Supervisors Symposium

Supervisors Symposium

December 8-9 , 2021

@ www.csbs.org � @csbsnews

CONFERENCE OF STATE BANK SUPERVISORS 1129 20th Street NW / 9th Floor / Washington, DC 20036 / (202) 296-2840

CSBS Supervisors Symposium December 8 ‐ 9, 2021| All Times Pacific Time US Grant Hotel, San Diego, CA

Day 1 | Wednesday, December 8 7:30AM ‐ 4:00PM

CSBS Supervisors Symposium Registration Presidential Foyer

Breakfast Palm Court

7:30AM

2021 Supervisors Symposium Welcome Remarks Presidential B&C Ed Gill Senior Deputy Commissioner California Department of Financial Protection and Innovation Setting the Stage: Exploring the 3 C’s (Climate, Crypto, Cyber) Presidential B&C

8:30AM

8:45AM

Charles Cooper Commissioner Texas Department of Banking Albert Forkner Commissioner Wyoming Division of Banking

Lise Kruse Commissioner North Dakota Department of Financial Institutions

Jim Cooper (moderator) Executive Vice President, Policy Conference of State Bank Supervisors

BREAK

9:45AM 10:00AM

Housing & Mortgage: Is There a Crisis Looming? Presidential B&C Kevin Byers Senior Director, Non ‐ Bank Supervision and Enforcement Conference of State Bank Supervisors Chuck Cross Senior Vice President, Non ‐ Bank Supervision and Enforcement Conference of State Bank Supervisors

Karen Pence Deputy Associate Director, Research and Statistics Federal Reserve Board of Governors

BREAK

11:00AM 11:15AM

Deep Dive into Cybersecurity Presidential B&C Charles Cooper Commissioner Texas Department of Banking

Tom Fite Director Indiana Department of Financial Institutions Kelly Lammers Director Nebraska Department of Banking and Finance

Chuck Cross (facilitator) Senior Vice President, Non ‐ Bank Supervision and Enforcement Conference of State Bank Supervisors

Mary Beth Quist (facilitator) Senior Vice President, Bank Supervision Conference of State Bank Supervisors

Committee and Working Group Lunches Supervisors Symposium Lunch : Palm Court CSBS Legislative Committee Lunch: Senate CSBS State Supervisory Processes Committee and CSBS Regulatory Committee Lunch: Congressional

12:15PM

The Future of Work in a Digital World Presidential B&C

1:30PM

Ryan Jenkins Speaker and Author Sync Learning Experiences

BREAK

3:00PM

Data in Supervision: Today's Realities and Tomorrow's Possibilities Presidential B&C

3:30PM

Roberta Hollinshead Director of Banks Washington Department of Financial Institutions Kelly Lammers Director Nebraska Department of Banking and Finance

Dharmin Patel Deputy Commissioner Arkansas State Banking Department Kris Rowley (moderator) Chief Data Officer Conference of State Bank Supervisors

Reception Palm Court

5:30PM

Day 2 | Thursday, December 9 7:30AM Breakfast

CSBS Supervisors Symposium Breakfast: Palm Court Fintech and Innovation Steering Group: Executive

Fireside Chat with FDIC Chairman Jelena McWilliams Presidential B&C

8:30AM

Jelena McWilliams Chairman Federal Deposit Insurance Corporation John Ryan President and CEO Conference of State Bank Supervisors

BREAK

9:30AM 9:45AM

Deep Dive into Climate and Related Risks Presidential B&C

Steven Rothstein Managing Director Ceres

Jim Scott Senior Advisor, Financial Institutions Ceres

BREAK

10:45AM

Deep Dive into Crypto Assets Presidential B&C

11:00AM

Angela Angelovska ‐ Wilson Co ‐ Founder, DLx Law Co ‐ Founder and Chief Legal Officer, Sila Senior Advisor, FS Vector Tim McTaggart Partner, McTaggart Law Firm General Counsel, Forethought Advisors LLC

Lunch Palm Court

12:00PM

Roundtable Discussions and Action Planning Presidential B&C

1:30PM

Sebastien Monnet Vice President, Learning & Development Conference of State Bank Supervisors

BREAK

2:30PM 3:00 PM

Innovations and Best Practices in State Financial Regulation – Peer ‐ to ‐ Peer Learning Presidential B&C

Melissa Sneed Chair, CSBSEF Performance Standards Committee Deputy Commissioner for Supervision Georgia Department of Banking and Finance Shauna Shields Vice Chair, CSBSEF Performance Standards Committee Bank Bureau Chief Iowa Division of Banking

Matt Comber Senior Director, Accreditation Conference of State Bank Supervisors

2021 Supervisors Symposium Closing Remarks

4:00PM

Presidential B&C

John Ryan President and CEO Conference of State Bank Supervisors

Housing and Mortgage: Is There a Crisis Looming?

Karen Pence, Ph.D. – Federal Reserve Board Kevin Byers – CSBS Chuck Cross – CSBS (Moderator)

Discussion Agenda and Topics

Introduction:

Chuck Cross

Market Overview:

Kevin Byers

Nonbank Mortgage Company Discussion:

Karen Pence

Closing:

Chuck Cross

Heading into 2022, rate increases have continued

Source: Black Knight/Optimal Blue LLC Daily Market Briefing: Mortgage Trends and Rate Activity, Nov. 30, 2021 (data as of 11/29/2021) https://www2.optimalblue.com/obmmi/

Housing Supply Remains Extremely Low at 2.4 Months

Nationwide home prices continue to boom

Source: Black Knight Mortgage Monitor Dec. 2021 Release (Oct. 2021 Data) https://www.blackknightinc.com/wp ‐ content/uploads/2021/12/BKI_MM_Oct2021_Report.pdf?

The U.S. Mortgage Market Grew to ~$11.3T during the pandemic, led by GSE securitizations

Source: Ginnie Mae Global Markets Analysis Report, November 2021 https://www.ginniemae.gov/data_and_reports/reporting/Documents/global_market_analysis_nov21.pdf Sourced from Federal Reserve Flow of Funds data.

Nonbank origination volume, while starting to slide, is still very high compared to past years

Nonbank servicing market share has continued to grow through the pandemic….

…and to segment among business models, with access to origination platforms fueling gains

Growth in use of subservicers reflects increased retention of MSRs by originators without inhouse servicing

Delinquencies overall remain elevated compared to pre-pandemic periods, notably in 90+ bucket

Source: Black Knight Mortgage Monitor, Dec. 2021 Release (Data as of Oct. 2021) https://www.blackknightinc.com/blac k ‐ knights ‐ october ‐ 2021 ‐ mortgage ‐ monitor/?

FHA Loans continue to show very elevated serious delinquency volumes and percentages across states

Source: HUD Neighborhood Watch, 10/31/2021 Release https://entp.hud.gov/sfnw/public/

FHA Loans in Foreclosure: Still Below Pre-COVID Levels but Increasing

FHA Loans 90+ DQ but not in forbearance: Increasing Rapidly with Wave of Forbearance Exits

Many Borrowers are Not in Forbearance or Loss Mitigation

Source: Federal Reserve Bank of Philadelphia Consumer Finance Institute “Examining Resolution of Mortgage Forbearances and Delinquencies” https://www.philadelphiafed.org/consumer ‐ finance

Delinquencies show marked differences by race/ethnicity

Source: Federal Reserve Bank of Philadelphia Consumer Finance Institute “Examining Resolution of Mortgage Forbearances and Delinquencies” https://www.philadelphiafed.org/consumer ‐ finance

NONBANK MORTGAGE COMPANIES AND MORTGAGE DEFAULT

Karen Pence Federal Reserve Board December 2021

Note. The views in this presentation are mine and not necessarily those of the Federal Reserve Board or its staff. Many thanks to CSBS for entering into a data-sharing agreement with the Federal Reserve and sharing its expertise so generously.

OVERVIEW

• Mortgage defaults were devastating during the global financial crisis • 7.8 million foreclosures occurred from 2007-2016 (CoreLogic); normally would expect perhaps 2.5 million foreclosures over a ten-year period • Defaults contributed to a 50% decrease in the number of nonbank mortgage companies from 2006-2012—nearly 1,000 companies (Bhutta and Canner, 2013)

• Reforms since 2008 make such a dire outcome unlikely today, but the mortgage system may still be vulnerable to defaults

• I will consider where default risk may be concentrated now, and describe some of the gaps in data and supervision

MORTGAGE DEFAULT RISK TODAY

MORTGAGE DEFAULT IS CONCENTRATED IN FHA AND VA LOANS

Source. Mortgage Bankers Association.

THESE LOANS ARE ORIGINATED ALMOST ENTIRELY BY NONBANK COMPANIES

Source. Housing Finance at a Glance, November 2021, Urban Institute

MORTGAGE DEFAULT PUTS STRAINS ON NONBANK MORTGAGE COMPANIES

• Liquidity strain • When borrowers don’t make their mortgage payments, servicers have to make payments on their behalf • Servicers are eventually reimbursed, but have to finance payments in the meantime • Ginnie Mae servicers advance payments until the default is resolved • Fannie and Freddie servicers advance for four months • Solvency strain • FHA andVA insurance does not cover all the losses associated with mortgage default • Fannie and Freddie guarantees cover most of the losses

NONBANK ACTIONS CAN ALSO INCREASE THE CHANCES OF MORTGAGE DEFAULT

• Nonbanks can affect the chance that a borrower defaults by the care that they take in underwriting and servicing loans

• Academic literature shows that originators and servicers with more “skin in the game” (such as more capital) put more effort into origination and servicing

SERVICERS WITH MORE CASH OFFERED MORE FORBEARANCE IN 2020

Source. Kim, Lee, Scharlemann, andVickery (2021)

HOW MUCH LIQUIDITY AND CAPITAL DO NONBANKS HAVE?

MORTGAGE REFINANCING HAS BEEN EXTREMELY PROFITABLE

Source. Fuster, Hizmo, Lambie-Hanson, andWillen (2021).

CASH RELATIVE TO EXPENSES HAS SOARED

Source: Calculations based on data from the Conference of State Bank Supervisors, Nationwide MultiState Licensing System & Registry. Data are for 100 largest for-profit Ginnie Mae issuers.

CAPITAL ADEQUACY DEPENDS ON TREATMENT OF MORTGAGE SERVICING RIGHTS

• Mortgage securitization creates a “mortgage servicing right” (MSR) • Present value of the expected servicing fees • Cash and MSRs are the two main assets of nonbanks • Leverage ratio used in current CSBS, FHFA, and Ginnie Mae capital rules treats cash and MSRs as equivalents • Proposed Ginnie Mae rule and current bank rule treat MSRs as more risky than cash

MOST NONBANKS HAVE CAPITAL ABOVE REGULATORY THRESHOLDS, BUT HOW MANY DEPENDS ON MEASURE

Source: Calculations based on data from the Conference of State Bank Supervisors, Nationwide MultiState Licensing System & Registry. Data are for 100 largest for-profit Ginnie Mae issuers.

NONBANKS WITH MORE EXPOSURE TO DEFAULT DO NOT HOLD MORE CAPITAL

Source: Calculations based on data from the Conference of State Bank Supervisors, Nationwide MultiState Licensing System & Registry. Data are for 100 largest for-profit Ginnie Mae issuers.

RISKS ASSOCIATED WITH MSRS

MSR VALUES DROP WHEN DEFAULTS RISE

• From 2020 10-K of Mr. Cooper: • “If delinquencies were significantly greater than expected, the estimated fair value of our MSRs could be diminished” • But Mr. Cooper (and other nonbanks) do not disclose by how much • This risk (unlike interest rates, which also affect MSR values) cannot be hedged • MCR data does not have any information on sensitivity of MSR valuations to changes in defaults or interest rates

MSR ARE OFTEN USED AS COLLATERAL IN COMPLEX FINANCING FACILITIES

• Nonbanks increasingly use sophisticated financial engineering to borrow against MSRs • Financing facilities have covenants and cross-collateralization provisions that make it difficult to see: • How much of the MSRs are truly unencumbered • How vulnerable the nonbanks are to margin calls if MSR valuations drop • MCR data does not specify whether MSRs are encumbered • Or provide information on terms and covenants of borrowing facilities

MSR TRADE IN AN ILLIQUID MARKET

• MSRs trade through brokers • When defaults rise, MSRs—especially Ginnie Mae MSRs—become hard to sell • Even in the mild stress of 2016, one MSR broker noted that: • “Ginnie Mae servicing offerings often receive between 0 and 3 bids” • Fannie and Freddie servicing received between 2 and 6 bids • “MSRs for low FICO borrowers are a challenge to sell”

CONSIDERATIONS IN SETTING CAPITAL FOR MSR

NONBANKS WITH CAPITAL BELOW THE GINNIE MAE THRESHOLD IN 2018

BELOW-THRESHOLD FIRMS PROVIDE LIQUIDITY TO MSR MARKET

BELOW-THRESHOLD FIRMS PERFORM KEY ROLE IN SERVICING OPERATIONS

Below- threshold

Above- threshold

10 15 20 25 30 35 40

34

Total UPB serviced

$1,704B $2,241B

14

13

Total UPB sub-serviced Number of firms

$892B

$58B

3

0 5

MSRs / Assets, 2018:Q4 MSR purchases, 2017-18 / Assets, 2016:Q4 Below-threshold Above-threshold

13

87

FINAL THOUGHTS

• Crucial to ensure that nonbanks have enough capital and liquidity to withstand a wave of defaults – and to not contribute to defaults • Although nonbank mortgage companies are currently flush with cash, their financial situation will likely worsen as mortgage refinancing subsides • MSRs are more risky than cash, and capital regulation of nonbanks should account that fact • MSRs are poorly understood, and the necessary data to analyze their vulnerabilities are not available • The operational aspects of servicing are concentrated in a handful of companies with significant MSR holdings

Supervision Technology: MCR Data Quality

Source: NMLS MCR Data

Delinquencies Remain Elevated with significant State-level Disparities, part 1: All Loans

Source: Black Knight Mortgage Monitor, Nov. 2021 Release (Data as of Sept. 2021) https://cdn.blackknightinc.com/wp ‐ content/uploads/2021/11/BKI_MM_Sept2 021_Report.pdf

FHA Mortgagee Letter 2021-18: Extension of First Legal Deadline Date through ~1/27/2022

 What Impact will this have on loss mitigation efforts and foreclosure filings?

CFPB’s 2021 Mortgage Servicing COVID-19 – Effective 8/31/2021 through 12/31/2021 1. Effective 8/31/2021 covering all residential mortgages secured by principal residences, excluding reverse mortgage and generally loans serviced by small servicers. 2. Prohibits foreclosure referral of 120+ delinquent loans between 8/31/2021 and 12/31/2021 unless specific procedural safeguards are met to give borrowers adequate opportunity for loss mitigation. 3. Specific exceptions apply: foreclosure referral occurs on or after 1/1/2022, borrower was 120+ delinquent prior to 3/1/2020, statute of limitations will expire before 1/1/2022.

https://files.consumerfinance.gov/f/documents/cfpb_covid ‐ mortgage ‐ servicing_final ‐ rule_2021 ‐ 06.pdf

Charles Cooper, Commissioner TX Tom Fite, Director IN Kelly Lammers, Director NE

What we want to accomplish today • State of the States – what are WE doing today for Cyber Supervision? • Supervision - where we need to go • Bank/nonbank • SSP and TSP Supervision • Risks & Opportunities • Ransomware – a national security threat • Table Discussions, Ideas and Commitments Capture for Plan

Keep My Bank Secure VIDEO Trey Maust and Phillip Hinkle https://www.youtube.com/watch?v=XoptEof29M8

Basic Cybersecurity Hygiene

Video Message: Focus on the Fundamentals (breaches occur due to failures here). 3 ‐ step process: 1. Board selects a framework and includes CIS Controls (which are uniquely focused on actual annual breach trends). 2. Board sets a budget of time and money (to adopt framework elements over time). 3. Board engages an audit (to measure bank’s success in implementing the Board approved framework). This is going to get increasingly complex

State of the States

State of the States

States that have TSP Authority 38

States participation in TSP Examinations • 37 exams in 2019 • 20 exams in 2020 • 14 exams in 2021

State IT Examiners

State IT Bank Exams • 1,256 in 2021 • 1,362 in 2020 • 1,459 in 2019

SSP Supervision 5 State CPCs (CA, KY MO, & NY)

389

(43 states reporting)

Using InTREx, FFIEC Handbooks, R ‐ SAT

(out of 13 SSPs)

State of the States: Nonbank Perspective

Does your agency conduct nonbank IT and cybersecurity exams - either as a stand-alone exam or as a component of a standard exam?

Internal Use Only

Please rank the following constraints to conducting nonbank IT and cybersecurity exams for your agency from 1 - 6.

MI Rank

1 2 3 4 6 5

Internal Use Only

Please rank where nonbank IT and cybersecurity supervision falls as a priority for your agency.

Internal Use Only

Has your agency used the Baseline Nonbank Cybersecurity Exam Program?

Internal Use Only

CSBS Board Approved Nonbank Cybersecurity Initiatives 2018

IT and Cybersecurity Training for Examiners

CSBS Model Data Security Law

Nonbank IT and Cyber Exam Programs

CSBS Model Data Security Law Board approved initiative from 2018 Complete and available on the CSBS website Based on the amendments to the FTC Safeguards Rule State adoption primarily on hold

Networked Supervision Public Priority 2021

For Nonbank Companies: Coordinate cyber-risk examinations for nonbank entities to ease regulatory burden, provide more consistency and enhance effectiveness.

CSBS Strategies for 2021

Awareness and Desire equal to or greater than 4 in 5 states for the Cyber Model Law language. Introduction of the law in 2 states in 2020 and 2 states in 2021.

15 states report using the Baseline Exam program in 2021.

Enhanced Exam program released for exam implementation in 2021.

Concerns Expressed by State Regulators

Examiner knowledge!

Examiner time and resources during exams

Scope Creep and Authority to Enforce

Overkill for smaller entities

How to address the Examiner Mindset Knowledge

Time

Authority

This Photo by Unknown Author is licensed under CC BY

Burden

CSBS Strategic Planning Priorities Recent Survey

What do YOU want the State of the States to be? Where are our greatest risk? 1. Bank Supervision 2. Nonbank Supervision & OCOE 3. SSP/TSP Supervision 4. Other – MSPs etc

IT Specialist- CSBS IT Advisory Team WE HAVE STRONG IT EXAMINERS IN THE SYSTEM

Opportunities & Action

INFORMATION SHARING AMONG THE STATES

TRAINING & RESOURCES

TECHNOLOGY

Ransomware Threat

Ransomware continues to increase

FinCEN Threat report

States Responded: 1. Issues Ransomware Best Practices 2. Industry Tool Issued: R-SAT Issued in 2020 banks 2021 nonbanks Less than 10 states mandated banks to complete

What have we learned from the R- SAT data?

Ransomware Campaign 2022?

Where do we go from here? What should our Strategic Direction be for 2022?

Table Discussion: What can your table and your department commit to in 2022 for enhancing State Cyber Supervision? Notes: • You have 10 minutes • Each table should come up with 1 – 3 items • DO NOT include any training items

&HUHV $FFHOHUDWRU &OLPDWH 5LVN LV %DQN 5LVN &6%6 %RDUG 0HHWLQJ 6XSHUYLVRUV 6\PSRVLXP

'HFHPEHU

.SMR EX WPMHS GSQ

6XVWDLQDELOLW\ LV WKH ERWWRP OLQH

$ERXW &HUHV $FFHOHUDWRU IRU 6XVWDLQDEOH &DSLWDO 0DUNHWV

&HUHV LV WUDQVIRUPLQJ WKH HFRQRP\ WR EXLOG D MXVW DQG VXVWDLQDEOH IXWXUH IRU SHRSOH DQG WKH SODQHW 285 0,66,21

7KLV DFWLRQ SODQ DQG SUDFWLFDO IUDPHZRUN LV RUJDQL]HG LQWR WKUHH PDLQ FRPSRQHQWV &ULWLFDO ,PSDFW $FWLRQV UHODWHG WR HQYLURQPHQWDO DQG VRFLDO ULVNV %XVLQHVV ,QWHJUDWLRQ $FWLRQV WR GULYH VWURQJ DFFRXQWDELOLW\ DQG JRYHUQDQFH RI VXVWDLQDELOLW\ DQG 6\VWHPV &KDQJH $FWLRQV WR GULYH SHUIRUPDQFH EH\RQG D FRPSDQ\ȆV RZQ IRXU ZDOOV &HUHV 5RDGPDS 2YHUYLHZ

3URWHFW :DWHU DQG 1DWXUDO 5HVRXUFHV

%XLOG D -XVW DQG ,QFOXVLYH (FRQRP\

6WDELOL]H WKH &OLPDWH

6WUDWHJLF 3ODQQLQJ DQG ([HFXWLRQ

%XVLQHVV $FFRXQWDELOLW\

7UDQVSDUHQF\ DQG 'LVFORVXUH

6\VWHPV &KDQJH

&HUHV 1HWZRUNV 2YHUYLHZ

&RPSDQ\ 1HWZRUN FRPSDQLHV DUH )RUWXQH

,QYHVWRU 1HWZRUN ,QYHVWRUV PDQDJLQJ PRUH WKDQ WULOOLRQ LQ DVVHWV

7KURXJK RXU SRZHUIXO QHWZRUNV ZH ZRUN ZLWK LQΉXHQWLDO LQYHVWRUV DQG FRPSDQLHV WR LQWHJUDWH VXVWDLQDELOLW\ SUDFWLFHV LQWR FRUH EXVLQHVV VWUDWHJLHV DQG VHL]H WKH RSSRUWXQLWLHV HPEHGGHG LQ D MXVW DQG VXVWDLQDEOH HFRQRP\

$FFHOHUDWRU 8 6 &DSLWDO 0DUNHWV 6XVWDLQDELOLW\ YLD 5HJXODWRUV )LQDQFLDO ,QVWLWXWLRQV

3ROLF\ 1HWZRUN %,&(3 OHDGLQJ FRPSDQLHV PRVWO\ FRQVXPHU EUDQGV

6867$,1$%/( ),1$1&( /HG ȉ%DQN $Ȋ WRZDUG LWV ELOOLRQ VXVWDLQDEOH ΈQDQFH FRPPLWPHQW DQG DVVRFLDWHG GLVFORVXUH HQKDQFHPHQW *29(51$1&( ',6&/2685( ,QIRUPHG ȉ%DQN %Ȋ %RDUG RI 'LUHFWRUV RQ *RYHUQDQFH 'LVFORVXUH &/,0$7( 5,6. 0$1$*(0(17 ,QΉXHQFHG ȉ%DQN &Ȋ RQ LWV LQGXVWU\ OHDGLQJ UHVSRQVH WR WKH UHFRPPHQGDWLRQV RI WKH 7DVN )RUFH RQ &OLPDWH UHODWHG )LQDQFLDO 'LVFORVXUHV 7&)' 0$7(5,$/,7< 675$7(*< 'HYHORSHG ȉ%DQN 'Ȋ ΈUVW 0DWHULDOLW\ $VVHVVPHQW LQFOXGLQJ LGHQWLI\LQJ SULRULW\ DUHDV WR IRFXV LWV VXVWDLQDELOLW\ VWUDWHJ\ JRLQJ IRUZDUG

&HUHVȆ ZRUN ZLWK WKH )LQDQFLDO 6HFWRU

&HUHVȆ ZRUN ZLWK )LQDQFLDO 5HJXODWRUV

$W WKH &HUHV $FFHOHUDWRU ZH PDNH WKH EXVLQHVV FDVH IRU HQKDQFLQJ WKH UHVLOLHQF\ RI WKH 8 6 ΈQDQFLDO V\VWHP E\ ZRUNLQJ FORVHO\ ZLWK EDQN UHJXODWRUV RQ FOLPDWH ULVN VDIHW\ VRXQGQHVV

7KLV WDEOH LGHQWLΈHV RQO\ WKH IRXQGDWLRQDO VWHSV WKDW IHGHUDO ΈQDQFLDO UHJXODWRUV VKRXOG SXW LQ SODFH WR DFWLYDWH FOLPDWH FKDQJH DFURVV WKHLU PDQGDWH

$SULO 6FRUHFDUG IRU ,QLWLDO 6WHSV RQ &OLPDWH &KDQJH $FWLRQ

&OLPDWH 5LVN LV %DQN 5LVN

&6%6 %RDUG 0HHWLQJ 6XSHUYLVRUV 6\PSRVLXP

%DQN 6XVyWDLQyDyELOyLyW\ :LOO FOLPDWH UHODWHG WUDQVLWLRQ ULVN SK\VLFDO ULVN LPSDFW WKH VDIHW\ DQG VRXQGQHVV RI WKH EDQNV \RX VXSHUYLVH"

6867$,1$%,/,7< )25 %$1.6

7UDQVLWLRQ 5LVN 5HSRUW

.(< ),1',1* 7KH FXUUHQW YLHZ RI EDQN WUDQVLWLRQ ULVN LV LQFRPSOHWH 'RHV QRW FRQVLGHU ΈQDQFLQJ RI LQGXVWULHV UHO\LQJ RQ IRVVLO IXHOV H J DJV PDQXIDFWXULQJ FRQVWUXFWLRQ WUDQVSRUWDWLRQ .(< ),1',1* 0RUH WKDQ ELOOLRQ KDOI RI PDMRU 8 6 EDQN V\QGLFDWHG OHQGLQJ LV H[SRVHG WR FOLPDWH ULVN SRWHQWLDOO\ OHDGLQJ WR VLJQLΈFDQW ORVVHV IRU WKH EDQNLQJ VHFWRU WKH EURDGHU HFRQRP\ $&7,21$%/( 5(&200(1'$7,216 L &OLPDWH ULVN DQDO\VLV LL &OLPDWH FKDQJH GLVFORVXUH LLL 6WUHVV WHVWLQJ DQG VFHQDULR DQDO\VLV LY ,QWHJUDWLQJ FOLPDWH ULVN LQWR UHJXODU GHFLVLRQ PDNLQJ WRROV ,1',&$7,9( $1$/<6,6 2) (;32685( 9$/8( $7 5,6. %DVHG RQ DQDO\VLV GRQH ZLWK FRQVXOWDQF\ &/,0$),1 XVLQJ SXEOLFO\ DYDLODEOH V\QGLFDWHG ORDQ GDWD )2//2: 83 5(6($5&+ 5HFHQW DQDO\VLV RQ WLPH KRUL]RQV DQG UHODWLRQVKLS EDQNLQJ XSFRPLQJ FRPSDQLRQ UHSRUW RQ SK\VLFDO ULVN

&ULWLFDO ,PSDFW $FWLRQV

7UDQVLWLRQ 5LVN 5HSRUW

2XU DQDO\VLV FRYHUV RQO\ D SRUWLRQ RI WUDQVLWLRQ ULVN DQG RQO\ IRU D SRUWLRQ RI EDQN SRUWIROLRV V\QGLFDWHG ORDQV 7KHUHIRUH WKH ΈQDQFLDO LPSDFWV ZH VKRZ DUH LQGLFDWLYH %DQNV QHHG WR H[SDQG WKH VFRSH WR FRYHU DOO ULVNV DQG DVVHW FODVVHV DQG GLVFORVH D GHΓQLWLYH HVWLPDWH

3RUWIROLR FRPSRVLWLRQ RI V\QGLFDWHG ORDQV RI PDMRU 8 6 EDQNV E\ FOLPDWH SROLF\ UHOHYDQW VHFWRU .H\ )LQGLQJ 'LUHFW ([SRVXUH

([SRVXUH GHWHUPLQHG E\ &/,0$),1¶V &OLPDWH SROLF\ UHOHYDQW VHFWRU &356 PDSSLQJ

.H\ )LQGLQJ /RVVHV LQ D &ULVLV

/RVVHV RQ V\QGLFDWHG ORDQ SRUWIROLR EDVHG RQ &/,0$),1 PRGHO

3K\VLFDO 5LVN 5HSRUW

.(< ),1',1* $ERXW WZR WKLUGV RI WKH LPSDFW RI SK\VLFDO ULVN FRPHV IURP LQGLUHFW HFRQRPLF LPSDFWV OLNH VXSSO\ FKDLQ GLVUXSWLRQV DQG ORZHU SURGXFWLYLW\ DQ DUHD WKDW KDV QRW EHHQ VXΊFLHQWO\ DQDO\]HG WR GDWH .(< ),1',1* 7KH ODUJHVW 8 6 EDQNV KDYH D 9D5 IURP SK\VLFDO ULVN RI PRUH WKDQ ELOOLRQ a RI V\QGLFDWHG ORDQ SRUWIROLR YDOXH HYHQ LI DGDSWDWLRQ PHDVXUHV DUH WDNHQ DQG LQ DGGLWLRQ WR WKH WUDQVLWLRQ ULVN HVWLPDWH .(< ),1',1* &RDVWDO ΉRRGLQJ GULYHQ E\ VHD OHYHO ULVH DQG VWURQJHU VWRUPV UHSUHVHQWV WKH ODUJHVW VRXUFH RI GLUHFW SK\VLFDO ULVN 7KLV ULVN LV HYHQ PRUH DFXWH LQ WKH 6RXWKHDVW .(< ),1',1* :KLOH SK\VLFDO ULVN JHQHUDWHV KXQGUHGV RI ELOOLRQV RI GROODUV RI ΈQDQFLDO ULVN IRU EDQNV LW DOVR OHDGV WR VLJQLΈFDQW RSSRUWXQLWLHV LQ DGDSWDWLRQ ΈQDQFH L H ΈQDQFLQJ QHZ LQIUDVWUXFWXUH

&ULWLFDO ,PSDFW $FWLRQV

3K\VLFDO 5LVN &UHGLW 8QLRQV DQG &RPPXQLW\ %DQNLQJ

Ɣ &RPPXQLW\ EDQNV DQG FUHGLW XQLRQV WHQG WR KDYH D VPDOOHU JHRJUDSKLF IRRWSULQW DQG IRFXV RQ D IHZ NH\ VHFWRUV VXFK DV UHVLGHQWLDO PRUWJDJHV FRPPHUFLDO UHDO HVWDWH VPDOO EXVLQHVV ΈQDQFLQJ DQG VPDOO IDUP DJULFXOWXUDO VHFWRU ORDQV Ɣ *LYHQ WKLV IRFXV FRPPXQLW\ EDQN DQG FUHGLW XQLRQ ORDQ SRUWIROLRV FDQ EH PRUH H[SRVHG WR FRQFHQWUDWLRQ ULVN DQG WKH SK\VLFDO ULVNV RI FOLPDWH FKDQJH EDVHG RQ WKH YXOQHUDELOLW\ WR DFXWH ZHDWKHU HYHQWV

6DIHW\ 6RXQGQHVV

8 6 H[SHULHQFHG VHSDUDWH ELOOLRQ GROODU GLVDVWHUV LQ

6RXUFH 12$$

6DIHW\ 6RXQGQHVV

)HGHUDO /HYHO 'HYHORSPHQWV &6%6 %RDUG 0HHWLQJ 6XSHUYLVRUV 6\PSRVLXP

&OLPDWH ULVN PDQDJHPHQW JXLGDQFH LV QRZ EHLQJ GUDZQ XS IRU EDQNV E\ WKH 2&& LQ FRRUGLQDWLRQ ZLWK LQWHUDJHQF\ SHHUV &OLPDWH FKDQJH SRVHV DQ H[LVWHQWLDO ULVNȏ UHJXODWRUV PXVW EHJLQ WR WDNH DFWLRQ QRZ $FWLQJ &RPSWUROOHU RI WKH &XUUHQF\ 0LFKDHO - +VX

5HXWHUV 6HSWHPEHU

7KH (XURSHDQ EDQN VWUHVV WHVWV DUH ȉSURYLQJ WR EH D YHU\ SURΈWDEOH H[HUFLVH ERWK IRU WKH ΈQDQFLDO LQVWLWXWLRQV DQG IRU UHJXODWRUVȊ VDLG )HG &KDLU -HURPH 3RZHOO %ORRPEHUJ 1HZV -XO\ 6HSWHPEHU

ȉ ([WUHPH ZHDWKHU KDV FRVW RXU HFRQRP\ RYHU ELOOLRQ RYHU WKH ODVW \HDUV :H KDYH WR PDNH VXUH RXU ΈQDQFLDO V\VWHP FDQ ZLWKVWDQG FOLPDWH FKDQJH DQG LV SUHSDUHG WR WUDQVLWLRQ WR FOHDQ HQHUJ\ Ȋ %LGHQ VDLG

&OLPDWH:LUH 1RYHPEHU

7KH 6HFXULWLHV DQG ([FKDQJH &RPPLVVLRQ LV SODQQLQJ WR LVVXH SURSRVHG UXOH RQ PDQGDWRU\ FRUSRUDWH FOLPDWH GLVFORVXUH IURP SXEOLFO\ OLVWHG FRPSDQLHV

3ROLWLFR 6HSWHPEHU

/LQN WR 6(& /HWWHU WR &RPSDQLHV 5HJDUGLQJ &OLPDWH &KDQJH 'LVFORVXUHV

)62& ,GHQWLΈHV &OLPDWH &KDQJH DV DQ (PHUJLQJ ,QFUHDVLQJ 7KUHDW WR 8 6 )LQDQFLDO 6WDELOLW\ 7KH UHSRUW KLJKOLJKWV WKH QHHG IRU EXLOGLQJ FDSDFLW\ ΈOOLQJ GDWD JDSV HQKDQFHG FOLPDWH GLVFORVXUH DQG WKH LQFUHDVHG XVH RI VFHQDULR DQDO\VLV DV D WRRO IRU DVVHVVLQJ FOLPDWH UHODWHG ΈQDQFLDO ULVNV

8 6 'HSDUWPHQW RI WKH 7UHDVXU\ 1HZV 2FWREHU

,QWHUQDWLRQDO 'HYHORSPHQWV &6%6 %RDUG 0HHWLQJ 6XSHUYLVRUV 6\PSRVLXP

7KH %DQN RI (QJODQG KDV WROG EDQNV DQG LQVXUHUV LW LV SUHSDUHG WR XVH LWV SRZHUV VKRXOG WKH\ IDLO WR SURSHUO\ PDQDJH FOLPDWH ULVNV 7KH FHQWUDO EDQN KDV EHJXQ D UHYLHZ RI SRWHQWLDO QHZ FDSLWDO UHTXLUHPHQWV VWUHVV WHVWV OLQNHG WR XQVXVWDLQDEOH DVVHWV

7KH *XDUGLDQ 2FWREHU

(&%ȆV FOLPDWH VWUHVV WHVW UHYHDOV WKUHDWV WR ΈQDQFLDO VWDELOLW\ Ɣ )DLOXUH WR WUDQVLWLRQ FRXOG KDYH VHYHUH FRQVHTXHQFHV IRU ΈQDQFLDO VWDELOLW\ DFURVV WKH EDQNLQJ DQG FRUSRUDWH VHFWRU Ɣ 6KRUW WHUP FRVWV RI WUDQVLWLRQLQJ DUH IDU H[FHHGHG E\ ORQJ WHUP EHQHΈWV RI D PRUH VXVWDLQDEOH ΈQDQFLDO VHFWRU DQG HFRQRP\

2ΕFLDO 0RQHWDU\ DQG )LQDQFLDO ,QVWLWXWLRQV )RUXP 1RYHPEHU

7KH 8 . ȆV ELJJHVW FRPSDQLHV ZLOO KDYH WR VHW RXW GHWDLOHG SODQV IRU PRYLQJ WR D ORZ FDUERQ IXWXUH DQG UHSRUW WKHLU FOLPDWH UHODWHG ULVNV XVLQJ 7&)' IUDPHZRUN VWDUWLQJ QH[W \HDU 7KH 8 . LV WKH ΈUVW LQ WKH * WR PDNH VXFK GLVFORVXUHV PDQGDWRU\

%ORRPEHUJ %%& 2FWREHU 1RYHPEHU

,)56 ZKLFK JRYHUQV ΈQDQFLDO UHSRUWLQJ LQ PRUH WKDQ FRXQWULHV DQQRXQFHG WKH IRUPDWLRQ RI WKH ,QWHUQDWLRQDO 6XVWDLQDELOLW\ 6WDQGDUGV %RDUG ,66% WR GHYHORS D JOREDO EDVHOLQH RI VXVWDLQDELOLW\ GLVFORVXUH ZKLFK ZLOO LQFOXGH 7&)' DQG 6$6% VWDQGDUGV

)RUEHV 1RYHPEHU

6WDWH /HYHO 'HYHORSPHQWV &6%6 %RDUG 0HHWLQJ 6XSHUYLVRUV 6\PSRVLXP

1HZ 1<')6 &OLPDWH 'LYLVLRQ (OHYDWHV WKH ,PSRUWDQFH RI

&OLPDWH 5HODWHG )LQDQFLDO 5LVNV LQ )LQDQFLDO 6XSHUYLVLRQ ȉ$V RQH RI WKH PRVW FULWLFDO LVVXHV RI RXU JHQHUDWLRQ FOLPDWH FKDQJH SRVHV ZLGH UDQJLQJ DQG PDWHULDO ULVNV WR WKH ΈQDQFLDO V\VWHP Ȋ VDLG $FWLQJ 6XSHULQWHQGHQW $GULHQQH +DUULV

1<')6 3UHVV 5HOHDVH 1RYHPEHU

:DVKLQJWRQ 6WDWH LVVXHV '), $OHUW RQ 6\VWHPLF 5LVN RI &OLPDWH &KDQJH )LQDQFLDO LQVWLWXWLRQV IDFH VLJQLΈFDQW SK\VLFDO DQG WUDQVLWLRQ ULVNV EXW KDYH DQ RSSRUWXQLW\ WR WDNH D SURDFWLYH DSSURDFK E\ HYDOXDWLQJ WKHLU RZQ FRQWULEXWLRQV WRȏ FOLPDWH FKDQJH ,I ZH IDLO WR SUHSDUH DQG DGMXVW RXU SUDFWLFHV FOLPDWH FKDQJH PD\ EH WKH FDWDO\VW IRU WKH QH[W ΈQDQFLDO FULVLV

:DVKLQJWRQ '), 3UHVV 5HOHDVH -XO\

8 6 3UHFHGHQWV )RU 8UJHQW $FWLRQ

&6%6 %RDUG 0HHWLQJ 6XSHUYLVRUV 6\PSRVLXP

+RZ 8 6 UHJXODWRUV KDQGOHG WKH < . WKUHDW Ɣ )LUVW ULJRURXVO\ VWXG\ WKH OLNHO\ LPSDFW RI < . RQ WKH EDQNLQJ V\VWHP DQG FDOFXODWH ZKDW LW ZRXOG WDNH WR SUHSDUH IRU LW Ɣ 6HFRQG ORRN DW \RXU RZQ LQWHUQDO V\VWHPV DQG HDUPDUN WKH QHFHVVDU\ UHVRXUFHV WR EULQJ WKHP WR D VWDWH RI UHDGLQHVV Ɣ 7KLUG DQG SHUKDSV PRVW LPSRUWDQW EHJLQ ZRUNLQJ ZLWK WKH ΈQDQFLDO LQVWLWXWLRQV WR KHLJKWHQ WKHLU DZDUHQHVV DQG HQFRXUDJH WKHP WR PRYH ZLWK DOO GXH KDVWH WRZDUG < . VROXWLRQV RI WKHLU RZQ Ɣ (QFRXUDJH VXSSRUW DQG FDMROHȏ ZH LVVXHG PRUH WKDQ LQWHUDJHQF\ SROLF\ VWDWHPHQWVȏ HYHQ WRRN HQIRUFHPHQW DFWLRQV LQ WKRVH FDVHV ZKHUH LW VHHPHG ZDUUDQWHG )RUPHU &RPSWUROOHU RI WKH &XUUHQF\ -RKQ ' +DZNH -U

< . 7KUHDW

+RZ 8 6 UHJXODWRUV KDQGOHG &\EHUVHFXULW\ WKUHDW Ɣ 2Q -XQH WKH )),(& ODXQFKHG D GHGLFDWHG ZHESDJH RQ F\EHUVHFXULW\ WR VHUYH DV D UHVRXUFH IRU ΈQDQFLDO LQVWLWXWLRQV DQG H[DPLQHUV Ɣ )),(& PHPEHUV DUH UDLVLQJ DZDUHQHVV RI F\EHUVHFXULW\ ULVNV DW ΈQDQFLDO LQVWLWXWLRQV Ɣ ,GHQWLI\ DVVHVV DQG PLWLJDWH WKHVH ULVNV LQ OLJKW RI WKH LQFUHDVLQJ WKUHDWV WKDW SRVH ULVNV WR DOO LQGXVWULHV LQ RXU VRFLHW\ Ɣ 7KH )),(& :HE SDJH SURYLGHV OLQNV WR MRLQW VWDWHPHQWV ZHELQDUV DQG RWKHU LQIRUPDWLRQ WKDW PD\ KHOS ΈQDQFLDO LQVWLWXWLRQV ZKHQ WKLQNLQJ DERXW WKH LVVXH RI F\EHUVHFXULW\ )),(& 6WDWHPHQW

&\EHU 7KUHDW

6DIHW\ 6RXQGQHVV %HVW 3UDFWLFHV

&6%6 %RDUG 0HHWLQJ 6XSHUYLVRUV 6\PSRVLXP

+RZ &DQ 6WDWH 5HJXODWRUV (QKDQFH 6DIHW\ 6RXQGQHVV $URXQG &OLPDWH 5LVN" Ɣ 3XEOLFO\ DΊUPLQJ FOLPDWH FKDQJH DV D V\VWHPLF ULVN WR \RXU VWDWH H J D '), $OHUW Ɣ $VVLJQ VWD· FRQGXFW UHVHDUFK KROG IRUXPV Ɣ /HYHUDJH \RXU 5HJLRQDO )HG *RYHUQRUȆV 2ΊFH DQG RWKHU )HGHUDO SDUWQHUV WR DFFHVV UHVHDUFK VKDUH EHVW SUDFWLFHV DQG DVVHVV UHJLRQDO FOLPDWH YXOQHUDELOLWLHV Ɣ ([SOLFLWO\ LQWHJUDWH FOLPDWH ULVN LQWR \RXU SUXGHQWLDO VXSHUYLVLRQ IUDPHZRUN IRU H[DPSOH &$0(/6 UDWLQJV Ɣ 7DNH DGYDQWDJH RI H[LVWLQJ GLVFORVXUH IUDPHZRUNV H J 7&)' WR DUP VWDNHKROGHUV ZLWK DFWLRQDEOH LQIRUPDWLRQ

%HVW 3UDFWLFHV

1H[W 6WHSV &6%6 %RDUG 0HHWLQJ 6XSHUYLVRUV 6\PSRVLXP

+RZ &HUHV &DQ +HOS 6WDWH 5HJXODWRUV

&HUHV ZRXOG ZHOFRPH WKH RSSRUWXQLW\ WR DFW DV D UHVRXUFH IRU &6%6 PHPEHUV WR SURYLGH Ɣ ,QIRUPDWLRQ RU UHVHDUFK LQ VXSSRUW RI \RXU H·RUWV WR PRQLWRU DQG PLWLJDWH FRQFHQWUDWLRQ ULVNV LQ FRPPHUFLDO UHDO HVWDWH DQG DJULFXOWXUH OHQGLQJ Ɣ :HELQDUV WR \RXU WHDP LQ VXSSRUW RI \RXU GHYHORSPHQW RI VXSHUYLVRU\ VWUDWHJLHV UHJDUGLQJ H[WUHPH ZHDWKHU FOLPDWH FKDQJH DQG HQYLURQPHQWDO MXVWLFH Ɣ ,QIRUPDWLRQ RU FRQWHQW IRU DQ\ FOLPDWH FKDQJH LQIRUPDWLRQ \RX ZLVK WR SURYLGH WR \RXU UHJXODWHG HQWLWLHV

1H[W 6WHSV

7KDQN

6WHYHQ 5RWKVWHLQ 0DQDJLQJ 'LUHFWRU &HUHV $FFHOHUDWRU IRU 6XVWDLQDEOH &DSLWDO 0DUNHWV

-LP 6FRWW &)$ 6HQLRU $GYLVRU &HUHV $FFHOHUDWRU IRU 6XVWDLQDEOH &DSLWDO 0DUNHWV

VURWKVWHLQ#FHUHV RUJ

MVFRWW#FHUHV RUJ

+HDGTXDUWHUV &KDXQF\ 6WUHHW WK )ORRU %RVWRQ 0$

&DOLIRUQLD 2ΊFH 3LQH 6WUHHW 6XLWH 6DQ )UDQFLVFR &$

6SHDNHU %LRJUDSKLHV

6WHYHQ 0 5RWKVWHLQ LV WKH IRXQGLQJ 0DQDJLQJ 'LUHFWRU RI WKH &HUHV $FFHOHUDWRU IRU 6XVWDLQDEOH &DSLWDO 0DUNHWV

7KH &HUHV $FFHOHUDWRU DLPV WR WUDQVIRUP WKH SUDFWLFHV DQG SROLFLHV WKDW JRYHUQ FDSLWDO PDUNHWV LQ RUGHU WR DFFHOHUDWH UHGXFWLRQ RI WKH ZRUVW ΈQDQFLDO LPSDFWV RI WKH FOLPDWH FULVLV DQG RWKHU VXVWDLQDELOLW\ WKUHDWV 6WHYHQȆV \HDUV RI H[SHULHQFH ZLOO EH FULWLFDO WR H[SORUH WKH PRVW H·HFWLYH VWUDWHJLHV IRU WKH $FFHOHUDWRU WR IRFXV RQ DQG PRYH FDSLWDO PDUNHWV WRZDUGV FOLPDWH VXVWDLQDELOLW\ +H KDV ZRUNHG DW ORFDO VWDWH IHGHUDO DQG LQWHUQDWLRQDO OHYHOV RI JRYHUQPHQW 6WHYHQ VHUYHG RQ PDQ\ QRQ SURΈW DQG JRYHUQPHQW ERDUGV +H KDV VSRNHQ DQG ZULWWHQ H[WHQVLYHO\ DQG ZRUNHG ZLWK SDUWQHUV LQ WKH FRUSRUDWH QRQ SURΈW JRYHUQPHQW DQG SKLODQWKURS\ VHFWRUV 6WHYHQ KDV D %$ ZLWK +RQRUV LQ 3ROLWLFDO 6FLHQFH IURP :LOOLDPV &ROOHJH DQG DQ 0%$ IURP 1RUWKHDVWHUQ 8QLYHUVLW\ȆV 'Ȇ$PRUH 0F.LP 6FKRRO RI 0DQDJHPHQW +H DOVR VWXGLHG DW 'XNH 8QLYHUVLW\ȆV ,QVWLWXWH RI 3XEOLF 3ROLF\ +H DQG KLV ZLIH 6XVDQ OLYH LQ 6RPHUYLOOH 7KH\ KDYH WZR JURZQ VRQV +H LV FXUUHQWO\ RQ WKH %UDG\ &DPSDLJQ IRU *XQ 6DIHW\ %RDUG DQG WKH 0DVV &LYLF /HDUQLQJ &RDOLWLRQ 6WHHULQJ &RPPLWWHH

6WHYHQ 5RWKVWHLQ 0DQDJLQJ 'LUHFWRU &HUHV $FFHOHUDWRU IRU 6XVWDLQDEOH &DSLWDO 0DUNHWV

VURWKVWHLQ#FHUHV RUJ

-LP LV 6HQLRU $GYLVRU IRU )LQDQFLDO ,QVWLWXWLRQV (QJDJHPHQW ZLWKLQ WKH &HUHV $FFHOHUDWRU IRU 6XVWDLQDEOH &DSLWDO 0DUNHWV ,Q WKLV UROH -LP ZRUNV ZLWK EDQNV LQVXUDQFH FRPSDQLHV DQG UHJXODWRUV WR DFFHOHUDWH WUDQVIRUPDWLRQ RI WKH SUDFWLFHV DQG SROLFLHV WKDW JRYHUQ 86 DQG JOREDO FDSLWDO PDUNHWV LQ RUGHU WR UHGXFH WKH ZRUVW ΈQDQFLDO LPSDFWV RI FOLPDWH FKDQJH DQG RWKHU VXVWDLQDELOLW\ WKUHDWV 3ULRU WR MRLQLQJ &HUHV -LP ZRUNHG LQ WKH JOREDO ΈQDQFH LQGXVWU\ DV DQ LQWHUQDWLRQDO EDQNLQJ DQG FDSLWDO PDUNHWV VSHFLDOLVW )RU RYHU \HDUV -LP DGYLVHG PXOWLQDWLRQDO FRUSRUDWLRQV PXOWLODWHUDO RUJDQL]DWLRQV DQG JRYHUQPHQWV RQ ΈQDQFLQJ DQG ULVN PDQDJHPHQW VWUDWHJLHV LQ ERWK GHYHORSHG DQG IURQWLHU PDUNHWV 0RVW UHFHQWO\ KH ZDV EDVHG LQ /RQGRQ (QJODQG DV &LWLȆV VROXWLRQV VDOHV KHDG IRU WKH &((0($ UHJLRQ FRPSULVLQJ &HQWUDO (DVWHUQ (XURSH 5XVVLD WKH 0LGGOH (DVW 3DNLVWDQ DQG $IULFD -LP KDV DOVR FRQGXFWHG EXVLQHVV ZLWK SULYDWH DQG SXEOLF VHFWRU SDUWQHUV LQ WKH &DULEEHDQ &HQWUDO $PHULFD &DQDGD :HVWHUQ (XURSH DQG WKH 8QLWHG 6WDWHV -LP LV D &)$ FKDUWHUKROGHU DQG D JUDGXDWH RI +DUYDUG %XVLQHVV 6FKRROȆV *HQHUDO 0DQDJHU 3URJUDP +H LV DOVR FXUUHQWO\ HQUROOHG LQ WKH +DUYDUG .HQQHG\ 6FKRROȆV 3/& SURJUDP ZKLFK IRFXVHV RQ 3XEOLF 3ROLF\ GHVLJQ DQG 0RUDO /HDGHUVKLS +H LV PDUULHG ZLWK WZR FKLOGUHQ DQG OLYHV LQ %RVWRQ

-LP 6FRWW &)$ 6HQLRU $GYLVRU )LQDQFLDO ,QVWLWXWLRQV (QJDJHPHQW

MVFRWW#FHUHV RUJ

CON F ERENCE OF S TAT E BANK SUPERV I SORS Presentation: Financial Technology Market Developments and Regulation

San Diego December 9, 2021

1

General Counsel & Founder Forethought Advisors LLC ‐‐ Partner & Founder McTaggart Law Office https://mctaggartlegal.com/ ‐‐ Advisory Board Member First State FinTech Lab ‐‐ Previously: • Partner, Washington, DC Office, Nixon Peabody ; Willkie Farr ; Pepper Hamilton • State Bank Commissioner, Delaware • Counsel, Committee on Banking, U.S. Senate • Chair, Trust and Fiduciary Services Subcommittee, Banking Law Committee, American Bar Association • FFIEC State Liaison, Conference of State Bank Supervisors • Research and Editorial Team Member, choosedelaware.com , FinTech Report, Joint Report by First State FinTech, Delaware Partnership for Prosperity and Biden Institute for Public Administration at the University of Delaware • Co ‐ Author, McTaggart and Freese, Regulation of Mobile Payments, The Banking Law Journal (June 2010) 2

Tim McTaggart E ‐ mail: timothymctaggart944@gmail.com Telephone: (202)412 ‐ 6610

Partner and Cofounder DLx Law ‐‐ CLO & Cofounder Sila Inc. ‐‐ Advisory Board Member FS Vector ‐‐ Previously: • CLO, CCO, and founding team member, Digital Asset Holdings LLC • Partner, Reed Smith LLP • Attorney, Latham & Watkins, LLP • Attorney, Fried Frank, LLP

Angela Angelovska- Wi lson

E ‐ mail: angela@dlxlaw.com Telephone: (202) 365 ‐ 1448

3

AI and Machine Learning: RegTech and

SupTech Timothy McTaggart General Counsel & Co ‐ Founder Forethought Advisors LLC ‐‐ Partner & Founder McTaggart Law Office

1 2 / 1 6 / 2 0 2 1

4

What is AI?

The term ‘Artificial Intelligence’ was first coined in 1956 by a Dartmouth University professor, John McCarthy. He defined Artificial Intelligence as: “Every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to stimulate it. An attempt will be made to find how to make machines use language, form abstractions and concepts, solve kinds of problems now reserved for humans, and improve themselves.”

SQLWebsite

5

S UMM I NG U P WH E R E WE A R E TO D AY • Many Bay area tech community consultants are not convinced that banks/financial services firms are fully committed to the bold use of fintech/AI products. • The concern heard is that banks will use the new technology products to push ahead with automation and to cut operating costs. • As a secondary theme, banks are seen as being interested in using predictive analysis from data sets to help combat fraud and improve compliance efforts. • There also is an acknowledgment that fintech/AI technology can assist banks in managing enterprise risk beyond the compliance improvements. • Regulators need to keep up on the technology deployed by industry. The marketplace today has RegTech firms in business throughout the world.

6

RegTech Fi rms Operat ing Global ly

Deloitte

7

ARTIFICIAL INTELLIGENCE: TRENDLINE FOR ACCELERATED USE • The cost savings potential for the use of AI tools in the banking industry is staggering. • Banks were projected to invest more than $5 billion in artificial intelligence in 2019, according to a September report by International Data Corp., up from $4 billion last year. Meanwhile, technological advances could cost the banking sector more than 200,000 jobs over the next decade, Wells Fargo Securities predicted in a report the following month.” Bank Dive by Dan Ennis. • An Accenture study from 2019 was cited showing $71 billion in savings to the financial services industry, with $12 billion in savings related to bank tasks that can be fully automated and $59 billion in savings by deploying augmented technology. • Dan Figgela, in a summary article for Emerj, an AI consulting entity (March 14, 2020), noted the deployment of AI by the seven top U.S. banks. He detailed this savings by J.P. Morgan’s “COiN” that also is mentioned by other industry reports.

8

AI : Accelerated Use Case

JPMorgan Chase invested in technology and introduced a Contract Intelligence (COiN) “chatbot” designed to “analyze legal documents and extract important data points and clauses” in 2017

Manual review of 12,000 annual commercial credit agreements normally requires approximately 360,000 hours

Results from an initial implementation of this machine learning technology showed that the same number of agreements could be reviewed in seconds

COiN has widespread potential, and the company is exploring additional ways to implement this powerful tool, although further information on the rollout is sparse

9

REGTECH/ SUPTECH

Is AI going to be a large part of the future of financial regulation for U.S. bank regulatory agencies with RegTech/SupTech? • Supervisory technology or “SupTech” is in its early stage of use by regulatory entities around the world • Canadian firm SQL Power Group Inc.’s website explains AI’s origins and the current pivotal point as financial regulators—not just bank regulators—around the world begin to use AI as part of the supervisory process

10

REGTECH/SUPTECH

S upTe c h ( c on t ’d ) No one i s s u g g e s t i ng t h a t A I i s g o i n g t o r ep l a c e huma n j ud gme n t i n conn e c t i on w i t h b a n k a nd f i n a n c i a l s up e r v i s i on . I n f a c t , i t i s j u s t t h e oppo s i t e : t h e r e i s a g r e a t d e a l o f c on f i d en c e i n huma n j ud gme n t d e v e l op e d w i t h ye a r s o f ex p e r i en c e t o eva l u a t e r i s k o f f i n a n c i a l i n s t i t u t i on s . T he r e i s a l s o g r e a t po t e n t i a l f o r A I a nd ma c h i n e l e a r n i n g t o con t r i bu t e t o t h e s up e r v i s o r y p r o c e s s by p e rm i t t i ng a n ex t ra o r d i n a r i l y f a s t r e v i ew o f comp l ex d a t a s e t s a nd t o d e v e l op s c r e en s o f t h e d a t a wh i c h may i d e n t i f y p a t t e r n s o f f i n a n c i a l i r r e g u l a r i t i e s o r o t h e r i mp r op e r condu c t

11

REGTECH/SUPTECH

SupTech (cont’d) • On its website, SQL notes other non ‐ U.S. financial regulators beginning to use AI and machine learning processes, including Australian Securities and Investments Commission (“ASIC”). • As noted by SQL: ASIC launched a pilot program in partnership with the Australian RegTech Association to apply cognitive learning tools and applications to accountant web pages to examine potential unlicensed or misleading conduct in relation to self ‐ managed superannuation fund activities.

12

REGTECH/SUPTECH

SupTech (cont’d) SQL also references a 2017 UK regulatory development with the Financial Conduct Authority (“FCA”), which announced it is looking into possible uses of AI and supervised machine learning to enforce regulatory compliance from analytics and detect financial irregularities. ● This provides additional support for the idea of using AI and machine learning to detect patterns that may be more apparent from a computer assisted scrub of data than would be apparent to a person reviewing the same data set.

13

B A N K S A N D A R T I F I C I A L I N T E L L I G E N C E

Why do Bank Regulators Need to Concern Themselves with AI Matters?

● An investment by a bank in AI to improve its risk profile to deter and fraud and effectiveness in using the technology ought to be a positive factor in evaluating the strength of management at the bank. ● AI is expected to be able to recognize patterns of suspicious activity in the AML area better and faster than human analysis of the same data. ● AI tools and technology can help banks and bank regulators combat “false positives” presented in the data reviewed, including AML matters, and to do so in a manner that helps to “clean up/scrub” the data base and maintain its integrity.

1 4

B A N K S A N D A R T I F I C I A L I N T E L L I G E N C E

Why do Bank Regulators Need to Concern Themselves with AI Matters? ● With respect to the “back of the house,” bankers and bank supervisors should be concerned whether the AI model and related technology is validated in a manner to avoid bias. ● Any such model involving the review of a consumer’s creditworthiness needs to comply with consumer credit laws to not discriminate when the bank is relying on the data analysis to determine whether to extend credit. ● If credit is not extended, or credit is proposed to be extended on different terms than requested in the application, then FCRA/ECOA and Reg B provisions may be triggered and require further notice and communication to the prospective borrower.

1 5

W H A T A B O U T N E W S T R E S S E S O N B A N K M A N A G E M E N T ?

Bank regulators are likely to see shrinking bank employment from middle manager employees who had handled underwriting processes, trading processes, and other quantitative tasks.

AI and machine learning will displace employees from those jobs so that there will be a smaller professional complement of personnel to review and be involved with oversight and judgment but far fewer bank employees.

By contrast, there will be a greater number of bank employees hired as IT/computer analysts, programmers, systems administrators and other tech related jobs.

Bank regulators always need to assure themselves that the management team in place is capable of assessing and managing the risk associated with the business plan of a specific bank.

1 6

W H A T A B O U T N E W S T R E S S E S O N B A N K M A N A G E M E N T ?

As a bank adapts to new AI tools and the workforce composition changes, bank regulators must be convinced that bank management can manage a newly configured bank work force.

Bank regulators need to closely review management and board executive talent to implement cybersecurity measures for a bank’s IT systems and processes. In particular, corporate secretaries are being asked to have ever greater corporate governance responsibilities along with an increased use of technology by corporate boards.

Technology by itself will never be a solution for a poorly managed bank.

1 7

How will Fintech and AI perform during a financial crisis and related significant recession or worse economic downturn? • One of the concerns for banks and bank regulators is that the fintech sector has emerged and grown largely during sunny economic times. • There are concerns that fintech and AI may work great during good economic times but work less well, or not at all, during times of economic distress.

ARTIFICIAL INTELLIGENCE AND REGULATION

1 8

How will Fintech and AI perform during a financial crisis and related significant recession or worse economic downturn? • A decline in cost savings can be anticipated for AI/fintech during a recession but the technology investment is likely still seen as a “positive payoff” because there likely will still be significant cost savings. • In short, the optimal use case for compliance purposes for AI/fintech may also be during a time of robust economic activity but the use of AI/fintech still will be valuable during economic stress times for compliance uses.

ARTIFICIAL INTELLIGENCE AND REGULATION

1 9

MODEL BIAS

Model Validation: Bias Concerns There is a premise that the AI process is pure and lacks bias. • Is that a reasonable premise? • Indeed, there are places where biases can creep into data sets and lead to conclusions that are based on bias rather than on judgment. Alison Jimenez in the American Banker on August 16, 2019, wrote an op ‐ ed entitled, “AI Use Carries Bias Risk for Financial Regulators.” • References a June 2019 hearing of the U.S. House Financial Services Artificial Intelligence Task Force, which looked at the benefits and potential drawbacks of using AI in the financial services sector. • Jimenez is the president of Dynamic Securities Analytics Inc. and she warned that the House’s AI task force should broaden its scope to evaluate the bias risk involved with AI and the use of algorithms.

20

MODEL BIAS

Model Validation: Bias Concerns (cont’d) Jimenez suggested that insufficient attention is being given to the potential for bias in supervisory algorithms used by regulators. • Financial regulators have used machine learning to detect fishy text in corporate filings, identify money ‐ laundering networks and discover tax cheats. • Bias can lurk in the collected Big Data, however.

21

21

MODEL BIAS

Model Validation: Bias Concerns (cont’d) Jimenez cites the 2017 FDIC National Survey of Unbanked and Underbanked Households: • 14% of Hispanic and 17% of African ‐ American households are unbanked. • In some instances, such households may not have a government issued ID card. • The upshot is that these households may rely heavily on cash and prepaid cards but bank AML algorithms flag frequent prepaid transactions as being suspicious even though there is no problem involved and similarly banks may file a disproportionate number reports to FinCEN for suspicious IDs used as substitutes for official government issued IDs.

22

MODEL BIAS

Model Validation: Bias Concerns (cont’d) Jimenez notes that FinCEN has found explicit bias in SAR reporting: • FinCEN has warned that filings should not be based on a person’s ethnicity, such as being of Middle Eastern descent. • Despite the potential for bias, FinCEN data is widely available to law enforcement agencies and FinCEN employs algorithms to screen filings to identify reports that merit further review.

23

MODEL BIAS

Model Validation: Bias Concerns (cont’d) ● This danger can be exacerbated if the high ‐ tech entities doing testing on such models are not populated with programmers/analyst teams with diversity and inclusion as part of their groups. ● Implicit bias can become part of these AI tools and analytics and it can be compounded if the data set being analyzed also has certain bias aspects due to imprecision in collection or cataloging of the information. ● There is a danger of discriminatory outcomes in the use of AI proprietary systems which leads to possible legal exposure for private firms. There are similar potential risks leading up to and including mission failure for public policy entities such as federal bank regulators if model bias is not addressed.

24