The FinTech Revolution

if you took a class with me you know I talk really fast and I'm gonna try and go through like 168 slides in one hour or so so fasten your seat belts and and don't don't don't blink so I used to study financial history back in the day and I'm used to love to go around and take a look at all these old Bank buildings like I grew up in Philadelphia and we had the first bank of the US second bank of the US and they're so imposing and they're meant to look imposing because they're meant to look like they last forever okay so on the Left we have Mellon Bank in in Philadelphia on the right we have Wells Fargo Bank here in San Francisco don't they look like they're going to last forever and so a lot of other companies wanted to imitate this architecture to convey the message of permanence one such company was Borders bookstore okay and you know you look at this you think wow looks pretty imposing and then you look down here and its final days right I mean who could have imagined that the bookstore would disappear from our lives I used to spend literally eight hours a week in bookstores until 1998 when Amazon was created and it went from eight hours to zero overnight right and so a lot of people think the banks will never die you they have franchises they have government regulation to protect them but a lot of people also thought the bookstores would never die and of course they have completely almost vanished from our lives now of course banking is a little bit different and so this is the question that we're gonna try and answer because of all the innovation and finance you can see the massive run-up in investment in FinTech and this is VC investment in fin tech startups now this is a small percentage of the investment because of course all of the legacy banking institutions are also heavily aggressively investing in in technology in order to transform their businesses but look out banks every single piece of functionality that you currently do well from mortgages to payments to credit provisioning right to checking to payments they are all under attack from all of these in vidual startups okay and so the future is for me fascinating as a financial historian I love watching this history unfold in real time if you like European banks we've got one of those too okay so I'm gonna borrow from the World Economic Forum they came up with this nifty diagram which walks through all of the different sectors of finance that are being transformed by technology and so we're gonna try and go through all these if we can as quickly as possible and if you want to go deeper we can go deeper you know figure out a way to come back to Haas and continue your education right so let's start with deposits and and lending let's go back in time a little bit to the 1919 80s I think it was nineteen early 1990s there was this Bank called Signet Bank okay and this was the 350th largest bank in the United States how did they go from being the three hundred and fiftieth largest bank in the u.s. to being the largest credit card issuer in the world in a very short period of time how did they do this how did they grow their business so dramatically and so quickly one guess big data okay big data now look what have lenders been doing lenders have been using data forever look finance is an information intensive business it's all about information if you studied finance you know about adverse selection you know about asymmetric information you know about delegated monitoring right it's all about information whoever has the best information wins so what are all these credit card companies doing up until the early 90s right they were all relying more or less on the same credit metric credit scoring model which came from fair isaac's FICO or something similar right from Equifax or from TransUnion okay and which took in a whole bunch of different characteristics of all the different borrowers okay and then you know came up with some predictive model now this is data science now we call it data science now but people have been doing this forever if you've used a spreadsheet and done a regression well then count yourself a data scientist I now Knight you write data scientist alright you've done it because all you're doing is you're taking a whole bunch of customers in your training data right and those are the rows and then you have all these characteristics of the customers that are sitting in columns like you know outstanding balance number of cards number of times you paid late etc okay and and then right with that data you then track the customer forward and look to see whether or not they default on a loan or fail to make a payment and that's all the FICO score is is doing right so if that's what they were doing then and that's what they're doing now what has fundamentally changed and this is it right so if anybody asks you if you were here last year you learned this as well or the year before if anybody asks you what data science is it's training and scoring give me some historical data train a model on it use that model to score new data and come up with a prediction about somebody's future behavior okay very very simple you get the bottom row and then you guess what fills in the final column which is the target feature super super easy lots of different methods okay I could teach you in an afternoon how to do this sort of thing okay so what has changed what has changed is the volume of data okay the number of rows and also the number of columns the kinds of things that we know about our customers okay now what Capital One did is add a whole bunch of rows so previously all of the other credit card companies simply refused to lend money to anyone whose FICO score was below a certain cutoff and so Capital One said you know we're guessing there are some diamonds in the rough we're guessing that some of these people who have never gotten credit are probably good for it but we'll never know unless we actually give them credit so they designed an experiment where they just blasted out credit cards to all of these supposed deadbeats and guess what happened most of them were deadbeats okay there was a reason why the banks weren't lending to these people okay there was a reason there that's stupid but they are lazy okay they are lazy because within this you know pile of hay there were a few a few needles but the only way to find out who the needles are is to actually go out and collect data on them and then track them going forward and see what happens now if you looked at the P&L of Capital One during the time of this experiment you would have seen that they were losing like Bo loads of money because of all these these deadbeats and so Wall Street was like these people are crazy they're not so we got to get rid of them the CEOs and lunatic you know we got to shut them down etc but they managed to hold out long enough to get the information which enabled them right to come up with a scoring model that utilized information that wasn't available or hadn't you know the model that the other banks have was not sufficiently rich to make predictions about these people okay so if you look at their P&L how would you characterize all the money that they lost with all of those customers anybody here a bank accountant okay or anybody remember accounting or do accounting right this is gonna be a loan loss this is bad credit this is a write-off it's gonna be an expense in the period right in which you expect it to occur however the better way to think about it is as Rd right you are investing money to acquire information that helps you to train up a better model okay and the companies that are willing to spend money on data acquisition are the ones that are ultimately gonna win and in this game okay so that's how they were able to expand their business by adding more rows right and more columns to their spreadsheet now let's fast-forward now the kinds of information that you can get about your customer is so much richer than the kind of information you could get back in the day right we all know about how Amazon and Netflix right and everybody else use big data to provide recommendations to the customers we all know about how faith book and Google use all sorts of information about you to provide customized news feeds and customized write search results and so forth so why shouldn't the banks do something similar so what kinds of data could banks look at that doesn't appear in a typical FICO score right typical FICO score how many times were you late let me give an example suppose you're coming out of Harvard Medical School or you know Haas School of Business and you have a couple hundred thousand dollars in debt and you're 27 years old what does your FICO score look like it doesn't actually look that great and yet would you bet on this person I bet on this person if you want a job come talk to me I I'm interested in hiring you okay because you know you have a ton of human capital which doesn't show up on your FICO score so there's a company called sofa social finance the way it got started was saying hey there are all these I'm sorry Stanford MBAs and they asked right the student loan people how many of these Stanford MBAs ever default on their student loans they were like well you know we did have a guy who died you know and and he didn't pay off the loans and so it's like alright so unless you're you know I mean the executive MPAs you got to worry about the sorry guys 30 year loan maybe not but for the full-time MBA I think we can bank on you know 30 years of lifespan right where they can pay this thing off all right they're not gonna unless they go join an ashram or something and disappear off the face of the earth they're gonna pay off that's what I guess more of a Berkeley issue than a Stanford Stanford issue okay but why isn't human capital incorporated into this model okay what other kinds of information right so look Netflix knows that standard classifications you know that they use in the media business like males 24 to 36 that's so crude that's so stupid that is so 1970s right we could do segmentation down to the individual level and identify the credit risk of an individual based on individual characteristics so all these FinTech startups are looking for new ways to evaluate credit okay you've seen some of them out there okay no then all around still a lot of them have failed a lot of them been acquired and so forth you know what do they do these are peer to peer they call them peer to peer lending platforms they're not really peer to peer in the sense that you think what they are our algorithmic lenders they're lenders that use write a platform that integrates and assesses different types of information and gives the lenders the ability to use different types of information to evaluate the credit risk and then you know of a partner bank that does much of the lending so what about for instance location data location data is used by Facebook to do advertising this is one of my days last summer right and I was at Facebook headquarters and they said oh yeah we've got this feature whatever I think it was two summers ago I was like oh that's interesting let me check it out so I looked up this is me one day starting off at Hoth you know going down do you have anybody who's ever commuted to the valley you know about this little dogleg that you know ways often sends you in weird places and then Here I am right I'm at LinkedIn headquarters I'm at Google headquarters Facebook headquarters you know I had lunch over here and I had hanging out here in the afternoon and then I went to K&L wine merchants spent some time shopping you know if you're a bank what do you think oh that's an interesting piece of information could be good I mean K&L does say you'd want maybe want to know are they buying the good Bordeaux are they buying their you know the the cheap one I don't know maybe that's a valuable piece of information okay maybe this guy's got a drinking problem who knows right like this could be this could be important information okay then BOOM okay the guy spends a couple hours at the opera house well is that good or bad do opera goers default with a higher probability or a lower probability I don't know but I bet it's different from the average person controlling for everything else why not incorporate that into our lending model now look there are regulations that tell you what you can and cannot incorporate but the amount of information in your facebook profile your LinkedIn profile is profound right here's a start-up in Africa that is using location data from your cell phone and activity actually in your cell phone to give credit to people who have no credit history because most people in Africa have no credit history but they do have digital footprint and digital crumbs that are everywhere okay so really interesting insight what about your personality your Big Five personality profile you guys know about this okay so conscientiousness is one of them people who are conscientious all being equal tend to default at lower probabilities so how do I know what your Big Five personality profile is I could ask you are you conscientious oh yeah heck yeah mr. banker yeah definitely then why are you twenty minutes late for the meeting oh well you know okay how do we know what your big five personality profile is how would you find out you could ask somebody's friends you ask somebody's family you know what they are not very good at assessing your personality you know who's really good at assessing your personality Cambridge analytic I know Facebook right and I'm not joking about Cambridge analytical right so if I take a look at your whet your Facebook profile and agreeableness is one of the big five okay if you see if you if if you well eat okay hae-sun zoo these people lo agreeableness okay maybe you have some friends who have liked I hate everyone okay not agreeable okay Book of Mormon agreeable okay it's a funny play all right I don't know so how do they know this because they've asked people at Cambridge actually Cambridge did this real cameras not cameras analytical right they actually asked people to fill out these very extensive surveys and then looked at their Facebook profile and match them and your Facebook profile does a better job the algorithm does a better job of predicting your Big Five than every other person in your life except for your spouse think about that that's pretty amazing okay so this is the site that was used research site at Cambridge which has been used by a lot now let me give you another example so there's a site called prosper at prosper in addition when you apply for a loan in addition to right answering all the questions about your income and so forth and they ask you to fill out a little essay why should we lend you money freeform answer you can write anything you want now do you think that the words contained and that have predictive value well we can just simply add that to our model and see if it adds value it turns out it does you can do text analytics right just like the words contained in an email are predictive of whether or not their spam you know you see the word Nigeria spam right unless you're in the oil industry I don't know right so what do you think of these words which of these words are more likely to correlate with a borrower who's gonna pay you back in the freeform essay what do you think of course if people if everybody knows this you know you can write a little book how to get a loan on prosper make sure you include these words right and you know dating sites do this right job application sites say do this and so it's gonna probably not work after a while but this study was published in 2016 which words do you think correlate just take a look these ones the other ones have negative correlations so if someone says I promise I will pay you when I get out of the hospital so help me God do not lend to that person okay they are not going to repay you okay now why I don't know like who cares well that's an answer for the psychologist this is data scientists are just looking for correlations now look still all we're talking about is fairly simple data you know the can fit no spreadsheet but of course big data is about all sorts of other things right that we could potentially use right like maybe you know when we use deep deep learning or when we use neural networks the models get very sophisticated and so rather than using like regression you're doing something much more complicated so how do I tell for instance the difference between a cat and a dog I basically start by having a bunch of humans go CatDog CatDog CatDog CatDog so I could do the same exact thing with a bank I could have a bunch of loan right evaluators credit risk evaluators come in look at these very sophisticated complicated profiles go approve approve don't approve don't approve don't approve and then I just extract from their brain whatever it is that they're doing okay and then I don't need to really understand it I just need to know that I can replicate it and that's what we're doing here right with neural nets you just have the images of the cats and the dogs people go CatDog CatDog CatDog boom and then you no longer need the people use that as your training data to generate a scoring model okay so you probably know this best from Silicon Valley the TV show if you are interested in learning about the valley and from your outside the area I always tell people just don't bother to take a class just watch this show you'll save a lot of money sorry Julie I didn't mean to steal any business from executive education but you know this TV show is free and it's really really good and this guy jinyang came up with the hot dog not hot dog classifier it's actually very good very accurate and you can access it for free on the App Store okay and so what else can we use this kind of neural net for right or deep learning what about MRIs we have doctors that spend all their time looking at MRIs like cancer no cancer cancer stroke no stroke stroke no stroke why do we need the doctors we don't we just have to observe the doctors for a while and then fire them all right because we can replicate whatever's happening inside their brains it's not that complicated and it's actually going to be more accurate and it's going to learn better and we're that's why we're seeing all of this talk of AI and the robots replacing the humans and indeed we should expect at some point that our surgeries will be done by robots and I don't see why bankers are any more important than surgeons okay so we now talk about robots taking over the world's finance jobs right banking sector ground zero for job losses from AI and robotics right technology will cut 30% of bank jobs says former city bank CEO the secret panda don't you bank once replace employees of robots what could possibly go wrong okay this sounds great that sounds wonderful if we could rid the world of the bank of it we could rid the world of bankers next stop lawyers right now it's kidding I'm also a lawyer so I'm not you know I don't want to get completely replaced so do you think it's possible at some point we will be able to completely eliminate the profession which has made so many of our alums happy and wealthy and so forth that you know really think well deep learning will never be able to completely replace the human there will always be role for human judgment and just to sort of confirm that we can see examples right the machine is not there yet your two-year-old would never try to eat a chihuahua I mean maybe yours would I don't know but like we have brains that are way more sophisticated than anything we have have yet or you know like no one's gonna throw a puppy into the fryer okay so we're still not completely there yet now this is funny but there's also some dark examples of how things like bias can get baked into algorithms and so forth this is a Google autotags all these photos I love this a close-up of a hillside next to Rocky Oh of tags giraffe where's the giraffe I spent hours looking at this trying to figure out where's the giraffe am I the idiot or is google vidiian right or this one a group of orange flowers in a field so these are sheep with little orange sweaters on them because the the algorithm is never seen right orange sheep so if it's orange and it's in a field it has to be a flower and there's one of the reasons why for instance venture capital will probably never right be completely automated now we're talking about run-of-the-mill kind of borrowers out on the street yeah great but again novel situations are gonna require humans I love this one right anybody know about these goats that live in the trees well Google doesn't I didn't know about it till I saw the picture I was like wow goats and trees Google looks at it and says giraffes they love giraffes for some reason I don't know okay all right so investment bankers may not your some of you out there your jobs are still are still safe okay now what about at the corporate lending side okay this is where things are really interesting right because you know corporations generate so much more information even than individuals now they might not have facebook profiles but you know what they have they have sales data they have lead data they have inventory data they've receivables data they have payables data right they have payroll data they have so much data and then what a banks do now they go to a company and they say give me your quarterly financials what the hell is wrong with that quarterly financials quarterly financials like when you walk outside and you're trying to figure out whether or not to bring a raincoat do you look at the quarterly Weatherly report from March 31st no right you want to know what it is right now and so what banks are now gonna do and this is what the wave of the future is they're gonna simply plug directly into your enterprise software get real-time data on all of your financials right we're not gonna have CFO's to pulling all nighters at the end of the quarter they're just gonna like push a button and print a snapshot of wherever the company is and the banks are gonna do exactly the same thing so we're gonna see real time interest rates real time credit lines that are built on real time financial data which you can get from the enterprise software now I mentioned banks but really why do we need banks at that point who's got access to the data the enterprise software companies so we need to look out as those guys as potential competitors if we're banks okay what about insurance well same thing every single truck in America now has some form of telematics in use why because not only do we want to know where all of our goods are we want to know where all of our trucks are now if we know where they are we also know how they drive we know right how fast they're going we know whether the driver is a good driver a bad driver we know if they're getting sleepy and they're veering off the side of the road so we can use this data not just to sort of provide guidance but also to price the risk right so if I have a fleet of trucks and I know how many are on the road and how they drive and what their contents are then I as an insurance company know exactly right the probability of a bad event and right the magnitude of the loss should it occur and that model is continually getting updated because whenever an event does occur it gets fed into right my data set in the cloud so at the end of it and again a lot of the stuff that we talked about in FinTech is at the consumer level my point is that most of the stuff that's really exciting is actually on the enterprise level so you as an individual there's a company started here in the information school it was called automatic' and what they manufactured was a little dongle that you could jam into your car if it was built anytime in this century and it will through your phone tell you right how fast you're driving and how well you're driving etc etc etc and you could provide this data to your insurance company so if you drive right fewer miles in the average you shouldn't pay the average if you drive more safely than the average driver you shouldn't pay the average now what do insurance companies do now they say you know look at your police report you know did you get caught speeding now getting caught speeding is not very highly correlated with speeding even much less anything else and so that's crappy data what we want is real-time data that's accurate okay telematics can provide that to us and so insurance companies are asking for this type of data and so I've made this point before but at some point we will be able to see our insurance rates vary in real-time so you know as you start driving right in a badly you know you're sleepy you haven't had your coffee and you're driving like an idiot you can see your insurance rates go up in in real time saying hey you know pull over take a nap now of course if you're driving to Vegas something else might happen right other insurance prices might be impacted okay well what about health insurance there's a company that I visited in South Africa many years ago Discovery Health and what they said is we can provide a bunch better assessment of your health risk if we have access to things that are predictive of your health status right so right now what does a health insurance company do or a life insurance company do they ask your age you know your your gender your marital status you know etc and that's that's really crude what would they really like to do what do we like to poke a hole in you and take your blood and all this other kind of stuff right but what else could they do you exercise or if you eat right those things should impact your expected health care liabilities or at least be correlated with your expected health care liabilities okay so how can they track that they can either ask you to self-report or they can ingest data from other sources right so all we're talking about is adding more columns to our model so what they did and they did this 25 years ago was they asked you if you join a gym to give them access to your gym data so you go in you swipe the card right you know and that goes right back to the insurance company you know she went to the gym now of course you know you could just like leave right after that but you know once you're there you might as well do something what about the food that you purchase right you know you go to Whole Foods and you buy it with kale you know it's like right so by the way just pay cash for the bad stuff you know right there's this there's ways that you can earn there's ways that you can you know game this right but what they found was that people who participated in these programs but had dramatically better health care outcomes and so they're able to offer them insurance premium discounts and so all the healthy people in all of South Africa flocked to Discovery Health and this helped them to become very profitable now furthermore they realized that there's high degrees of correlation between your health and your driving behavior people who buy kale tend to also drive really well okay you know people who buy I don't know you know potato chips and beer you know maybe they don't drive so well okay so we actually they were actually able to adjust driving prices and life insurance prices based on the information they have so look when you have lots of information this also allows you to go into multiple product verticals as a financial institution okay and so there was an insurance company in England that wanted to look at your Facebook profile to adjust your insurance rates and this created some fluffin and some you know some some controversy and so forth as you might imagine but people seem to be willing to share this data including their exercise date how many of you have fitbit's now if you're an insurance company wouldn't it be great to know what's going on with that Fitbit so I have a friend who signed up for one of these policies you know there's there's these companies out there that say you know give us access to your Fitbit data so I know this friend of mine he actually signed up for this thing and then he took his Fitbit and he put it on his dog he went to the dog park had the dog run around there smoking a cigarette you know so look big data doesn't always mean accurate data you got it you know you got it you gotta worry about this this sort of thing right but look you know health care institutions they want to know if you comply look if you're on an $80,000 a year hepatitis treatment plan you know we want to know if you take the pill okay and if you don't take the pill then we have to run you through another one of these things next year so how do we know if you take the pill they have connected pill boxes right you open up the pill box and then doctor gets a message a lot of times if you have a senior citizen parent who's little you know whatever you need to find out are they taking their medicine this thing will tell you but you still don't know if they swallowed it I mean they could be like where'd that pill go right and and that's me actually and so what we really want is a pill that tells you if it's been swallowed and we have that now right it does ingestion detection okay and remember everything is gonna be connected sooner or later okay what about investment man how many of you use Robo advisors okay look I mean I've been teaching in financial engineering for a long time algorithmic trading program trading right at the institutional level through high frequency and through like you know long/short equity stuff and you know global macro a lot of that stuff has been automated but at the retail level we're now starting to attract individuals to these these Robo advisors like personal capital like wealthfront and betterment now I've always been puzzled cuz I learned back at my own you know business classes in nineteen you know ninety or eighty eight or whatever it was you know how to do mean variance portfolio optimization I know some of you just like at shivers like oh ho I thought I put that behind me right but it's like Harry Markovitz invented this stuff back in the day it's like a little standard deviation a little mean he's our he's looking he's like don't call on me like covariances you know correlations this is easy math okay this is not that hard why do we pay people you know 100 or 200 basis points to do the kind of math that a smart eighth grader can do I don't know because I could do that with a couple lines of code now look you know what you really want is you have some sales person come in nice tie oh I'm taking care of you I'm making sure you're safe for your Dharma and blah blah blah right you want to give us 300 basis points that clown no maybe you are that clown Wow like I say when people come to me and they want to learn finance I say look don't even take any fun hands class just take a marketing class because that's what you really need if you want to be a financial advisor okay because now right if you're a financial adviser you can forget about the kids taking over the business with you because they don't want this nonsense they want to let the machine do the trading for them okay but these models betterment wealthfront etcetera they're extraordinarily they're extraordinarily simple they're not very complicated you know you talk about Facebook and you talk about Google and cetera those are complicated data science initiatives what is wealthfront do what is betterment do it's basically like 60/40 you boom you know a couple ETS it is not that like I said a few lines of code now what you really want to do is start incorporating you know information about the individual customer okay so that every customer has their own unique portfolio that matches up not only with their individual preferences attitudes towards risk which you can't get from a simple survey best way to find out someone's attitude towards risk banks now ask you how much risk do you like that's like the dumbest question ever right look at their Facebook profile if they say I love bungee jumping okay boom but not the answers the question much better than asking someone face to face okay so we want to know about your risk but we also want to know the rest of your portfolio because your taxable account is such a small chunk of your portfolio so I have a cousin he works at Apple he's got most of his money in Apple he's got a house in Los Gatos right and he works at Apple so human capital real estate financial portfolio highly correlated even in the S&P 500 which he would like to have right pure S&P 500 he's still not diversified because his house and his human capital are tied to the Silicon Valley ecosystem okay do any of these incorporate that no of course not because we don't know what the mean variance characteristics are of human capital is an asset drill down to individual si si codes or company levels okay but now we're starting to collect that kind of data and that's what these companies will be able to do segment all the way down to one right like Facebook like Amazon okay now what about capital-raising okay we have an alum couple alums who created this wonderful site called IndieGoGo Kickstarter these were originally designed to help finance films and so forth now they're being used to finance companies okay as you may know now originally this was the idea was you can get some rewards or you can prepay for your product but now you can also access equity okay you can now access equity through these platforms sometimes called peer-to-peer platforms enabled by the JOBS Act and so there are some famous examples of companies that have been started with crowdfunding for products and rewards like the pebble smartwatch but this pales in comparison with what we're seeing in icos which are effectively a way for companies to jump start by selling either tokens which are like kind of products or equity which is being shut down by the SEC the crazy thing is that last year there was more money raised in certain quarters to ico issuance than all the venture capital funding across the board this is probably going to go away but amazing what happened here and it was enabled because of this crowdfunding we'll talk a little more about that finally payments payments is usually the first thing you think about when you think about innovation in fin tech because it's the one that you encounter pretty much every day so right what does payment payment is transferring value okay transferring value from one person to the other and usually what you do is you do it through some kind of payment system right and there are lots of different payment systems out there credit cards checking you know cash is a payment system wire transfer and so forth now let me give you a financial innovation in a microcosm again I study financial history I watch it happen in real time do you remember the days when you used to write a lot of checks okay I remember I love talking to alumni because they actually remember some of the things I remember and I'll you know like cable TV and stuff like that right or like you know desktops and things so it's great I remember I used to write a whole bunch of checks now the way this would work is you'd go and you'd give somebody a check in a store and the check that that person received they would give it to their bank and then their bank would have to somehow get it to my bank so that my bank could get it back to me and at the end of the month I'd get an envelope with a stack of checks does anybody remember actually getting a packet of just like big fat envelope full of checks at the end of the month let's tell my students about this there I would and they're like professor why do checks bounced and I said well back when I was young they were made out of rubber you can tell him anything right no no right one of my students asked me I was talking about when I was assigned the Rubik's Cube in college as a weekend assignment and one Mike one of my students was like why don't you just google it right like they don't remember a time before Google it's kind of amazing right okay so this physical check had to make this journey to a clearinghouse right and the Clearing House might be for instance somewhere where that you know at the Fed and then that check would you know they were all come together and there'd be someone who would be like with it little you know visor entering all the data and like clearing everything back and forth okay and so a lot of data entry had to take place now because of this huge volume of checks when I was in school Citigroup used to actually have put the checks in these big sacks and fly them to Ireland where you know people in Ireland overnight would enter all the data and then the checks would fly back in the morning back to New York okay now if you were trying to figure out how to make this process kind of more efficient what would you do why do you have to actually move the checks around why can't we simply move the data around well the data has to be entered how do we enter it well why don't we send instead of write checks to Ireland sent because I only used to be low wage back in the day why don't we send them pictures of checks and they can enter and then we can send it to India or it's even cheaper and then they can enter the data from the images but then we can apply OCR and we can get rid of the Irish and the Indians right okay but again the first step is you have to make it possible for the images to circulate and not just to check so there's a technological piece there's also a legal piece because there's what's called payment systems law where a negotiable instrument right has to actually be present and so they need to change the law so that an image could be a negotiable instrument okay now of course you can essentially pay with a check and the clearing it right in the registry all that stuff happens almost immediately because they can scan the cheques can the data at the point-of-sale and now when you get a check do you take it down on the bank and deposit it no you take a picture of it and you deposit the picture of the cheque and not the check itself now as a consumer this is what you see you're like oh wow FinTech revolution like I get to put my check in with a camera on my phone but this came after long after on the back end at the institutional level right you have the institution of the legal and the technological stuff happening okay that's where all the big money is not at the level of the consumer so how do banks make money one of the ways that make money is by moving money okay by moving money and where do they make the most money you know we think about things like checks and cash but in reality the vast volume of this is in wire transfers this isn't incredibly sophisticated and yet archaic system that banks use to move their money around these are all the different ways right in terms of payments non-cash payments that you see at the consumer level right and have been a lot of disruptions in payments and the question is is this sector about to get majorly disrupted by this thing we call the blockchain now before we get to blockchain you know people think oh I know payments have been disrupted because we now have you know Google pay and Apple pay which is amazing right you know you just go ahead and gene and pay and so forth okay and even Starbucks has gotten into the act and I can pay with my Starbucks app now how many of you think this is disruptive innovation this is not disruptive innovation okay this is not disruptive innovation because all of these payment systems are still working through this open loop these rails which include the transaction acquire the issuer processor and then this intermediary right it has to go through this system the same system that it did before so when we think about like you know Apple pay this is like you know saying wow this is so transformative I have a motorized mounting block that enables me to get on my horse easier okay not exactly transformative because you're still using the same pipes you're still using the same intermediaries you're still using the visa and the MasterCard system right to make those payments now of course there's gonna be negotiation over you know who gets the bigger piece of the pie and so forth and so on and there's also quite a bit of data capture that can happen at the point-of-sale so companies like square for instance they're actually collecting data right every time somebody makes a transaction square makes no money on those transactions why do they do this to get the data that can then use right to provide other services to the merchants that they work with so now of course squares getting into lending and that's because when these small if I go if I have like a you know Vietnamese band me truck out on Bancroft and they go into a bank they ask for a loan bank will be like yeah right they go to square and square C's oh you're selling like 3,000 bond means a day you know sure like we'll give you some money okay because again they have now access to a whole column of data that the bank's can't see okay okay so when do things get really disruptive they get really disruptive when you disintermediate that payment system that sits in the middle okay where you have write a closed system now PayPal is kind of like that if I venmo you some money as long as you keep it in your PayPal account your venmo account and you don't try to turn it back into your bank account then that's just sloshing around in the venmo system it never has to go through right visa or mastercard now that's still pretty small percentage of transactions now but where in the world is this taken off China China Alibaba in $0.10 are more disruptive than pretty much any company in the United States and the financial services sector and the consumer financial services sector okay and if we look at the number of mobile payments happening in China you almost can't pay with cash they laugh at you right I mean if you go to buy like a stick of gum of some merchants in fact if you go to a beggar on the street he'll be like we checked you know no they'll only take WeChat if you want to pay you want to help out some poor beggar on the street there look at you show up with cash to be like you know who are you tourists you know right so you know take a look this is an amount of mobile payments happening in China compared to the United States and a blue line that includes right Apple pay and Google pay right and all the things that many of you guys use I mean this is just completely night and day okay and if you look at the amount of what the merchant pays for this okay it is nothing compared to what you have to pay with some other products and so this is now what this does is now right if I'm if I'm Alibaba I have all this information right from all the merchants and that means that I now can assess the credit quality of that merchants better than any bank okay and if I can get that merchants to purchase its supplies from a supplier using Ali pay then I have all this information about the merchants and now I can back out right the inventory levels of that company right and then if I have all my consumers buying it I know how much money they spend find out how much money you spend then I have a good assessment of what your credit quality is so they're building out these financial super stores you know the fantasy of the bank in the 90s and they're doing it built on better data than anybody else has you know when Armageddon happens in China and all the banks go belly-up ten cent and Alibaba will be left standing over the dust because they have the best data in all of China okay now we even in Kenya have something like an paisa right and this is more sophisticated than most of the things that you use okay when you make payments okay here in the US and so who needs to watch out right the banks need to watch out for companies here in the states like Amazon which is already in the business although currently in partnership with banks now the most disruptive potential technology for payments is is Bitcoin why potential I said why because it doesn't have to go through anybody this is actually a pure P - peer transfer funds I don't have an Alibaba I don't have a ten-cent I don't have a venmo I don't have a Visa MasterCard Chase or anybody it's just you and me we transfer value okay so if the internet was all about transferring information right the blockchain or Bitcoin is meant to be about the transfer of value online now III could do a whole lecture on of course on on crypto currencies and I won't but just this is why everyone was so excited because you disintermediate everybody so this is the promise the world's biggest bank with no actual bank no cash so Bitcoin has the promise of being the Airbnb the Facebook to uber of payments okay and this is one of the reasons why there's so much investment in the blockchain space because of this potential okay and the idea is that we do not have some central trusts that have thought now in reality of course everyone puts their Bitcoin into a wallet in which case means you're bringing back banks and so forth more or less but we'll just forget about that for now and try to figure out where the applications here are now if you are not a drug dealer or a tax avoiders why would you use this because at the end of the day it still is pretty darn cheap to use the current system here in the US right not that complicated you get a pretty good value anti-fraud all this other stuff which you don't get with Bitcoin okay you can reverse transactions I don't think Bitcoin will ever replace anything here in the US domestically the bigger opportunity might be internationally think about remittances people pay their relatives back in other countries tons of money flowing from the US and other developed countries back to these developing countries okay remittances is at the same level as FDI massive flows of cash and how does it happen it happens through the traditional banking system now the traditional banking system once you go across country boundaries becomes extraordinarily expense and fee driven and in and slow so this shows you like for instance South Africa the average cost of permitting money from South Africa if you're a labourer there is about sixteen and a half percent that's insane alright that's crazy from Japan eleven percent okay so this is an opportunity potentially and the reason why it's so expensive is because in order for me to transfer money I have to go to my bank and somehow get the money into the bank account of someone in another country now if those banks don't have accounts with one another they have to go to another bank which might have accounts with both banks and if that bank doesn't have an account with both banks then they have to go up another level right and so what you have are these interlocking balance sheets or Ledger's and each step of the way you have to have a common banker who can move funds from one bank account to the other just like the Fed whenever a check would roll through the Clearing House would move funds from one bank account to the next and so the more levels and layers you have to go through not only do you have to pay fees right in transaction costs because of the sterile balances that are sitting in these accounts the liquidity that's needed to make it all work out but you also have to wait till end of day for all the accounts to clear so time is long payments are long and so that's why everybody's like look at this Bitcoin it's free wonderful okay sorry to disappoint you but it's not gonna happen okay it's not gonna happen this is the promise of Bitcoin a lot of startups in this area I've met a lot of the founders lots of promise the problem is Bitcoin is an issue Bitcoin will never be a successful currency because it does not have the three fundamental characteristics you need for currency right it has to be a good store value it has to be a good meeting of exchange has to be right a good unit of account this shows you the volatility who wants some their money sitting in Bitcoin even for a couple days unless you're a one of these you know speculators you see it a meet-up okay look at the volatility of Bitcoin this is compared to the volatility of the pound okay and that's brexit I don't want my money in this in this thing what about unit of account this is the price of a gallon of gasoline denominated in Bitcoin okay that's why nobody actually denominates anything in Bitcoin when said we take Bitcoin what's your price well six dollars right and then you convert it back to Bitcoin okay as a medium of exchange the world's largest Bitcoin conference stopped taking Bitcoin for registration fees okay and of course the amount of energy that is needed this is the Bitcoin this amount of energy Bitcoin uses right now it's it's somewhere between Switzerland and Czech Republic that's a lot of energy and this cut means that you have to actually pay right for the energy that's consumed so now instead of being free you actually have to pay transaction costs to get your transaction to the front of the queue otherwise you might have to wait right and then you might even have to wait like an hour two hours three hours before you feel fully confident that that money has gone through so look the number of transactions that Bitcoin can handle very very small and furthermore you can't spend it you send it back to your relative in Mexico it's a Google I got Bitcoin and then you go into the store and you want to buy right some tortillas and they're like we're gonna take Bitcoin so you got to get it back into pesos somehow okay and so it's this movement in and out it's going to cost you and it's going to cost you a lot more than simply running the money through a bank wire right because of the spreads just simple lack of liquidity you know so this is an actual transaction fee if I go to a Bitcoin ATM and try to withdraw it and in fiat money I want to pay anywhere from 6 to 10% and that's not including the spread okay so look Western Union isn't going to go out of business any time soon they're gonna adopt different technologies okay but this is where the big money is it's not in remittances it's in these business-to-business payments across border and that's why we have these kinds of companies like ripple and stellar this is all about right business-to-business back-end big volume transactions which are currently you know very expensive so if we look at the payment system they're not sitting at there like you know you're sending money back to your uncle right it is right at the bank level so the banks are making transactions with other banks now nothing's changed here you're still doing something like correspondent banking except now because the data needs can move around more quickly then the money moves around right we can settle these things more quickly so the difference between a Bitcoin and a ripple in a nutshell again I'm not too much time to elaborate is that what Bitcoin uses is an open permissionless digital mean distributed ledger system right where anybody can jump in and do a transaction right that's kind of the beauty of it I can go buy Bitcoin you can go buy Bitcoin we don't need to ask anybody's permission and there's no middleman that can you know accept or deny approve or reject it's distributed the consensus around validating these transactions what's different about systems like ripple and most of the blockchain systems that will prevail is that they are permissioned meaning that you have to be you know a member and membership has its privileges and so you have these different nodes and the nodes basically know one another okay and then if I want to give money to my uncle back in in you know in France let's say and I have to go to one of these banks that are a member of the ripple Network and then my uncle will go to one of these banks that's a member of the ripple Network and then those banks will do their transacting through this distributed ledger okay okay so we've talked about payments very quickly using blockchain what about other kinds of S clearing and settlement if you understand the logic and I don't expect you would after that very brief description of how blockchain works for for cash once you understand that this will essentially by giving everybody access to a single source of truth by facilitating right the movement of assets or what we call the internet of value why restrict it just a cash cash is actually a very very small part of the economy so this is how security is currently clear and settled extraordinarily complicated system takes a long time you've got a lot of different players okay anybody who's participated in this world you know how all the different players right you have custodial agents you know you have brokers you have the exchanges you know the regulator's all of them have their own Ledger's all of them have their own way of tracking the movement of these assets I do a transaction it might be cleared at the end of day it gets settled you know a couple days later when the actual asset moves around okay we can lay waste to this whole thing potentially right with distributed Ledger's okay what about real estate okay and again the places where I think this will happen the most quickly are the ones that currently have the highest transaction costs right now meaning the system right now has a lot of frictions and the nature of the transaction is very simple so real estate may have a lot of transaction costs even though the assets aren't super simple so in India for instance this is what it takes to transfer a title in India and I pick India because India might actually potentially leapfrog the u.s. in this regard if all the players get together this is what a title office looks like in India good luck trying to figure out who owns what okay and by the way if you want to collect on a debt on a mortgage or some other debt in India good luck you know this is makes jarndyce versus jarndyce look like a walk in the park okay sue somebody in India and we for 20 years just to get your first appearance before a judge okay that's very difficult and it's not just the assets itself it's really much more about the security interest right so in the u.s. we call these article 9 security interests you know where I will give you a loan and then you pledged some asset whether it's real estate or a car or you know business inventory or a factory or something like that meaning that Maya writes in that asset are limited okay how do you keep track of this stuff now right with these liens oh by the way insurance same thing this is actually New York Life this is how they keep track of all their insurance policies right going back a hundred years my title insurance it's a mess I have a friend that's got to start up right now in title insurance I remember when I sold my house in North Carolina the buyer didn't you'd get a mortgage so the buyer gave me a check for the whole property value they didn't even know about my mortgage now this before hospital I guess is why I got the job but I got this check in I'm like huh the mortgage company will chase down the buyer when the payments stop coming in should I take the whole check or should I tell the buyer hey you forgot about the mortgage what do you think what would you do now if you said you'd take the whole check then we can revoke your degree for a character violation right of course this reason why this lawyer didn't know that I had a mortgage on the property is because if you want to know about security interests you've got to actually you know go and and do a lot of research it's extraordinarily complicated okay and it's only partially digitized now take this take go take one more get you add it to a couple thousand mortgages and then create a whole bunch of you know mortgage-backed securities this is what you get okay this is a candidate for blockchain okay so look what our contracts contracts are simply if then probably if you do this I'll do this if you don't do this I'll do that anybody know anything about computer code same thing okay and so what we really are looking at now in the most exciting area I think for blockchain that I'm fascinated by is this idea of self enforcing contracts which I've been working on you know intellectually for over 20 years okay here's an example of a self enforcing contract right where you give two keys to the safe deposit box no individual can access the box right or the pirate treasure map you rip it in half and this forces cooperation so this guy Nick swabbo wrote this article called smart contracts back in the 90s probably one of the most influential articles that you've never heard of and what he said is you know what if we could figure out a way to minimize the use of all these lawyers and get rid of a lot of this litigation in lawsuits now first of all you'll never happen the lawyers well you have jobs but the idea is like how do I create a contract it's like a vending machine a vending machine is you go put the money in and the this stuff comes out okay and if they don't have stuff they don't take your money it's a contract but you don't sign anything okay so think about the repo man if you understand this by the way if you understand this slide and they kick us out in a minute you will learn everything you need to know about smart contracts this is an article 9 security interest if you borrow money to buy a car the lender has an article 9 security interest in your car they can repossess your car if you fail to pay okay you can use what's called self-help okay they don't need to sue you you need to get the sheriff and so forth they just hire the repo man now the problem is the repo man has to find you they have to find your car they have to break in your garage they have to you know mace your dog okay they've do all sorts of things which is why they're kind of unsavory can have you seen this TV shows about the repo men they're not savoury people okay how could we eliminate the need for repo man because what we're trying to do is we're trying to write a contract as long as you make your payment's you get to drive the car if you stop making the payments you get to drive the car what if we just baked into the car an ignition which shuts down when the payments don't get made okay that is a smart contract payment stops car stops okay so you're driving down the road look this is new better get home by midnight you know make that payment it's like not pulling over dude make the payment from your phone because you're not gonna make it home all right okay cars that have these kinds of systems people pay people pay enforcement costs are much lower and therefore more people have access to credit that wouldn't otherwise have it so the closest example I get to give in finance to a vending machine is an SPV a special purpose vehicle essentially it's a vending machine right so what you have is your money flowing in money flowing out right so all the CDOs CDO squared right SP V's right every kind of instrument you can think of is essentially a vending machine it's on autopilot if you want to make money without doing any work become a trustee of an SP V I have some lawyer friends that live in Hawaii they're ester trustees for like 500s Peavey's and they do nothing okay cuz everything is automated now this is what it looks like is what the Master Agreement looks like it's a dumb contract and there are literally millions of master agreements for all these SP V's let's replace the dumb contracts smart contracts okay if we do that the if-then proposition is if the money hits the account the money goes out of the account it's all on autopilot it's all on Robo pilot okay okay and so obviously you'll still need lawyers but it's it's pretty amazing now if you go back to your core finance class and may have learned about these arrow-debreu securities they're just essentially if-then clauses if-then propositions and where do you sim this if you're interested in blockchain and you want to make money forget about sending money back to your uncle in France this is where you should be thinking this is the size of the entire economy of the world right here GDP this is the notional principal of all derivatives okay this these are all forwards futures and options these are all derivatives except interest rate swaps this right here interest rate swaps biggest universe that most of you have never heard of millions of lawyers millions of traders and millions of accounts all that can be put on the blockchain okay and all this clearing system settlement system which we took you know dodd-frank is like this thick this much of it has to do with clearing and settling of swaps okay it's a it's a nightmare it's a mess and we can take this whole system and simplify it with distributed Ledger's okay I'm gonna talk about supply chain but supply chain you kind of get the idea right you've got goods flow in if I have inventory and I want to get financing for my inventory I go to a bank or lender the minute the item comes in the loan amount increases to match that inflow the minute item goes out the loan amount drops so the T financing of the supply chain mirrors exactly the movement of the goods through the supply chain how do you do that you do it with sensors right sensors track the movement through the supply chain so auditing for those of you who work for big auditors Deloitte whatever these guys not that well not to consultants but you know the auditors UI whatever your job will no longer be to audit accounts that whole business will be wiped out your job will be to audit the data inputs right if I've got a sensor it tracks my inventory you're gonna go and take a look at that sensor and make sure there's no tape on it that's your job okay to make sure that the data comes in it comes in authentic because after that the way the data is stored the way it's distributed that will not be an issue okay so we're starting to see these guys so net takeaway started off with this circle of innovation and we walked through kind of what's now traditional FinTech which you know is only 10 years old okay we walked through starting with Capital One all the way forward organization of checks briefly touched on blockchain people think blockchain is about payments okay and they think it's about potentially market provisioning it's actually also about deposits and lending in fact it's it's gonna be affecting every single part of this circle distributed ledger technologies okay which are not the same as blockchain which is not the same as crypto which is not the same as Bitcoin think distributed ledger technologies so thanks for coming in sorry I'll stick around I'll be here for lunch anybody has any questions I'm happy to to discuss privately [Applause] you

Channel: Berkeley Haas Alumni Network

Views: 89,606

Rating: 4.9185753 out of 5

Keywords:

Id: EC1V9MvzX4Q

Length: 69min 5sec (4145 seconds)

Published: Mon May 14 2018