I think it's a very difficult topic to be concrete. It's refreshing, I thought, also to listen to the commissioner, because with all these secular stagnation stories, a sort of pessimism, now we have a more positive message. But still, it's very challenging,

because what is not addressed in that sort of conference is trying to put together all the elements that make that, why is it that Italy, for example, productivity start to stop growing already somewhere in the mid 90s already? And there are very interesting studies trying to explain that, but when you try to explain that, you have

to combine many, many different things that combine together and lead to that situation. So this session is essentially on advances in measurement of innovation and entrepreneurship. We have three panelists, and I will, I mean, they were very great, because I got their CVs

and it was terrible, because it was too much, I wouldn't spend too much time by reading the CVs, so they were very nice to say, do it very simply, let's gain some time. So we will start with Fred, United Nations University, Merit, I knew it, I knew, by the way, Merit. Yes, and then Christian Ketels from Harvard Business School,

you are going to follow. And then finally, Scott Stern from MIT Sloan School of Management, you will finalize the 10 minutes introductory comments, and then we go into first maybe a change reacting to each other, and then we open the floor to the audience. And so let's start when you are ready.

And you are ready? You are ready? OK. So we will start. Most of the words you are about to hear have been spoken by previous presenters from the very beginning until most recently from the commissioner. What you will get in this presentation

is a discussion of why we try to measure these things that people have been talking about. So we will ask the question, why broaden the definition of innovation? Perhaps you didn't already know that it was a limited definition, but don't worry about it.

The outline, if I can press the right button. Ah. Yes, on the right, very good. The outline is very simple. The issue here is if statistical measurement, that's what people do at Eurostat, this being Europe, is to support the development of innovation policy, PAWS,

and the monitoring and evaluation of implemented innovation policy. And notice I've just slipped a word in there, which is implemented. Ministers can talk about policy. They like to do this. But the real question is, what happens

when it is implemented, made to work, or at least made to try to work? Who is going to do the monitoring and evaluation? This is where you get the odd statistician who can be easily dispensed with when it doesn't give the right answer. Then measurement of innovation must,

must be made in all economic sectors. And you may think it is now. Well, it isn't. That's the summary of this talk. So measurement of innovation in an economic sector includes measurement of the linkages between the sectors. And we heard a lot today already

about how sectors ought to be linking together, how they ought to be collaborating on innovation and research and development and other things. So how do you measure that, and what does it mean? And the policy implications then are broadening of the definition is what I will end on. So this is a rapid tutorial, which you don't have to take.

Down the, talking about sectors, down the left-hand column is what you will find in the System of National Accounts manual 2008. Not every one of you may wake up in the morning and read that manual. But it is actually quite interesting in places.

Down the right-hand side, you get the words that I will use in the course of the presentation. Where we're going to spend some time, perhaps even two minutes, is on systems approach. And as we are dealing with innovation,

well, the actors that we're going to talk about are universities, government departments, business enterprise, private nonprofit organizations. And they engage in activities. And those activities might be training the workforce to deal with innovation, might be making capital expenditure

to support innovation. It could even involve research and development about which we keep hearing as an activity of innovation. These people, institutions, are engaged. And notice I said people. I will get to Eric von Hippel at some point,

are engaged in linkages. They connect to one another. And those linkages can go both ways, or just one way. But that's where the material, the energy, the finance, the knowledge, the information, flow back and forth and have influences on what the actors do, outcomes.

Now, we just heard about jobs and growth being a good thing. Every innovation policy I have ever looked at down at the bottom of it has jobs and growth as a priority. So nothing new there, leading to long-term impacts. And that might well be well-being if you're lucky. But innovation doesn't always work the way

you want it to work. I won't talk about innovation in the financial sector in the United States pre-2008. We'll talk about something else. Innovation for measurement purposes. This is where we are now. This is a little frightening, but it is where we all are now.

So humor me for a moment. It is the implementation. And I'm reading to you from the Oslo manual. This is the manual which governs innovation in the business sector. It's a joint Eurostat OECD manual. This governs your life when you worry about measurement.

It's not so big. It's not so big. This is true. It's not so big. It's about to get bigger. We're in the process of revising it. It will become quite heavy and will have greater impact when it is dropped on the foot of the minister. So let me get to what it is, a new or significantly

improved product, or a new organizational method, business practices, and so forth. What the word that is critical here is implementation, that word again. And a new or improved product is implemented when it is introduced on the market.

Business enterprise introduced on the market. And then processes are implemented when they're brought into use in the firm. All of this you can read in the Oslo manual, which you can download and read at your leisure. So that's where we are now. But where are we going, is the question. Now, there's too much on this slide

to absorb in the time I've got. So the only thing I want you to look at is made available to potential users. Previous slide, we put it on the market. This slide, we're changing the definition a little to say we're going to make the product, which

is a good or a service in national accounting language, made available to potential users. That gets rid of the market problem. And we can now be a university, or a hospital, or a government department producing a new good or a service and putting it out there for potential users

to use. This liberates us considerably. So I can now talk about firms, public sector institutions, Eric von Hippel's households that make new or significantly changed products available to potential users. Now, all this is interesting.

And if I were coming out of the business sector, I would say, OK, there's no change. Because the way in which we make product available to potential users is we put it on the market. Not quite. You've all got an I thing or a smart thing in your pocket. And when it's not talking to Langley or GCHQ telling people

about you, you will have apps on that thing for which you have paid nothing. You are using email addresses for which you have paid nothing. You've all got a little corner of the cloud for which you have paid nothing. So there are products out there being put out,

made available to potential users. That is you, which are quite different from market priced products. So we could end up with two kinds of innovation discussion. So that is worth thinking about. However, if we do that, we're going to have to have more surveys. If there's a statistician in the audience, you'll recognize that surveys cost money.

And therefore, if we're going to do any of this, we need a bit more money. Well, that gets us to outcomes. Internationally comparable definitions of innovation for all system of national account sectors, support for policy development in public and the business

sector, and for monitoring and evaluation of the word, again, implemented policies. So then, stimulus or more analysis, keyword, analysis. We don't do enough of it, of innovation in all the sectors and the interactions between them. Now, I won't bore you with non-linearity, not

that kind of meeting. We will move on. One of the little problems that we have, and we've just had it, is that nobody really cares about the propensity to innovate of a government department or a firm in country of your choice being 55%.

That does not interest our political masters. What interests them is jobs and growth. So OK, everybody is innovating. That's what the surveys say. What is it doing for the people, the jobs and growth issue, or sustainable, green, inclusive, pro-poor?

Take all those words and string them together, and you end up with what is called restricted innovation. So it isn't just a case of taking the definition and applying it, getting the answer. You insist that it give rise to the political objective.

And that requires subsequent surveys, some social surveys even. This will frighten some people. We could look at the impact of technologies and practices, and there is artificial intelligence, the cloud, the digital economy. We keep hearing about then. Mutual distributed ledgers. In this room, you will know mutually distributed ledgers

by a different name. And somebody will say it at some point if you're in FinTech. New materials. Yes, blockchain. Yes, and that is all over this community. But I will move on because my time is about out. So to measure restricted innovation

requires additional surveys, including social surveys conducted at different times. We've just expanded the time scale. And that's a whole different analytical issue for survey statisticians. So conclusion. Innovation happens everywhere.

Definitions are not standard outside of the business sector. We're working on that. Work has been done in the public sector and the household sector. You've already heard about Eric. And the definitions need to be standardized. So that is the message I bring to you.

Broadening the definition is a step in this direction. Then we can do policy and impose restrictions and have a better analytical understanding of the system. So that is the issue. Now, if you want more information, not for the partially sighted, I can't see it from here, you can have a look at that paper, which

you get by clicking on the URL. There you get all the references. And that's it. Thank you. Thank you. Thank you. Thank you. I mean, it's a little bit worrisome,

because we are nowhere. I mean, I understand the definition, which is quite broad. But how far are we in doing that? So it depends, perhaps, of areas. Can you give us some example of where we could implement that sort of measurement?

Well, we have some examples in the Nordic countries, which have been studying public sector innovation. Public sector innovation, that's what you mentioned. That's right. And that's the closest to having its very own manual. So that could come back in the discussion. But we're seeing a little bit concrete, because I think the broadening of the definition

is quite interesting. I can follow you. But then I say, yes, but can you give me examples where this definition is really useful today? And I got the impression we are not very advanced yet, because you are still at the level of defining what you want to measure. And in a definition which is not so easy to implement.

And then not opening the debate, but just to get a quick But I won't respond at this point. Well, I can give you a very quick response. All this is based on empirical experience. So our Nordic friends are measuring public sector innovation. The definition doesn't quite align with other definitions. That's how we got into this 25 years ago,

built on 15 years of experimental surveys. We learned how to do it. Then we codified it. We wrote it down. This is the third edition we continue to learn. Painful though it may be, we continue to learn. And we are now producing the fourth edition.

And there will be some reference to this material. So it's worth reading. I mean, I would have a look. Read it before the end of the year, because then there will be a new one. OK, Christian. Well, thank you very much for the invitation. So Fred made the point on your last slide that innovation is happening everywhere.

And that's, of course, an optimistic view. We also heard from the discussions earlier today that innovation is very spiky. And I want to talk about two dimensions, I think, in which we hopefully understand now better how the environment in which innovation and entrepreneurship is occurring is actually quite different

across the overall universe of the economy. One dimension is geography. So I'll talk a little bit about regions and what we learned about different regions. The other one is about sectors, because I think what we've also seen is that innovation and entrepreneurship happens in specific sectors and, in fact, often at the intersection of related industries

and related technology fields. There has been quite an advancement, I would say, in terms of the data that's available. And I'll talk a little bit about that, but then also want to drive it towards what are some of the emerging policy implications out of that data that is emerging. Now, if you talk about geographies and you talk about related industries,

that's really what clusters are all about. And clusters is not a new idea by any stretch of the imagination. It has been in the literature for hundreds of years. Mike Porter's book came out 26 years ago. But there has been a decisive shift in the discussion. And I think we moved, really, from these case studies that looked at Frankfurt Finance or Southern German Automotive

or, of course, Boston Biotech and other examples, to really trying to look at comprehensive data sets that help us to look at the entire economy. And I think the specific, important insight is that we now start not to learn only from the top performers.

How did Silicon Valley get to where it is? But we understand much more about the average performers. What are the challenges, but also the opportunities that they are facing? How do we get about that? And this is actually a lot of work that Scott Stern and his colleague Mercedes Delgado at MIT have been doing. Well, we first looked at the geographic footprint

of industries. This is a little bit the old traded local. But it's actually a little bit more than that. It's about how do you view industries in terms of their geographic distribution? Which are the industries that tend to concentrate versus those that are broadly distributed across space? We then looked among those traded industries.

Which are the ones that tend to be connected? And we looked really at the revealed evidence of co-location of input-output locations and also of skill linkages that exist between those. And the third step was to kind of look at what's the interaction between these cluster groups, these groups of related industries, what's emerging

at the boundaries. The reason that this is important is that we see that the competitive dynamics are really very different between the traded and the local. And I think Scott will probably pick up on that. Entrepreneurs in the traded sector have an opportunity to serve a global market. Entrepreneurs in the local industry can also be great.

But their growth potential is initially limited by the local market. Many others have looked at linkages between firms. I think what we feel is that looking at the revealed evidence of co-location and the other linkages actually gives you a sense of something that's very irrelevant for companies. Because these are the other types of sectors and industries

that they aren't acting with. The downside of our data approach, and that is why these grouping clusters are so important, that it has a tendency to be backward looking. Because the geographic footprint is kind of the cumulative effect of all the things that mattered in the past. Now, what we want to look at in innovation and entrepreneurship is we want to look forward.

And so for that, I think we've seen that really understanding cross cluster linkages is quite important. But we wanted to move beyond just stating that industry A and industry B is related. What we instead look at is where are there already today signs of weak linkages that might be coming something more

in the future. And here's an example of digital industry, some work that we've done for the European Commission. Each of these colors is kind of one traditional cluster. So a set of related industries. But we see that this group here in digital industries shows evidence of being related to each other. So what we're studying is whether or not

there are new combinations of industries emerging that form the nucleus of new clusters, trying to understand how that's going on. So that's kind of giving us, if you will, a language. It gives us a set to look at the data that exists about national and regional economies. What's the type of data that we now have available? Well, for example, we can look at all of these cluster

categories. Here's one example, production technology was a little bit more where Europe is strong and already has an established position. But we can see what's the geographic footprint like. And if you are a region within these darker areas, you can figure out who are my peers, who I'm really competing with. What are the relevant business environment conditions

on which I'm competing with another location in Europe? What's the footprint that we see there? You can also look at individual regions. And I picked here one, mid-Uland in Denmark, which is right north, somewhat north over the northern to the German border, which

is kind of a good example of a region that's doing quite good, but it's not Copenhagen. It's not the leader in the country. And I think the picture that we see in these cluster portfolios is not untypical. They have two areas where they're very strong, production technology and livestock processing. So pig farming in Denmark is still alive and kicking,

is a very successful industry. But then they also have some positions in other areas. And so for them, the challenge is really, what do we do out of this mix of some positions that we could develop, some already strong? How do we move forward from here? There is some more data evidence that they can rely on as they move forward.

We also can see within the data across the entire European economic geography, what are the hot spots of emerging industries? And what we looked here is really, what are the regions that already have strong positions in these type of industries that have shown strong growth and that are kind of overlapping, that really have linkages?

And what I see here is actually a little bit the challenge that we're facing, because you see the usual suspects. You see the strong regions already. And so they are, just by their structure, in the best position also to take advantage of the next opportunities of innovation and digitalization moving forward. If you would look only at traditional clusters,

the map actually looks quite different, because we have some regions that are quite good in a number of more traditional clusters, but do not have that broader portfolio that helps them to move forward in the kind of traditional innovation and entrepreneurship space. So what do we learn from the data? So this was just describing what we see

in terms of economic geography. Well, first of all, we looked at what we call the strong cluster. So those are those regions that are in their cluster category within the top 20% of European regions. And we usually use nuts regions as kind of where the data is being collected. And what you get is 3,000 clusters roughly in Europe

that kind of are in the top 20% on one of those criteria, so in terms of specialization, size, productivity, and or growth. And so we gave some stars for that. There are different heuristics to do that. The interesting observation is that these clusters account for 50% of all payrolls, so all wages

paid within traded industries. So remember, these are only 20% of the locations, 50% of the value creation in traded industries in Europe. So we do see this concentration. Now, what I'm not showing here is that there's also a lot of churns there, about 10% to 15% of these clusters that over a period of five, six years

either move out of the group of leading clusters or move into this group. So there's both legacy effects. If you're in this group, you have a better chance of succeeding, but it doesn't destined you to failure if you're not there. You just have to see both of these effects at play. We can also then relate the presence

of these strong clusters to kind of economic performance indicators in which we're interested. And there's a lot of work that Scott and Mercedes have done on US data. We have started to do a little bit more work on that in Europe, and we do see there is a positive relationship on many of these indicators that matter.

In particular, since we're talking about entrepreneurship and innovation here, what we do see is that there seem to be higher rates of entry and higher rates of firm growth of these new businesses as we move forward. And so how this played out in the European data is that we see that about 40% of all the gazelles, and we use a little bit wider definition than usually,

so it's companies that are five years or less old and that have been growing for at least 10% annually in terms of their employment number. But we see that about 40% of those in Europe are in clusters that are strong. So again, 20% of the locations, 40%

of the gazelles in those locations. And interestingly, the gazelles that are in these strong clusters have, on average, 50% more employees at the end date that we register. So there is at least some indication here that clusters provide an environment, not necessarily actually for the new idea to start, because that could come from anywhere, but to turn that new idea into a business that

can kind of grow and that is sustainable over time. We also note that there are huge differences across regional performance, where policymakers at the end are interested in, and kind of the strength of the cluster portfolio. And here again, we looked at these emerging industry strengths, and it's quite striking, again,

how unbalanced this distribution is, that kind of the top regions in terms of these indicators are doing so much better, particularly on patents, one of the kind of favorite innovation indicators, but also on some of these others. So understanding better the cluster portfolio and how that relates to regional prosperity, I think,

is an area where we get some more evidence, but clearly need to do more work. So towards the end, how does this matter for policy? Well, first of all, we think that clusters in itself and cluster data is not the answer, but it's a part of the intelligence that you need to make the appropriate policy

choices in the right type of locations. And that starts with diagnosing the problem. What are the locations that are really behind, because they have a poor composition of their economy? They're insufficiently specialized. They kind of have something everywhere, versus those that are specialized, but they are weak in those strong sectors,

and there's kind of a structural problem there. So you can get a much better understanding of what the real problem in your economy is. Then as a policymaker, I think you can become more effective in targeting and challenging your policy interventions. So if you do something, let's say on entrepreneurship in your region, what are the companies that you should work with so

that you really get leverage from these type of interventions? Knowing the cluster portfolio in your region is very important. How can you upgrade? I think through this notion of related clusters and development path, you can be more targeted, rather than say, no, we need new sectors if we're a region in Greece. You can say, well, where are the areas where

you might have a little bit of capacity already to build on, so where an extra investment really can make a difference? And I think for Europe, or for national governments for that matter, if you start something in a new area, let's say renewable energy, you can be smarter about what are the locations that actually have some of the assets to turn those type of investments

into something that creates real benefits. Now, I think the discussion has shifted quite a bit in terms to clusters, at least, from a discussion that was a lot about creating clusters. Clusters are good, and that's what the data shows, to how can we use clusters? How do we deal with those? And there is a broad range of efforts that have been used.

And I think what we see is that unfortunately, some that you see here to your left with the red are things that work very much in the short term, paying subsidies, intervening in the market, but have a very poor track record in the long term. While the other ones are kind of works more slowly,

but those are really the things that affect productivity, the big difference is really that effective cluster policies leverages clusters. Clusters emerge naturally through the forces that shape our economies and our economic geography. But how you work with them then makes a difference, and they can impact how you work with policies.

My last slide, what does it mean for entrepreneurship policies? Again, as you run entrepreneurship policies, I think becoming smarter about where you do certain things, entrepreneurship is not sector neutral. There are some things that are effective for all entrepreneurs, but how you then do it,

you always work in a certain industry and with certain partners. Use clusters as an organizing principle. So I'm very saddened to see that so far, I think we still have very different communities. We have the cluster folks on one end, and they do their stuff. And then we have entrepreneurship people, sometimes in the same region, that do their stuff,

rather than really think about how can we integrate those, and how can we make sure that the great new companies find buddies in existing large, mature companies that can provide the bridges to a government. Final point, somebody has to do this. And we talked a little bit about the role of clusters. So there's often organizations that need to organize collective action around which clusters

emerge and become strong. And here, I think, with the cluster organizations that we especially have in Europe, we have around 2,000 or so. Maybe too many, but we have invested in that. This is an asset that we use, and we should use it also for entrepreneurship policy. Thank you. Thank you. Please, I know. Yes, first.

Christian, I was struck by your map about the geography. If you take a country like Italy, we see a country where productivity growth has been very slow, almost stopped. And you have a north and a south, basically. I mean, you see it very clearly.

Have you been able to identify in a comprehensive way, of course, the factors in the internet? Because they have basically the same institutional environment, legal environment, and all this. So what explains this big difference? Because not only history of clusters. Did you go into this?

So I think we're starting to go much more into using this data. What is very important is that, of course, clusters are not the only show in town. There are many other factors that matter. And in some ways, clusters are almost more an indicator that there is a combination of mutually reinforcing things.

And you see in Italy, I think, a lot of divergence across different regions, as you see in Spain, and to a good degree also in Germany and other countries. But I wouldn't claim that we have the answer. But I think we can help with the data, I think, to look in better places. We will discuss that when we finish the presentations.

But that's a key issue. Because how do you interrelate all these factors into policy action, and then draw the policy action? I mean, Italy is a very interesting case, I think. When you look at the diffusion of information technology, and you see what are the factors,

the institutional factors that slow down. I was struck by the World Economic Forum, sort of mapping countries, where you look what are the hindrances to the diffusion of ICT technology and all this, and how you can bring that to your cluster, which is a very micro sort of story that you have.

And still, I don't see the two together yet. But you're not yet there. That's what you just said. You are. That's a net state. So not the entrepreneurship. OK. So first, thank you, and thank you for both all the speakers and for this panel. And I'm going to build very naturally,

both on the discussion that we've already had, as well as very specifically on Christian's presentation. That's why I went after you, Christian, I think. So it's been hinted at already. But let's make sure that we understand. And I think Peter sort of has already, in some sense, hinted at what's the policy action.

Well, I think this is, at the end of the day, a lot of how people start this conversation, I think not in a monetary policy sense, but in a regional policy or even national policy sense, is they would like to be the next Silicon Valley. And this is a picture of Santa Clara Valley in the 1920s.

And they have a notion that somehow, that's a long time ago, but not infinitely long ago. And that is now, right that very location is the location for the new Apple headquarters. So you get a sense of change. And what's the challenge with that? And I think Christian already talked

about some ineffective policy making. And I think Fred also hinted at that. Our colleague at the Harvard Business School, Josh Lerner, wrote a book a number of years ago that was called The Boulevard of Broken Dreams, Why Public Efforts to Boost Entrepreneurship Have Failed. And it's a really good read.

This I really do recommend. It'll keep you up at night if you're in this business. And the reason is that he really documents that much of the effort to jumpstart regional growth and to link to those jobs and growth goals through entrepreneurship initiatives and even to a certain extent through innovation initiatives that are sporadic in nature, have not

delivered the goods in the kind of macroeconomic way that we'd like. And so that really kind of frames the challenge that Fiona and myself and other colleagues around MIT addressed a number of years ago. And Fiona is going to talk a lot more about this tomorrow.

But I'm just going to sort of talk about how we ended up thinking about measurement here is essentially what's the plan? How can we grow through the acceleration of innovation-driven entrepreneurial ecosystems beyond writing a report, beyond a sporadic initiative? How can we learn from that pretty long record of failure and actually try to transform that

into a more systematic program of actually impact that is also potentially measurable? That is at the foundation of something that Fiona is going to spend, I think, a certain amount of time talking about tomorrow, the Regional Entrepreneurship Acceleration Program. I'm not going to talk about that in any long way except to say that one of the things we do in that program

is we try to bring together the notion that we've talked about all throughout today, stakeholders, the stakeholder orientation that Marty Schmidt sort of started out with. But recognize that if you just bring people together and they can't agree on the as-is state, their ability to come up with solutions and even a diagnosis

is going to be limited. And so that leads to Fred or to Christian or to others to say we have to have a systemic analysis. Interestingly, that turns out to be key, the ability to connect real understanding by stakeholders of what the problem is and the challenges and the opportunities of a region

allow you to develop prioritized strategy. So far, so good. We do that in a kind of systematic way with a framework and so on and so forth and develop a strategy, but I want to kind of jump all the way to a very specific challenge, which is that as the minister made clear,

innovation and entrepreneurship take time. And so when we had our first REAP cohort, I was talking to one of the members of one of the stakeholder teams, a senior member of that team, Champion. And she asked me, she said, Scott, suppose we do everything that this approach is about.

Suppose we really implement this in our region. Suppose we really do focus on innovation-driven entrepreneurship, growth entrepreneurship, and really use cluster analysis, the whole thing. How would I know that it was succeeding, even if it was? And I said something about jobs and growth.

And she said, I will be dead by that point. And that really changed my view. It led, building on some of the work that Christian mentioned, a few other things earlier today that were in my mind and our minds, Fiona's mind, the question was, can we develop meaningful and actionable, real-time metrics

for the assessment of these ecosystems for the purposes of policy, action, and assessment? And the gentleman on the right here is Jorge Guzman, who is a MIT doctoral student. Actually, he just got his, no, Dr. Guzman, just a few weeks ago. And he really pioneered in his dissertation

a new way of thinking about this problem. So the first insight was that to grow firms, they have to register. That's probably a pretty obvious point. But interestingly, the business registration records, both in Europe as well as the United States, are interestingly disconnected from national statistical

agencies, by and large. That's not so true in the Nordic countries, but in many. The second problem, as was already hinted at, though, is that if you go look through a business register, almost every single company is Vlad's Pizza. I'm looking at you, Vladimir Bulovic.

Is some guy named Vlad who wants to sell some pizza. And let me be clear, I'm sure Vlad could make a nice pizza. But these are small, local businesses, or even businesses of an intermediate variety. However, those firms that have either the intention or potential to grow do different things

right from their founding. So if you look at Jeff Bezos at Amazon, he called himself the world's biggest bookstore. He registered in Delaware. He had a patent and trademark. He did a whole bunch of things right at founding that said he was trying to grow. Now let me be clear, most firms even that are trying to be those growth firms,

those catalytic firms fail because entrepreneurship is risky. But some do succeed. And then what you can do is therefore create a mapping. Just think of it as a prediction. You can create a mapping between those firms that ultimately grow onto those digital signatures of their entrepreneurial quality at founding

against the full population of startups. And you can do that in a systematic way. You can do that in a systematic way. And so just to give you a sense, this is based on a data set that's essentially 80% of the US economy over 30 years.

So just to give you a sense, what's the chance that you have an IPO or a big acquisition? We've also done this on employment and some other metrics as well, but based on some things that you might do right around the time of your founding. So if you name the firm after your founder, 70% less likely to grow. So sorry, Vlad, too. You didn't name your successful startup, though, after yourself.

So good going. OK? Firms that get trademarks, 500% more likely to grow. Interestingly, many of you know Delaware registration is something that the firm can choose. Any firm can choose to register in Delaware. It's more costly. It gives you access to basically being open to additional financing.

Firms that both apply for a patent in their first year and register in Delaware are effectively 20,000% more likely to grow. There's an incredibly skewed distribution of outcomes that's grounded in the skewed distribution of initial entrepreneurial quality.

Just to give you a sense of that, this estimate on the full population, or 80% of US firms, the top 1% of our distribution accounts for more than 50% of all startup growth outcomes in the United States over the past 30 years, and our top 10%

is about 75% to 80% of those firms. And why would that be? Why would that be? So because they're trying to because those were. I would do that. I would register that because. No, no, it's not causal. So these are predicted. It says it right there. Prediction, not causal. So it is not because you register in Delaware.

It's that it's the firms that are trying to do, that are trying to grow, do different things at founding. So just to be really clear, this is just a predictive analytic. You can be more fancy in that predictive analytic, but it is just a prediction, not a causal estimate. You're going to be able to do causal things with that estimate in just a second.

So once you have this estimate and you predict, you now have the ability to have a prediction at founding of every single firm that's formed, including firms that have been recently formed. So effectively, you can use this for real-time analysis of every firm that's

starting up within a given system. We use that approach to propose three new types of statistics. One is essentially a measure we'll call entrepreneurial quality of basically the average of a group of firms. We'll call that EQI. We could then say, well, we don't want to just have one good firm. We'd like to have a lot of good firms.

And so we have quality adjusted quantity. We're going to call that the regional entrepreneurship cohort potential index, or RECPY. And then going very much to what Fred was saying about we can actually measure, do the firms actually succeed that we're supposed to succeed, we can actually over time mark ourselves to market in terms of seeing,

did that ecosystem give you the performance it wanted over a longer time frame in an ex-post study? So let's be clear what kind of things we can do. So this is incredibly, it turns out to be a fairly fine-grained diagnostic tool. So for example, this is a map.

Many of you have been to the Bay Area. This is the Bay Area. This is roughly Silicon Valley. If you look over so few things, there's no entrepreneurship at the airport. Excellent. Second, if you look at the black blob over here, that's Stanford. But there are students in Stanford. And there are a few faculty houses.

And that's where the individual faculty member might have an LLC for consulting. But right around Stanford, so the nomenclature here is the size of the bubble is the quantity of firms. The color of the bubble as it gets darker, higher entrepreneurial quality, higher potential. You see this wall of entrepreneurial quality around Stanford. Individual street addresses in San Francisco,

which were in 2012, 2013, exactly where we see the arrival of particular accelerators and registering hundreds of firms in individual locations. New insight from this, for those from an inclusive innovation perspective, inclusive entrepreneurship perspective, the East Bay has a lot of handymen and plumbers

and accountants and pizza shops, but not a lot of growth entrepreneurship, even at founding. So this, Marty, I think will like this one, right, is this is the ecosystem just around MIT. And what you can see is that there's on the one hand,

for those of you who are MIT alums, you remember Central Square and the endless amount of eating and drinking that can go on in Central Square. I see a few people remembering it fondly. But then this right here, this is the Cambridge Innovation Center. And it is, in fact, the highest single address of entrepreneurial quality in the United States.

We can then use that to actually give a real-time, roughly real-time metric on a zip code level of the average entrepreneurial quality by location and also quantity. And just so you understand this, so places like Miami, right,

that just have low registration fees, right, those are places that have tremendous quantity of entrepreneurs, but relatively low entrepreneurial quality at founding. On the other hand, what comes up in this graph, as you can see, the total black just means we don't have that state's data, but we have for most of the other states,

we have 34 in here. What you can see is Silicon Valley comes out, the area obviously around Boston, so on and so forth, Austin, Texas, so on and so forth. Interestingly, when you use this, you actually get new answers to questions that I think macroeconomists do pay attention to.

You actually revise your view on the business dynamism debate. Right? President Draghi said one of the three planks for pessimism or concern was the declining rate of business dynamism in the US and Europe. That's a quantity-based measure. It turns out that once you adjust for quality in the United States,

you get a much more cyclical story and one that actually started increasing relative to GDP from the beginning of the crisis. Now, lots more work needs to be done. We would like to do this as well in Europe, but I think that's a sort of signature that this might be a little bit different. I'm going to get it done in just a few minutes.

On the other hand, then you can look at what happens in terms of given their quality, what's their performance? And here what you see was it really was a golden era in the United States in the late 1990s where that innovation-driven entrepreneurial engine kind of worked. Companies were founded of high quality, and then they scaled.

And they have these transformative impacts. And what we see by and large in the data is it's not a problem of great entrepreneurs. And so this, I think, goes, I think, to some of the comments that have been made. On the other hand, there does seem to be, at least in the United States, a failure to scale.

To be clear, there is no regional relationship between the quantity of entrepreneurship and economic growth, but there is an association between not causal, I don't want to claim, but between kind of 10-year economic growth and measures of entrepreneurial quality. You can also use this for program evaluation.

You can go into regions that have gotten these various initiatives and look at how has that changed their business environment? Very often, these most initiatives are going to be relatively micro and are unlikely to shift the quantity number very much because the dry cleaners are going to still be opening and closing. On the other hand, here was an example

where we were able to look at a targeted intervention in Oregon in the stimulus package in the United States. We have done an exploratory analysis, as well, of Spain. This can be done on European data, and you can see kind of dynamics over time. Oh, yeah, look at this. Peter's decided we're going to stay on Spain for a little bit. In principle, you could scale this.

Just to be clear, the key is not just to write academic studies around this. The key is by agreeing on the as-is state, agreeing on the opportunities, can we use this as a driver of policy and acceleration, in part because innovation and entrepreneurship are not

controlled by any one agency or person. They involve the stakeholder approach, and that stakeholder approach within an ecosystem can use that shared understanding as a driver of coordinated activity. Ultimately, can we leverage advances in measurement data, the kind that we've been talking about,

to foster innovation and entrepreneurship in the U.R. area? And on that, I will conclude. Thank you. You want to react to your colleagues here? Is there anything you disagree or you want to emphasize?

I think you're quite on the same line. I don't know you. Yeah. I would certainly support what Scott just said, that everything reduces down to a systems problem. And if you don't understand the linkages in the system, you don't understand the problem and the outcomes.

That may not be quite what he was saying, but that's. But I agree with that. Good. In that case, we have a core, almost consensus. Yeah, maybe just to underline one point that Scott said, and that I think sometimes the discussion gets lost.

I mean, we all want higher quality entrepreneurship. But I think the key is not that we agree on that. The key is that we, I think, for Europe, need to agree that there are many pathways to get there. There is not one industry. There is not one policy approach. But we really need different kind of context-dependent strategies to do so.

That doesn't mean that the European level has no role to play. But it doesn't have the role to play to say, OK, we identify one region that did a good job, and now we roll that out throughout Europe. I don't think that's going to work. Success, and I think that's really a lot of the lesson from the cluster data is through differentiation and heterogeneity.

It is not through kind of benchmarking and once done. Can I just build very directly on that? So for example, people often look at venture capital, and was even mentioned before, the deep well of early-stage financial markets in the United States. And that is associated with 55% of all equity growth outcomes in most of the big firms,

say a bare majority of the big firms that we see emerge in the United States. But you know what? There's another 50%. And interestingly, they tend to be in a much more diverse range of sectors. Their roots to growth turn out to be quite different. Interestingly, and this is just a more recent paper,

the one we've written, it turns out that their growth, the kind of simple choices that we're looking at here, their intention to grow turns out to be similar, but their sectoral composition and their path to growth are quite different.

I have a question regarding the predictive power of those regressions. I mean, I can see how not naming after your name,

the company, and doing the LLC, that those seem like naturally with predictors. But I would think that there's way more than that. And so I wonder, what's the R-square of that regression? So it's, once again, think about it that in an out-of-sample test, our top 1% give you 50% of the total outcomes.

No, no, no, that's not the question. The question is, what's the R-square of the regression? I want to focus on the top 1%. I want to focus on, if I take a company, a random, and I try to forecast the evolution of that company based on where they are incorporated, the LLC name.

Yeah, so I can send you a whole bunch of papers. So it is, so taken collectively, so it is actually a fairly predictive regression. Think of it as R-squared in the range of 30% or so. But that R-squared on individual companies, as you saw in the micro-predictions, we wouldn't be getting that micro-geography

unless we were basically predicting fairly well on average. And I could show you a whole bunch. I feel like maybe for this group, we could talk more. But the micro-geography that we're able to establish is a pretty reasonable way of thinking about the predictive power of the regression. But why is Delaware so coming out?

Because if you register there, it must be some clever entrepreneur. So it's a sort of proxy for many other things. Basically, there's roughly a, so in the United States, but as I said, in the work we've done in Spain, which is a very different system, we apply a very different, we find things that are true to that system

and run a regression, run the prediction model there. In the United States, Delaware registration is an uncontroversial way to organize yourself if you are interested in becoming a growth firm. And the reason is it has a consistent body of corporate governance and law that goes back like 130 years that

allows it to basically be a good place to do growth startups. It would just register, yeah, yeah, yeah, yeah, yeah. Yeah. Yeah. You choose here. Very good, very good. I'd like to come back to this notion of a cluster of portfolio

that you need to have within a region and also tie that to this discussion on creative destruction. So if regions are already strong in one particular area, they will have a cluster in that area. If they want to grow to new areas, that will develop a new cluster. So how do the linkages between the clusters

actually need to be, and what would be a good region that can make these transitions, have that creative destruction? So that's not just at the cluster level, but that's how you aggregate at the regional level this cluster portfolio. You're on the same scenario? Yeah, I'm on the same issue. When you talk about the regions and clusters, one may get the wrong impression

that everything is there, OK, and so they grow, they develop, or they don't. But in fact, when you think about the Silicon Valley on the one hand and the cluster at MIT and Harvard, they pull essentially the talent from the whole planet.

I mean, it's like black holes, you know, vroom, you know, they absorb, OK? And so it's not that only that there is something there, but it's because of an initial success. It creates a dynamic that the whole world is feeding that initial success.

So the comparisons across regions in that sense are not really informative, because this initial difference completely skews the entire dynamic of what's going on. Maybe you answer to the two questions. Yeah, maybe I'll start, and then Scott, you read.

I think you're entirely right that I think focusing exclusively on Boston and on Silicon Valley can be misleading, but I think that's where the new data that's now available really can help us, because we can much more look at the entire universe of regions and try to track how are they developing. And what we do find is that indeed that both all regions

specialize in certain ways, and there is a systematic relationship towards performance. Now to Reinhilde's question, I think that is indeed, I would say, kind of the bleeding edge, because what we know to do quite well is activate existing clusters. And I think that was the kind of the traditional model

of cluster policy in Europe, strengthening strength, and that works relatively well. What we know is needed is structural change. And I think what we've learned so far is two things. One, that this is way more risky. So we can't use the same models that we had for the mature industries, where it's just

kind of incremental investments. But the second one, which seems to be coming out of the data, is that as you move into new sectors, it's important to understand what legacy assets you have. And so your likelihood of success is higher if this is something that's not totally unrelated to where you are.

But if you're automotive and you move into digitalization of automotive industries, or use some of the competence from engineering that you have, that's where something new can happen. Think about wind energy in Denmark and some of the coastal regions. They were actually using some of the manufacturing capacity that was still there from the maritime and wharfed industry.

But then that was applied to a totally different field. Of course, a lot of new knowledge was needed. But they could build on something that they had. And I think that's where we are at the moment. Still very risky, but a little bit more targeted than just saying, let's figure out what globally looks like an interesting industry, and let's go all after that.

Yeah, so first let me completely agree with Christian. I think you're going to be hearing from Donna Shizemann and Lourdes tomorrow, who there's a real agenda around how you do. What does it mean to have new innovation and new entrepreneurship that builds on historical strength?

What does that actually mean? It means that there are many regions that invest in very big, shiny programs that are pretty unrelated to anything that anyone in that region does. And if any of them get it all successful, guess what they do? They move to Silicon Valley or to Boston, Massachusetts,

to Cambridge. And the challenge is that if you start and build on something that builds on your unique things that are physically there in your region, that also are very often in the subtle human capital of the region. Raj Chetty, the economist at Stanford,

did this fascinating work on patenting. And one of his many results is that even no matter where they live, children who grew up in Minnesota, which is the historical home of the medical device cluster, are today leading medical device innovators,

independent of whether or not they're still in Minnesota. And there is this very, right, because that is the area that they learn from. And so part is that building on strength gives people a way of saying what they're about, what is the potential competitive advantage of the companies

that they form, and that will itself reinforce and sort of build that actual revealed comparative advantage of the region. And I think that that piece of the puzzle has, I think, been missing from too many policy discussions at both the regional level and, to be honest, as well, I think, as people try to aggregate this up to the macroeconomic level.

Thank you so much. I wanted to tie this discussion in with the previous discussion on R&D, because it was really very noticeable that it wasn't much mentioned in this discussion.

And I think most policymakers need a very simple sort of formula for innovation. And it continues to be that R&D is the key measurement for most policymakers that they associate with innovation.

And I think that's very important to understand because if that's what we have, that's what they think will get them to innovation. They see it as a very linear path. And I have sat on discussions with Romania or Bulgaria, where huge focus on R&D as if that was the first thing that they necessarily needed.

And my point is that can also divert the attention. I wanna give two examples. I remember when Nokia started getting into trouble around 2007, 2008, and a lot of people in Brussels would say, don't worry, they are one of the top 10 R&D spenders in the world, they'll be okay.

And of course we know that wasn't the case. The other example I wanna give is I used to be an innovation advisor to Flanders. Flanders is a very successful region. And they measured innovation with R&D, where they were actually quite good. And the numbers of PhDs they produced.

So the market was flooded with these PhDs. And then they said, we actually wanna turn these people into entrepreneurs. So they gave them entrepreneurship training. With the result, it was completely useless because once you have a PhD, you don't really wanna become an entrepreneur. But it is an example of policy that has gone terribly wrong. So my point is, what we strive towards

is where we are going. And if that doesn't lead us to innovation, we have a problem. So if you could comment on this a little bit, because your models are great, but they are way too complex for most politicians. So they need very simple formulas. And I wanna suggest a couple. Productivity growth, business churn,

internationalization of firms, and digitalization, where we have very poor metrics on where we are right now on digitalization within firms, or where we are striving, where are we trying to go. So I think that would be maybe also quite useful for the discussion. Thank you. Let's take Fred first, and Scott. Yeah, don't.

By the way, I agree with what you said. I mean, but Fred and Scott. Yes, between the two of us, we will leave you in a different state. The first thing I'm going to do is take a document from the European Parliament. And I lived in this part of the world long enough to know that the Parliament

doesn't talk to the Commission. And we operate in different ways. But the role of innovation is to turn research results into new and better services and products in order to remain competitive in the global marketplace. So innovation is to pick up the R&D

and turn it into something useful. Now, one of the little problems in life, if you measure this stuff, is that more firms innovate than do research and development. And this is particularly true when you get down to SMEs, which are all clustering together and doing interesting things in Christian's world. And we'll hear what they do in Scott's world in a moment.

And the question is, how do we formulate policy which helps these firms which don't do R&D because they don't have the capacity, they don't have the linkages. And within the EU, you will find voucher programs, for example, which allows them to go along to the local technical college and say,

I've got this money from the government, I've got this problem, can we get together and solve it and move on? And if we do this, then we have a possibility of growing those firms so that they can get to the point where they can build the capacity to do research and development and make the cluster scale up to something enormous.

This is not easy. And the difficulty, and this is not a criticism if there are any in the room, that policymakers have is trying to understand the soundbite description of what we are doing. And that you put your finger on that.

And the only thing I can say is admit that innovation isn't just R&D, it requires other things. And we should be looking at policy in those areas. And that will be the challenge I would throw back. We'll now see how Scott solves the problem. And Kristin, Kristin.

So no, thank you for your comment. And so let me make two different points. And of course, Kristin as well should join in. So let me make two points. So the first point is, I actually think that while R&D is an important, just so we're clear, R&D is in fact an important input to innovation

as are the accumulation of scientific and engineering professionals. Maybe not PhDs necessarily, but some measure of the innovation workforce. I think that there are tremendous potential to actually building on Fred's definition,

to actually come up with metrics that really are about what new products there are. Which is, let's be clear, different than measured productivity growth. I think that the surveys in Europe, the Community Innovation Survey, which gives you a basic metric that you can disaggregate to a reasonable level.

What share of revenues come from new products? That's actually, or services. That's a pretty good metric. Now, once again, I'm a big fan that dashboards need to be simple. So on the innovation side, I think there's output-oriented metrics as opposed to input-oriented metrics.

And then we can then figure out the relationship between those outputs and inputs in more systematic ways. And there's a large academic literature about that. But even in a policy making sense, sometimes R&D matters a lot. And in the pharmaceutical industry, R&D matters. In the app economy, not sure we'd call that R&D. I think with the kind of work

that Marty's been doing some leadership on with the MIT engine, what we would traditionally call the R&D is gonna matter a lot because there's big physical investments and big skilled scientific and engineering investment. The other part is just on the key part, just to be clear, once again, I have tremendous both respect for,

agreement with, violent agreement with, the move towards thinking about young firms rather than small firms in terms of business itemism and the role of churn. But just till we're clear, the main secular decline that we see in both the Europe,

certainly in the US, as I can tell you, in that decline in business dynamism is the decline of mom and pop stores. But that probably didn't contribute much to productivity growth anyway. There's no relationship between churn and regional economic growth. There's just a fact.

Quality adjusted turns out to be a much better predictor of actual jobs and growth. So you know what I'm saying? So I don't wanna overstate that, but I do think that it's important that having some measure of different ways that businesses are being organized and what the ambition of those founders are matters. Yeah.

Quick reaction also. We have three more questions here. Yeah, no, very quick. I mean, I have a lot of sympathy for the different types of indicators and maybe we can work more on this together. But what makes me nervous is that the reaction that I've also seen in Southern, Eastern Europe and in many other regions is that as soon as you have one indicator, change becomes one directional.

And everybody says, okay, we're all running in the same direction. We need to find ways to create indicators that focuses on coherence and says, yes, you have the right plan for your region and not you're just on one line. And I must admit that's difficult and I'm not sure we have the solution for that yet, but maybe that's a worthwhile task to take on.

Thank you. I have one question here and then here. No, later, you'll come later. One, two, three, four, five. My name is Jorgen Sarakinos from the European Patent Office.

I have a question for Professor Kettles and Stern. It has to do with this Delaware thing connected to the map you showed, very interesting map of hotspots. If I looked carefully at your map, it looks like Southern Germany and Switzerland appear to be the hottest spots

for innovation in Europe, more or less. Would you say that there is a Delaware relationship there, like if a company starts there and they have big ambitions, that it's a pretty good predictor that in the Munich area and in Switzerland, they will also make it big? Thank you.

Yes, maybe. So look, can I clarify, Leo? Quick response. Delaware registration, those firms are located elsewhere. That's just a place, yeah, just, yeah. Yeah, so the answer is no. I think what we are picking up here is something different. We're not measuring patents. We can relate patents to this and see where are the patents registered.

But what we see here is more that these regions in Southern Germany and also Northern Italy and some other parts are actually diversified regions that have strong positions in a number of these strong clusters, and that is what this indicator is picking up. So they have a lot of opportunities at the intersection of different industries.

While many other regions might be good at this or that, but they have kind of limited opportunities from those islands in which they are strong at. But we can talk a little bit more how that is related to patents offline later on. Yeah, thank you. My name is Oliver Stahl. I'm a serial entrepreneur now in the field of robotics and alumni at MIT Sloan Fellows Program.

I'm from Munich. I think you call this a five-star region. Still, we have challenges to innovate in the area of policies and regulation. It takes months, years, and setting up a company still takes two months sometimes. If you think back when China started its rise to the economic power they are today,

they implemented special economic zones, mostly along the coastline, mostly around ports. Is that something you could consider as a kind of cluster? Think about the special economic zone we established in Europe, maybe around technology clusters, but where setting up a company is just maybe a matter of weeks,

where venture capitalists and business angels like to go because they have special tax policies and regulations. Is that something you have seen and considered this special economic zone, so call it the special tax zone? I think you get the point. Thank you.

I'm very skeptical about that. I think we're at a level of economic development when we need these reforms, we have to reforms for the entire economy. In fact, also that it's easier for the pizza parlor to kind of set up those businesses. For emerging economies, I think it's a whole different discussion, but I don't think this is a solution for us.

Paul Romer, who is both being economist but also the chief economist of the World Bank has spent the better part of a decade trying to implement exactly that solution, I think to not tremendous success. Otherwise, we wish him well at the bank and whatever,

but that particular piece of his agenda, fundamentally, you have to get people, cities are real things. You live in Munich for a reason. Would you move to a special economic zone? Yes, I would. Okay, okay, yeah, okay. No, but there are reasons you don't. Okay, yeah, absolutely, absolutely. Excellent. And I don't wanna, yeah, yeah. In addition, yeah, let's continue.

Let's thank you, by the way. Yeah, this is Enrique, I'm with MIT. The measure of entrepreneurial quality seems to kind of hint to being skewed towards the 1% in certain specific CIPCOs within the United States, for example. What does that say about the talent pool

available in a country? And is that a natural distribution? Could it be like 10%, not just 1%? Is it La Pareto principle applying there? So we have, I can send you a different paper. Our estimates of entrepreneurial quality

are incredibly skewed. They follow an accelerated power law distribution that if you had asked me that before we did this research, I would not have anticipated that. And it really gets to something I think that MIT and the Innovation Initiative and Marty and his team have really dealt with, which is a very small number of new companies

end up being the carriers of a lot of the innovation impact. And seeding that pool is extremely important. Once again, there are many ways to grow. It could be more diversified. We could really think of an inclusive innovation model.

But it is also nonetheless true that when we conflate apples and oranges and we hope to do better by quantity, that those efforts I think have had a less robust success metric. Thank you. Hello, good afternoon. My name is Jose Pozzo. I'm the tech director of EPIC,

the European Photonic Industry Consortium. I was very excited about the model presented by Professor Stern. But my question is to Professor Kettles. So imagine that we have the fantastic model from Professor Stern with the improved version of Van Medler. We identify some regions, like for example the region of Eindhoven

that is gonna create the next generation of devices for the next generation of data centers. Or Dresden, who is focusing now on the rapid manufacturing and laser-based manufacturing of Copenhagen or Grenoble or Switzerland. And we identify those regions. And you had in one of your slides that we had to support cluster initiatives in the long term.

Taking the, using the fact that we have in front of you people from the European Commission, like Van Medler, people, VCs, what could you say to them? What should be the things that we can do to support cluster initiatives and foster entrepreneurship in those clusters?

Well, I mean, there is a big policy agenda that the Commission has been using for quite a number of years now. And I think rightly the Commission has refrained from directly intervening into these cluster organizations itself. So it has focused on data, like the cluster mapping that we talked about. It has focused on the excellence of these organizations.

And I think that's exactly the direction to go. I think it can maybe do more in helping EU member countries and regions to think about how they can deploy these type of initiatives in the right way. I think the critical issue here is that we need to move away from a model

where we try to use analytics from top down to see what are the best places in Europe and we channel money to them for photonics or something else. What we need to realize is really what Scott said, that it comes down to the strategy, the people, and then of course the assets that are underlying these type of regions.

But that you get in a process where you maybe have competitions in which some countries used to really get the combinations of assets and willingness to do something to move together. But to me, that has to happen much more at the national level rather than at the EU level to kind of have the connection to what the real local issues are to move things forward.

Okay, we have one question there. Thank you. My name is Wolfgang Unger, also an alumni from MIT from a long time back it seems. I do have one question. Not being part of the policy world of entrepreneurship and innovation,

but as an outside spectator, reading about computerization and robotics in particular, it seems a lot of the success of new companies that have grown successfully and dramatically has been in the computer world. And artificial intelligence is of course becoming more important. Amazon has launched its Echo,

which sort of brings artificial intelligence to the normal consumer. And all of a sudden, eyes are opening about what this technology can do. And there's a study I think in 2015 of Oxford University about the reverse employment impact of some of the innovation that we are now talking about as being generally positive, which is good.

And there's a book by Martin Ford, Rise of Robots, which I guess is quite worth reading about some of the implications. And there are some worries about how many jobs are created and what does this growth in a particular sector mean for the macroeconomics and regions as a whole?

And how would you actually now bring those two perspectives together, the positives of growth, and of course, the destructiveness of potential impacts of those successful firms? And how is that being quantified or measured at this point?

Who wants to answer? I have an answer. They might want to answer. No, you do it, you do it. Okay, well, you know, so let me say two things. So one, I think I would divide your comments into two parts. One is the software eating the world hypothesis

that just, really, this is all about IT. And I will say that, while I think our IT is an important sector and I think Ryan Hilda's earlier presentation sort of spoke to a bunch of that, interestingly, the gentleman from BMW earlier today, I think, said it extremely well, right?

When you buy a nice car today, when you buy a BMW, you're buying a lot of software. But you're buying a lot of software to make a cargo. Do you see what I'm saying? So software in that sense is, the large scale impact of software eating the world has been to make real things, the physical world, more effective.

Obviously, there's also a digital world and that's, you know, social media and things like that. So there's a mix of that. And some of those, and Christian hinted at that in saying that if you apply IT to the shipping industry and you're really good at shipping, that's a different strategy than either trying just to build more boats

or trying to be an app economy. Do you see what I'm saying? The second on robotics, I mean, clearly, there are many people here from MIT who can speak to that far more than I. I think a very, one of the real advantages I think of MIT, and at least that I've been able to absorb,

is that you really, there are, you know, which has a leading role in this area, has been the ability to really start bridging some of the gaps between the economists and social scientists and the engineers. So having, you know, CSAIL actually talked

to the economics department and the business school and we had a group that tried to do that for a few years and what I took away from that group is that at some point, well after we are all gone, someone will vent a thinking machine that can do cognition. And I don't know when that'll occur. That'll be some other breakthrough. But what we're talking about now,

and then let me clear, that's, once that happens, that would, you know, a fundamental scientific advance like that that's unrelated to today's efforts, that's a different thing. But except for that, basically you're in a situation where prediction models, what we're calling artificial intelligence, are very good at doing prediction.

And what that means is that human skills like judgment have just become a lot more important. And guess what? As best as I can tell, it's not like we're really close to the optimal level of human judgment currently in this world, at either a down home level or at a macroeconomic level.

And so I think we have a pretty long way to go. And we could talk maybe a bit offline, but thinking about how much more human judgment can be applied when prediction is made more clear is I think a very important area for thinking about the employment situation. There's a last question. So I'm Jonas and I'm from the UN.

So I have two questions and one comment. My comment is, is it possible for Harvard to issue a retraction on the cluster thinking as a whole? Every time that we go out and we sit with politicians and city officials, they tell us, and we are there to help them come up with some sustainable innovation,

they want to have a cluster. They want to have not one cluster, they want to have 10 clusters. They want to be the global leaders for ICT, they want to be the global leaders for medicals and everything else. And it all has its fundamental roots, basically in Porter's popularity of the term cluster thinking. And then when you ask these politicians if they can mention one cluster

that they would like to copy, they all fall back to Silicon Valley or Cambridge basically. So that's for you. The other one is basically one issue in Europe and in the developed world as a whole in the US, 2% of the population start companies.

Out of them, only 9% are women. Is that something that can be measured? And if so, how are you considering entering that into the work you're doing right now? And finally, disruptive. On the map that Scott showed us, we see employment, how that intensifies in growth companies.

How can we measure disruption? I mean, if we look at Google, they were not the first search engine. If we look at Facebook, that was not the first social network service. So the disruptors were the first entries and they are normally not successful. So how do you support that and how do you measure it?

Thank you. Last one, Chris. Well, so believe me, I've thought a lot about this particular problem because I feel that we have often the wrong friends as well as the wrong enemies. Indeed, I think the perception sometimes of Porter's work

and he doesn't sit here, but has been too often that, okay, clusters are good, let's create a cluster. That's not what's in his book 25 years ago and that's certainly not what we're advocating or what our work is about with different governments. But I accept that we need to do a better job in communicating on how does this inform

more effective policies. And so watch that space, hopefully we can do something more in that direction. It's easy. Now I'm asking him if we do the, there is an opinion poll. So I ask, is it easy to do because it's in my notes. You wanna answer on the other one? Yeah, go ahead. I think the gender thing,

the gender thing is extremely important in the third world and I also come from a UN organization and spend a lot of time in Africa. And the education of women changes things considerably and technology also has an impact on how they are empowered

and I will offer you telephone banking. Not necessarily M-Pesa which allows you to move money around, but full telephone banking which you find in South Africa. Even the illiterate people are able to memorize patterns on the cell phone keyboard and they can access

their accounts and they can turn their little garden into something that produces value and they can bank the money, avoiding the husband taking the money and going down to the pub and consuming it. Very good example. Yes, well it is technology which changes people's lives. So if we had more examples like that,

maybe we could persuade the countries from asking for a cluster from Harvard or using that literature and broadening their interest in empowering the people. But this is something that could go on for a long time

so I will stop there. Still it's nice to hear. I'll be short because we're over. Your question was roughly when Fiona and I and Bill Allot sat down to design the Regional Entrepreneurship Acceleration Program, what we, and you know, right, that was the question that we had in mind

was how do we go beyond slogan and sort of oh we need a cluster to really undertaking that in a more systematic way. Fiona's gonna talk about the work, you know, our work in that and her work in that and I think that the, that that has been a model

that I think we've both learned from about how to try to work with stakeholder teams to move beyond that kind of overly quick model. So there is here in my notes, I should announce now an audience poll, the Mentimeter. You did it before?

Okay, so you know how it works so I can give you the question. Yeah, you have it there? Well, I didn't write it but it's the, how important is it for Europe to enable creative destruction? Well, depends how, I didn't write it but. And then I will conclude after.

Yes, very important. Who said no? Ah, well it's going up, it's going up. Somewhat important, somewhat important, important, very important, not important. I mean it's not so, yeah I didn't write the question, I said it.

Take the incumbents that are around today and have startups take it over. You know, kind of competition through innovation. No, maybe in line with your 20th, well, not important. Did you push the button now? Two.

No, let me, I mean there's no conclusion. I mean the first point I think it was a great initiative. I learned, I learned that I don't know much about these things. My view was the question of diffusion of technology, basically, diffusion question. So I'm more, what I did in the past,

much more than usually, is look at rankings of countries about, you know, I mentioned the World Economic Forum or the World Bank sort of indicators, try to spot, you know, what are the weaknesses across countries and the reinforcing aspects, you know, of the institutional environment because that's what is interesting in these studies

is a little bit in the cluster spirit is all these things that act together and not piece by piece that you consider them. So I thought, for example, if I was in the past fascinated by some of the studies on Italy, for example, Zingales was one of the persons that I thought was quite interesting to try to explain, you know,

why productivity growth stopped in Italy by the mid 90s already and trying to link, you know, many different things together in the results. For example, there are many tiny firms. I mean, the Pizzar company, which can also innovate a lot, by the way, because the definition that you use, I think, is a broad definition.

I think it should be like this. The marketing, for example, how you pay, you know, for the product, how you deliver the products and many of these things and not so much also the question of robots, you know, and all that other discussion on inventions or close to inventions, I think. So the diffusion is probably the main thing, has to do with,

and it's relatively easy to spot, you know, the major weaknesses. When you go, and I liked it also, this approach when you go into regions, because the regions are quite interesting because they are in the same institutional environment and to some extent, to some extent, the same, more or less, cultural environment. Oh, well, language is not the only thing I know.

But I found it fascinating why in, I mentioned Italy, but there are many other cases, why in the north of Italy, you have a lot of innovation, your love of the diffusion of technology is real, why in other parts of the country with the same sort of laws, the high taxes, and why you have a totally different picture. And I think we need,

I mean, it was a little bit what you said, I mean, that we look at figures like R&D and all that, and I think that misses the point. I mean, it's always important to have money for research and development, but that's not the key issue, as you rightly said. And I thought it fascinating, for example, in a study on Italy, that the fact that you had a number of thresholds in terms of,

I would say, labor protection, but I'm not criticizing these law, per se, but they were thresholds, you know, the bigger you are, the more the pressure you have in terms of having a consent entreprise, bedre of serat of, you know, this.

And that you see in Italy, you have this huge amount, proportionally, of very tiny firms, and there is a threshold, but you don't really grow beyond that. And then you link it to family enterprises, and why are these family enterprise not innovating very much, not introducing information technologies, which would be relatively easy, theoretically.

There is a lot of things due to the fixed costs that you have, to the education, you know, and these things, and why are they there? There's a lot of barriers to entry also, and very difficult to explain, because the competitive environment, a priori, the institutional environment, is more or less the same, the legal environment, I say it's the same.

So I think, if there is a comment on this, you're welcome. But it's very interesting to go, what I was a bit worried about your presentation, is that it signaled, I mean, to me, that we are very far from understanding and having a good system when we can bring solution rapidly, because I don't think the voters

are going to wait so long before we can redress a little bit the productivity challenge that we have. And so I was a bit worried about these presentations, and of course, the other questions about the implications of the diffusion of technology on the labor force, on labor in general, and the distributional aspect, but that's another, that's for another conference.

So in short, I think it was a very good initiative, refreshing, we have been, as you know, as central bankers, quite innovative and very often criticized because we have been innovative, but we have been challenged by the lack of classical instruments where our rates went to zero. So we had to innovate,

and we could do that in that institution. We tend to think very successfully, but I mean, that's not always what people say, but we think we have been, you want to comment on my, no, you're welcome if you want. Yeah, yeah, please, please, yeah, go, yeah.

North or south. No, but I'm just saying that because of the regional dimension of Christian also, I thought it quite interesting when you have, you can control

for the formal institution environment, the formal, of course, not the informal, the formal institution environment, why? Because then you control for one aspect like taxes and things like this, and I thought it interesting to within a sort of formal institution environment, the same one, the differences you have. And so, well, thank you.

It's not, the conference is not finished. It's finished for me, but it's quite good. Let me step in just with a few logistic issues. First, thanking the panel, thanking Peter for sharing his support, all panel members.

Now we move to the dinner. The dinner is now following seamlessly, directly. So just follow, let's see, the people leading you. We'll be gathering from here, taking the elevators, and then going to other part of the building. I'm told to advise or to encourage you to bring your coats, gear, stuff so that you don't have to come back here later

because it might be closed, whatever. Though tomorrow we start at 8.30. Very important, come sharp at 8.30. I'm also told to remind you to bring your valid ID. Some people could not come in or had to come in late because they don't have a valid ID. It's not enough a driver license. You have to bring either your passport or national ID. With all that, thank you very much,

and let's continue dinner.