[SQUEAKING] [RUSTLING] [CLICKING] GARY GENSLER: Welcome
back to fintech. We're getting together
for our fourth class, and we're going
to get going now. Thank you for tying in online
via this modern technology. Today we're going to move to
important customer interfaces-- user interfaces and
user experiences that have been greatly enhanced
by open application program interfaces. Now, application
program interfaces is not new to finance, and it
didn't start in M&A finance. It's a basic concept
of how one computer is talking to another computer. And again, this is not a
computer science class, but we're going to spend
a little bit of time on what application
program interfaces are and what this
concept of open API is and some of the
companies around it and how companies have
organized themselves, both in terms of finance,
and a little bit even outside of finance and some
of the challenges with regard to that. What API has also led
to is a movement of and a greater concentration
around companies called data aggregators. And we'll get into that as well. As I've sort of worked through
with the class already, it's my thought that
certain technologies have been moving into the
finance technology stack, and are really part and parcel
to what finance is in 2020. And I think open API, with its
movement around the globe-- and some countries have
moved further in it, some countries have moved
less, some companies have moved further, some companies
have moved less-- that this is a
trend that will be fully incorporated into the
technology stack of finance country by country, if it
hasn't already been so. But it's been a major disruptor
in the last five, seven years. And it will continue to
be a bit of a disruptor for the next handful of years. By the end of this
decade, we'll look back, we'll say that that's
probably already happened, and there'll be
enhancements around it. But we'll be fully inside of it. Maybe not as dramatic as
artificial intelligence and machine learning
to our minds-- thinking of machines
extracting correlations better than we can. But I think it's
fairly dramatic, in terms of the
advancement of finance and all that we
have in front of us. So that's why I decided to
take one class and one lecture on it. I also want to thank everybody. Romain put out a survey. About a third of you took it. It was a voluntary survey. And it's helpful. And what I took back from
the desires of students were could I do a little bit more
talking about the companies in the fintech space? And maybe add, if I might say,
a little less, whether it's on the broader technology
or broader policy side. I'll do my best. But I also want
to share with you why I think there's
a reason to have some foundation-- some
foundation of what the technologies are,
or even some foundation on the public policy
side, as we move into the latter
part of this class. The latter part of this class is
really to talk sector by sector what the competitive landscape
is, which companies are up, which companies are
down, and so forth. But as we've already
talked about a bit today, the companies really matter. And I'll sort of interject
that more as we go forward during this semester. One other question is whether we
could have some guest speakers. I'm kind of going to have
to apologize in advance. This is kind of what you get. But over the next
handful of sessions, if I come up with somebody
to bring in-- it's a little hard at this juncture
to bring in guest speakers. But I certainly
appreciate the request. And with four weeks
to go, you never know. But with that, I'm going to
start to share some slides. I don't know, Romain
if you have anything you took from the feedback,
which was very helpful. PROFESSOR 2: Thank you. Thank you very much for
filling out the survey. That was great. Just two small
announcements from my side. One, a reminder for
your team formation. Friday is the deadline. Please register your teams
in Canvas in the People's tab using the Groups function. If you don't know
how to do that, I'll be happy to show you. So please register your teams
of three to four students for the group assignments. And then second reminder,
as per the syllabus, have one of your team
members just send me an email saying what is the
topic that you wish to study for your group assignments. Please don't forget to do that. The deadline is also
Friday, April the 10th. Thank you. GARY GENSLER: Thank you. And part of that is
just to really help you engage in this subject. To Romain and myself, it doesn't
matter which sector you pick or whether you choose
to sort of write your paper from the perspective
of incumbents, to big tech, or to a venture capitalist. But it's really to
help you sort of engage in the world of fintech,
that which we're studying, and to pick one sector. And are you taking it from
the entrepreneurial strategy of the disruptor,
or are you taking it from a big firm, either
incumbent or a tech firm's perspective in this regard? So as I said, we're going
to talk a little bit about application
program interface, how this fits into the
history of finance, deeper dive in open
API and open banking. You were all asked
to prepare and sort of look at the Bank of
International Settlements report. So I'm not going to dive in. I'm going to hopefully
assume that you've had a chance to engage with that. And then talk a little bit
also about data aggregators and a movement amongst many
in large financial firms to have a standard
setting group-- a financial data
exchange on open APIs. And then try to close out a
little bit on robotic process automation. So a lot to cover. But I think we'll keep it going. And I ask again,
use the Participate button to raise the blue hands. Use the chat function. Romain's going to
keep interrupting me. And we'll have hopefully
a good session together. So the readings, again, Bank
of International Settlements, this consortium of 60 or so
central banks around the globe, stands up reports
and studies from time to time on new challenges
they're seeing. So just that they pick this
topic to write a report that came out this past
November of 2019 is indicative right
there in and of itself that it's a movement
around the globe that they thought they needed to study,
and an important aspect. I picked one article
out of the many articles that you can grab off the
internet about fintech for the next decade. And this was one article
that I thought was relevant, because it really
did also include this concept of the interface--
the customer user interface and what was going on
with regard to APIs. And then one
article, again, just to give you sort of a primer
on robotic process automation that I think is
relevant as well. So again, just if anybody wants
to give a little short brief as to what are the major
trends on the marketing side that's affecting financial
services right now. And we can go broader
than API, of course. But Romain anybody want
to sort of contribute now? PROFESSOR 2: Let's
see, who will be our first volunteer of the day? Michael. AUDIENCE: Not sure if this
is exactly marketing side, but I think customer
trust is maybe more-- must have trust in
connecting your personal data to a third party. I'm thinking something
like mint.com, where I add my bank account. I can connect it
and feel comfortable giving that information to a
site that's not necessarily my bank or my credit card. GARY GENSLER: Absolutely. Absolutely. So one of the major changes-- and this has been
a number of years, as companies like
Mint and others, where you'll give
your data to a company that you've personally
never met a human being. You've downloaded an app
and you're giving your data to that company,
to give you advice or to give you a better rate. Have you used Mint, by the way? And do you mind just saying what
product were you shopping for, or what service were
you looking for? AUDIENCE: I was just trying
to track my spending because I was spending way too much. GARY GENSLER: All
right, so there are also sites where it can help you
keep your personal budget, if you wish-- like personal
accounting and so forth. And you give them permission
to look into your bank account. But this trend of using-- and there many
competitors of Mint that we'll dive into in
the next couple weeks. But there are many competitors. It's basically
saying, all right, I am going to
permission a website-- a data service provider,
fintech company-- I'm going to permission them
to look inside my bank account. I have to give them my password
and my username in some way and have them linked in. Big change. It takes away a little of the
comparative advantage possibly of the big incumbents, because
you and I can give somebody else permission to go in
and look at that data. PROFESSOR 2: [? Iqbal? ?] AUDIENCE: Hi. Morning, professor. I just want to
highlight that there's an avenue to market or to
showcase financial services. It's starting to be on
a lifestyle platform. So if you look at, for example,
Alipay, it's all about access to [INAUDIBLE] online shopping. And WeChat's about
access to its gaming. But there's a lot of these
conflicting financial services and key services
that people want. So it's nothing new. In fact, Walmart and even
now [? Star Market ?] has banks opened up
within its grocery stores. So I think there'll
be a trend that will be for the [INAUDIBLE]. GARY GENSLER: And so what
[? Iqbal ?] has raised is not new-- decades old-- that even big retailers,
bricks and mortar retailers, thought, well, should we
have a counter for you to do some finance? Should we embed finance
into the retailing business with the trends with
in-store and branded credit cards for decades? But now online
embedded finance inside of Alipay, which was mentioned
just then, and other companies, to embed the financial
aspects right in their product offerings. And some of that
embedded finance is an Amazon credit card or
an Alipay payment mechanism. Some of the embedded
finance can even be that if you're buying
a mattress from Casper-- I think I may have
referenced this already-- but if you're buying a
mattress from cash online, that they might offer what's
called point-of-sale financing. Now, you and I don't
think of it that way. They just say, you can pay
over four installments. But in fact, they're lending
you money for those four installments, and it's
embedded right in the product. And Uber and Lyft actually
embed some finance right in their app. Now, they've built it on top
of credit card companies. But they could
choose to build it on top of their own sort
of financing mechanism with their drivers and the like. So big change, embedded finance. Multi decades in
coming, but now it's embedded finance in
a technological way that it looks
seamless to the user-- or can look seamless. PROFESSOR 2: Nikhil. AUDIENCE: The other
piece that I thought was interesting was the article
talk about first the uncoupling of financial services. So a lot of fintech incumbents
decouple the product offerings. But then there's a resurgence
of bundling products again. So when you look at
Wealthfront, they started with just
asset management. But now they're offering
cash management. Robinhood's the same way. It came out trading,
but again, they're trying to pivot to
cash management. So they focus on very niche
products to begin with, but there seems to be
bundling happening again. GARY GENSLER: Right. And I think this is
unbundling and rebundling happens in many industries. Some of the proponents in
the most optimistic side of financial
technology say, we can have a flattening of finance. We can have a full
unbundling, where you can get each of your
products separately. And while the
technology can and does facilitate that, I
think our tendencies in markets and as humans,
and our behavioral science tells us that bundling happens. So even if this transitional
period of time-- even if this late 20-teens
and early 2020's allows for a lot of unbundling, I
believe that a lot of this will stay bundled and rebundled
within particular websites, particular product offerings. I don't know of Leonora
is trying to speak, because your name's popping
up here on the screen. But if not-- how would
open API initiatives influence all of this? I've sort of said it
a little bit myself, but anybody want to chime in? PROFESSOR 2: Geetha? AUDIENCE: Yeah, I think I
speak from the incumbent side. I think open APIs have made
it so easy from a banking perspective, in terms of ability
to track what is being accessed of customer data. Because before APIs
and before some of the standouts
like OAuth and OIDC, they were pretty much
dependent on screen scraping. And it was very difficult
for us to track. And the customers pretty much
gave their user ID and password to these vendors. And they could pretty much log
in and screen scrape anything, and we won't even know. GARY GENSLER: So you're saying
Gita, from the incumbent side-- from the big financial firm
side, these new startups-- the Mints of the
world, so to speak-- were able to get folks'
usernames and passwords, and come into the
incumbent systems, and grab the data
through a method called screen scraping, or
even through a robotic process automation, which we're
going to chat about. And you're saying you feel from
an incumbent side it's better to manage all of that-- those interfaces, which
could be disruptive. You'd have more awareness
and more security around that interface. AUDIENCE: Correct. GARY GENSLER: And then
how does it influence it from the consumer side? Anybody want to throw an
idea from the customer side? I agree with what Gita's saying. PROFESSOR 2: So I'm not
sure it will be related, but Yi had his hand up. GARY GENSLER: Sure. AUDIENCE: Yeah, I
think the benefits is more for the fintech company
rather than the incumbent. Because it basically
lowered the entry barriers. So I guess from the
consumer perspective, now you have many more options. Previously, you can only work
with those like large banks. But now you can work with many
more companies that are there. So the market is more
competitive, basically. GARY GENSLER: So if I
permission a company like Mint, Mint can give me advice around
my cash flow management. If I have permission
a company like so SoFi, which is, frankly,
competing with the banks-- competing with lending products. But if my SoFi to
see my accounts, then I get more
options as a consumer. SoFi might compete
with the banks, but also give me a better user
interface-- user experience as well. So it's far broadened out
the world for the consumers. And then I put a
couple of readings in about robotic process
automation and screen scraping. Gita's already
raised this a bit. And it's being used. It's also being used by
the financial sector. So anybody want to add how you
think about robotic process automation in a positive way? In a way from the incumbents
and the financial firms, how they're using it. It's certainly also being
used by the disruptors trying to get information
out of the incumbents. They're using it as sort of
an automated screen scraping to grab information
from the incumbents. But does anybody
want to sort of say how maybe the
incumbents are using it in their business models? PROFESSOR 2: Hassan. AUDIENCE: Hi, professor. Yeah, I just have a
question about the APIs. I don't actually totally
understand what are these. I mean, is it like
an operating system? I mean, what's an API exactly? I've tried to google it, but
I've not quite understood. GARY GENSLER: So I'm
going to dive into it. Can I hold that, Hassan,
just for a minute? Because I'm going to dive into
it in the next six minutes. But fundamentally, an
application program interface is a method that two
computers can speak to. And the traditional
way, and the old way, it was just how one system
can go in and get information, get objects out of
the other system. And then we'll do a
little cursory dive into what it looks like today. Because it's moved a lot
from that traditional method. On a web-based API system, you
can actually sort of almost-- it feels like on your computer,
on your laptop, so to speak, that you're actually operating
that system at a distance. So you sort of think of it like
it's not only just grabbing some data or objects or
records, but it's actually sort of grabbing
some application, as if it feels
like you're really operating it on your system. Open API is this
conceptual framework that it's a public
interface, rather than a private interface. And if you think that
you could literally have just a restricted private
interface that you give one other computer, whether
it's an affiliate of yours or a customer of yours,
a contractual private interface to do that which
they wish to do, or is it a public interface, where
broadly it can be interfaced? And so that's sort
of a thumbnail. But let me sort of develop
that a little bit more over the next half hour. Does that help? AUDIENCE: Yeah. Thank you very much. PROFESSOR 2: So to answer
your question, Gary, we have Ivy and then Andrea. GARY GENSLER: OK. And then I'll move on. Thanks, Ivy. AUDIENCE: Yeah, I just wanted
to answer the question. Because I thought the
Medium article did a good job of talking
about the different ways that are banking use cases. And I've actually
gone through a lot in terms of just
on the banking side like actually having to do
customer onboarding and KYC. And like we actually have
what's called a KYC analyst. So all this is actually
very, very manual typically. And going through
compliance checks, we'll be working on a deal,
and at the last minute, you can't get compliance
checks through. And it's usually a huge pain. So I was excited to see
that they could potentially change the system,
because it ends up being a huge headache
for everyone. I think my question, though, is
like, when I read about this, obviously it seems like-- I'm sure the
technology is there. But I think about just like
any bulge bracket banks-- their IT infrastructure
has been in place for like 50-plus years. And changing that
technology is like removing one piece of a really
big complicated puzzle. So I almost just question
how feasible that is. Like the article does
a really good job of laying out what we
can actually automate. But I guess I am a
little bit skeptical as to like when that
time will come. GARY GENSLER: And Ivy,
may I ask which bank or what country you were
working in that was doing this? AUDIENCE: I worked at
Merrill and Barclays. GARY GENSLER: So Merrill-Lynch
or Barclays, you're right, they're incumbents. And they can
automate many things. The technology is in advance
of what they are actually doing right now. But robotic process automation
can provide both incumbents and disruptors. But the Merrills and
Barclays are already using it to some
extent to automate that which humans are doing. Some of it, as we'll
dive into later on, is about reading documents. Just literally an object
reading a document. You could say that one form
of robotic process automation that we already have is when
we can deposit a check by using our mobile phone. Now, that also uses natural
language processing, and it sort of reads that into
the computer, and so forth. But you're absolutely right
that each in incumbent is dedicated to ensuring that
if they're moving from a legacy system to a new system, that
they do that in their own way, so that it doesn't
break in the middle, and that they actually
do save money. And they have to schedule it. And they might not get to
that project for 18 months or even five or seven years. And sometimes that's what
gives the fintech disruptor the opportunity-- that the incumbent can
use a new technology. In this case, RPA. They can use that technology. But they can't schedule it for
three years or even longer. Or if they're
scheduling it, they have to do it in a
very controlled way. And the disruptor
then is coming along and having an opportunity
in the meantime to attack that business model. So I thank you. Romain are there others,
or should I move on? PROFESSOR 2: We finally
had Andrea, who I believe has experience on this topic. So we could all benefit from it. AUDIENCE: Yeah, but I wanted
to add a similar point that we just discussed with Ivy. So we can move on definitely. But what I appreciate about
the RPA in the incumbents is that it helps
them to increase the efficiency in operations. But at the same time, it also
brings in the standardization. And people can focus--
or employees of the bank can focus on more
value-adding services. For example, the advisory
to the client when it comes to more
difficult products, such as mortgages
or consumer loans. But at the same time,
as you mentioned, the time and the resources
that you need, especially in huge incumbents like
Merrill or Barclays, it takes a lot of
time to get there. GARY GENSLER: Right. And so what's interesting about
each of these technologies-- and this is true if
we broaden it out-- open API, robotic
process automation, they're both a tool
for the disruptors to produce services
and new business models to the consuming public,
And, they're also tools to the
incumbents, depending upon on how they can
invest in those tools, how they can update their
business models, and the like. And in particular in open
API, we'll see there's a real challenge, that the
large incumbents-- the JPMorgan Chases and the
PNC Banks and so forth-- are sharing data now with
a lot of the disruptors. They're providing greater
services to their ultimate end users and customers. But then also they're
concerned about, are they giving up some of their
inherent competitive advantage of controlling that data? And so there's some
real trade offs, particularly in
the open API world, as to how banks look at this. So I wanted to go back to the
question of like, what is API? I just was able to find
a slide of TechTarget that I thought was helpful,
just to think about this. An API, just think
of two systems that are linking together. And sort of the original APIs
was just a local application program interface that sort
of could connect two computers together. Web-based API, which went
further-- and web-based APIs have been around for 30 years. This is not a new technology. But web-based are
designed really to represent such that
they could be widely used. HTML now and later
something called REST API, which is more
recently, these are protocols. These are protocols that
multiple parties on the web can come into your
service in some way. And then we talk
about public APIs, which really sort of
broadens it out further. And you could even
think of this as sort of private registered APIs
to broadly publicly available APIs. And what many businesses
are trying to use-- and they manage it. And it's a very active
part of strategy, is to manage that interface
to allow the developer community to interface
with your platform. Facebook needs to do it. Instagram needs to do it. Any network-based or
platform-based service actually wants to encourage developers
to build applications on top of the platform. So now you bring
this to finance. Does JPMorgan Chase, does Bank
of America, does Barclays, does Merrill want to do the same-- encourage development
of applications on top of their platform? Recognizing that they make
a lot on their platform-- their balance sheet, their
funding models, their data, their analytics. Do they want to encourage the
development of applications on top? And it's a little bit of a push
and pull, because they're also encouraging development
of applications that might compete
with them, but also enhance what they're doing. So I sort of found this
other spaghetti slide that I thought was
really helpful. Now, if you think
about thousands of financial firms dealing
with thousands of public, this spaghetti wire
gets even worse. But application
program interface is a classic problem of math. It's that if you only have two
computers by two computers, you might need four
lines of connections. But all of a sudden
when you have thousands, you could have this
spaghetti wire of interfaces. So what's happened in the midst
of this in sector by sector is that there have
been companies that really have an API
management layer in between. And in many sectors,
you've seen a consolidation around either one standard
for APIs or companies that are really service
providers in the middle. And here are some
examples in non-finance. And whether you know
underneath the hood about Twilio or others-- Twilio probably has,
I'm told, something close to an 80% market share
in voice and messaging. That they're sort of that
layer in between the disruptors and the service providers
and the classic companies, providing a layer in the middle. A layer in the middle. Checkr does it on employment
checks when we want a check. Each one of these companies
were startups over the last 20 years. Salesforce is a
little bit earlier. And they really took-- even if they didn't
know it at the time, they took the concept of open
API and API and they built it. And in fact, there's these
words they call API first-- that you can build
a strategy saying, we're going to be the
best providing APIs. Now, who's done this in finance? Who's done this in finance? One company that you might
think of as a payment service provider, Stripe,
did it very well. Stripe basically said, we
will be that API provider in the payment space. What company did it really well
most recently that we talked about last class was Plaid. And we're going to come
back into Plaid as well. But Plaid did this very well. Now, they didn't
start by saying, we're going to be
only an API company. But they did start with
being an API company. It was two entrepreneurs,
in Plaid's case, about six or seven
years ago that said, there really isn't a good
API in the payment space. Now, there was Stripe. And I'm not sure about
their foundational stories, why they didn't feel that the
Stripe APIs were good enough. But they created an API. And of course, Plaid just
recently sold for $5 billion plus to Visa, or
announced the sale. So we talked a little
bit about finance and this history, the stack. And I'm going back to the
customer interface stack. We don't need to go back
to bricks and mortars and electronic payments. But again, sort of the
conceptual framework of this class and
how I think about it, the most recent things
added to the stack-- the internet, mobile phones,
contact cards, and the like-- those are all fully engaged and
part of the finance technology stack. And now we're adding
open API and chat bots, which we talked
about last class, and natural language
processing, robotic processing. We're adding that to the
finance technology stack. And that's sort of the
conceptual framework. And I think those are where
the opportunities are. In five or eight or
10 years from now, it will be taken for granted. Of course, there'll
be enhancements. And every change,
there's an enhancement that creates some
opportunities as well. So what do we mean by open API? To [? Iqbal's ?]
earlier question. Open Application
Program Interface provides developers
access to and an ability to get permissioned
customer data. That means you and I
give a password, you and I give a username. Open banking
initiatives-- open banking is unique to the
financial world-- are initiatives usually
promoted by the official sector. Usually promoted either
by the central banks or by the banking
regulators, and sometimes promoted by
competition authority, to say you must
share certain data or you must create some
standards for these open APIs. And again, as I just
mentioned, open APIs are in every field
of online services, from voice to
employment to retailing. But the competition authorities
and the central banks around the globe
said, well, maybe we should lean into this-- the official sector-- lean
in and promote competition for this sharing of data. And these initiatives,
which you've got a chance to read about,
do have policy trade offs. And why I focus on
these policy trade offs is they're also
about competition. Incumbents are not
going to really want to embrace
competition like this. Having said that,
many of the incumbents recognize that it might
be that they're shifting right now-- that their
business models are shifting, and they have to think
like platform companies. And to the extent that the
JPMorgans or the Barclays think more like
platform companies, they want to encourage
the developer community to develop applications
on top of their platforms. If they think it's simply
losing market share to the SoFis and to the new
banks, then they're going to be more
closed to open API. If they think more like
platform companies, they're going to
want to promote it. But there are also trade offs. There are also trade offs
that the big incumbents would say about cybersecurity. If we just let anybody
into our platform, they can also bring down
that platform possibly. And so these are real
public policy trade offs, but they're also commercial
trade offs as well. [Romain I don't know if
there was a question. I see some activity in the chat. PROFESSOR 2: We have a
question from Geetha. GARY GENSLER: Please. PROFESSOR 2: Gary, I've been
thinking a lot about incumbents becoming platforms. Other than an ability to allow
access to their customer data to lenders, I'm
not able to think of examples of what kind of
platforms incumbents can be. Like, they are not a
specific payment company. They're only servicing
their own customers. So I wonder if you
have any examples. GARY GENSLER: So I think
many of the major banks are already platforms. They don't necessarily think
of themselves this way, but they are bringing together-- and this is true for centuries--
they are bringing together borrowers and
lenders, or savers. So they have a balance
sheet that brings together, in some cases, millions of
savers and tens of thousands or hundreds of
thousands of borrowers. And they're bringing
that together. And when you consider that sort
of the modern large financial firm with $1 trillion or
$2 trillion balance sheet, many of the largest [? were ?]
[? all-- ?] and so forth. And so they're
bringing that together, and they're managing risk. And they're very good
at it, or some of them are very good at
that risk management of those millions of
depositors and tens or hundreds of
thousands of borrowers. That in itself is a platform. That creates a data
advantage that they have about risk management. And it could be a real one. And they can offer activities
and products on top of that. The fintech startups
can be, in a sense, a front end distribution
channel to bring more in. Because a lot of
the fintech startups don't have that same
balance sheet capability-- that same ability to in
essence earn a spread between savers and borrowers. Now, you might contend,
well no, they're losing their
competitive advantage, because those startups
sort of have that customer interface, that direct-- one might consider it the
last mile with that customer. But the technology companies
have taken another approach in saying, we like developers
developing applications on top of our platforms. We think it brings
more activities. We think it brings more
interface, more data. And so they're sort of a
network that is bringing data and activity on the platform. But I agree with you. It's a real strategic trade off. It's a real strategic challenge. And the incumbents are
still sorting it out. PROFESSOR 2: Camilo? AUDIENCE: Yeah, hi. I just wanted to ask, in
one of the other lectures, they tried to classify
the open banking between [? prescriptive, ?]
to facilitate market-driven, and in-process. But there were many countries,
as for example Colombia, who didn't have any classification. What's the meaning of that? It means that it's forbidden. I mean, is it impossible
to scrape the data from that from the bank? GARY GENSLER: So
it's a good question. The Bank of International
Settlements paper I think is to which you refer. And what they did was country
by country review where it is. And in fact, that's
probably good leading to the map that was
in that report that I'll show. In many countries, the
central bank, bank regulators, and competition authorities
have not said much. I'm not close enough
to Colombia to know. But in those countries, you
can still technologically go in and scrape. An alternative,
as I lay out here, is actually scraping
from the screen. And so I think actually
it's the next-- I grab this slide from the Bank
of International Settlements reports. And thank you, it's a
great setup, Camilo. That Europe passed
something called the Payments System Directive. And it was actually
the second time that the European parliament
came together and said, we need to promote in
the payment space-- not necessarily in
the lending space, but in the payment space-- that there's more
availability from merchants and from fintech startups
to access your data and move money. That's Europe. And as the Bank of
International Settlements says, it's more prescriptive. In the countries that are gray
on this map-- most of Africa, many parts of the
Mideast, and so forth, parts of Latin America-- the Bank of International
Settlements, I should say, only covers about 60 countries. So many of the
parts that are gray are countries that Bank of
International Settlements didn't survey. But those countries
are probably, like the US, technically
market-driven. There is no federal
law in the US, and there's no
banking rule, that prescribes that you must
have open API in the US. So I think that in Colombia
it might be a technically market-driven, where
the startups are using some form of screen scraping
and some form of robotic process automation. But what we found in the US
is companies like Stripe, companies like Plaid,
others that I'm going to list in a moment,
like Yodlee and others, just built a business model
out of creating public APIs. And the incumbents
possibly in Colombia are dealing with the same. I don't know. AUDIENCE: Yeah. Well, I don't know. In Colombia, the banks are
happy to not let create APIs, because they don't believe
themselves as platforms. They prefer to charge
any type of fees for whatever they can do it
in the small print letter. So they get more the
defensive approach, rather than what you're saying,
seeing themselves as a platform and monetizing over many
layers of application. GARY GENSLER: Yeah, I think
that in many countries and in many incumbents,
they're taking a sort of guarded approach,
and sort of saying, no, no, no, this is dangerous
for our business model. This is dangerous for
our business model. But to that incumbent that
does that, often they lose out. It sometimes takes
years to lose out. But it's that
somebody finds a way to encroach on that
business model. And what we're finding
around the globe is that more and more
countries are promoting this. Brazil, for instance,
is taking it to their national legislature. Now, with the corona crisis, I
haven't stayed closely enough. It looked like it was
going to pass this spring. And maybe one of the students
knows whether they did. But the central bank was
trying to, in essence, insert more competition, so
that E2O and others in Brazil had to sort of take a
little bit more competition in that environment. So my prediction would
be, even in Colombia, it would be hard to guard against
this for multiple years. Maybe a few more years. But they'll be guarding a
business model that will have some encroachment on it. The Bank of
International Settlement also sort of did
this chart, which sort of said what's mandatory. Australia and Europe is
fuller to the mandatory. That's why they're at
the top of this chart. And you see the US,
China, Argentina, probably Colombia, is technically market
driven towards the bottom. It's also what products? Is it all about
just funds transfer? Europe was more about payments. Was it also about the
underlying product information? Because if it's
forced upon the banks from the official sector-- in the US it's not. But in Europe, in Australia,
Hong Kong, it has been. If it's sort of by
the official sector, then it might well
be that you have to include the product
information, as well as the information underneath it. So back to the
market-driven side. So the market-driven side is
what's happened here in the US. And what's happened here
in the US-- and I list-- this is not just the US,
but also Europe, Australia, and some places. I list about a dozen firms that
are very active in this place. This is not forced by
the official sector. This is basically the
private sector saying, let's avoid that spaghetti wire. These are companies
that are going in between all the startups and
the financial firms and saying, we can provide a standard
way to do these interfaces. Plaid that I talked
about just announced it's being sold to
Visa for $5.3 billion. What do these firms compete on? They compete on a lot of things. First is coverage. If I'm Plaid, do I cover
all the banks that you need? If I'm a new startup,
and I'm thinking, should I use Plaid or Yodlee? And Yodlee and Plaid
are deep competitors. Maybe one can stay just as
competitive with each other as Goldman Sachs
and Morgan Stanley. Yodlee was acquired years ago
by a company called Envestnet. But Yodlee and Plaid, deep
competitors to each other. Well first, what
coverage do they have? As a startup, if I use
Plaid instead of Yodlee, can I get to all of the banks? What services do they
give me as a developer? Obviously, what price? Are they reliable? What's their security,
and the like? So all of these are important. And one of the sort of
incumbents in this space, Fiserv bought CashEdge. And Plaid was being
bought by Visa. So some of these get caught up. SaltEdge is in 55 countries. And then I did a little
research on a paper that was written three years ago-- or it's really a website
that keeps a list. And I sort of put together--
this list is three years old. These are the
account aggregators in all of these countries. And I'm sorry, Camilo, I
don't see one for Colombia. But all of these
companies in some way built up a business
model to say, we will serve in these
various countries to be between the
incumbents and the startups. We will build some
business model in between. Now, what can you
do once you do that? You get a lot of data. And so what used to be
just a software company-- Plaid, for instance,
started as, in essence, we can provide this software. Started to have remarkable
amounts of data. And that remarkable
amounts of data in Plaid-- founded in 2013, sells
for $5.3 billion to Visa, or announces the sale,
only 450 employees. And Forbes estimates it's only
had $109 million of revenues. So that's nearly
50 times revenues. Quite a multiple. You might wonder why
Visa's doing this. This is from the Visa slides
that Visa announced in January. This is how they
sort of picture. They look at Plaid. They say, Plaid is connected
to 2,600 fintech firms. That's 2,600 fintech
firms connected to Plaid. And then Plaid on the
other side provides an API into 11,000 data sources-- 11,000 financial firms. 2,600 fintech firms
connected to Plaid, Plaid connects to 11,000. Again, that's the classic
place you want to be. Plaid is at the bottleneck. Plaid is at the
gateway, you would say. And they can collect data
on 200 million users. Visa thinks there's
a value in that. We'll have to see. We'll have to see
whether that ends up being a good acquisition
for Visa or not. But these companies
in the middle-- the Plaids, the Yodlees,
have right in the middle created value. Romain I'm going to
pause there and see if there's some discussion
that we can get going. PROFESSOR 2: [? Iqbal. ?] AUDIENCE: Yeah, professor. Can you shed some insight as to
the relationship between Plaid and the financial institutions? What's in it for the
financial institutions to face Plaid in reality? Were there fees involved? Because in the example
we looked at some for some Latin American
countries, such as Brazil, the incumbents are reluctant
to participate in open API. And it would be [INAUDIBLE]. GARY GENSLER: It's
a terrific question. And I'm going to
apologize, because I'm going to slide back
here on the slides a bit to go back to something. This is what they want to avoid. Even the incumbents don't
want thousands of folks to enter into
contractual arrangements, enter into secure private APIs. Even the incumbents
want to sort of avoid this spaghetti wire of mess. Now, it might have been less
strategic and more incremental that when you think about all
these companies on this list-- these 10 or 12 that have built
up these remarkable businesses, or around the globe all these-- that many of these started
as small software providers. And they were saying
to the big incumbents, we can handle that
spaghetti wire. We can ensure that
people aren't just screen scraping and using RPA. It's going to happen anyway. You need somebody
to help manage. And so every modern
online company in finance and
outside of finance has to have an AP
management strategy. You don't normally think about
something in the CEO suite, this chief executive
officer who's managing this. But here we are at MIT, I'd
like to bring this technology. This is an important
technology question that you need to be
asking is, how are we managing our
developer community, our public interface? Are we going to
do it internally, or are we going to hire a Plaid? And I can tell you that
the banks now-- this has happened in the last
four to seven years. And many of the banks
are sort of saying, this has gotten out of hand. This has gotten out of control. Is this something
that we really want to allow the Plaids and
the Yodlees and others to commercialize our data? But I think that in part, it
was incrementally decisions about how to manage
broadly this interface. And it's also possible
that it wasn't far enough up in the C-suite, and
it was sort of managed down in some regards. But one reaction is this
financial data exchange that's happened here. And I've listed the
major founders of FDX. But Financial Data Exchange--
and this is straight from their own white paper-- they want to create this group
of remarkably large banks. I think the members of FDX have
$2 trillion of market value, if I remember. Because many of these actually
have $2 trillion balance sheets. When you add up all
of these companies, they've come
together and said, we want to create standards
for financial data sharing and secure APIs. We want to basically not sort of
have the Plaids and the others commercialize our data. Even though that's
not in FDX' mandate. But in some regards, the
cat's already out of the barn. And now they're
grappling with it. But the JPMorgans and the
PNC actually, I believe-- which is a large
mid-sized bank in the US-- has stopped having Plaid have
access, if I'm not mistaken. So it's sort of an interesting
challenge back and forth. PROFESSOR 2: You now
have Geetha and then Lin. GARY GENSLER: Geetha's from the
incumbent side going to tell us why they're using the Plaids. AUDIENCE: Yeah, I just wanted
to make a comment on Plaid specifically. We allow access to a number of
fintech lenders to our data, because then we have no choice. Because the customer is
providing access to their data, and we have to allow
them to access it. To somebody's question
earlier, the biggest problem with Plaid-- and I'll probably
send some articles on kind of issues between
Plaid and Capital One. There was a period of time
when we stopped giving access to Plaid, because
Plaid wouldn't comply with any of the standards
that we imposed. For example, we
needed fintech lenders to come from certain IPs,
because they looked like bots, right? Because they do all
these batch accesses, and they look like bots. But Plaid wouldn't comply
with those standards, and it became very difficult
for us to differentiate Plaid from a bot. And even though we had APIs,
they would do screen scraping. Sometimes they would come
through our mobile channels. Sometimes they would come
through our web channels. They would come through any API. It became really hard for
us to differentiate Plaid from fraudsters. So there was a period of
time when we blocked Plaid. And then they came to the
table, we had conversations, and then we let them in. GARY GENSLER: And Geetha,
you were at Cap One? AUDIENCE: Yes. Yeah. GARY GENSLER: So for those
who might not be familiar, Cap One is one of the largest
credit card banks in the US, competing head to head with Bank
of America and Citi and Chase, Discover, and American
Express and so forth. I sort of call it the Big Seven. And Cap One's the largest
that isn't in and of itself one of the big four
banks in the US. So it's a remarkable
organization, and it's cut out a big
chunk of its business model around serving customers
in the credit card space. And here you're
hearing Gita say, but we didn't really have a
choice, is what you're saying. Romane, others? PROFESSOR 2: Let's go
with Lin, and then Carlos, if we have time. GARY GENSLER: Please. AUDIENCE: Yeah. So in the example for Plaid,
it seems like most of the data, if not all of the data,
is financial data coming from financial institutions. I'm just curious to
hear your thoughts on the market moving towards
using non-financial data. So customer data in
the retail health sector to inform or build
financial services upon, and whether there are
any people doing that. PROFESSOR 2: GARY GENSLER: No, it's a
really good observation that the data aggregators-- I'll go back a slide. This whole cacophony
from three years ago around the
globe, or the data aggregators here in the US-- these big eight or 10
in the US and Europe-- these data aggregators
are largely about financial data for now. So their business
strategy-- these dozen companies, their
business strategies was sort of API first. They built an API upon which
they got a network advantage to get a lot of financial data. But it is also true that
they could add other data. And in fact, Plaid
gets bought by Visa. Yodlee got bought by Envestnet. And Envestnet is also in asset
management and fund management and so forth. I believe you're right. Your observation is right. A lot of these "API first,
data aggregator second" are also going to be adding
other non-financial data. Alternative data, as
we've talked about, can come from finance,
like whether you're paying your utility bill or not. It can come from finance
and building full cash flow underwriting models. But it also can be your
social media footprint. And it often is. Alternative data is
about your social media input and your
footprint, I should say. Your footprint. So I think you're right. Their comparative advantage
is around this financial data. Plaid, that's their advantage. I want to note one company
that's on this list here, I think I included, was Galileo. Did I? Yeah, I included Galileo. Galileo has done a
recent fundraising round in the middle of 2019. It's still private. I think their valuation
was $1.2 billion. Might have been a little higher. I can't recall the figure. So it's a unicorn. Not many of us have
heard of Galileo. But two of their big customers-- Chime is one. I'll just say, Chime is
one fintech company that's one of their big customers. Last year, Galileo had a glitch. And Chime went
down for 24 hours, because Galileo's standards
in API had a glitch. So sometimes, you're
that fintech startup relying upon a Galileo,
in that instance. And if Galileo software has a
problem, you have a problem. So there's that challenge
too in the midst of this. Marqeta has done a
fundraising round. I think its valuation
is about $2 billion. Again, still private, still in
the venture land, and so forth. So there's significant
value that has been built, API first, data
aggregation second. And these companies
are all somewhere between the incumbents and
the developer community and the startups. Back to the class, Romain? PROFESSOR 2: Carlos is up. AUDIENCE: Hi. So I just want to give
an example of a shout out to Plaid,
really, for something they're doing at the moment. So as part of the SBA PPP
program for emergency funding for small businesses with
the whole corona crisis, there's $200 billion being
disbursed over the course of three months. And actually probably
this Thursday there's going to be an
additional $200 billion approved. $250, sorry. So what Plaid is doing
is they're actually helping, because of their
role as a data aggregator, to populate the payroll portion
of these loan applications, because a lot of banks are
trying to do it manually, which is completely
infeasible, given the massive size of the
rollout that they're proposing. So thanks to Plaid and
what they're trying to do, at least, a lot of
loans can be approved, or at least applications
submitted exponentially faster. And that can make a huge
difference for businesses to survive. GARY GENSLER: And I thank you
for bringing that information in. Do you work with Plaid? Or have you worked there? Or is this something you're
reading about online? AUDIENCE: So I've
been following, because I work for a
fintech-focused VC that actually invested in Quovo,
which was acquired by Plaid. So I know the model pretty well. GARY GENSLER: These companies
have brought a great deal more financial inclusion. These companies have opened
up the banking sector, to some extent. And now the banking
sector was probably going to be opened up
by competitive pressures in any regard. But with the Plaids,
earlier Stripe, the Galileos and the Yodlees and
others, what they've done is they've brought
some standardization to that spaghetti wire
that I put earlier-- these interfaces. They're in the weeds. To some of us, it might
just be mundane connections. But they're not. They're really, really
important connections. And I would say if you take
anything out of this lecture and moving forward, if
you're at an incumbent or you're at a startup,
that you sort of think about your API strategy. What is it? You might not manage it, but
it's good to know what it is. You might know what your
competitors view of it. But I agree with Carlos that
in this time of crisis-- the corona crisis in the US, the
program that he's referencing, was of this $2 trillion package
that Congress just passed, there was $350 billion
to fund small businesses. They have to be 500
employees or less. But small businesses. It's called a Payroll
Protection Program, PPP. And through the Small
Business Administration, it guarantees the
banking sector can make loans that are forgiven. So they're forgivable loans
up to $10 million a piece. They're a loan,
but they're really a grant if you can show
in your application that the money is going
to support employment. And there's various
conditions and conditionality. And so what Carlos
is talking about is in the loan applications,
the loan applications were first being filed
I think last Friday, but it might have
been over the weekend. And these loan applications had
to go through your employment and show who you had earning
less than a certain amount of dollars per employee. So a lot of data. A lot of data. And it sounds like
Plaid is actually helping facilitate
that, in some regard. So thank you. Romain. I think I'm getting close to
that 15 minute warning time. But is there another
question, or are we just going to move
on to robotic process automation for a minute? PROFESSOR 2: I think
we can move on. And indeed, we have
17 minutes left. GARY GENSLER: All right. So let me just
pause for a minute again on financial
data exchange. I'm not close
enough to it to know where they are in
their development. But basically, a large set
of big incumbents-- and I've just put a visual of
their principal members. But then they have behind
this another 50 or 100 second tier members. And by second tier,
this is the group that's primarily running its
board of directors and sort of coordinating. And it has all the big US
banks on here, the largest-- Quicken Loans,
which is the largest in mortgage lending
in the US, and CIFBA, which is the trade association. They're coming
together to say, we need to standardize financial
data sharing standards for secure authentication and
how this user experience works. But part of it-- well, part of it is
the incumbents sort of trying to reclaim a
little bit of the territory. I don't think the cat's going
to be able to be pushed back into the barn, so to speak. Or you might say the
horse is out of the barn, and I won't use the cat. But I think that this is an
interesting dynamic that's going on right now. Now, you put that in the
context of what we talked about, robotic process automation. Robotic process automation
can be used in a positive way by the incumbents and
by the startups to do account opening, onboarding,
as we talked about earlier, loan processing. Anything that you
can take and automate that which humans are doing. And I'm not sort of
moving into chat bots and the conversational
interfaces that we've already spoken about. I'm talking about the more
standardized documents, reading the documents,
and opening the accounts, and, as was mentioned
earlier and I maybe should have put
on this list, also compliance for any
money laundering and Know Your Customer. Those are very positive ways. But it's also a way
that challenger banks and challengers can use
it to scrape the data and pull the data out
of the incumbents, as long as a customer
permissions it. So it's sort of both of those
sides in a very positive way across the entire
field of finance-- a very positive way that
we can automate that which humans are doing in the past. This is frankly just the same
as when the tractor came along and we were automating
what humans and the horse and the cow and the oxen
were doing in the field. We can automate that
which the human does. But it changes the way that-- the sort of dynamic between
incumbents and startups. And then I want to pause and
see if there are questions. And then I wanted to say bit
on assignments, and then close. PROFESSOR 2: We have a
question in the chat for you from Alex, a Sloan
fellow, who tells us that one of the readings
one of the authors ends with consumer will access
financial services from one central hub. But then another author
in another reading speaks of another trend
of embedded fintech. So which ones of
these perspectives would you tend to agree with? GARY GENSLER: Well Alex,
I think that's the debate. And I think you probably could
guess that I'm at neither end. I don't think that we'll end
up with truly decentralized and we can sort of disaggregate. We had this discussion
a little earlier. Will it be completely unbundled? I don't think so. I think there's still going to
be economic benefits to scale. Sometimes that's driven
by balance sheet. The balance sheet financing
side leads to scale, though securitization and
asset securitization as a way to rent somebody
else's balance sheet and to sort of lay off
that risk to somebody else. But I think that there's a
lot of reasons that scale and network effects-- balance sheet being one of
the network effects, data being another network
effect, customer user interface another
network effect-- that there's benefits
to some scale, that there will be
some concentration. Will there be one portal? Will we go all the
way to one portal? Will it be that
Amazon will sort of be the one place I go to for all
of my online financial needs? Jeff Bezos might have
a strategy that way. You might write papers that way. I would caution that
I think that there will be some
competitive pressures and economic pressures
on the other side-- that we'll still have
some disaggregation. For instance, in
country after country, for decades, there's
been the debate whether insurance products will
be sold by banks and banking products by insurance firms. And for some reason
around the globe by and large, we still
see this separation. Some of it's just
legacy and history. Some of it is regulation. But it's also how we
think of the products, that we think of buying
our life insurance or our household insurance
or auto insurance a little differently
than we think about funding the same things. Funding our house is called a
mortgage and funding our car is what's called an auto loan. Now, it's true in some
countries, it's come together. But by and large,
it's not one portal. But Alex, do you
have a point of view? Did you did you feel strongly
one way or the other, if you're there still? AUDIENCE: I think the embedded
fintech is quite insightful, because nowadays
that people talk about customer-centric
ideas even more than before, because we have more data to
provide relevant services. But the thing is that when we
talk about the jobs to be done, financial services are
not actually the job. The ultimate job you want to
get done, you want to buy house, so you need a loan. Which means that maybe it makes
sense, some of the services will be the second layer. I mean, be embedded into
those jobs that really need to be done by the customers. GARY GENSLER: I want to
say this about embedded finance and this concept
that was in the article. And I'm glad you reminded
me of this, Alex. I do think that we're
going to see more what's called embedded finance. And by that term,
what I take from it-- and it's not a
well-defined term. You can't look at a
Webster's dictionary and say what's embedded finance. But what I take from that term
is that non-financial firms will embed some
aspect of finance in their product offering. Not a new concept. Even the local grocery
store in the 19th century was saying you can buy your
groceries and pay tomorrow. You know, that local merchant. And that led over 50 or 70
years to the modern credit card being invented in the 1940s
and taking off by the 1960s. So we've had some concepts
of embedded finance. But I think that the
modern technology will provide more alternatives
to embed a payment mechanism or a lending mechanism. And those will be the
two principal things. Some aspect of a loan or
some aspect of payment right in an app. In a sense, like
Uber and Lyft does. In a sense, like of
point-of-sale lending-- that I can buy
something online and pay on installment over
four or six payments, maybe over multiple months. I think embedded
finance is a trend. I also do think that because
of this API movement-- and this is a way
to close on APIs-- that startups can
provide services in a way they couldn't in the past. That a Robinhood
could come along-- and we'll talk about
Robinhood more when we talk about capital markets. But a Robinhood could
come along and say, we offer that you can
open brokerage accounts on this handheld device. You can scan your documents
using robotic process automation and natural
language processing. You can scan your documents
by taking a picture from your phone, open
a brokerage account, and trade automatically,
interfacing-- there's probably dozens of APIs that stand
between your Robinhood app and the New
York Stock Exchange. And it's not just
Robinhood directly to the New York Stock Exchange. There's multiple
interfaces in between. So I think it's facilitated a
remarkable shift in finance, that we can do those things
right on our cell phones. And embedded finance and
these APIs relate to it. I'm just not quite
there that it will all be bundled up into
one service provider, or that it will
all be fragmented. I think we'll have a
little bit in between. I just want to mention
something on assignments, because the April 8th. Again, assignments--
particularly at this stage, where we're pass emergency,
no credit emergency, or sort of colloquially
pass/fail-- assignments are a great way
to engage in this subject. I think of that even when we
have A's and B's and things like that. I just really ask you to
engage in this subject. Pick some subject that
interests you, that excites you, and you want to write
to an incumbent, to a big tech firm like
Amazon or Andreessen Horowitz. Which Andreessen Horowitz owns a
slice of a lot of the companies we've been talking
about already. I mean, they're a remarkable
venture capital firm. And just think of some
sort of strategy around it. The individual
papers are just a way to break up the
assignment in a way that one person
would write about it. Overall market
analysis, somebody doing the technological trends. If you pick that you're writing
to Amazon about, let's say, the consumer lending space, what
are the technological trends really doing to that space? Who are the
traditional competitors and who are the big
tech competitors? And just breaking it up three or
four pages, 900 to 1,200 words. Again, we're just helping you
kind of engage in a subject. PROFESSOR 2: Perhaps you
can answer another question that I've received from a
student who does not understand how Plaid makes money. What is the business model? GARY GENSLER: With only $110
million in revenues-- now, that's a Forbes estimate-- you sort of go, wait a minute? How does this work? But how the space makes money-- not just Plaid, but
how API companies make money in finance and outside of
finance is they do charge fees. I apologize, I don't know if
it's a penny a transaction or what it is. And maybe somebody on
this call actually knows-- Gita or others know their
actual revenue model. But charging a very small
fraction of money per trade. And then what they do
is data aggregation. And with that data
in other fields, then you try to see how
you can commercialize that data by cross-selling
either directly. Or remember, a lot
of these companies like Plaid and Galileo
and others are in between. They're behind the scenes,
between one side of the market and another side of the market. But when you create
that like Twilio, Twilio is an 80% market share
between voice and messaging and things like that--
in the US, at least. So all of a sudden,
then you can either enhance your fee
structure, because you're such dominant
oligopolies, sort of. You can collect economic rent. Or you commercialize the data. And it's the data
aggregation side-- I think Visa's paying what
might be 25 or 50 times revenues for that data aggregation side.
