Return to innovative investment and policies that supportive innovative investment in a distinct size in just a second are no promises. Hi. And and I'll kind of come back to this part as well, that somehow a lot of what's happening there is through a very particular route in terms of the economic impact is through the domain of entrepreneurs. Okay. I'm going to try to convince you the reason we believe that is not necessarily very sturdy, but it is real time. And sort of like what we're going to try to do is kind of think about kind of that there's going to be a first order set of issues of how we systematically go about and measure what we're going to call an entrepreneurial ecosystem. And then we're going to start to link those to the impact of policies and institutions and look at the impact of the police and startups or maybe what comes around them. And then next week, we have all that together as part of a little framework that's going to say, Ah, here's what's actually happening in places and how the impact of that okay as to we plan and have no idea if it's gonna work. It's a lot to do, but well, we'll get we'll get to Okay. I'm sorry,
could you upload the slide by any chance?
Um, yes, Luca, can you go into the Dropbox and thank you, so yeah, yeah. Thank you. I'm sorry. Absolutely. So so let's get started. So this picture is from about 100 years ago. And it is literally the area in Santa the Santa Clara Valley in California. And if you look very carefully about a mile this way, is the new Apple headquarters. Like this is the heart of Silicon Valley 100 years ago. And so you What do you see there? Not much. And I don't even need to say like, it's like, you guys are proud of Gidwani like, Oh, my goodness, it was just farms. And now it's this great place, Silicon Valley. We're all going to study innovation. And there's a lot of innovation takes place in Silicon Valley. Right? Right. In fact, I like this least gives you a sense of this. Right? Kind of you look the concentration of patent activity United States, incredible peak in Cisco San Jose areas around Silicon Valley. Now imagine we actually went to Silicon Valley and we wanted to see the entrepreneurial ecosystem. So this isn't literally the geographic center or whatever you would call Silicon Valley. This is the area called El Camino rial. Okay. Would that indicate to you in any way that you're in one of the most dynamic innovative places in the world right it's Haskell somewhere in there. But like, besides says, impoverished, right, we all kind of agree. It's actually kind of results a little bit of a challenge, which is that very often our view of places or industries or activities that are new and innovative or doing all this great innovation driven up first and second, are actually quite hard to kind of capture what that is. Like, where is it? What is it? How would we actually measure it? And unclear the fact that you can't measure it has lots of consequences? Because if you're, for example, directing policies, or you're trying to understand kind of, you know, what's happening to VCU returners or whatever it might be, that's all you do, right? Well to filters, all that is going to be shaped at some deep level on what you put in your sample. Right? And after you actually see so far, so good. I'm actually gonna do this even more dramatically. So this location is basically where we have never had a chance to disguise it, right. You've never seen this. I mean, you've seen it, but yeah, okay. So this zoom, remember, this is really good. You guys can't recognize where this is. Yeah, that's the Charles River. Okay. This is Kendall Square. Oh, only 40 years ago. This is like from 1970 up. Okay. And what you can see is that basically right this just to be clear, that's the Longfellow bridge. Right and what you can see is the decision was nothing there. Right? And that's what that looks like the same picture to that. Right. And when we clear that transcript like this pretty shocking, isn't it? Yeah. Now one last question. Who do you think likes the shift from that picture that picture? What kind of people might be interested in land owners? Absolutely. You're going into a lot of that life. Everyone was led very happy. Okay. Let's start with that. Who else? policymaker, right, you're the governor of Massachusetts. That's the Massachusetts circle right there. Right MIT, which is separately a landowner but also obviously is adjacent to it. Right. Right. It used to basically not really interact, right, even though we all talked about, you know, location and networks and right. Basically, this was a common barren area, but Boston used to pay Cambridge to dump the stove Very good. Okay, so let me clear the way though this can write an interesting the way this transformation occurred, was actually through. Not like it wasn't like some big company moved in or the government built some big research facility. Right. It was actually through the first real company that was big. There was company that most of you wouldn't have never heard of, called Lotus software basically, was the second company will produce the most successful spreadsheet of Lotus 123 precursor to excel sheet, just things like that. And Biogen was early, and then there's kind of a whole Euro countries with x, y. These are some more recent, right. Reason I sort of say that and there's going to be kind of a way of getting there is to sort of highlight the fact that this picture is reflecting a real thing. Like there's literally hundreds of billions of dollars of economic activity piled up you know, both not just the physical infrastructure but the market cap of the companies. that are associated with the probably trillion dollars. Right. This is a huge number. I should probably go calculate. Or just the companies that got headquartered here are probably worth a trillion dollars. It's just a huge amount for basically something right. Did you see what I'm saying? And for reasons large and small, what I mean by that is, the large reason might be that we want to kind of come policy the small reason as we we want to study particular ways in which it works, right academically, small not being prevented, beautiful. Design, right? If you basically need to somehow capture that phenomenon. And we're gonna want to understand its dynamics. We're going to understand the linkage between sort of kind of where does that come from? Strategy Performance, how to institutions matter, someone's ever okay with the agenda. Okay. So now comes the problem. Okay. There's like this dirty secret of the study of innovation entrepreneurship. Okay. Was that till till recently? No one cared. Nobody cared at all. Okay, so I'm gonna be very serious about that. During the 1980s and 90s. There came a very and this work you're going to have seen already in this course. Where are we? 12? Right. They we felt we felt we were we know. But you should imagine that people about, you know, in the 80s and 90s, they started writing about this object. They called the national innovation system and they're kind of a variety of different contributors there. Interestingly, literally three books came out in one year, and they all have the same type, national innovation systems. So with a wonder to highlight that, we'll call them you know, studies thereof or whatever, always by 2001 vote of freedoms. Is this comparative analysis? Yeah. And they were all but actually, it was part of a broad movement. I mean, the you know, kind of rivals as well. They were not traditional economists. They were more illusionary Congress, people, management schools, kind of policy people. But they all became completely fascinated with the idea that there were in it that it seemed right that relative to traditional economic activity, innovation, either measured as new products or some firms that you thought were very innovative Labs is something that we're very constantly particular locations relative to the distribution or economic wasn't just some places were rich and some places were poor. It was in certain places produce a lot of innovations, particularly in certain kinds of customers in areas and other rental. Okay, and just be clear, in the early days of this fraud movement. There were very, you know, kind of thoughtful study some of what you sow in this class, who did you see who sort of contributed you like, like, he didn't think he was doing that at the time. We didn't even know who these people were pretty much. But right. So Adam Jaffe studies, right of the role of location from universities are not studies of entrepreneurship, your studies of innovation, and more maybe human studies of the impact of your patent patent to suddenly break down
the details, but you're just gonna give you I don't think you've seen this exactly. Before Devil is Marian Feldman. And Marianne, for example, had a very nice set of papers. We'll come back to some of those, but that kind of start to characterize the literature that kind of looked at, you know, how unequal in an industry was innovation as a function of where universities, their GDP of production, you know, they kind of tried to understand and characterize these innovation systems. You know, myself by quarter Jeff bourbon, you know, we've spent a lot of time on this, right? Like what what we'll just be clear, one incredible frustration that I had when I was a junior faculty member, you know, I was like going to these conferences I found them incredibly frustrating, because, because basically the use case study after case study of like, how the German system operated, and that person would go and be very fascinating, like it was interesting on its own terms, and then somebody would come up and do the Create system. And then somebody would come up and do the US system, and you sit there with a certain set of facts and you had no idea what to do with any of it. And I understand that I should be a more thoughtful person, but I wasn't because I was very impatient. Sup. So we were like, well, there was the emoji and we weren't the only people who there's been we did to make progress in this area, was we sort of said, what would be a framework that was saying that not every case is unique, but also it's not simply just like GDP per capita in r&d. And so we sort of put together three ideas. In the paper, the Germans national innovative capacity, what we did was we sort of combined rover bass growth, which at that time was a relatively new thing, but we put that at the center, right, because locational specific location specific Romer base growth that was kind of like, I drive what we call this kind of common innovation infrastructure, the kind of innovation systems language thinking about specifically kind of linkages and interdependencies. In the world institutions, and then specifically building out a lot of Mike Porter for competitive advantage nations, right thinking about kind of the role of specific industrial customers. My core hypothesis, that we're not really covered what maybe we've covered this a little bit but that that a lot of this operates with a kind of Indus at the look at the kind of national or regional industry level rather than at the global industry level or across all industries. And that led to a little equation, which if you might notice, was the equation that I put up on day one of this course, right, because this is this idea that you could study innovation systematically within places within the economies whatever, by looking at the things that determined the inputs, right, we spent a lot of time in this course on out to this point, things that determine that right how current knowledge production is dependent on past knowledge production with a huge amount of time, and how those are affected by parameters and policy. Alright, so far so good. Okay. Let's be clear, we sort of went to town on that, right? Like we said, Okay, how do these things matter? These are basically elasticities. And then once you have, of course, a predictor of how many times per capita and you try to predict that out of sample or across time here and we created this kind of innovative capacity index, and that became a way that that's very popular methodology now of using a kind of scoring model in the first stage, to kind of think about and then out of sample, kind of look at how you do it. Okay, this is a completely reasonable set of slides. Yeah, so far. Maybe not. Okay. What is missing? From that whole story? So who's missing? What kind of grid model? So let's imagine this is sort of the government. Over here. This is Brian. If you look at the group crashing, we have something about universities and universities are somehow the linkages. And then maybe this is like industry, government, industry and university. And usually industry we're thinking of as existing players. So Right. So you have IBM was the big player. Then you have you know, MIT is the university and then you have government provided support to the NSF for the computer. Okay, who is missing? startup entrepreneurs, right? Literally, there just was no and I just said Just so we're clear. This and some of you are existing policy contacts, this fall national integration systems like it was invented and then it as I said, you feel the basically were didn't exist before the late 1980s. And then all of a sudden, it became very popular and then basically, you know, now there's like, in the EU, right? They have whole government ministries that are basically organized. This, they have whole teams of people this is what they do with their lives. They measure this assess policy of Aramis and this is the Bible they look. Right. So far, so good. And the question was, where is entrepreneurship? Essentially nowhere to be found. Okay. Now that might have been because if you think about the industrial district through the 1980s, maybe startups are the most important player Yeah. So who knows? Maybe that's wrong. But But that's one theory you could have is that that was the era of big business of, you know, big rd labs and the r&d lab was working, you know, like, it'd be like, you'd be in Norway and there'd be like stop oil, and then they'd have relations with the government. And then the university do some pockets for us.
So on the side, and I can tell you that, you know, I'm sure that both x and Zoltan Rodrick would agree with us is so this became this dominant movement, right? All this all this stuff, right? We did some work around that. But then I can work there were people who studied entrepreneurship and you have to give credit to resulted X and David odd direction. They sort of said, Oh, we're missing this big part of the economy. And I can tell you because I was guilty of it, but I'm just telling you, no one listened to like that. Like if you look at where the papers appeared, it wasn't very good journals. It was like, this was not top tier, you know, you're very smart, sharp guys, and they work very hard and all their perks are for the time and age of what it was. They were just as good as the others. I can guarantee you that time, which is people right, but they realize we're missing is that the role of small firms. Okay, so they went out. They were not the only people knew but they really took the bull by the horns. And they tried to document it and what they said is small firms are more entrepreneurial, and therefore more innovative. And there were like, What you saw a little bit up here. He finished last week but did not get the Sabrina house. Did you hear about the SBIR program, Mr. Speaker, we'll come back to that. But you had a reading for last year. It's also on this weekend in Vienna. So it's a really good read, right? You've heard about the Small Business Innovation Research program. Right? So there were policymakers it wasn't like no one was thinking about this, but like that was just considered some like real sideline relative to me. Okay, right.
And whether or not you like this research or not, I can guarantee you the one thing they sort of put facts on the table they introduced the term like skin cells at some point and everything like this for a while, you know, it motivated to development of systematic evidence about the incidence and impact of entrepreneurship. So we agree right so far, right? We've agreed this is probably right, right. This stuff is probably important somehow for right. The big thing is, of course, innovation and economic performance. But we also agreed that entrepreneurship was now fair. And here's our best candidate to go and measure entrepreneurship, which is the role of small firms. Alright, so far, so good. Okay, so for example, I guess I would say, I think salted X was more young people, emerging people, but he was like that, the leader of the movement, you know, for example, founded something called the Global Entrepreneurship Monitor that became an annual report where they went around to different countries around the world and they would go and survey about oh, come up here. Yeah, absolutely. Right. They go and survey say 2000 people got in the United States. They say, are you an entrepreneur? Okay. And then they will try to say, Okay, I want you know, you know, what's the business and whatever. And so, you know, and so what, and then they would measure something and then they would try to, they've done various adjustments over time. And once again, I'm not going to beat up on the Global Entrepreneurship Monitor. What it is honestly, like, they they do adjustments that are more true to you know, sort of almost like trying to create an index that isn't really from their data, but their data is based on a survey of individuals trying to get Are you an entrepreneur, and your attitudes towards entrepreneurship. And let me be clear, if ever you want good cross sectional data across countries about attitudes towards entrepreneurship, the gentleman it uses. Yeah, right. But it's very good. It's very, very good on intention. You know, I'm sort of attitudes like people in Brazil say being in it, you know, if you fail, you know, if I had questions like, if your brother fails in entrepreneurship, you consider a failure in society. And like in Silicon Valley, people crave failure. And then you know, like Korea, people are like, ah, your parents will never talk to you that kind of thing. Okay. But this is from their own data, guidance, but then the main thing they do is they take these 2000 people side, how many people are entrepreneurs who produces people who operate a small business, they look at how many people maybe they didn't multiply by two started the business more recently. And then they, for example, measure, here's what they call the total entrepreneurship, right. They have other measures that I'll come back to in just a second. And here is the United States, right? There's no real scale here. This is basically on some that really scale too much. And so if you look at the top is Chile. Robot right behind that is Guatemala. Okay. Here's my question. And I know you care. I'm pro Guatemala. I wish well for Guatemala, I want Guatemala see cash. Do you think that Guatemala is the second most entrepreneurial country in the world? Like what must that mean? There were there are a lot of entrepreneurs in the world well why there are a lot of people
okay, what? Okay, so that's it. That was very true. What do they get? So what's the job
okay, do you think that that okay, now over here though, right, the United right. Okay. So over here is like Israel and the United States? Is that the only quarter? So? So you might say, Oh, well, United States at least is on the top half year. We certainly have a lot of impactful companies. Do you think they're capturing here in the United States metric all the startup, the Silicon Valley Fever methodology, for example, 2000 people who've read the undecideds gradually, I have 2000 people, how many are likely to be entrepreneurs at any moment in their lives
you think most people own a business at some point in their lives? Like 90s random people are just gonna say no, I'm not on firm run it should be on my VA for entrepreneur like they might be like, Go Elon Musk, but like I got my job. Right. Like, you know, I'm saying I love my trip. And here I am at the Sloan School as a professor of management with my very secure job right right. You see what I'm saying? Right? Most people work for larger organizations, larger the better with stable jobs like that's supposed to be common. Even among our 10% of entrepreneurs, how many do you think in a sample of 2000 would have recently started? We're just telling you, there's only 10% of population is ever entrepreneurs, or is that a moment in time running? A small business? Let's say that's 10% at best How many would be recently started? Let's say character set, how many people might have started with 2000 people with 220 people? And now Oh, boy. 20 people what do you think is the average thing even in the United States that they do? I took a sample of 20 new entrepreneurs in the United States. I just looked at the most 20 recently registered businesses like that happened to that. What would be the average thing that people would do? If you're a restaurant owner, they'd be a plumber. They would be opening up a small consulting firm, the Italian market will a single person lawyer, right? Nothing wrong when we've got all of this juice. Okay. So you might now have out of 20 people who are even new entrepreneurs, maybe one or two, right, who are like the beginning slides, I have to save your time. So if you just take a survey even reasonably well, there's nothing wrong with the gem survey as a survey. And it actually does a pretty good job of saying our people run for businesses, because remember, we found out read the law, right? We found out from this research that small businesses there were some small businesses that were likely to have a big impact on some small businesses or produce it and patents that are very important to decide. But most entrepreneurs and most small bit like just because some small businesses ultimately have a lot of impact. It is not the case that most small businesses have impact. That's a real challenge. Okay. So he just dives anyone? This is a good test. Does anyone know who this guy is? Right now well, that Steve Jobs that's actually a crappy old photo of Steve Jobs that's clearly with a poor font or for resolution, but that Steve Jobs so we can agree. If you ask most people who is an entrepreneur, who do they say, Steve Jobs, Elon Musk and maybe Mark tech separately, right. But this guy over here i There's Barack Obama just so we're clear, looking with a lot less gray hair, so that was when he was running for president. And this guy over here, I swear you actually actually like have like, the like Saturday Night Live skits. This guy was Joe the Plumber. Okay, and the great things of Joe the Plumber was an entrepreneur who did not like Barack Obama's. And he came and he said, I can't expand my workforce. I'm Joe the Plumber. And I can't expect you know, if I get hit with these high taxes, I'm going to be worse. off because you're going to go with that. It turns out there were three things that were true about your the plumber. First, his name was not Joe. Second, he wasn't a plumber, and he didn't know his own business. And for even whatever business he didn't know when would have actually had no more additional factors. But does that matter? No, it does not okay. For the purposes of the story, right? We all love Joe the Plumber, he turned out to you know, then run for Congress and whenever he became before about there was during the 2008 presidential election, there was about three weeks that were completely dominated by what every single politician in America was going to do for Joe the Plumber. I swear to you look it up online. Okay. Okay. Is he assuming he really did own his own business? Is he any less an entrepreneur than Steve Jobs? Right, what is an entrepreneur? What does that mean? So and we know one owns language, but probably a pretty good definition of entrepreneurship that we talked about. Right? Right. And I think it was class two, or whatever it was, was an entrepreneur, somebody who starts and runs a business and takes financial and personal risks for their own game. That sounds like right. Sometimes people will sort of put in maybe not for their own game for social enough social entrepreneurship might be a broader definition but own game here find a market right. Okay.
And so the question is, clearly there's something right so so just to kind of make sure we know where we're going. We started with this definition, which sort of reflects a lot of the implicit assumptions of our last 11 weeks of this course, which was right up till now. It's all been about innovation. And sometimes the entrepreneurs have kind of kind of been swept up in the in the in the mad rush of things, but not really. And then everyone sort of agrees that these entrepreneurs somehow matter. But for a long time, and when I say a long time, I mean, in many cases, if you look through the literature to this death, like the gym report, is the most influential International report on entrepreneurship. And what they measure, right, you show the power at best. Like, like, like the governor, they then go through activities that they then throw in, like the r&d budget of the country, right but that's sort of like from other people's data. And maybe that's, you know, if you scale the number by r&d, and then you're measuring r&d, that's not necessarily entrepreneurship.
So what do you have to do? How are people going to write in this this puzzle people when I say people, I mean me, but I think the literature for a long time, right, there were a bunch of people who you heard from Adam, and obviously your people, right? And there were people like Josh Lerner who I'll come back to in just a second who said entrepreneurial finance but they were starting very narrowly entrepreneurial finance is very special group of companies, you know, that was a particular model. Good for that. Really was very little was us. And so then what happened, right and this is, you know, so on the one hand in the background right, we've studied this a little bit was, you know, people think back to Schumpeter, you should pay to be an entrepreneur and then separately, and then in particular, complementary to the rover model that we emphasized in class, the backyard and Peter how its growth model of a Schumpeterian theory of revenue. And their model was very much grounded in the idea that the motive to innovate was to leapfrog the established firm. First, good, right. But that's a theoretical model. And once again, Schumpeter is more aspirate desperation. Words, but you know, kind of atmospheric, right. Then there was this great book by Steve Davis, John Paul Langer, and Scott Davis, a whole bunch of great guests. And building off a variety of studies they did, in which they figured out and this kind of update to this data, right, this really was became John Holt Wagner's defining contribution. Like there are many other applications to our whole time are brought to the table, but this is the like, the conceptual leap that he made that a lot, right. They did a tremendous work in helping to build massive databases that he led. But what they did is they realized that the reason we have focused on small firms was because we were really interested in were young firms that were only small, because they were young. Does that make sense? And that seems like a very elementary point, but it's really important. Right? That in some sense, right. So this is taken from the 2010 paper right, along with co authors, but mostly at census. I think Davidson is also on this paper. What they show is is that this is basically
right, this is so this is looking at right does size matter? after controlling for aid. Okay. And basically what this shows is that once you get the age controls what seems to be this big size effect is actually not so big. And it really is just for the very, very smallest firms. But that if you look at essentially once you look even controlling for size, age is the driver of that employment growth there it's not clear that all you care about is net new job growth. I mean, I guess if you're a politician, maybe that's what you're interested in the year before the election, but like why you care about net new job growth? I mean, you're right now the US economy has 3.5% unemployment, net new jobs, new job, whatever, but that having been said, very, very right, kind of grounded in kind of models of reallocation models are almost like a Ivanovic model of industry evolution. What you end up with, from this is the idea that there's lots of turnover and the kind of you know what you're interested in are these young firms that had a turnover. The OSI got missing on a staff. reallocation, just kind of like trying out new things, who knows who's gonna win, and it's these new firms that are going to somehow that growth, they that add lots of new jobs and new jobs. So everybody okay with that?
Okay
so the first big thing in our measurement of entrepreneurship, that really changed was a shift from small firms to young firms. Except one thing, even if I focus on young firms, what is the average particular who is the average on firm what kind of firm Joe the Plumber right. Now, that doesn't mean that that's right the firm's that grow a property that Joe is Joe the Plumber, massive growth would be called Finding Your Plumber never wanted to be so massively scaled. But nonetheless, right? Most of these people, you're still capturing the quantity of entrepreneurs even with a young an age definition rather than a size definition are going to be smallest. Why do I sort of slightly hit on that? That's going to kind of lead us to a big shift in how we might think about these issues. If so, for example, once you get into this world, you might then define as a metric that you're interested in. Right and many good papers excellent papers basically created this metric because you might then go and say, Oh, can we get a population of every firm to file a tax return to the United States, a cola object, the longitudinal does that and then we're going to look at the number of firms that are introduced database by state one by industry, whatever you want per year, vertical, that firm entry rate. So these are firms that newly file a tax record. And then we're going to find other right and then other firms are going to stop paying taxes. Now, some people don't pay taxes now because their business is bankrupt, because, you know, they're friends with you know, they've titled Trump or something like that, right? But that's besides Right, right. But people who stopped paying taxes are based on closing your business, they don't have revenue, or to call those people and then we could imagine we take the gap, right between the third entry rate and from exit rate as relative to basically over the existing soccer firms. And we're call that a percentage and then we're gonna call that object, that empirical object, this this dynamism and then you could imagine if you're a young, ambitious researcher, what would you do? You could put it on the left hand side of a regression, put it on the right hand side of the regression, we're ready where you want calibrate the macro model. Great, great. So a huge amount of work was done, kind of looking at this firm entry minus firm exit equation for entry minus for exit divided by the stock Okay, and then get once again, most people so once you get back here, right, yeah, once again, right? People aren't caring about entrepreneurship. Then life went on. We had the.com Boom, we'll come back to that in a second. But then this work really started to get attention and in particular, when they did start to get attention is what is the major pattern you see in this data through 2011? Like if you had to say one big fat like like if you were asked to describe the major fact about business dynamism, what would it be? Yeah, that just right, that exit kind of stays the same and then during the financial crisis increased, and that the entry rate had a secular so fair, that if you believe, somewhat inspired by champagne or is read by your vanity genetic eon, that this business dynamism, rain is the be all end all to economic performance, because you have this fact that it's these new verbs are creating all the jobs. This was really bad. They're gonna get that and it's bad both and it's not interesting, theoretically. Good. Often that's implicit in these papers and for sure front and center foreign policy right, okay. So, suppose you have found yourself in exactly this situation. Right. So you observed this inquiry right, you were at census where you an economic policy maker in the United States, and you were you would see many of carefully done. Presentation, read much papers by Walter miner and his many right Ryan Decker, who was a graduate student, excellent, excellent people, but sort of people who are producing papers like this. What imagine you were kind of now appearing before Congress. What would you propose as a solution for America's economic woes during the financial crisis? What do you want?
Well, you're gonna cheat because you already weren't the client. Right? Okay, what is more fair trade, you want more new firms? Right shots on? Right. The idea is, we can argue with the dope details, we'll make progress, but we'll begin to spotlighting. He was the head of Economic Studies at Brookings. He was the head of entrepreneurship at Tauchman foundation. For many years. And Bob was one of the great, you know, both economic policy makers and past 40 years one of the greatest advocates and both in my own career but also John Porter and many, many other people was very influential in really creating field with empirical entrepreneurship, you know, through his leadership within that, but here's his thing is we're going to agree what drives economic growth wants me to encourage startups to buy startups he very specifically meant that we want to move this natural. You want more net view? We want more business scientism, as the primary metric might spark up your private stock. Okay, okay. Why might that be that? What might be challenging, or what might be good or bad about such? So think great. I mean, most of you were probably in elementary school or some such thing when that becomes a maybe early high school from you, but sort of right when the financial crisis was going on, but it was a big thing. Like most of you have some memories that occurred and, you know, right, we certainly didn't like the big banks and thing right Occupy Wall Street, right. Why would this, what would be both? Why would that be a good idea or bad idea?
We don't want any sirens right? You want Okay, so one would be that this is how you could move this number quite a bit. Right? By going squat a voc if you will. So the water model and what's gonna be loved by model but the people my mom who are serving you know, basically works, you know, if you I can guarantee you one way to to have a lot of entrepreneurship, collateral big firms and pet lot of unemployed people who take care of themselves, right. That's right. As well, you might want it to be, but just to be clear, you know, like all else equal, if you could avoid that problem. Certainly, this doesn't look good. It certainly seems worrying. And there's nothing in this that says it's not leading to a loss and what you care about in fact, it's consistent. Okay. So
in some sense, what you're trying to do, is like remember, we went from we started out with this innovation systems approach. Right? Which is incredibly influential and forgotten about entrepreneurs. And then for about 20 years, there was a way of thinking about entrepreneurs first, its small firms. And that turned out to be not like that was helpful for some purposes, but in some sense was conflict definitely complete this moment. But then II would go into the young firms. Right, there's something about the heterogeneity of firms. And so that is where this little thing comes in. This was Natalia family, and she was student here at she was going to pursue one of two metrics. Allergy and so on the one hand, she was thinking about opening up a gluten coming out of flown with an MBA and opening up a gluten bakery, a gluten free bakery that she's like, I'm going to really impact or she separately and she was like a PhD student and professionals are absent, right? And she feels like, or I could do off the device to sort of solve gluten problems forever. I'm actually gonna play the video we're not complaining. We're complaining people either. Either way, I should have mentioned reminded myself of her name was okay, this is nothing. Okay, either way. And so Bill a lot, trust her, but Fiona sort of wrote a small little article that, you know, not really sure how excited but definitely, me affected other people as well because they just said they kind of made a very simple point. But a very helpful point that was sort of grounded in the literature, but it's useful to sort of like they needed to kind of reduced it to its assets. They said, it is really, really useful. Right? To make a distinction between SMEs small and medium sized enterprises, right, in some sense, which is sort of simultaneously the old x and Audra to small firms thing, but it's also like all the retail entrepreneurship, even new firms that were they called innovation driven entrepreneurs when we were just starting, which I'll talk about next week. The regional entrepreneurship acceleration program, and that became like, right into the fabric or program right from the beginning. Okay, I just right. And one point they need almost immediately is that these are really different objects. And the woman kind of soften them and make it more continuous in just a second. But if you think about it, the cut you know, like, if you decide that you care about small business people, the support that you would give them would be a very different portfolio of policies and programs than if you're trying to encourage innovation. Thank you said, Okay. Moreover, the dynamics of these firms are going to be very different. Right? If you open up a restaurant, and you are unprofitable for your first six months, like in terms of your marginal probability, guess what's happening, right? It's basically you and your money, you run out of money and guess what happens? You close. Right? If you have some innovation underlying your business, there's some I write a kind of right now we go back to class one or see everything from class one onwards, if there's an actual innovation that has potentially many paths, and potentially could scale as opportunity to create this new thing. Maybe we display people more more or less. Right, Database Driven entrepreneur. You could imagine a dynamic where actually you go to these people, let's call them venture capitalists, they give you money, knowing you're gonna lose money for quite a while. As you do whatever technical derisking find the customers whatever it is before you start scaling and once all that investment is made up of this opportunity for expansion. Okay. Just to be clear, like I'm gonna sort of focus on that distinction that brought that brought like, like, like, I should have done this in a slightly before, but turned right, that was kind of represent it sounds like it was maybe maybe it was representative representative of a broader shift in our study of young firms. Okay. So there's first and obviously, this is their key paper from a basically a 2011 or 12 paper, and what they looked at Oh, I think 2012 Right. What they worked out was actually counseling firm survey, which kind of smartly sampled, not on floors, not the full population. And they're like, Okay, out of this whole group, what was the percentage that did activities that would be even vaguely associated with the kinds of things you would expect an innovation driven entrepreneur? Do they use those terms? Okay, but I actually think the concept is actually very helpful. Think about it very clear. And you get things like 3% of the firm's applied for a patent 8% or top, you know, got even a copyright on their like, website. Like and when you get it you have a right it's pretty simple to get a copyright on your you actually today. So you automatically get right if you might remember from Janet. Janet structure kind of number we took out by copyright comes almost for free and everything that you do, I know that you got a copyright
but you can apply for right, you know 12% Apply for a trademark on their name. So even even a medium sized restaurant will apply for a trademark a few $100 as part of right once the business gets going nicely. But the point is that out of right, even reasonably, you know this is kind of a sample of selected these are real businesses. These are independent contractors. These are not just people working for themselves, not just self employment economy. You got to have a company with employees. Even on that, almost all the firms are not so engaged in these innovative it in choices. or activities that reflect some earlier Okay.
Could you come back one slide please. Um, do you think it is a gene of ID intrapreneurship for to have exponential growth.
So it's so if you want to pay back these people, this better a car. I mean, if we net profitability
linearly grew it also can somehow make us net profit. True,
but discounting will kill you. Yeah, I mean, I mean, obviously, it would depend on the slope. I mean, if you made it, right, but and once again, this is what the idea would be that if you think of it as as that a network that most VCs would use, looking at the performance of a venture and say you're three to five would be what was our person like once we slowly get product market fit of some sort? Over? Right, but once we started getting wrapped, right, once we're out of once we're out of red zero, right, they would say, oh what's our year on your growth as a percentage? He's got a tank. And so if you want, you know, this is this is actually this even imagined constant percentage, and obviously, what literally happens, tail off and ultimately, Microsoft, or something like that, you know, has limits. You know, I'm saying but by and large, right, you know, a characteristic. Yeah, tell us an implicit characteristic of what we call a startup of successful of these kinds of small number of ventures that seemed to have a disproportionate impact is that they would have a disproportionately a disproportionate increase in that might be realized by something like an exponential. rate. So far, so good. Now I have a question about IDs. So this guy does, it's not like when you register a firm or like when you started listing. This article is two pages long. Yeah.
I was thinking that they could be confirmed even without losses.
Because it's Yeah, absolutely. And in fact, to Fiona and I, so get into this, but if you've tried to restart actually a lot of work over the past decade, looking at the different tools like whole menagerie of animals now, not just the report itself as a whole that she loves to talk about. Well, I would go button some others, because there's just much more and that's going to be very consistent with what I'm gonna show you in a second. But there's but we basically don't know what but actually to your point and two, both of these points. Like this article is 10 years old, which is like long in some absolute sense, but sort of like short and even short and academic sense. Right. And so it's sort of like an even this paper, you're just establishing the basic facts of how many of these young entrepreneurs even would seek to grow at all. It's like a relatively new fact. And so I would say there's much less known in this area you're playing it's sort of solid. I'll try to cover a few things that we you know, in a little bit, but there's much less known in this area than than for example, the dynamics of you know, if you might remember when Adam and I think Pierre did this a little bit as well, but you probably seen three presentations that I've sort of careful characterizations of the citation profiles over time of papers or patents. You said I'm saying these dynamics are kind of like, there's just big debates about this. And just to go a little bit to your point, there's actually which is actually actually a macro point. There's this kind of completely fascinating paper by a guy at NYU. That basically argues that the big problem that people don't understand is that growth itself is linear, rather than exponential, like at the economy level. I don't want you I'm saying so you know, I'm saying like, that's a whole separate lens. Okay. Why does all this matter? Because this is where, you know, myself and Jorge story here. Right? Is it's hard to describe how much confusion that like we've now taken an hour on this, most people have an intention standard, like a millisecond and basically a huge number of actually consequential policy discussions, papers in the literature, like your from academic literature, to policy discussions to how you teach system whatever is completely computer. So let me just sort of start with for example, so let me go back. Let me just do this this way. We go back to this. In your sense of history, you think there was a boom in entrepreneurship during the.com era of like the popular impression of that thing? Does everyone agree that during the late 1990s, that was a great time to be entrepreneur, you got like Google, you know, anything from you know, Netscape and Google and, you know, like, all these companies kind of, you know, Yahoo and Yelp didn't that hackathon and they everybody loved it and Netflix, right. So far, so good. What happened according to the official government statistics around entrepreneurship during the late 1990s? Okay, just to be clear, that's it. Next tells you write that you're not capturing the thing you want. Or you're catching one that does that, even if you make exactly what you want, but then you didn't do the other people on this other thing. Right. And so there was kind of like, you know, and I think the data but it's, I think useful, right? So for example, even just this contrast of interestingly. Right, so here's this decline in entrepreneurship in the United States. What was happening in venture capital investment in the United States during the same era, right when this huge decline was, what companies were founded right during that great recession or just before after companies you know, today took Facebook was 2004. Uber, wow. Airbnb, right. Yeah. Like all the company many of the big platform social companies or you know, mobility, food, right, Open Table, all these different companies were basically getting started. Right, and in fact, what you see is a huge increase, right, right during this period, where this is like taking a nosedive, actually, you're seeing an increase. And so that leads right to Marc Andreessen, who you can think of whatever you want, but he certainly, you know, Caitlyn has good quote, right. Marc Andreessen had this great tweet. X wouldn't call that whatever it be. But there's two other ideas, whatever life you could read in the same newspaper, and when I say like in The Wall Street Journal or the New York Times, or at the NPR macro angle conference, you'd have a paper that was about there's too much VC. Everyone entertaining these companies are are not making money. The business models unsustainable, whatever it would be. But it all be that everyone is blessed to be an entrepreneur in our universe at MIT. And then hotwire and his team would come up. Ah, no entrepreneurship, economy stalling out you know, time to just abandoned America the same time and it's not the case that one was right and one was wrong. Like, it's not like always, it's not the case that like anyone was lying, or even making a large you know, there's no calculation. This wasn't like a scientific error. This was a conceptual distinction. I just to be clear, on I guess I got rid of that one. Um, that
reads up to the question in a previous slide. Yeah. So does that indicate or impacts it move ahead, who haven't differentiated given that a quality based measure indicates that for an entry, traveling, the left si graph, this
kind of, right, well, it would sort of say Just so we're clear what people you know, when I say people, I mean, like the academic side, I think more generally as well policy. They just took away from this. Yeah, that there was this complete loss in business dynamism. And they sort of like didn't pay attention to this number. It like you're thinking, Oh, everyone's spent a lot of time thinking how to how to resolve this. The reason why Marc Andreessen is quote was so nice for us, is because people just weren't putting the dots together, that you can't have both things to be true in the same lesson. People the macro Congress for saying, that's it. Entrepreneurship is tough. When it for example, MIT we went from having less than 5% of our undergraduates join startups to having more than 25% Join startups. And just due to a tax incentive, you think MIT students have lots of opportunities at the undergraduate level, and just kind of idea that everyone's not ferocious. That's an interesting point, young growing firms. Kind of interesting, right? So does that help? distinguish it? Like I think you're asking for like, this exit is rising the exit rate. That's the exit you I would say yes, that would predict something about some dynamics, but I think you're taken into more groups. Okay.
So the right graph shows the differentiate and the left graph. And like, the, my understanding, right, right grasses indicate all of the bases.
This is just venture capital investment. This is just how much the VCs expense will hold quality. I know. The thing on the right hand side is not a bug if you get the quantity of venture backed up, this is the total investment in venture. You can also do a quantity graph at this time and it shows the same exponential increase was rise, walk combinator TechStars. All these accelerators, the center time, real focus on real ecosystems. That whole shift was occurring at a time. They're looking at it and I'm looking to this day, I triple A huge I'm gonna bite. But I'm telling people about results. Conceptually, empirically, theoretically. I'd like to say Oh, our evidence was last word. But it's not
it's a total there was no Yes. Like, like, I can't quite describe how you and once again, John, the O'Brien Decker and an apple team on the brand and our accent. They looked at their numbers. This is what the you know, they were there they are. They're flexible, like, you know, how to be very upright. But at least the idea of business dynamism is a very powerful concept in part because of this you know, like, listen, being a policymaker doesn't know which industries you know, who knows, you know, Airbnb is a hospitality industry or their software companies without Lowe's, and be hard for the government to know in advance that it's gonna be. Okay. That was a very reasoned view. So for example, let me give you a fact, at least in the United States, at the census, they're not allowed to call individual companies like they can't do case studies. So like they could they want to talk to you and talk to the economy. They have like a public media on the record, and you really means is they have relatively limited interaction with the real nothing wrong with that. I'm just saying that that's for sure on time. Okay. So then I met Okay, and then like, right there in the midst of this, Josh Werner comes out with this book, the Boulevard of Broken Dreams. And he's like, yeah, there are these policy efforts. To promote entrepreneurship, but they all fail. Okay, and this idea that, you know, somehow there's something wrong, and it's just, you know, people will intention, but their total, just, you know, just a lot of Boulevard in the title. It's the boulevard broken. So I'm gonna talk about this next week. So there was one response to that, which was we founded program here at MIT that I'm gonna talk a little bit, trying to link to policy and initiatives and programs called the REACH program, some of you know, some more than others, that we now work with at more than a regions around the world trying to solve this problem at a core level and right has been very I think, has and what I would say is that riepe is an engine is it provides a framework and a theory of things that you could build whole academic literature. Like I think there's just tremendous opportunities going forward to connect innovation driven entrepreneurial ecosystems to systematic study, in a way and they're kind of deep insights from the reprogram that have been talked about a little bit next week as part of the second thing, right, this is back to this. Right here. You see that? We cleaned it up. That's just for you. I did that right now. Okay, I believe you have that blue line and they say it's all very good. Okay. All right. We've cleaned up the graphics a little bit. Right. But I really but but basically, you know, Jorge took this course. And we sort of like and coming out of reap where people were interested in what were the and people just completely confused about this. And they said what are the metrics that we should pay attention? Right, if we want to encourage, surely encourage, encourage more. What programs should we do? How should we validate how should we know if there's making any progress? Okay, so that was the thing that led to that final cut. Okay, so actually put Jorge and collaborators and so let me just kind of start theoretically, kind of what we're trying to capture in this is going to come closer to I think what you've learned. So let's imagine
Oh, I'm missing a slide. No, this is old slide.
Yeah, but I'm still missing. Yeah, that but I'm missing that slide. This one. Yeah. Oh, well. I'll come back. Oh, you are okay.
So let's imagine you have a population of uncomfortable so this is actually excited to write you and imagine that there is a quality to each. I think of the quality as the potential progress. Okay. Let's imagine that growth itself and this is why I focused on exponential wherever is itself so highly skewed. Okay, but the underlying quality, right is some you know, value, observable time entrepreneur but not to you the adults so the government doesn't really right. But the but the entrepreneur does, and once again, you could fuzz that up quite a bit, but just imagine the entrepreneur knows more than you do. Now. Imagine that when establishing their firms. So this now gets to exactly the question about or that came up just recently and how you measure these. Let's imagine that you faced some choices. And let's imagine that they all had the same one for a second. Imagine you face some choices, or call those choices. Ah, okay. So as you're thinking about how you organize yourself, so you could imagine we saw some of them already. Like, is it worthwhile to get trademark? Well, if you're running your own little business, you don't really need a trademark, you could infringe you. Stop Stern. You know, Scott Stern's entrepreneurial consulting firm doesn't really need a trademark because you know, it's not really happening. Right. But if you call yourself thermonuclear biotech, right, you might want to trade martagon case, you're kind of crazy experiment, actually. Right. Then and there some performance outcome. And you could think about this very carefully. But that's observable several years after filing what's called the number of years s. Okay, and imagine its history. You either end up growing above some threshold amount or not. And let's imagine just kind of in the back of your mind, which is fine. I will say that the distribution of burden boasts itself okay. Okay. Now, let's imagine that you a dot that the cost of each of these governance options involve sunk costs, okay, so that any one of your H's is sunk in time, with some error, you're saying so there's some idiosyncratic cost as well. And let's imagine that the value is either kind of is weakly positive. So the the returns to adopting the better governance the higher levels of governance are, but I'm not going to do this closely. But just imagine for a second, that they are weaker the value of that thing is deeply increasing. The value of trademark is higher, if in fact, I have a higher IQ, because it's more likely I'm going to be some big firm and then I'm needing some intellectual property protection. There has ever been so far, so good with that, okay. So under essentially what I just said enacted that the average cost the costs are uncorrelated things like that, right. The expected you know, kind of dimensionality if you think it was almost like that, this kind of how many H's you choose? Okay. Is weakly increasing. In everyone, so far, so good with that now, let's imagine that I reverse that process. Now let's imagine I observe a dataset that is only composed of the following things. I observed all the firm's I observed their choices of non derivative qualities. And here their choices they make it dynamic, that's age, and I observe J, did they actually get the outcome or? Okay? And let's imagine that I then create an inverse mapping that map's G as a function of h. Okay, and basically recovers the expectation when we call that theta hat, basically, right, which is kind of the underlying probability that you end up with G remember that to 01 variable as a function of age.
Then if this is true, then this is true, that the exploitation of of this object your estimated quality level in bird per page as mapped through the universe mapping from G conditional H to in fact, the path is weekly. Okay, so what that means is that now and once again, you know, there's actually some very nice properties of this, that I'm gonna come back to for that on the site, which is interestingly, what that means is that the distribution of underlying quality is actually only going to be wider the actual true distribution of volume is always going to be wider than what you estimate in the law. And the reason is, because basically the observables H only capture a certain amount of the underlying variation. You see, I'm saying like they're, they're the degree of distribution of variation among quality is itself bounded by what you can sort of like what the ages project into. Right? And that's going to be sort of a strict subset of the true distribution, the full rate department right, so far, so good. So of corporate governance choices, ah, sacrificing crossing property, basically, you only choose it if you're above certain level, you're saying then you deal with certain costs, that early corporate governance signals and the underlying theta and the mapping from governance was to the growth outcomes last recover the probability growth conditional on corporate governance choices. Now. If I then do a regression of growth, like I have the good growth outcome as a function of did I get a trade market family? Should I interpret that as causal? No I should not. Right? It's a truth by this is just the this is basically the broader logic books, right? Is people make choices, right? Those right, better than industry coach choices are super, super informative. Right? Because they reflect all the information that a person has said and it doesn't tell you the impact of choices. It says, give it to people who have a particular view on the world, make some choices, and read those views or correspond to reality then you can kind of trace out the impact. So far, so good. Okay. And so armed with roughly that insight, it's because a little while we got there, Jorge then kind of discovered his dissertation right a way of systematically estimating entrepreneurial quality I skipped right and combined a bunch of this with me, but sort of right here is the engine behind this is kind of like combined, three core ideas. So one, you could go to the census, but what's the problem with tax records? In general? What do you need,
get access to people's tax records.
You need approval or to be in Denmark. Right? Right. But But basically, that right, the great thing is one thing Jorge discovered that I remember the moment he was in my office, and I remember just was like, Ha, you know, somebody who's themselves smoking always has an LLC is the way I'm like, Absolutely, it's true, right? Is it every single company that has ever grown, has to register their business. It especially true the antennae distributor wearable. Sometimes it's directly connected to tax records. Sometimes it's not. But this is the good part of opening a company. This is where you get limited liability. It's like how you have a board of directors how you get different tax status, date and time. Right. No one likes paying taxes. Everyone likes being loaded from the library. So are you it's good to look closer, just registering the business itself, right closer to the time that you say Oh, I'm gonna go into business. Right. Then immediately, what he discovered was that there are differences among these firms. Right. And he knew this from the Hearst and Pugsley data, right, that firms just do very different things at not just over their time, right, which is hurting puzzling kind of whenever you proud whenever you were alive with these very different things. In terms of racism, we'll come back to that in a second. Okay, and moreover, some of those choices like how you govern the firm. Whether or not you apply for intellectual property, are highly observable. They would take work to put together dataset in a systematic way, but there are observable other interests. Okay. Okay. And then finally.
Right. The success stories, the ones that drive the Decker and hotwire results are theory small number of firms basically 5% of the big job brokers are the most extreme examples. They are the ones who make all the difference in terms of net income. And so if you just focus on the extreme outcomes, you see what I'm saying and the great thing about extreme outcomes. Is their suit super observable. Like if you aren't, you knows, I don't care what the metric is. If you want to observe every firm in Massachusetts who's done a public offering, or have big acquisition, or has more than 1000 employees or whose revenue is greater than a billion dollars, I guarantee you all those are super terrible. You don't want because everyone talks about that. You've all heard a majority in a way that you have not ordered pizza. Pizza Yes, okay. So far, so good. Okay. First, this is a bit of an advertisement. I will say and some other people started to this. I'll tell you, for example, business registration records are still remain an absurdly underused resources. Like literally you can tonight yeah, I think you download for $50 every business in ever incorporated Massachusetts since like 1850, like MIT is in the dataset. Okay. sec elecard to good yields and it's comparable. Its comprehensive. And true, basically, across almost every jurisdiction in the world. Like basically, you know, it doesn't capture the self employed self employed people in the United States that are on tax record, that poor person, you know, in sort of like, who's just picking apples or whatever they're doing, maybe they're, you know, they're not you're not you know, they're part of the informal economy. You're not going to capture that. But if you're trying to premise a study, so So there might be other types of entrepreneurship to study that aren't captured by this, but for sure. Everyone who's in the sample of firms that are doing innovation and growth, ultimately have to register your
name, their telephone number like that, and, importantly, their precise street address. In some places you have to put like who the board of directors are, and some places you have to put the industry. Some places you have to put like, prescription business varies a little bit we're going to use and we're going to use street address and most importantly, what kind of company you're going to be. And specifically do you want to be a limited liability company, sole proprietorship or a corporation and incorporation. You want to be what's called a Delaware corporation. It's kind of higher level of corporate governance that's associated with Paceman. Remember when Elon Musk having all the trouble? That's because I was Delaware court very strong corporate governance system, okay. So, second, is just to kind of highlight that there is meaningful variation in choices. And once again, your earlier question you know, when Jeff Bezos got you know, here's the Harvard bookstore that's never gonna change. Actually, you know, what's funny is that they will change eventually, but Right, but like, you know, basically the Harvard bookstore, founded in 1938, hasn't changed much since right. Time. Right, when Jeff Bezos got into business with Amazon, he didn't name it Jeff's bookstore. He didn't register as a local business. He registered the business in Washington State as Amazon Incorporated, right made it a Delaware corporation, and within the first four weeks apply for a trademark, applied for a patent and did a whole bunch of other things that were associated with, you know, Jeff Bezos might have visited a family, but he was at least had the intention and some notion that he might grow. So he had private information and private in his Bayesian approach was he's getting ready so far, right.
So we've then you know, I'll sort of show you various versions. But Naresh imagine, once you have, this is a good idea. You write a few like little test papers, you can then kind of game right so we have a dataset that basically from 1988 to 2014. And then another one, that's basically most of the states were 2014 16 was a little bit less than 40 billion firms, which is a pretty big number, and basically can see you can see the exit review stage. Basically speaking, see the entry of lots of firms. And in all those states essentially, you can see there St. Louis. So in our research, right, so imagine that what we do is we just do a simple logit you can imagine this being much faster, that we've done much fancier things, but interestingly actually think that doesn't really do justice to the whole thing. So imagine what you do is you say, Okay, I'm just gonna put a few variables on the right hand side that are basically the startup playbook. Like they're not fancy, they're not like some clever insight on our part. They're like with every entrepreneur the Trust Center The our dialogue Who do you want your talk to read samples of useless you did X, Y, and Z. Now, of course, if you're, if you open up a pizza shop, and you're Scott's pizza, and your goal is to sell pizza, you don't follow this playbook, because it's costly. It's not super costly, but it's the few $1,000 that you otherwise run out. Okay. So for example, firms that get a trademark in their first year are about 300% more likely to grow. Once that growth yard is defined, I'll show you some robustness as acquisition. You're very lucrative acquisition or IPO for six years. Really smart
homes. Intention to go Oh, no. It's not causal. Right there on the slot. The whole point is selection. Okay, we are predicting we are selecting the needle sideways that on purpose. But then we're gonna do things but then once we have it, we're gonna be able to predict and then we can use them eventually because I'm just trying to make sense of the numbers. So so your you have 40 million firms and then how much how many of them want less than one so it's basically your to what so basically, should be in this file dataset. It's about a 115 100 chance that it's really rare to grow. Super rare. No matter how rare you is. It's rare than you think. Like because it's basically it's that's a precise statement. It's actually like an accelerated power law. So like no matter how skewed, you think the distribution of outcomes is, if even more skewed than that. Okay, there's a new devastating fact for many things. Like you'd say, it means that, like, you might say, Oh, maybe maybe we could get rid of Jeff Bezos and his employment impact and replace it with like lots of, you know, smaller firms. The problem is you need like, several million of them to replace one analyst okay. So what that means is one small differences on the outcome level, have macroeconomic yellow shape. Okay, so just to kind of give you a sense, right, so once again, we're starting with one in 1500. And then on the backside of that, and once once again, we saw from the other numbers, right, things like patents are like, less than 1% or even less of the population. trademark is a little bit more cooperations around 30%. Delaware corporations only about 2% sales tax. So they're all these rare outcomes. And it turns out, they all voted. So for example, if you apply for or acquire patent in your first six months and you wouldn't register your firm in Delaware County, you're at 300% more likely. Right? And so most of you in this class for classes or for early research or a project, you have done empirical research, that most people Yeah. Okay. It is very rare, right? We we publish papers all the time, where we show Oh, this effective things 10% Like that's a big dump. That is a much bigger number than 10%. Right? And it basically allows you to basically just rip out right right so this is sort of right, the out of sample right distribution. So what we do then, is we write we have all of our, we take this model, we do have some sectoral controls, we have your controls and things like that state controls and then what we have is we can then out of sample right, we can predict how our prediction does in predicting the successes. And so what you get is that like the top 1% of our s estimates, predicts 37% of all success outcomes in the United States over 25 years. And the top 10% include about safety with a with a pretty darn stupid not a great mom. Like you can do much better than that. If you like did machine learning and you brought immediate you like, like like you could do better than that. But the point was to show you could go a long way with incredibly simple intuitive question time
