How is Silicon Valley changing Wall Street?


(bright jingle) – Hello, welcome to Faster FinTech. I’m delighted you’re all here tonight. Thank you for coming. Faster FinTech, today’s
discussion is a discussion jointly sponsored by the
School of Engineering and the Graduate School of Business. Before we get started, I
would love to remind you all to please turn off your cellphones. And, throughout the evening
you will see two people walking up and down the aisles. They are collecting questions
for the Q&A portion. Please signal to them when
you see them coming through. Now it’s my pleasure to
introduce tonight’s program. The Intersection Programs
consists of a series of conversations in which faculty from different Stanford
schools explore areas of common research interests from their different
disciplines and perspectives. To talk about the future
of financial technology, we have brought together
three experts in their fields. Balaji Prabhakar is VMWare founder’s Professor of Computer
Science and Faculty Member in the Department of
Electrical Engineering and Computer Science. His research interests
are in computer networks, notably in data center networks and cloud computing platforms. He has also worked on societal networks, networks vital for society’s functioning, such as transportation,
electricity and recycling systems. Darrell Duffie is the Dean
Witter Distinguished Professor of Finance at Stanford University’s Graduate School of Business. He is also a Senior Fellow
at the Stanford Institute for Economy Policy Institute and an alum from Stanford Engineering, with a PhD in engineering
economic systems. Darryl’s research focuses on the design and regulation of capital markets. And our moderator for this
evening is John Markoff, a research affiliate at Stanford Center for Advanced Study and
Behavioral Sciences. He reported on science and technology for the New York Times for over 18 years. And was nominated for a
Pulitzer Prize five times, and with his team, was
awarded the Pulitzer in explanatory reporting. He is currently researching
a biography of Stuart Brand, the creator of Whole Earth Catalog. Please join me tonight in welcoming them. (audience applauding) – Thanks everyone. Can you hear me in the back? How does? Okay, we’re good, great. And Stuart Brand was a Stanford student, so there is a connection. So it’s been a half
century, since the advent of packet-switched networks. It’s been 48 years since the
microprocessor was invented. And so we’re partway through
this interesting transition from a commerce system, which
is rooted in paper and metal, to one that’s rooted in bits and glass. And, the question I hope
we can pursue tonight, is whether this is a one-time transition like indoor plumbing. Are we sort of at the
end of this transition? Or is it something else? Are we at the beginning
of something that’s gonna remarkably transform the world over the next couple of decades. And to get us going I wanted to ask Darryl if you could start by
maybe painting a picture. Maybe not a decade in the future, but there’s all these
forces based on technologies that are merging largely
from this part of the world that are changing the commerce system. What might it look like in
the relatively near future? – We already have,
actually, a pretty good idea just looking into China. They’re getting a head
start on the United States. They have a mobile payment system that’s roughly a factor of a
hundred more transactions than in the US. And 500 million users swishing money around
on their mobile phones to make payments using QR codes. There are even infamous videos
of beggars in the street, with rather than a hat, they
actually have their QR code ready to collect payments
by WeChat Pay or Alipay. Just so that I can kind of gauge the level of familiarity. How many people here
tonight have ever used, or have on their iPhones,
Venmo or WeChat Pay, or Alipay. (John laughs)
Wow, okay so. That’s a sign that it’s not 10 years off, it’s happening right away. And the vision is, what
I call faster payments. When you pay today, most of the time it’s going against your credit card. It might be going against
your bank account. But the person that’s
getting paid is actually not getting paid until tomorrow,
or perhaps the day after. And a lot can happen in a day or two. The money gets held up. It goes through deposit
accounts in the banks where it holds very little,
it gets very little interest. And vendors are getting
delayed on their payments. Meanwhile, in the big
banks on Wall Street, they need to make their
payments much faster just because there’s
so much money involved. And they are also developing
new payment methods. One of the visions is that rather than going through
the bank payment system where all of these devices
are connected to now, all the people in the
audience are getting paid one way or the other through the banks. That may happen through other methods, like cryptocurrencies. How many of you have ever actually purchased a cryptocurrency? Wow! Okay, so we have a pretty
sophisticated audience. Okay, so first, let me
disabuse you of the idea that this could ever be done with Bitcoin. Because Bitcoin is just not suitable for making fast payments. It’s very slow. And its price is very volatile. So you couldn’t count on the value that you wanted to send getting
to somebody in a proper way. But there are new methods
called Stable Coins. Which are cryptocurrencies,
which if done properly, will always be worth $1.00 per coin. And they’ll be able to be
transmitted very rapidly. And there are lots of
entrepreneurs in Silicon Valley, and elsewhere, that are
developing these right now. And they’re gonna be
ready to deploy anytime. – There’s still, I think,
a fair amount of skepticism about some of these new
kinds of cryptocurrencies. Tether in particular
is one that seems to be in the gun sights of skeptics. Are the currencies that you
believe will solve the problems of the first generation
of cryptocurrencies, do they already exist, or do you think they are to come? – They’re just nascent. JP Morgan, the bank, has just come up with an imaginatively
named JP Morgan Coin. (audience laughing) Which is fully backed dollar for dollar with deposits at JP Morgan. And that will work. – Okay, so take me through this. Are we gonna see a proliferation
of payments systems and how will I as a consumer
make sense out of this? Or will there quickly be a monopolization? – It’s kind of like the subway system or the telephone system. There’s only a certain number of these that can actually survive, because of the massive
economies of scale and scope. It wouldn’t work if I wanted to pay you with JP Morgan coins, and you wanted to pay me with another cryptocurrency like Tether. So we’re gonna end up most likely having a small number of
these that are widely used, if the authorities allow that. That’s another issue entirely. – There’s this thing called fiat currency. – Yeah.
– Will they be, how will they? – Okay, so let me put my thinking
on as potential regulator. So if I were at the Federal Reserve, which is the central bank
of the United States, I would be fine with this as
long as it didn’t get too big. And then I would start to worry that people are using this other
currency than the one that I, the Fed, controls, then I won’t be able to do my job of controlling inflation and promoting economic growth. And if some operational
failure were to occur, that could bring down
the financial system. I think the central banks of the world will allow these to sit on the fringe, maybe be used for consumer payments, or for specialty purpose big
bank settlement of trades. But they won’t be the
general payment currency of the economy. – And then, what does this do? This kind of the emergent
payment system world, you’re talking about due to
our existing banking system. Do we need banks in this
kind of an environment? Are they at some risk? – I just gave a talk at the New York Fed a couple months ago and it was titled, Disruption is Coming. And I was speaking about
the impact on the banks. Some of the banks are gonna
try to get ahead of the curve. But the ones that don’t, they’re gonna see those
massive amounts of deposits sitting in their accounts, for which they’re paying
almost no interest. Those are just gonna go way down. And they’re going to lose
out on the technology wars. JP Morgan alone is
spending 16 billion dollars a year on tech, of which
a lot is devoted to these kinds of initiatives. And there are other banks
that are doing the same, but a lot of other banks
are gonna get left behind. Any bank that’s big enough,
or not nimble enough, is gonna lose out in this disruption. – So, there have been reports
in the last couple weeks, that Facebook again is
working on its own currency. I think are Zuck-Bucks by some. (Darrell laughs)
(audience laughing) – The Amazon, Facebook, Google, these are the analogs of
what’s happening in China. Those of you in the
audience that have Alipay and WeChat Pay, you’re holding the analogs of what in the US would be Amazon Pay, or the new Facebook coin,
or a possible Google coin. And it doesn’t have to
be a cryptocurrency. It could be a fast payment system. But those tech firms could
eat the banks for lunch if they really got going on this. Because they have all
of the network in place. They don’t have to develop a network. – Yeah. And so in that model also, what happens to things like ATMs? – Okay, so how many in the audience actually use paper money today. Raise your hand. Wow, a lot of paper money users, too! Well I haven’t carried
paper money in a long time. But I think ATMs are gonna be few and far between pretty soon. – And then so there’s been
this model and this transition. ATMs were supposed to kill bank-tellers. But in fact, because of the
falling cost of computation and the explosion of computer networks, banks put branches on every street corner and they began to use
tellers as salespeople. In the new world you’re talking about, does this get reset one more time? – Absolutely. It’s going to be a lot of. Well actually there was a
story in your former newspaper just last week I believe, about small, community banks in America not having any success. And the large banks reducing the number of branches dramatically, to the point at which, you
may have to go searching for an ATM machine in 10 years, like you now have to search
for a telephone booth. And the number of employees in banks that are actually doing customer service is also gonna go way down, I think. – So I want to get to speed quickly. But, before we get to speed. What about the people
who can’t participate? What about the people who
don’t have bank accounts, who are being pushed through the cracks? In this kind of a world are they pushed farther off the edge? – Yeah. This is a serious concern. China has just passed a law, requiring that vendors take paper money, because they’re so concerned
about the un-banked people, particularly in the countryside, that don’t have mobile
phones and bank accounts. Now China actually has
pretty good penetration. But let’s take a country like Vietnam. Only 30% of people in Vietnam
actually have a bank account, which is kind of a prerequisite for using any of these things. So, if you don’t have paper money, and vendors won’t take paper money, that could be a serious problem. Sweden has remarkably
little use of paper money. I was there just a couple months ago. I couldn’t actually use my
paper money at my hotel. And they are worried. And they’re probably going
to introduce some form of electronic money that
will be a substitute. – That’s not the case. Doesn’t the United States
have a kind of a cash economy, an underground economy
that trades in cash? – Yeah. – Thriving? The people who sort of
that work on the books, and then they’ll work
off the books for cash? That’s called the underground economy. – Yeah, if you think about
the amount of paper currency that’s circulating right now, it’s about a trillion and a half dollars. Probably only half of
that is used legally. – [John] The rest is in
hundred dollar bills. – The rest is smuggling money, or underground, untaxed economy. India recently tried to
remove cash from its system, or a large portion of
the cash from its system, in order to try to reduce the amount of this kind of black
economy use of paper cash. – So if you had to predict
a decade from now in the US, less cash, no cash? – The US will be one of the last to go. It’s a very conservative
country relative to some of the other
countries that we’ve just been speaking about. So I think cash will be
here for a long time. I don’t know how long. But we still have a $1.00 paper bill, which is kind of a sign
that Americans don’t like giving up their paper money. – So, to make this work you
need these very powerful computer platforms, back ends, markets, that allow these transactions to happen. Balaji, how did you come to get interested in the problem of time and computing? – Well actually if you
look at how computing, let’s say in the ’60s, bifurcated from centralized sort of things that were single machine,
or a set of machines that were the high speed connect, look like a single monolithic system. Then it went to this
whole super-computing, high performance system,
sort of trajectory. And the other thing that began
was distributed computing. Lots of cheaper, less powerful machines connected to the interconnect
that was not necessarily super high performance
and that evolved into now sort of essentially the cloud
computing sort of power line. At the same time on the
networking side, communications, we had circuit switching
and packet switching. And, what circuit switching
guaranteed was 100% real time, nothing is buffered anywhere. And so the circuit is
dedicated for a connection. Packet switching changed all
that and introduced jitters. What I send to you may not get to you the same speed every time I send it. It could take more or less time depending on who else was there. So that jitter network which
we call best-effort service, so as not to promise anything at all. Went and became this
global internet, ethernet, all the data center stuff. I think we’ve given up
essentially both conceptually and in fact, given up
on time and real time. In the sense that if you have to take one of these computer systems
and make them look like a single monolithic system, which they do. Like databases carrying
consistencies of various kind. They will do this by exchanging messages. So we’ll stop doing some work for awhile, then read the nodes of
the database will agree that this is the order in which
transactions shall appear. Commit to that, and then proceed. – [John] This is like a trade. Like a financial trade
would be a good example? – That’s a good example. For example, foreign exchange trading, many people are concentrating
putting it on blockchains, especially the settlement
and clearance side of it. And it’s important to have things in a single, common timeline if possible. Or, a non-controvertible,
nobody disagrees with it. Everybody’s commonly agreed
to this succession of events. And, that was then. Because we as a community gave up on. We starting building systems
with a clock-less assumption. And that has been the ethos
of distributed systems. And, we learned how to cope with this. But now imagine that we actually
can revisit this question of can clocks be synchronized
across a large network where there are random delays possibly, between any two pair of nodes. That’s the problem that more recent, about a year and a half, or two ago, we starting working on. – [John] And you solved right? – Yeah.
– For all practical purposes? – More or less at scale. – Good enough that a market like NASDAQ could deploy it commercially. – Yes. – Would you talk about
why they chose to deploy? Was it simply about front-running this issue of being able– – Yeah, so there are two things here. Front-running is a, I’ll come to that separately. Front-running is sort of
information arbitrage of I know something ahead of somebody else. But keeping that aside for a second, just consider a single
financial exchange on its own. There are market participants, they’re all sitting,
accessing certain gateways where the financial
exchange officially begins. So gateway one, gateway two,
gateway three, like that. And then this transactions are, the bids that market participants
place across these gateways, maybe different gateways for
different market participants, make their way through
this matching engine. Where if you’re looking
for something to sell, and I’m looking for something
to buy, we meet over there and the matching engine performs a trade. So long as the prices are
how we would like, okay. And in this world moving
gateway to matching engine, there’s a need for what’s called fairness. Fairness is that the order
in which transactions cross the gateway, no matter
what gateway they arrive at. That is exactly the order in which they should be executed. So if you saw a price that’s favorable for your strategy or trading, and you chose to transact then, and you cross the gateway ahead of me, then you should be transacted. Your transaction should go through first. Unfortunately, in the world where they are networks of jitters, this is not easy to guarantee. So occasionally it does happen. Now these networks are
very carefully engineered. The measure of the delays and sometimes the length in the wires. If one path is shorter than the other they would make it longer. And so it’s very carefully
manicured networks are what these financial
executions look like. In fact, we were taking a
tour of one of the exchanges, and the flooring is transparent. And the goal is to show
the market participants, look all the wires are the same length. (audience laughing) We wouldn’t have known. They told us that this is why the floor, the tiles are transparent. – In their world right
now, what kind of precision do they need? – So they’re at the
nanosecond level, pretty much. – So that’s 12 inches. That’s how far light goes in– – No. It’s five nanoseconds per meter on fiber. And you know it’s a pretty, if you have, because transactions
come in thick and fast. The time taken going from tick to trade is down into a few tens
of nanoseconds now. – [John] And you’re able
to give them the precision just in software now, that was sort of the breakthrough?
– That’s right, that’s right. – [John] And it’s called Hoigan’s is that correct?
– That’s the algorithm. – [John] You did this with Google. – Google was a collaborator. We worked the summer internship. Student, whose thesis work was this. (speaking foreign name) I see Shieldly was another
collaborator on that work. And (speaking foreign name)
are the three students, and Google researchers. – In the past you’ve
actually created companies to create technologies and then sold them. When did you choose not to
go that direction this time? – We are. (audience laughing) But let me say why. First of all it is time for now. The thing is that sometimes the software has to be produced for
production readiness. And it’s not easy to take an
advanced academic prototype, which may be good for a
proof of concept, or a pilot, that somebody can install and have it run in production environments where
the network interface cards are variously different. Or operating conditions are changing. And so that’s one of the things. But I think time itself
is a foundational piece. This is sort of, you have the option to revisit the choices
made a few decades ago. But why go this way verses that, and then say, if time is
solved, what can we do? – I got this sense. You’re now slicing time so thinly. It reminded me of you know Zeno’s paradox of the turtle bumping his
nose against the wall. You’re kind of pushing up
against fundamental limits? – There’s still pico. Where nano, there’s femto. People actually are
talking about such things. The thing is, I’ve hears CERN physicists say that we’re in
relativistic times scales. This is the speed of light. Speed of light is three
times ten meters per second. We’re in that relativistic terms. But there are phenomenon
that are lower time scales. But for practical purposes
in today’s technology I think we’re where we should be. – Yeah, and was there a significant? Michael Lewis wrote a book in 2014 which was very popular and
very controversial apparently. Was there a real front running problem that these people had to solve? – Well, there is a front running problem. There is. One way you can think of it as that if, sorry, is the case that if you’re trading with other people’s money, you’re required as a broker acting on
behalf of somebody else to look for the national
best better offer, BBO, that requires this person to go check with multiple exchanges as
to what is the best price. But the very act of checking
for a hundred shares of some company’s stocks. You’re close to one exchange,
and the other exchange that you’re checking is distant from you. Now in both of these places, you’re getting into them on wires. It’s possible for someone
else to hear your bid here and take an air link, like
a microwave link over there, and get ahead of you and wait for you. That’s the front running problem. And they know exactly
what you’re looking for because they already
heard you say so here. – Right. – The way of us. One way of solving this
problem is to put speed bumps. Speed bumps are, you slow
down the speed of trading. You don’t have to do it as frequently as people are doing it. That’s one way. The other way is also just for the trader to not place orders until a
certain time in the future. Place them simultaneously. The clocks are synchronized. You could do this. – Yeah. I wanted to asked. Now, you’ve controlled time. Google talks about this notion of a planetary scale computer. These platforms that are smeared out. Are there other obvious
financial applications, capacity would be one, obvisouly, but are there other things
do when you can precisely control time the way you can now? – Oh, yeah. A lot of the distributed systems that people have built like databases, Google has its global
database called Spanner, which guarantees what’s
called external consistency, as seen by an external observer. Events are sort of in a linear order. As it happens to an external observer. That is done through clocks
they have called true-time. Similarly in our group
we’ve done other work using accurate clocks. Let’s go revisit whether
distributed ledgers of the blockchain type could be sped up. Whether consensus
protocols, like famous ones called Paxos and Raft, could they be made to be higher throughput and lower latency, very accurate clocks? And the idea here mostly is that if you can agree on time, then it becomes easier
to agree on other things. That how time acquires
primacy in these ventures. – Yep. Darrell, you brought
up the specter of China in a really quite striking way, and we sort of just glossed over it, but you basically said China’s ahead of us in the deployment of this
financial technology. If you’ve been to Chine recently particularly in urban China, I don’t know how spread
out it is across country. It’s really quite striking. WeChat, as a social and economic tool, is so far ahead of the way we as Americans are using Facebook, for example. I guess I want you to assess
where China is as a competitor, either financial or technological, take that where you think it will go. – You’re hitting on one of the questions of the last few months. Because, as you know,
the U.S. administration has been engaged with China
over issues of this sort, technology transfer, and the
use of sensitive technology. In the finance area, I don’t think the technology transfer
issues are so critical. There was a recent
attempt by a Chinese firm to acquire a stock exchange in Chicago, which the United States turned down. Each of those two countries
have an authority whose purpose is to accept or deny a transaction that would involve technology transfer. That’s how it’s currently
being controlled. But, right now, because of
these trade negotiations between China and the United
States, this is all in play. How much will American
firms be permitted to do in China without transferring technology. How much trade involving technology will go in either direction? That’s all being worked out. It’s a question that’s
extremely pertinent right now. – I have to tell you. As a reporter, if you go to Beijing and you’re in the 3rd Ring, you feel like you’re in Silicon Valley. Except that the air quality is much worse, and they don’t seem to
worry about this thing called democracy. – Yeah. – Culturally, it’s so similar it’s eerie. I teach a course on
China’s financial system for our graduate students
here at Stanford, and one of our speakers
is a venture capitalist who works in both the
United States and China. This is Richard Lin. He described Beijing as
second only to Silicon Valley in terms of the vibrancy of the
tech entrepreneurship scene. They’re moving very rapidly there. They’re government is heavily involved. Their venture funds, a
significant amount of that is directly coming from the government. – A general question about
these next generation financial technologies. Do you think as rule they
are either stabilizing or destabilizing in the sense
of the kinds of problems that our financial
markets face on occasion? Can you make any generalizations? Do they give us guard rails? Do they strip away the guard rails? – I think so. From my view, I’m gonna take
a view of a technologist. Used to sort of seeing software upgrades go from once in two
years to once in a year. And now it’s all happening continuously. Certainly hardware and firmware updates are all reasonably frequent. Financial inter-trading platforms or financial institutions
for regulatory reasons or even just having so many participants and they all need to get ready to move to some new thing, have
moved slowly in the past. But, increasingly, I think,
the disruption is coming. Is correct. One of the advantages of
freeing these networks of the worry of accurate
clocks is that it can move them into the cloud, for example. That’s one of the big
motivations and goals. Because now you don’t have to be worrying about a small number of gateways. The cost of participating in these markets goes down dramatically. Because it costs a fair amount of money, a few tens of millions
of dollars to be present at all of the venues
where the exchanges are, and get yourself one of these cages, so your machines can sit
there and participate. I think that disruption will happen. And now you’re at the pace of cloud. If cloud providers make
available more tools, which they have, then you can provide the trading strategies more frequently because the market data
is available right there. This is sort of what’s exciting in a way. That the whole ecosystem
could change fairly quickly. I think in that sense I can see it. Disruption of commerce. – I wanna add to that. I think Balaji is exactly
on the right note, which is that the demand
for speed and volume is growing by leaps and bounds. The technology just needs
to stay ahead of that. All of the case, not all,
but many of the cases of financial crashes that we’ve had are cases in which the
technology was not able to keep up with the demand for speed. Today, everyday, there
are 17 trillion dollars of transactions on the
world’s large payment systems. Were it not for the various
stable vast technologies enabling that, global commerce
would just have to slow down. That would not, People would not have as many jobs. They wouldn’t have as much to eat. The whole economy rests on developing a financial system
that’s capable of staying ahead of that demand. – It’s another kind of
green revolution, I guess. – Yeah, (laughs) good one. – I don’t wanna leave China. We talked before about
this social credit score that the government of
China has instituted. I was thinking about it. I use this service called
Mint, which is a free service, and it shows me my credit score every time I go there. I thought there was only one word missing. It’s a cultural difference. The credit score may already be being used like a kind of quasi social
credit score in America. Do you, are you worried? What would be the impact
of that kind of technology? – Yeah, just to make sure we’re
all on the same plane here. John’s talking about the
development of a system in China that basically
monitors individual behavior and keeps track of who’s
been doing good things and who’s not been doing good things. – [John] According to the government. – According to the government. That’s a very centralized system. It’s very controversial. It’s very effective if
you want to be in charge. It has other short falls. I don’t think it would go over very easily in the United States. On the other hand, a friend of mine visited my office two days ago. This is a banker, who
just bought a database for only $250,000 that
allows him to monitor everyone in the room, everyone in the United
States through their GPS. They can see. If you’ve signed up for
any kind of services on your iPhone, chances
are you’ve checked the box without it knowing it,
that allows that service to monitor your location and
certain types of transactions that you’re doing. So when you walk into Walgreens, or you go into the, perhaps
to get a driver’s license, or go to the bank, many
of those kinds of things are being monitored. This bank analyst is gonna
use this database for. He’s not gonna use it the
way a social credit score would be used, but another person could keep very close tabs on you. He said when he bought this database and realized that he
can monitor individuals right down to where
their home location is. – [John] So he.
– He decided to turn off all of his, uncheck
all the checks on his. – [John] This is anonymized
data that is not anonymized. That’s right?
– Not really because it’s anonymous, however, it is able to
monitor your location including the residence of your home, basically where you are. They can basically get. By merging with other
databases can find out exactly who you are. What you’re doing at
most times of the day. – In Europe, there this
things called the GDPR. Would that prohibit this kind of? – Yes, you can request
that you get unchecked from all of this. And you can do that. He did it himself. My friend did it himself by basically going back through all
of his iPhone connections and turning them all off. But, yeah, that probably would keep that in check to some extent. I don’t know. I’m not an expert on
this particular system. – There’s also a school of thought that blockchains, which are
supposed to be immutable, but design would also
run afoul with the GDPR. Have you looked at the
problems that they would face? – No. Balaji, any ideas on that? – No, you’re right. It’s immutable. But still anonymous.
– Interchangeable. – Anonymous, nevertheless. You could potentially preserve
some aspects for anonymity. Because that’s the other
aspect of blockchains’ time. That is interesting and useful. Like fait, hard currency. You can destroy it. So longs as they’re immutable. But it’s anonymous. The whole fit can change. It doesn’t stick to you in anyway. I think that feature
exist with blockchain. It’s a question of how sticky
something can be made to be. As far your personal
information is concerned. But, just one thing. The disruption of when I hear Darrell talk about payment and currency systems. The idea that a mobile
phone in a person’s hand or people’s hands. A service offer from the cloud. This is sort of a big. For me, as a networks and
the cloud type person, this is a big deal. That’s changed everything. If you look at us hailing for taxes and now you hail on your phone. That’s a phone in my hand and
some service in the cloud. Retail, Amazon’s done that to retail. We used to go to the
department stores to buy, but department stores
are becoming like banks. No need for them in the physical location. Are not relevant anymore. Physical stores are not
relevant necessarily. So, I think this. Something that is a service offered to you for which you don’t need to
have cash payment directly. Mobile phone payments are adequate. Then you get a real service back. Okay. I think that disruption as applied to the financial services industry itself is essentially what’s
getting out of the way now. – Well, that raises on
interesting question and maybe theoretical. But, I’d like to ask your reactions. There’s this mysterious
absence of inflation in the American economy right now. There is a particular school of thought that it’s because of Silicon
Valley’s technologies that there is no inflation. I was wondering if you guys either have a view point about why is there no inflation. – Well, actually, there was a study just completed about a month ago that said that’s not the hidden reason. It’s not like. That’s not the black matter in the economy that’s keeping inflation down. In fact, the added technology
is not contributing to GDP as much you might think. A lot of it is being used for
personal ledger activities like playing games on your iPhone. (John laughs) But there is. It is an interesting question
why inflation is so stable despite all of the rollercoaster ride that the economy is taking. It’s been a big debate. It’s a mystery among central bankers. Some people ascribe it to a broken theory known as the Philip’s Curve,
which is supposed to show a relationship between
unemployment and inflation that has stopped working. And people are completely
mystified about why it’s not working anymore. – I have one small claim to
fame in the fintech world and that is that my
uncle invented the ETF, which is a very small claim but. – Wow. Uncle Nate invented the spider
at American stock exchange some years ago. ETFs have become a very
significant part of the economy. There is also an argument
that hidden in the ETF world is the kind of leverage
that got us trouble in 2008. I was wondering. I know this is all speculative. But I was wondering if that’s a concern. – Yeah, I was for three years. I was on the board of
directors of iShares, which is the largest ETF complex. We were getting these queries about is it really true that there’s an explosive problem here where people try to cash in
their ETFs at the same time that the underlying assets
will have to be sold, that the economy will
head into a tailspin. Of course, you might
second guess my opinion since I was at iShares, but it’s simply not the
problem that people suggest. The ETFs don’t add to the problem, in fact, mitigate the problem because in order to cash in an ETF, unlike a normal mutual fund, nobody has to sell the
underlying stocks or bonds. You can just sell the ETF. It would be like taking your mutual fund then selling it directly rather than forcing someone to
sell all of the assets in it. It’s not really the problem
that it’s made out to be. Now that’s not to say that
there won’t be hiccups in the system if there’s
a bad run of events and people decide they all
want to sell their ETFs. Yeah, the prices will go down a lot. Then other investors may
step in and buy them. But it’s not gonna cause
a financial crisis. – I also wanted to ask you
guys about the financial system and the future of the work force. Because there is a kind
of a narrative about 2008 and the great recession. That one of the consequences is these large corporations contracted. Then when they came back,
largely through business practice re-engineering, they came back and they grew without jobs. There was a layer of
semi-skilled white-collar labor that went away, millions of jobs that were the real casualties of that recession. In the same boat, we had this
mystery around inflation. We have a mystery about employment. We have the lowest, despite
all of the forecasts in Silicon Valley that the
robots are gonna get us. There’re more people
working in this country than ever in history right now. There’s lot of caveats, but still we have pretty low, We have reasonably low unemployment. I wondering if you guys. If we can now look
forward and see this wave of financial technology
that’s sort of spreading out through the economy. Do you have a view one way or
the other about its impact? – I have a game. I think it’s a reallocation. If you look at what
happened to transportation. Detroit as it was losing jobs and layoffs were taking place. Suddenly, Uber and Lyft are
picking up a large workforce. It’s sort of a different kind
of job in the same industry. Re-definition of the
industry was underway. I think if you look at, I’ve been speaking to a number
of Wall Street type firms, I keep hearing them say things that I’ve been hearing in the Valley like, we are a tech company that
happens to be trading, or happens to be, or is it. Amazon is a tech company
that happens to be in retail This a mantra that is now
being repeated in Wall Street. We’re a tech company that
happens to be a bank. It must had been a bank, etc. What that means also is
that the amount of tech uptake into this country, not just consumers of technology
with there being forever, but creators of technology,
like some new tools that can be more
ubiquitous and widely used. That’s what they’re heading towards. If they require the bank
strength to do that, that some of these tech
companies you have, I think that’s going to be
interesting re-definition. – [John] Thoughts unemployment? – I’m quite an optimist on this. You just look at the progress
of human civilization. The extent to which we’ve
hit the ceiling in terms of our ability to find new
and exciting and productive things to do, we’re
not even close to that. So as new technologies develop
they’ll be leap-frogged by other technologies. Many people of the audience,
especially the younger ones, will be coming into the workforce to either develop or exploit
these technologies. There’s no particular reason, at least not in economics, to believe that this a, at some
point, because we’re all gonna be out of work because
we’ve gotten so productive that there’s nothing else
left for us to do ourselves. I don’t think that’s coming. – Along those lines, do
either of you have thoughts about the future of financial advice? I was looking today at these startups like Betterment and WealthFront, and you look at assets under management, you got several billions dollars. Then I went over and
checked Vanguard and Swab, and they have many trillions of dollars. What’s the future of the
automated financial advisors? Is that something that
will be as disruptive as? – It’s getting quite popular. It’s a very inexpensive way
to get financial advice. I think it can be used for, for example, if you want to help the
people in your company to manage their retirement savings through your pension plan. Personally, I don’t think
it’s good value for your money to have each one of them
go off to an independent human financial advisor and
get individual financial advice much of which will not be that affective. I think having a more machine learning or a more algorithmic approach to this that can be cheap is the way to go. – Can you imagine a world in which Siri and Alexa and what have you will be the pathways, conduits
for that kind of conversation you’ll have with your financial advisor? – Could be. I think it literally
could be done your iPhone. On your smartphone. – I’m going to be more positive. He’s from business school, and I’m from school of engineering. So, yes. I think algorithms are
gonna start advising. They’ll probably do a good job. They’ll probably pass
the Turing Test before. An algorithm is gonna be a
financial advisor before. In that sense, it’s coming. It’ll be faster. Yeah. – Let me ask you about inequality. In the last two decades,
there’s been a dramatic increase in inequality
in the United States. Do you think technology in Silicon Valley is to blame at all? Why are we seeing this
tremendous spread of wealth and poverty in America? – I don’t know who is to blame, but it’s definitely, it’s happening. Both are happening. It’s related. It’s too much to say one
is causing the other. I don’t if Silicon Valley’s
upward is big enough. I think manufacturing is where
a lot of it is occurring. One thing I’m noticing
is the amount of things people don’t mind reusing, it is going up. It is a very interesting
generational thing that I see a lot of. Ownership of things is not as
important for younger people, which could see as sort of
tapering off on manufacturing. – [John] The car industry. – The car industry is an example. – [John] Yeah. – But others, too. People are staying in other people’s homes for a lot longer than they used. People wanted to own homes early soon. Quickly as possible. But that’s not the case anymore. Not necessarily. That’s also happening. That’s not similarly inequality. But that represents a certain
kind of demand going down. – It’s a question of. You spoke earlier, John, about a portion of the middle class. The phrase that economists use is the hollowing out of the middle class. I think you have seen
a lot of that happening in the last 50 years in the United States with the extremes of income inequality and wealth inequality in this country. To a certain extent,
it doesn’t need to be. There could be a much more
progressive tax system that redistributes some
of the income and wealth without necessarily
killing off Silicon Valley and all of the other
sources of wealth creation. You could create a lot of wealth and then distribute
the wealth more evenly. It doesn’t have to be
in the form necessarily of direct payments. It can be through healthcare,
minimum wages, housing. There’s a lot of difference views about whether that’s an
affective approach or not. We’re seeing a much fiercer debate now in the political spectrum
about redistribution, as you’ve probably noticed, with some of the new
democratic congress people advocating much higher tax rates. Personally, I think the income inequality in the United States is out-of-whack. And it should be fixed. – We’re gonna have
questions from the audience, but I wanted to ask you one other area. There a interesting current
debate in Silicon Valley that plays on it, business strategy, it lives on top of the platforms we’ve created here in Silicon Valley. I was started by Reid Hoffman who was advocating this business strategy which he calls blitzscaling which is embodied by Uber and Lyft, as you raise a tremendous
amount of capital, then you dominate a market. You get a monopoly as soon as possible, as a business strategy. It’s probably not new in Silicon Valley. I think Larry Ellison said, it’s not enough to win;
every one else must fail, which is kinda an example of blitzscaling, I would imagine. Recently, Tim O’Reilly, who runs a publishing company, and he’s an investor, has sort of criticized that as a strategy because it might not be
the best thing for society. Amazon, for example, grew
without taking a lot of capital. They borrowed money, but they didn’t take investment capital. That was a different approach. O’Reilly’s company grew organically over a long period of time. It might be argued that having a non-monopolistic situation
in a particular market might be good for an economy. If you had many competitors,
it might benefit consumers. Although, I guess, Google
is a counter example. Perhaps. Have you looked at that argument? Do you either of you have thoughts on the Silicon Valley strategy-dejour? – We come back full-circle
to where I started which is with WeChat and
Ant or Alibaba in China. Those, that’s a duopoly. Blitzscaling is a fantastic word for describing the competition
between those two firms to get to scale as fast as possible because of the enormous
economies of scale and scope. If you just think about, why does anyone wanna join WeChat? It’s because everybody
else has joined WeChat. And the faster you can get to scale, the greater your chance
of dominating the market. Uber verses Lyft is a similar example. These are all a social,
economic enterprises that thrive on scale
and network economies. You should expect it
in the payment system. You should expect it telecommunications, social networks of all types. Yeah, we should expect
monopolies or duopolies to dominate ever more as the economy becomes more integrated through faster technology like fintech. – I actually think that in
one of the things I worked on for about eight or nine years, as you know, in transportation. There are many routes in a bus system or a metro system that are not running at full capacity and are,
in fact, not profitable. Where as a requirement of a public service like public transport, that it takes some routes with loss, because there’s still some people that need to be serviced on those routes. It’s just simply the way a city is. You happen to have a household there. There’s not much traffic going over there. But you need to get home and get to work. So your route is running at a loss, but the system as a whole
will subsidize that route. This is true for the airline industry. It’s true for many things. I worry that if systems become monopoly, and not just a duopoly, that the neglect that
some areas may suffer from are really going to be
difficult to explain. Or to, for a city to bare. This is already happening. Sometimes, when you’re
traveling abroad or whatever. Be in India, and you say, You hail somebody, and they show up, and if it’s not on Uber or somebody, you order on your phone, who already knew where you were going, often they listen to where you want to go, then, sorry, I’m not going there. You lose some 20 of these wackos that are just passing you by after they’re not going
where you want to go. So this is an example of something that’s very tough for public systems. I think that’s not necessarily desirable. When people burn money, what they do is, this is the sort of loss
that they kinda take on. The system can take it on. – So, there’s some good questions here. Let’s let the audience participate. This is for Darrell, the first one. “How do you think the digital economy “is going to hurt the jobs
dependent on the cash economy?” – Okay, well, it used to be that marijuana was a cash economy. That’s probably not the case anymore. – (laughs) It still is. They don’t let the bank yet. – [Audience Member] By regulation. – Ah, that’s. I’m trying to think about what. It makes sense. As cash becomes more and more scarce, it’s gonna be hard to run
your business on cash. And people with cash
won’t be able to use it. But I’m trying to thing about what significant parts of
our economy that are legal can’t be done without cash. – [John] A lot of them have been dragged into the
gig economy, haven’t they? – Yeah, I still pay cash for my haircuts. $17. (John laughs) But, I suppose, if there were no cash, they would have to take a credit card. (laughs)
(John laughs) I don’t know. That’s a good question, but I can’t think of
any parts of our economy that depends so heavily on cash that we we can’t do without it. – Yeah. Here’s a question for Balaji, first one, “Blockchain and related systems require “intense electricity, server
equipment, and land use. “Who should pay for this?” – That’s the way it is now. But it is one of those first generation, second generation sort of thing. What I’m sure we’ll
read five years from now is the same thing we read about
the Princeton ILLIAC machine taking a big room and being
less than some kilowatts. I’m not even sure it got up to kilowatts. Whatever the number was, right. Some amount of memory
it had, it was so small. It’s going to be like that. The inefficiency,
technological inefficiency. People will find a way
to overcome and solve. As to whether the puzzle’s
solving is required, which is why all of this is happening, that’s not clear, meaning that other ways one can conceive of
potentially doing things. People will conceive of them. – The second one is someone who basically didn’t take your message on stability to heart. “We have seen U.S. stock market crashes “because of robot computing. “If we speed up robot
computing, is that better?” – Right, so this isn’t about that. It isn’t about speeding up. The faster fintech he refers is not to what the market
participants are doing, how fast they come back. That has come down as a number. This is more about if
you have an exchange, could you get transactions from one side of the exchange to the other at about the same time
or exactly the same time. This is about removing the jitter intrinsic to packet-switch networks. But doing is, by adding an enormous delay to the whole process. One could do that. And that is an example of how
one could add a speed bump, as it’s called. There’s a trying to do it without
having to do a speed pump. – [John] Yep, yep. Another question, “When I buy stocks, “I usually put in a quote,
market order, unquote. “Is that good or bad? “And by doing so, does the…” booker, Booker? “…broker…
– Broker – [Audience Member] Broker
– “…make most money “in the spread?” Anybody, wanna take it?
– You? – So, if you’re in a hurry, you can put in a market order, and you’ll pay a very small fee, because somebody has been waiting for you and willing to buy or sell
based on your market order for a small profit. But if you’re not in a hurry, and you know exactly the price
that you’re willing to pay, you can also put in a limit order. It’s pretty easy to do. I’ll bet many people in
the audience have done it. Both strategies work depending on whether you are in a hurry or not. – This one I don’t quite understand. “How do you feel about
payment from order flow? “Does it ultimately help
or hurt retail investors?” – Order flow. – I think I know what
that person has in mind. You might think that you
would need to pay the exchange a fee in order to execute your order, but in fact the exchanges are so anxious to get a large number of orders that they will pay you to put limit orders on their order book. Because that creates more liquidity, and makes your exchange more attractive. It’s a big controversy. But, to tell you the truth, no one has been able to show one way or the other whether this is a good idea or not. There’s not a lot resting on this one. – Okay. Here’s something. I talked about glass, but I
didn’t talk about wireless. How 5G, the much faster
wireless communications, could it create opportunities in fintech? So this new global technology standard, will it meaningfully change fintech? – It’s a way of communicating. Unless you’re taking your subway train or your metro and you
wanna trade from there. It’s not clear that that might change. It does change something which is the number of
connections you can get into, let’s say, the internet. You don’t have be tethered or wired. You don’t have to be. Because a lot of these things, it’s about approximating,
not in a physical sense, but actually in ether-sense, how far are you from some place where you should be transacting. It could change fairly dramatically. How much of the city
could very quickly get to what ever other part of the
city it needs to get to. That’ll change. Now is that finance is the
only main beneficiary of this? Yeah, it’s one of the beneficiaries. I’m not sure it’s, ah. – Do you think that 5G is a
threat to wifi as a technology? – Some perhaps. But I think wifi will continue to exist in the sort of really
small local area sense. 5G can carry a lot more data. But it is intrinsically gonna be more of circuit switch type of thing in the main because
it’s gigabits per second of data it can give you. I see more as a fatter pipe, and a really high speed one at that. One thing I can say is that the clocks need to synchronize to get
that kind of benefit of 5G. If a base statistician’s
clock is well synchronized as being down to a few nanoseconds, then the amount of
throughput that you might get from a 5G network will be
somewhat proportionately less. That’s one of the things
we’ve become aware of, that 5G is looking for clocks. – Once again to China. One of the numbers that I saw today was that the rate of
new cell towers in China is an order of magnitude faster, no? – Yes. – It is?
– I completely agree. – Yeah.
– It is staggering. It’s just popping up. – So if there is an advantage to 5G, they’re gonna get there first. – Yeah, infrastructure is critical. – Yeah. This is for Darrell. “What are your views on the tension “between distributed,
i.e.; fragmented markets, “and the quality price discovery process, “especially given greater
number of exchanges “in regulated…” N-M-S – Reagan M-S. – Reagan M-S. – [Audience Member] Reagan M-S. – Reagan M-S, “…in the U.S.” Does that makes sense? – Yeah. Whoever wrote this question is a deep aficionado of financial markets. Let me unpack it a bit. It actually goes back to something Balaji was raising just a minute ago. Right now, if you send an order into the, let’s say, the
New York Stock Exchange. If it’s gone through your broker, it’s actually searched for
the best price available on every exchange in the country. That tends to consolidate the available demands and supplies of the
stocks in the United States in a way that gives more
efficient execution. But that’s not the case
everywhere in the world, particularly in Europe. There’s a lot of concern
about fragmentation across different exchanges. Because, if I send a large
order to one exchange, it may get a very bad price relative to what would had been the case had all of the exchanges merged into one, similar to, but not quite,
like the U.S. NMS system. I actually was just in Basel,
Switzerland a month ago on this exact issue debating with a number of market participants and regulators about how serious a problem this is. And I think it’s a serious problem. I have a doctoral student
and I are actually working on this one right now. How inefficient is it to
have market fragmentation? I think it’s relatively inefficient because it means you’re breaking up the sources of liquidity
that are able to absorb your demands and supplies
of trades into bits. You have to go put them
back together again in order to get a good trade execution. – So here’s what we
talked a little bit about before we got started. “How fragile are the
fast financial systems “to mass hacking by state
sponsors or rogue tech groups?” – Wow.
(John laughs) This is something that three
of us were just discussing before we sat down here tonight. That’s pretty interesting. Okay, I’ll start, and then. This is a serious problem. Many people cyber risk
to the financial system as the biggest source of
financial stability going forward. In other words, before it
was the big bank collapsing was the biggest threat, and now it could be a cyber
attack on the financial system. I’m gonna be in Washington,
D.C. in two weeks talking about this problem and what would happen if a major bank had a cyber attack. Would there be a bank-run? And, if so, how would you handle it? Because large depositors
are not just gonna sit and watch their money
frozen into their accounts even if it’s not gone. If they can’t get it right away, they’re gonna panic and try to take it out. That could be very
damaging to the economy, if everybody tries to move their money out of the banking system all of a sudden. It’s a very serious problem. – Do we need better protection that the circuit breakers we have? In that kind of situation,
what would happen if the circuit breakers were tripped? – You don’t want to do
the circuit breakers. You want to do the opposite. You want to keep everything
as open as possible so that the money can flow. – [John] Who makes that decision? – Unfortunately, the decision may be made by the depositors and our own users of the financial system. If they pull back, then that’s bad news. – On the engineering side, many financial companies that do trading derive time through GPS now attached to roofs of the buildings, which locate it with the exchange. This is a fairly common thing to do. In other words, if you
have good access to GPS, you know your location accurately but also you can get fairly accurate time. Then you run a wire. Figure out how long the wire is precisely. Then you know what the time is here if the time over there was blocked, okay. Now, this is very easy to be jammed. One big fear is if I
just jam all of the GPSs, then this is gonna be a huge problem. Already in anticipation
of things like that People, companies have
been taking dark fiber from time authorities. For example, the national
physical lab in the U.K. supplies dark fiber based time to a number of the financial institutions, directly. That’s an example of how. Canada, as well. They’re doing things like this. That’s one way. The other reason some of the companies are looking to move to the cloud is that cloud had to look at security. They constantly do it. They probably have very good tools, certainly, potentially much better tools than financial institutions have for protecting against attacks. So that’s the other reason. I see access management
being the other one, big thing that the cloud is good at. I can see how it is definitely fragile. But, at the same time, people are aware. Counter measures are being adopted. Think of it as one of
those things that you just. The good guys get ahead a little bit. The bad guys get ahead. It’s a business to be in forever. – We have a question on the
other side of the ecosystem, actually, other end of the ecosystem. By that same token, “How confident are you
that startup fintech “companies can protect
client data and assets “from hackers and security intrusions? “Will they still protect client privacy? “Or, will our privacy evaporate?” – This is probably better for you. – Well, I. Sorry, if you host yourself in the cloud, Coinbase has been operating
in the cloud for awhile now. Startups like RobinHood,
etc., probably operate mostly in the cloud. Then you’re about secure as cloud is. It’s pretty secure. But nothing is 100% secure, in one sense. There’s also this other side of how much money evaporate compared to any other alternative. It’s not about fintech startups. It’s about any other way in
which a service can be offered. All your credit cards. Transportation companies, for example. Or elsewhere. Those are secure or not. – That raises another question. Just on the side. “Would workable quantum
computation be a huge threat “to market security?” – It would threaten a lot of things, not just market security. At the same time, again, for every one of these technologies that you can come up with something to try and break it, the same thing would give you away to protect yourself against it. Then the question is how can you move from an existing thing
to this quantum world? You’re talking about regulations and how would it operate over there? The human components of this will have to be upgraded. Our ways of policy setting. – Just to tag on that, on the issue of privacy and data. The events of the last year, particularly the testimony
of Mark Zuckerberg in congress and his decision yesterday to announce that Facebook is now going to prioritize privacy. These are signs that the mass collections of personal data are being
perceived as a threat and that we are making a fairly sharp turn in the use of technology
towards greater prioritization of protecting individual data. There’s gonna be a lot more on this coming in the next few years. You’ll see either legislation
or large tech firms trying to get in front of legislation by demonstrating that they
can protect individual data and that they’re not gonna pass it on as freely as they have
been in recent years. We’re coming close to the end. Let me try and combine
these two questions. We’re back on the China
theme, as a competitor. The first one is, “Is there a chance for Chinese currency “to become the reserve currency, “encroaching on the U.S. hegemony?” And the second half of that is, “Do you see Alipay or
WeChat Pay participating “in the U.S. market?” – Those are related but not that closely.
– Exactly. Let me take the first one. Because of China’s own concerns about capital leaving the country, The use of their currency, the Renminbi, in global markets has actually gone down in recent years, not up. It’s a very small fraction of the usage of the U.S. Dollar on global markets as a reserve currency. It’s not even close. Having said that, if China keeps growing at its current GDP growth rate, history shows at some point
it’ll become a reserve or the reserve currency. But we are far, far from that now. As far as WeChat Pay, it’s
in the United States now. You can use Alipay to buy
United Airlines tickets now. I see on their website. WeChat Pay is commonly
available around the world. I read yesterday, a
Norwegian touring company has signed up a deal
with one of those two. I forget which. So that when you go to
Norway to do tourism you can use their
payment system in Norway. Yeah, it’s going global. If the United States, providers of these kinds of services don’t
get in gear quickly, that will be the defacto
mobile payment system around the world. – So Facebook is about competing
with WeChat Pay globally. That’s what’s going on? – Good question. I’m not sure what they intend. Everybody’s speculating
what Facebook has in mind. If they go that way, they could do quite a good job in the U.S. – Thank you guys. I learned a lot. Why don’t we give them a hand? Thanks, everyone.
(everyone clapping) (everyone clapping)

Leave a Reply

Your email address will not be published. Required fields are marked *