "Charge-off"

Data on Collateral

Data on Borrower

Data on Loan

Type

Description/Location

Value

How do you currently use analytics in managing your loan portfolio?

Credit Worthiness

Income/Debt

Assets/Liabilities

Employment

Essential Data

Structure

Dates

Amounts

Rates

Loan Id Number

Loan Type Indicator

Origination Date

Original Loan Amount

Unpaid Balance

Unpaid Balance - Senior Loan

Lien Position

Maturity Date

Days Delinquent

Charge-Off Amount

Charge-Off Date

Current Credit Limit

Original Credit Limit

Interest Rate

Loan Term

Original Collateral Value

Collateral Address

Original Credit Score

Updated Credit Score

Updated Score Date

What is the most difficult part of analyzing a portfolio?

Key Points:

-Not all data is equally informative

-There are diminishing returns to accuracy

-Know what data you need, why, and how it will be used

-Would the data gap still exist for new loans?

Using Analytics Effectively

`

The "Right" Data

Best Practice #1

An organizational commitment to invest in the right data:

Tools & Know-how

-Invest: return through driving revenue, controlling costs, mitigation risk

Business Need(s)

Review

Business Needs(s)

1. Regulatory requirement

2. Big questions

3. Competitive advantage/Relevant

Regulatory Requirement

Best Practices:

#1-An organizational commitment to invest in data

#2-Use Analytics to prove to Examiners you are deliberate risk analysis and decision making

#3-A commitment to solve the big questions with data and analytics

#4-Create and report against a "Model Portfolio"

#5-An organizational commitment to use analytics to compete and stay relevant

#6-Use the right metric, presented in the right way

#7-Use basic analytics for consistent portfolio reporting and monitoring

#8-Use advanced analytics for periodic reporting and strategic planning

What do you expect will be the focus of your next exam(s)?

Tools & Know-how

Real World Example #1-Fair Lending

1. Race/Gender not generally known

2. How to know if a rate/fee is "fair"

Determining Race/Gender

Step 1. Match Address to Census Data

Baseline likelihood of being

a certain race -- 43% P(Hispanic)

Step 2. Match last name to name/race database

Apply Bayes Law for conditional likelihood of being a certain race -- now 95% P(Hispanic)

Is the "Pricing Error" correlated with a certain race/gender/age group etc.?

Determining Expected Rates

Step 1. Regression analysis to determine influence of borrower attributes

Interest Rate = b1*FICO + b2*LTV + b3*Market Rates + Error Term

Step 2. Calculate the difference between expected and actual rates

Actual Rate-Expected Rate=Pricing Error

>The FI is responsible for indirect

>Intent does not matter

>Awareness does not matter

Basic Analytics/Reports

Best Practice #2

What makes a good metric?

-comparative

-understandable

-changes the way you behave

Use Analytics to prove to Examiners you are deliberate in your ability to identify, measure, and monitor all types of risk and in your decision making.

1. New Production (by Grade and L.O.B.)

2. Concentration (by L.O.B. and Geography)

3. Grading (Overall, LTV, Credit Score)

4. Performance (Growth, Delq, C/O, Profitability, Stats)

5. Migration (Credit Score, LTV)

6. Exposure/Negative Equity

7. Watchlists/Rules-Based Lists

8. Static Pool (aka Cohort/Vintage)

9. L.O.C Monitoring/Decisioning

Thank you.

Ian Dunn/Taylor Nadauld

ian.dunn@visibleequity.com

taylor.nadauld@visibleequity.com

888-409-1560

3 Keys to Presentation:

-Don't try to present too much

-3 to 4 metrics per analysis

-Show analysis in context (trends, policy/model, peer, etc.)

Types of Metrics:

-Qualitative vs. Quantitative

-Vanity vs. Actionable

-Exploratory vs. Reporting

-Leading vs. Lagging

-Correlated vs. Casual

Advanced Analytics/Reports

Big Questions

1. Probability of Default Modeling

2. Expected Loss Modeling

3. Portfolio Optimization

4. Stress Testing

5. Fair Lending Analysis

6. Application Analysis (What-if)

7. Marketing Analytics

What are some questions in your organization you wish data/analytics could help you solve?

Competitive Advantage

How to build a winning team the A's could afford?

Real World Example #2

-What should my portfolio look like?

Mathematical Optimization

Portfolio Optimization

Interest Rate Env./

Economic Conditions

Maximize Net Interest Income (NII)

NII=Rate-(Cost of funds)-(Allocation of Operating Expenses)-(Default Risk)

Risk Culture/Profile

If no constraints then move all money to most profitable loans!

But constraints do exist:

-Concentration Limits

-NII Change Limits (Interest Rate Risk)

Where are we in the business cycle?

The Model Portfolio should reflect the institution's risk philosophy, culture, and profile.

Model Portfolio

>Product Mix

>Planned Delinquency/Charge Off Rates

>Profitability Targets

Market Demand

Policy Limits

What type of loans are in demand?

Policy Limits should inform and guide the Model Portfolio.

Do we have or can we acquire the expertise to effectively manage all products in the Model Portfolio?

Questions:

-Does he really not watch the games?

-Why did he share the "secret sauce"?

-Do they still play moneyball?

Expertise

Transforming Financial Institutions:

1990s-early 2000s: IT/Systems

Next Frontier: Data Analytics

Data/analytics infused in every major decision related to:

-Driving Revenue

-Controlling Cost

-Mitigating Risk

Best Practice #5

An organizational commitment to use analytics to remain relevant and push the boundaries of what’s possible

visibleequity.com

CUNA Lending Council, Nov. 4, 2013