View original document

The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.

For release on delivery
3:45 p.m. EST
November 16, 2017

Where Do Consumers Fit in the Fintech Stack?
Remarks by
Lael Brainard
Member
Board of Governors of the Federal Reserve System
at
“FinTech Risks and Opportunities: An Interdisciplinary Approach”
a conference sponsored by the University of Michigan

Ann Arbor, Michigan

November 16, 2017

The new generation of fintech tools offers the potential to help consumers manage their
increasingly complicated financial lives, but also poses risks that will need to be managed as the
marketplace matures.1
In many ways, the new generation of fintech tools can be seen as the financial equivalent
of an autopilot. The powerful new fintech tools represent the convergence of numerous
advances in research and technology--ranging from new insights into consumer decisionmaking
to a revolution in available data, cloud computing, and artificial intelligence (AI). They operate
by guiding consumers through complex decisions by offering new ways of looking at a
consumer’s overall financial picture or simplifying choices, for example with behavioral nudges.
As consumers start to rely on financial autopilots, however, it is important that they
remain in the driver’s seat and have a good handle on what is happening under the hood.
Consumers need to know and decide who they are contracting with, what data of theirs is being
used by whom and for what purpose, how to revoke data access and delete stored data, and how
to seek relief if things go wrong. In short, consumers should remain in control of the data they
provide. In addition, consumers should receive clear disclosure of the factors that are reflected in
the recommendations they receive. If these issues can be appropriately addressed, the new
fintech capabilities have enormous potential to deliver analytically grounded financial services
and simplified choices, tailored to the consumers’ needs and preferences, and accessible via their
smartphones.2

1

I am grateful to Kelvin Chen for his assistance in preparing this text. The remarks represent my own views, which
do not necessarily represent those of the Federal Reserve Board or the Federal Open Market Committee.
2
See Lael Brainard, “The Opportunities and Challenges of Fintech,” (speech at the Conference on Financial
Innovation, Washington, D.C., December 2, 2016),
www.federalreserve.gov/newsevents/speech/brainard20161202a.pdf

-2Consumers Face Complex Financial Choices
When the first major “credit card,” the Diner’s Club Card, was introduced in 1949,
consumers could only use the cardboard card at restaurants and, importantly, only if they paid
the entire amount due each month.3 Today, the average cardholder has about four credit cards,
and the Federal Reserve Bank of New York estimates that American consumers collectively
carry $785 billion in credit card debt.4
When signing up for a credit card, consumers face a bewildering array of choices. Half
of consumers report that they select new cards based on reward programs, weighing “cash back”
offers against “points” with their credit card provider that may convert into airline or hotel
“miles,” which may have varying values depending on how they are redeemed.5 In some cases,
rewards may apply to specific spending categories that rotate by quarter and require that
consumers re-register each term, and the rewards may expire or be forfeited under complicated
terms.6
In some cases, the choices may be confusing. Let’s take the example of zero percent
interest credit card promotions. A consumer may choose a zero percent interest credit card
promotion and expect to pay no interest on balances during a promotional period, after which
any balances are assessed at a higher rate of interest going forward. But if a consumer instead

See, e.g., Damaris Olaechea, “Who Issued the First Credit Card?” Nerd Wallet (February 12, 2014),
www.nerdwallet.com/blog/credit-cards/issued-first-credit-card/.
4
See, e.g., Consumer Financial Protection Bureau, The Consumer Credit Card Market, Executive Summary
December 2015, http://files.consumerfinance.gov/f/201512_cfpb_report-the-consumer-credit-card-market.pdf and
Federal Reserve Bank of New York, August 2017,
www.newyorkfed.org/medialibrary/interactives/householdcredit/data/pdf/HHDC_2017Q2.pdf.
5
See J.D. Power, “Attractive Rewards and Benefits Drive Credit Card Selection, Satisfaction and Spend,” press
release, August 20, 2015, www.jdpower.com/sites/default/files/2015137_U.S._Credit_Card_Study_PR_Final.pdf.
6
See, e.g., Jeanine Skowronski, “Why 5 Percent Cash Back Is an Overrated Selling Point,” American Banker,
November 6, 2013, www.americanbanker.com/opinion/why-5-cash-back-is-an-overrated-selling-point Consumer
Financial Protection Bureau, The Consumer Credit Card Market, section 7, December 2015,
http://files.consumerfinance.gov/f/201512_cfpb_report-the-consumer-credit-card-market.pdf.
3

-3chooses a zero percent interest private-label credit card with deferred interest and has a positive
balance when the promotional period expires, interest could be retroactively assessed for the full
time they held a balance during the promotional period.7 Even sophisticated consumers could be
excused for confusing these products.
As it turns out, it is often the most vulnerable consumers who have to navigate the most
complicated products. For instance, one recent study of the credit card market found that the
average length of agreements for products offered to subprime consumers was 70 percent longer
than agreements for other products.8
The complexity multiplies when we go beyond credit cards and consider other
dimensions of consumers’ financial lives.9 The Federal Deposit Insurance Corporation has
found that nearly a quarter of the Americans that don’t maintain bank accounts are concerned
that bank fees are too unpredictable.10 Even though mortgage debt is over two-thirds of
household debt, nearly half of consumers don’t comparison shop before taking out a mortgage.11
Student loans now make up 11 percent of total household debt, more than twice its share in

7

See, e.g., Consumer Financial Protection Bureau, The Consumer Credit Card Market, section 6, December 2015,
http://files.consumerfinance.gov/f/201512_cfpb_report-the-consumer-credit-card-market.pdf.
8
Id at section 5.1.2.
9
In 2007 researchers at Carnegie Mellon estimated that it would take the average consumer 244 hours a year just to
read the privacy policies of the websites that he viewed that year. Aleecia McDonald and Lorrie Faith Cranor, The
Cost of Reading Privacy Policies,” I/s: A Journal of Law and Policy for the Information Society (2008), pre-press
version available at http://lorrie.cranor.org/pubs/readingPolicyCost-authorDraft.pdf.
10
Federal Deposit Insurance Corporation, FDIC National Survey of Unbanked and Underbanked Households,
(2015), www.fdic.gov/householdsurvey/2015/2015report.pdf.
11
See, e.g., Michael Corkery and Stacy Cowley, “Household Debt Makes a Comeback in the U.S.,” New York
Times, May 17, 2017, www.nytimes.com/2017/05/17/business/dealbook/household-debt-united-states.html; See
Alexei Alexandrov and Sergei Koulayev, “No Shopping in the U.S. Mortgage Market: Direct and Strategic Effects
of Providing Information,” Consumer Financial Protection Bureau Office of Research Working Paper No. 2017-01
(Washington: CFPB, April 7, 2017), https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2948491. (“[M]any
consumers do not seem to realize that there is price dispersion. . . . In one of our counterfactuals, we show that
eliminating non-price preferences results in savings of about $9 billion dollars a year.”)

-42008.12 Over 11 percent of student debt is more than 90 days delinquent or in default--and
researchers at the Federal Reserve Bank of New York estimate that this figure may understate
the problem by as much as half.13
Today, consumers navigate numerous weighty financial responsibilities for themselves
and their dependents.14 It seems fair to assume they could use some help managing this
complexity. In the Federal Reserve Board’s annual Survey of Household Economics and
Decisionmaking (SHED), more than half of respondents reported that their spending exceeded
their income in the prior year.15 Indeed, 44 percent of SHED respondents reported that they
could not cover an emergency expense costing $400 without selling something or borrowing
money.16
New Tools to Help Consumers Manage Their Finances
Given the complexity and importance of these decisions, it is encouraging to see the fastgrowing development of advanced, technology-enabled tools to help consumers navigate the

See, e.g., Michael Corkery and Stacy Cowley, “Household Debt Makes a Comeback in the U.S.,” New York
Times, May 17, 2017, www.nytimes.com/2017/05/17/business/dealbook/household-debt-united-states.html.
(“Student borrowers today owe $1.3 trillion, more than double the $611 billion owed nearly nine years ago. About
one in 10 student borrowers is behind on repaying the loans, the highest delinquency rate of any type of loan tracked
by the New York Fed’s quarterly household debt report.”); Jesse Bricker, Lisa J. Dettling, Alice Henriques, Joanne
W. Hsu, Lindsay Jacobs, Kevin B. Moore, Sarah Pack, John Sabelhaus, Jeffrey Thompson, and Richard A.Windle,
“Changes in U.S. Family Finances from 2013 to 2016: Evidence from the Survey of Consumer Finances,” Federal
Reserve Bulletin vol. 103, no. 3 (September 2017), www.federalreserve.gov/publications/files/scf17.pdf.
13
Federal Reserve Bank of New York, Quarterly Report on Household Debt and Credit, August 2017,
www.newyorkfed.org/medialibrary/interactives/householdcredit/data/pdf/HHDC_2017Q2.pdf. ( “...[D]elinquency
rates for student loans are likely to understate effective delinquency rates because about half of these loans are
currently in deferment, in grace periods or in forbearance and therefore temporarily not in the repayment cycle. This
implies that among loans in the repayment cycle delinquency rates are roughly twice as high.” Id. n.2 ).
14
This can be particularly challenging for older consumers. See, e.g., AARP, “Financial Innovation Frontiers,” April
2017, www.aarp.org/content/dam/aarp/home-and-family/personal-technology/2017/04/Financial-InnovationFrontiers-AARP.pdf. (“50+ Consumers Face Unprecedented Financial Complexity.”)
15
Board of Governors of the Federal Reserve System, Report on the Economic Well-Being of U.S. Households in
2016 (Washington: Board of Governors, May 2017), www.federalreserve.gov/publications/files/2016-reporteconomic-well-being-us-households-201705.pdf.
16
Id.
12

-5complex issues in their financial lives. These tools build on important advances in our
understanding of consumer financial behavior and the applications, or “app,” ecosystem.
Researchers have invested decades of work exploring how consumers actually make
decisions. We all tend to use shortcuts to simplify financial decisions, and it turns out many of
these can prove faulty, particularly when dealing with complex problems.17 For example,
empirical evidence consistently shows that consumers overvalue the present and undervalue the
future.18 Researchers have documented that consumers make better savings decisions when they
are presented with fewer options.19 They have shown the importance of “anchoring” bias--the
tendency to place disproportionate weight on the first piece of information presented. This bias
can lead consumers either to make poor financial choices or instead to tip the scales in favor of
beneficial choices, as with automatic savings defaults.20 Similarly, “nudges” can help consumers
17

See, e.g., Krista Tippett, On Being, radio interview with Daniel Kahneman, October 5, 2017,
https://onbeing.org/programs/daniel-kahneman-why-we-contradict-ourselves-and-confound-each-other-oct2017/.
(“…[It] is actually completely not possible for a finite human mind to be rational or to obey the axioms of
rationality. You’d have to know too much . . . [T]he cognitive rules are, to a large extent, simplifying rules. They are
shortcuts.”)
18
See, e.g., David Laibson, Andrea Repetto, and Jeremy Tobacman, “Estimating Discount Functions with
Consumption Choices over the Lifecycle,” National Bureau of Economic Research Working Paper 13314
(Cambridge, Mass.: NBER, August 2007), www.nber.org/papers/w13314.pdf.
19
See, e.g., John Beshears, James J. Choi, David Laibson, and Bridgitt C. Madrian, “Simplification and Saving,”
National Bureau of Economic Research Working Paper 12659 (Cambridge, Mass.: NBER, 2006),
https://sites.hks.harvard.edu/fs/bmadria/Documents/Madrian%20Papers/Simplification%20and%20Saving.pdf.
(“Many financial decisions that individuals face are complicated and daunting for those who are not financial
experts. .By collapsing a multidimensional set of options into a binary choice between the status quo and the preselected alternative, this intervention increases participation rates by 10 to 20 percentage points among affected
employees.”); Sheena S. Iyengar, Wei Jiang, and Gur Huberman, “How Much Choice is Too Much?: Contributions
to 401(k) Retirement Plans,” Pension Research Council Working Paper (Philadelphia: University of Pennsylvania,
The Wharton School, PRC, 2003),
https://pdfs.semanticscholar.org/04f0/7b37fc9deb167e56c729e1f35e052998ba4a.pdf. (“Using data from nearly
800,000 employees, …[our] results confirm that participation in 401(k) plans is higher in plans offering a handful of
funds, as compared to plans offering ten or more options.”)
20
See, e.g., Daniel Navarro-Martinez, et al., “Minimum Required Payment and Supplemental Information
Disclosure Effects on Consumer Debt Repayment Decisions,” Journal of Marketing Research, vol. 48 (November
2011), http://journals.ama.org/doi/pdf/10.1509/jmkr.48.SPL.S60?code=amma-site. (“They find that while presenting
minimum required payment information has a negative impact on repayment decisions, increasing the minimum
required level has a positive effect on repayment for most consumers.”) See, e.g., Bridgette C. Madrian, “Matching
Contributions and Savings Outcomes: A Behavioral Economics Perspective,” National Bureau of Economic
Research Working Paper 18220 (Cambridge, Mass.: NBER, July 2012),
https://sites.hks.harvard.edu/fs/bmadria/Documents/Madrian%20Papers/Matching%20Contributions%20and%20Sav

-6in the right circumstances or instead backfire in surprising ways.21
These behavioral insights are especially powerful when paired with the remarkable
advances we have seen in the technological tools available to the average consumer, especially
through their smartphones.22 Smartphones are ubiquitous. The 2016 Federal Reserve Survey of
Consumer and Mobile Financial Services (SCMF) found that 87 percent of the U.S. adult
population had a mobile phone, the vast majority of which were smartphones. 23 Smartphone use
is prevalent even among the unbanked and underbanked populations. Survey evidence suggests
we are three times more likely to reach for our phone than our significant other when we first
wake up in the morning.24
Some evidence suggests that smartphones are already helping consumers make better
financial decisions. The 2016 SCMF found that 62 percent of mobile banking users checked
their account balances on their phones before making a large purchase, and half of those that did

ings%20Outcomes.pdf. (“Indeed, automatic enrollment is an extreme form of simplification; individuals who want
to save need not do anything. Psychologists have long recognized that choice complexity can affect decision-making
outcomes. One result is procrastination--individuals put off decision making as choices become more complicated.”)
(Internal citations omitted.)
21
See, e.g., Dean S. Karlan, et al, “Getting to the Top of Mind: How Reminders Increase Savings,” National Bureau
of Economic Research Working Paper 16205, (Cambridge, Mass.: NBER, July 2010),
www.nber.org/papers/w16205. See, e.g., Jialan Wang and Benjamin J. Keys, “Perverse Nudges: Minimum
Payments and Debt Paydown in Consumer Credit Cards,” Penn Wharton Public Policy Initiative Issue Brief, vol. 2.
no. 4 (Philadelphia: University of Pennsylvania, April 2014), https://publicpolicy.wharton.upenn.edu/issuebrief/v2n4.php.
22
See, e.g., Lael Brainard, “Where Do Banks Fit in the Fintech Stack,” (speech at Northwestern Kellogg Public–
Private Interface Conference on New Developments in Consumer Finance: Research and Practice),
www.federalreserve.gov/newsevents/speech/files/brainard20170428a.pdf.
23
Board of Governors of the Federal Reserve System, Consumers and Mobile Financial Services 2016,
www.federalreserve.gov/econresdata/consumers-and-mobile-financial-services-report-201603.pdf (Washington,
Board of Governors, March 2016). The survey found that 40 percent of the unbanked population and 70 percent of
the underbanked and banked populations had access to a smartphone.
24
See Bank of America, Trends in Consumer Mobility Report, 2015,
http://newsroom.bankofamerica.com/files/doc_library/additional/2015_BAC_Trends_in_Consumer_Mobility_Repo
rt.pdf. (Finding that 35 percent of survey respondents reported that they reach for their mobile devices first thing, as
compared to 10 percent for their significant other.)

-7so decided not to purchase an item as a result.25 In addition, 41 percent of smartphone owners
checked product reviews or searched product information online while shopping in a retail store,
and 79 percent of those respondents reported changing their purchase decision based on the
information they accessed on their smartphone.26
And those use cases just scratch the surface of what is possible. First of all, the
smartphone platform has become a launch pad for a whole ecosystem of apps created by outside
developers for a wide variety of services, including helping consumers manage their financial
lives.
Second, the smartphone ecosystem puts the enormous computing power of the cloud at
the fingertips of consumers. Interfacing with smartphone platforms and other apps, outside
developers can tap the computing power of the leading cloud computing providers in building
their apps. Importantly, cloud computing offers not only the power to process and store data, but
also powerful algorithms to make sense of it. Due to early commitment to open-source
principles, app developers have open access to many of the same machine-learning and artificial
intelligence tools that power the world’s largest internet companies.27 Further, the major cloud
computing providers have now taken these free building blocks and created different machinelearning and artificial intelligence stacks on their cloud platforms. A developer that wants to
incorporate artificial intelligence into their financial management app can access off-the-shelf

25

Board of Governors of the Federal Reserve System, Consumers and Mobile Financial Services 2016,
www.federalreserve.gov/econresdata/consumers-and-mobile-financial-services-report-201603.pdf.
26
Id.
27
See, e.g., Blair Hanley Frank, “Meet Google’s Cool New Natural Language Tool, Parsey McParseface,”
Computerworld, May 12, 2016, www.computerworld.com/article/3070059/artificial-intelligence/meet-googles-coolnew-natural-language-tool-parsey-mcparseface.html. See also Google Research Blog, “TensorFlow--Google’s
Latest Machine Learning System, Open Sourced for Everyone,” November 9, 2015,
https://research.googleblog.com/2015/11/tensorflow-googles-latest-machine_9.html. Cf. Tom Simonite, “Facebook
Quietly Enters Starcraft War for AI Bots, and Loses,” Wired, October 9, 2017, www.wired.com/story/facebookquietly-enters-starcraft-war-for-ai-bots-and-loses/.

-8models of cloud computing providers, potentially getting to market faster than by taking the
traditional route of finding training data and building out models in-house from scratch.
Third, fintech developers can also draw from enormous pools of data that were
previously unavailable outside of banking institutions. Consumer financial data are increasingly
available to developers via a new breed of business-to-business suppliers, called data
aggregators.28 These companies enable outside developers to access consumer account and
transactional information typically stored by banks. But aggregators do more than just provide
access to raw data. They facilitate its use by developers, by cleaning the data, standardizing it
across institutions, and offering their own application programming interfaces for easy
integration. Further, similar to cloud computing providers, data aggregators are also beginning
to provide off-the-shelf product stacks on their own platforms.29 This means that developers can
quickly and easily incorporate product features, such as predicting creditworthiness, determining
how much a consumer can save each month, or creating alerts for potential overdraft charges.
Researchers have documented the benefits of tailored one-on-one financial coaching.30
Until recently, though, it has been hard to deliver that kind of service affordably and at scale, due
to differences in consumers’ circumstances. Let’s again consider the example of deferred
interest credit cards. It turns out only a small minority of consumers miss the deadlines for
repaying promotional balances and are charged retroactive interest payments, and they typically

See, e.g., Lael Brainard, “Where Do Banks Fit in the Fintech Stack,” (speech at Northwestern Kellogg Public–
Private Interface Conference on New Developments in Consumer Finance: Research and Practice),
www.federalreserve.gov/newsevents/speech/files/brainard20170428a.pdf.
29
See, e.g., Envestnet Yodlee, “Envestnet Yodlee Unveils Personal Financial Wellness Solution Powered by Data
Intelligence,” news release, June 12, 2017, www.prnewswire.com/news-releases/envestnet--yodlee-unveilspersonal-financial-wellness-solution-powered-by-data-intelligence-300472018.html.
30
See, e.g., Brett Theodos, et al., An Evaluation of the Impacts and Implementation Approach of Financial
Coaching Programs (Washington: Urban Institute, October 2015), www.urban.org/research/publication/evaluationimpacts-and-implementation-approaches-financial-coaching-programs.
28

-9have deep subprime scores.31 Similarly, for consumers that opt into overdraft products on their
checking accounts, 8 percent of consumers pay 75 percent of the fees.32 Up until now, it has
been hard for consumers to understand those odds and objectively assess whether they are likely
to be in the group of customers that will face challenges with a particular financial product. The
convergence of smartphone ubiquity, cloud computing, data aggregation, and off-the-shelf AI
products offer the potential to make tailored financial advice scalable. For instance, a fintech
developer could pair historical data about how different types of consumers fare with a specific
product, on the one hand, with a consumer’s particular financial profile, on the other hand, to
make a prediction about how that consumer is likely to fare with the product.
The Evolution of Financial Autopilots
Since the early days of internet commerce, developers have tried to move beyond simple
price comparison tools to offer tailored “agents” for consumers that can recommend products
based on analyses of individual behavior and preferences.33 Today, a new generation of personal
financial management tools seems poised to make that leap. When a consumer wishes to select a
new financial product, he or she can now solicit options from a number of websites and mobile
apps. These new comparison sites can walk the consumer through a wide array of financial
products, offering to compare features like rewards, fees, and rates, or tailoring to a consumer’s
stated goals. Some fintech advisors ask consumers to provide access to their bank accounts,

31

Consumer Financial Protection Bureau, The Consumer Credit Card Market (Washington: CFPB, December
2015), http://files.consumerfinance.gov/f/201512_cfpb_report-the-consumer-credit-card-market.pdf.
32
Trevor Bakker, Nicole Kelly, Jesse Leary, and Eva Nagypal, Data Point: Checking Account Overdraft, Consumer
Financial Protection Bureau report (Washington: CFPB, July 2014),
http://files.consumerfinance.gov/f/201407_cfpb_report_data-point_overdrafts.pdf. Likewise, one out of five new
payday loans are rolled over six times or more. Kathleen Burke, Jonathan Lanning, Jesse Leary, and Jialan Wang,
Data Point: Payday Lending, Consumer Financial Protection Bureau report (Washington: CFPB, March 2014),
http://files.consumerfinance.gov/f/201403_cfpb_report_payday-lending.pdf.
33
See, e.g., Phil Patton, “Buy Here, and We’ll Tell You What You Like,” New York Times, September 22, 1999,
https://partners.nytimes.com/library/tech/99/09/biztech/technology/22patt.html?mcubz=1.

- 10 retirement accounts, college savings accounts, and other investment platforms in order to enable
a fintech advisor to offer a consumer a single, near complete picture of his balances and cash
flows across different institutions.
In reviewing the advertising, terms and conditions, and apps of an array of fintech
advisors, it appears that many of these tools offer advanced data analysis, machine learning, and
even artificial intelligence to help consumers cut down on unnecessary spending, set aside
money for savings, and use healthy nudges to improve their financial decisions. For instance, a
fintech advisor may help a consumer automate savings “rules,” like rounding up charges and
putting the difference into savings, enabling these small balances to accumulate over time or
setting a small amount of money aside every time a consumer spends money on little splurges.
The early stages of innovation inevitably feature a lot of learning from trial and error.
Fortunately, as the fintech ecosystem advances, there are useful experiences and good practices
to draw upon from the evolution of the commercial internet. To begin with, one internet adage is
that if a product is free, “you are the product.”34 In this vein, fintech advisors frequently offer
free services to consumers and earn their revenue from the credit cards and other financial
products that they recommend through lead generation.
Of course, many fintech advisors are not lead generators. Some companies offer fee-forservice models, with consumers paying a monthly fee for the product. Other companies are paid
by employers, who then provide the products free of charge to their employees as an employee
benefit. In these cases, they likely have quite different business models.

See, e.g., John Lanchester, “You Are the Product,” London Review of Books, vol. 39, no. 16-17, August 17, 2017,
www.lrb.co.uk/v39/n16/john-lanchester/you-are-the-product. Cf., Monica Anderson, “Key Takeaways on Mobile
Apps and Privacy,” Pew Research Center FactTank, November 10, 2015, www.pewresearch.org/facttank/2015/11/10/key-takeaways-mobile-apps/.
34

- 11 But for those services that do act as lead generators, there are important considerations
about whether and how best to communicate information to the consumer about the nature of the
recommendations being made. For instance, according to some reports, fintech advisors can
make between $100 and $700 in lead generation fees for every customer that signs up for a credit
card they recommend.35
In many cases, a fintech advisor may describe their service as providing tailored advice
or making recommendations as they would to friends and family. In such cases, a consumer
might not know whether the order in which products are presented by a fintech assistant is based
on the product’s alignment with his or her needs or different considerations. Different fintech
advisors may order the lists they show consumers using different criteria. A product may be at
the top of the advisor’s recommendations because the sponsoring company has paid the advisor
to list it at the top, or the sponsoring company may pay the fintech assistant a high fee,
contingent upon the consumer signing up for the product. Alternatively, a fintech advisor may
change the order of the loan offers or credit cards based on the likelihood that the consumer will
be approved. Moreover, in some cases, the absence of lead generation fees for a particular
product may impact whether that product is on the list shown to consumers at all.
There appears to be a wide variety of practices regarding the prominence and placement
of advertising and other disclosures relative to the advice and recommendations such firms

See, e.g., Nathaniel Popper, “Automated Assistants Will Soon Make a Bid for Your Finances,” New York Times,
December 7, 2016, www.nytimes.com/2016/12/07/business/dealbook/automated-assistants-will-soon-make-a-bidfor-your-finances.html.
35

- 12 provide. Overall, fintech assistants have increasingly improved the disclosures that explain to
consumers how they get paid, but this is still a work in progress.
The good news is that these challenges are not new. The experience with internet search
engines outside of financial products, such as Google, Bing, and Yahoo!, as well as with other
product comparison sites, such as Travelocity and Yelp, may provide useful guidance. As
consumers and businesses have adapted to the internet, we have, collectively, adopted norms and
standards for how we can expect search and recommendation engines to operate. In particular,
we generally expect that search results will be included and ranked based on what’s organically
most responsive to the search--unless it is clearly labeled otherwise.36 Accordingly, when we
search for a product, we now know to look for visual cues that identify paid search results,
usually in the form of a text label like “Sponsored” or “Ad”, different formatting, and
visually separating advertising from natural search results.37 Even when an endorsement is made
in a brief Twitter update, we now expect disclosures to be clear and conspicuous.38
As fintech advisors evolve to engage consumers in new ways, disclosure methodologies
will no doubt be expected to adapt as well. For instance, some personal financial management

Cf., Federal Trade Commission, “Mary Engle Letter to Search Engine Providers,” n.4, June 24, 2013,
www.ftc.gov/sites/default/files/attachments/press-releases/ftc-consumer-protection-staff-updates-agencys-guidancesearch-engine-industryon-need-distinguish/130625searchenginegeneralletter.pdf; Federal Trade Commission, “FTC
Consumer Protection Staff Update Agency’s Guidance to Search Engine Industry on the Need to Distinguish
Between Advertisements and Search Results,” press release, June 25, 2013, www.ftc.gov/news-events/pressreleases/2013/06/ftc-consumer-protection-staff-updates-agencys-guidance-search; Federal Trade Commission,
“Commercial Alert Response Letter,” June 27, 2002, www.ftc.gov/enforcement/cases-proceedings/closingletters/commercial-alert-response-letter.
37
Cf., Federal Trade Commission, “Mary Engle Letter to Search Engine Providers,” n.4, June 24, 2013,
www.ftc.gov/sites/default/files/attachments/press-releases/ftc-consumer-protection-staff-updates-agencys-guidancesearch-engine-industryon-need-distinguish/130625searchenginegeneralletter.pdf.
38
Cf., Federal Trade Commission, “The FTC’s Endorsement Guides: What People Are Asking,” September 2017,
www.ftc.gov/tips-advice/business-center/guidance/ftcs-endorsement-guides-what-people-are-asking; Federal Trade
Commission, “.com Disclosures: How to Make Effective Disclosures in Digital Advertising,” March 2013,
www.ftc.gov/sites/default/files/attachments/press-releases/ftc-staff-revises-online-advertising-disclosureguidelines/130312dotcomdisclosures.pdf.
36

- 13 tools now interact with consumers via text message. If consumers move to a world in which
most of their interactions with their advisors occur via text-messaging “chatbots”--or voice
communication--I am hopeful that industry, regulators, consumers, and other stakeholders will
work together to adapt the norms to distinguish between advice and sponsored recommendations.
The Data Relationship
While the lead generation revenue model presents some familiar issues that are readily
apparent, under the hood, fintech relationships raise even more complex issues for consumers in
knowing who they are providing their data to, how their data will be used, for how long, and
what to expect in the case of a breach or fraud. Let me briefly touch on each issue in turn.
Often, when a consumer signs up with a fintech advisor or other fintech app, they are
asked to log into their bank account in order to link the fintech app with their bank account data.
In reviewing apps’ enrollment processes, it appears that consumers are often shown log-in
screens featuring bank logos and branding, prompting consumers to enter their online banking
logins and passwords. In many cases, the apps note that they do not store the consumers’
banking credentials.
When the consumer logs on, he or she is often not interfacing with a banks’ computer
systems, but rather, providing the bank account login and password to a data aggregator that
provides services to the fintech app. In many cases, the data aggregator may store the password
and login and then use those credentials to periodically log into the consumer’s bank account and
copy available data, ranging from transaction data, to account numbers, to personally identifiable
information. In other cases, things work differently under the hood. Some banks and data
aggregators have agreed to work together to facilitate the ability to share data with outside

- 14 developers in authorized ways. These agreements may delineate what types of data will be
shared, and authorization credentials may be tokenized so that passwords are never stored by the
aggregator.39
It is often hard for the consumer to know what is actually happening under the hood of
the financial app they are accessing. In most cases, the log in process does not do much to
educate the consumer on the precise nature of the data relationship. Screen scraping usually
invokes the bank’s logo and branding but infrequently shows the logo or name of the data
aggregator. In reviewing many apps, it appears that the name of the data aggregator is frequently
not disclosed in the fintech app’s terms and conditions, and a consumer generally would not
easily see what data is held by a data aggregator or how it is used. The apps, websites, and terms
and conditions of fintech advisors and data aggregators often do not explain how frequently data
aggregators will access a consumer’s data or how long they will store that data.40
Recognizing this is a relatively young field, but one that is growing fast, there are a
myriad of questions about the consumer’s ability to opt out and control over data that will need
to be addressed appropriately. In examining the terms and conditions for a number of fintech
apps, it appears that consumers are rarely provided information explaining how they can

See, e.g., Finicity, “Finicity and Wells Fargo Ink Data Exchange Deal,” press release, April 4, 2017,
www.finicity.com/press-release-finicity-wells-fargo-ink-data-exchange-deal/; Wells Fargo & Co., “Intuit Signs New
Data-Exchange Agreement with Wells Fargo,” press release, February 3, 2017,
www.wellsfargo.com/about/press/2017/intuit-agreement_0203/; Intuit, “Chase, Intuit to Give Customers Greater
Control of Their Information,” press release, January 25, 2017, www.intuit.com/company/press-room/pressreleases/2017/Chase-Intuit-to-Give-Customers-Greater-Control-of-Their-Information/; Wells Fargo & Co., “Wells
Fargo, Xero Agree on New Data-Exchange Method,” press release, June 7, 2016,
www.wellsfargo.com/about/press/2016/new-dataexchange-method_0607/; Silicon Valley Bank, “Xero and Silicon
Valley Bank Partner to Offer Innovative Companies Next-Generation Financial Management,” press release, July
16, 2014, www.svb.com/News/Company-News/Xero-and-Silicon-Valley-Bank-Partner-to-Offer-InnovativeCompanies-Next-Generation-Financial-Management/.
40
See, e.g., Penny Crosman, “Data-Sharing Debate Grows Contentious as Fintechs Vent Grievances,” American
Banker, August 15, 2017, www.americanbanker.com/news/data-sharing-debate-grows-contentious-as-fintechs-ventgrievances.
39

- 15 terminate the collection and storage of their data. For instance, when a consumer deletes a
fintech app from his or her phone, it is not clear this would guarantee that a data aggregator
would delete the consumer’s bank login and password, nor discontinue accessing transaction
information. If a consumer severs the data access, for instance by changing banks or bank
account passwords, it is also not clear how he or she can instruct the data aggregator to delete the
information that has already been collected. Given that data aggregators often don’t have
consumer interfaces, consumers may be left to find an email address for the data aggregator,
send in a deletion request, and hope for the best.
If things go wrong, consumers may have limited remedies. In reviewing terms, it appears
that many fintech advisors include contractual waivers that purport to limit consumers’ ability to
seek redress from the advisor or an underlying data aggregator. In some cases, the terms and
conditions assert that the fintech developer and its third-party service providers will not be liable
to consumers for the performance of or inability to use the services. It is not uncommon to see
terms and conditions that limit the fintech adviser’s liability to the consumer to $100.
Traditionally, under the Electronic Funds Transfer Act and its implementing Regulation
E, consumers have had protections to mitigate their losses in the event of erroneous or fraudulent
transactions that would otherwise impact their credit and debit cards, such as data breaches.
Those protections are not absolute, however.41 In particular, if a consumer gives another person
an “access device” to their account and grants them authority to make transfers, then the
consumer is “fully liable” for transfers made by that person, even if that person exceeds his or

See 12 CFR section 1005.2(m) (1). (Excluding from the definition of “unauthorized electronic fund transfer” any
“electronic fund transfer initiated… [by] a person who was furnished the access device to the consumer’s account by
the consumer…..”)
41

- 16 her authority, until the consumer notifies the bank.42 As the industry matures, the various
stakeholders will need to develop a shared understanding of who bears responsibility in the event
of a breach.43
Shared Responsibility and Shared Benefit Moving Forward
So what can be done to make sure consumers have the requisite information and control
to remain squarely in the driver’s seat? Establishing and implementing new norms is in the
shared interest of all of the participants in the fintech stack. For instance, in the case of credit
cards, mortgages, and many other products, it is often banks or parties closely affiliated with
banks that pay fees to fintech advisors to generate leads for their products, pursuant to a contract.
Through these contractual relationships with fintech advisors, banks have considerable influence
in the lead generation relationship, including through provisions describing how a sponsored
product should be described or displayed. Banks have a stake in ensuring that their vendors and
third-party service providers act appropriately, that consumers are protected and treated fairly,
and that the banks’ reputations aren’t exposed to unnecessary risk.44 Likewise, some of the
leading speech-only financial products are currently credit card and bank products.45

Comment 2(m)-2 clarifies that if “a consumer furnishes an access device and grants authority to make transfers to
a person . . . who exceeds the authority given, the consumer is fully liable for the transfers unless the consumer has
notified the financial institution that transfers by that person are no longer authorized.”
43
See Kaitlin Asrow and Beth Brockland, Liability, Transparency and Consumer Control in Data Sharing
(Washington: Center for Financial Services Innovation, September 18, 2017),
http://cfsinnovation.org/research/liability-transparency-consumer-control-in-data-sharing/. (“There are differing
interpretations of how regulations, such as the OCC’s Third Party Risk Management guidance, the Gramm-LeachBliley Act’s privacy and safeguard rules, and Regulation E liability limits, should be applied to data sharing between
financial institutions and third parties like data aggregators and financial technology applications.”)
See also JPMorgan Chase, “How You Can Protect Yourself,” (last visited October 18, 2017),
www.chase.com/digital/resources/privacy-security/security/how-you-can-protect; Liz Weston, “Why Banks Want
You to Drop Mint, Other ‘Aggregators,’” Reuters, November 9, 2015, www.reuters.com/article/us-column-westonbanks/why-banks-want-you-to-drop-mint-other-aggregators-idUSKCN0SY2GC20151109.
44
See, e.g., Board of Governors of the Federal Reserve System, “Guidance on Managing Outsourcing Risk,” SR
letter 13-19/CA letter 13-21, December 5, 2013, www.federalreserve.gov/supervisionreg/srletters/sr1319a1.pdf.
45
See, e.g., American Express, “Introducing the Amex Skill for Amazon Alexa” (last visited October 18, 2017),
www.americanexpress.com/us/content/alexa/.; Leena Rao, “American Express Debuts Its First Amazon Alexa
42

- 17 Accordingly, banks have incentives to invest in innovating the way they disclose information to
consumers, as they also invest in new ways of interacting with them.
As for consumers’ relationships with data aggregators, there’s an increasing recognition
that consumers need better information about the terms of their relationships with aggregators,
more control over what is shared, and the ability to terminate the relationship.46 We have spoken
to data aggregators who recognize the importance of finding solutions to many of the complex
issues involved with the important work of unlocking the potential of the banking stack to
developers. And while there are some difficult issues in this space, other issues seem relatively
straightforward. It shouldn’t be hard for a consumer to be informed who they are providing their
credentials to. Consumers should have relatively simple means of being able to consent to what
data are being shared and at what frequency. And consumers should be able to stop data sharing
and request the deletion of data that have been stored.
Responsibility for establishing appropriate norms in the data aggregation space should be
shared, with banks, data aggregators, fintech developers, consumers, and regulators all having a
role.47 Banks and data aggregators are negotiating new relationships to determine how they can
work together to provide consumers access to their data, while also ensuring that the process is

Skill,” Fortune, May 11, 2017, http://fortune.com/2017/05/11/american-express-alexa-skill/; Capital One,
“CapitalOne is on Amazon Echo” (last visited October 18, 2017), www.capitalone.com/applications/alexa/.
46
See, e.g., Consumer Financial Protection Bureau, “Consumer Protection Principles: Consumer-Authorized
Financial Data Sharing and Aggregation,” October 18, 2017,
http://files.consumerfinance.gov/f/documents/cfpb_consumer-protection-principles_data-aggregation.pdf; Kaitlin
Asrow and Beth Brockland, Liability, Transparency and Consumer Control in Data Sharing, Center for Financial
Services Innovation, September 18, 2017, http://cfsinnovation.org/research/liability-transparency-consumer-controlin-data-sharing/.
47
Cf. Kaitlin Asrow and Beth Brockland, Liability, Transparency and Consumer Control in Data Sharing, Center
for Financial Services Innovation, September 18, 2017, http://cfsinnovation.org/research/liability-transparencyconsumer-control-in-data-sharing/.

- 18 secure and leaves consumers in the driver’s seat.48 In many cases, banks themselves were often
the original customers of data aggregators, and many continue to use these services. According
to public filings, more than half of the 20 largest banks are customers of data aggregators.49 The
banks have an opportunity as customers of data aggregation services to ensure that the terms of
data provision protect consumers’ data and handle it appropriately.
Regulators also recognize that there may be opportunities to provide more clarity about
how the expectations about third-party risk management would work in this sector, as well as
other areas experiencing significant technological change. Through external outreach and
internal analysis, we are working to determine how best to encourage socially beneficial
innovation in the marketplace, while ensuring that consumers’ interests are protected. We
recognize the importance of working together and the potential to draw upon existing policies,
norms, and principles from other spaces. Consumers may not fully understand the differences in
regulations across financial products or types of financial institutions, or whether the rules
change when they move from familiar search and e-commerce platforms to the fintech stack.
Consumers, as well as the market as a whole, will benefit if regulators coordinate to provide
more unified messages and support the development of standards that serve as a natural
extension of the common-sense norms that consumers have come to expect in other areas of the
commercial internet.

See, e.g., Penny Crosman, “Data-Sharing Debate Grows Contentious as Fintechs Vent Grievances,” American
Banker, August 15, 2017, www.americanbanker.com/news/data-sharing-debate-grows-contentious-as-fintechs-ventgrievances.
49
Envestnet, Inc., Form 10-K, March 17, 2017, www.envestnet.com/report/2016/download/Envestnet2016AnnualReport-Form10-K.pdf.
48

- 19 Conclusion
The combination of technologies that put vast computing power, rich data sets, and
artificial intelligence onto simple smartphone apps together with important research into
consumer financial behaviors has great potential to help consumers navigate their complex
financial lives more effectively, but there are also important risks. I am hopeful that fintech
developers, data aggregators, bank partners, consumers, and regulators will work together to
keep consumers in the driver’s seat as we move forward with these new technologies. If we
work together effectively toward this goal, the fintech stack may be able to offer enormous
benefits to the consumers they aim to serve, while appropriately identifying and managing the
risks.