 We will now start the panel discussion on innovation, investment and productivity chaired by Peter Pratt. And during the two panel discussions today and starting now, you'll have the possibility to actively participate in the debate by answering live some questions posed by the chairs. So just to make sure that you have your iPads in front of you, if you go on the polling menu, you will have one minute to answer the questions after which the results will be shown on the screens to further the debate. And we will be doing that during this and the next panel. And I now would like to give the stage to Peter. So this is the first panel session. I have been asked to introduce the participants here very briefly. So let me do that. This is a very atypical panel for me because I don't know you actually. I never interacted with you in panels, which is usually not the case. And I'm very curious about what you are going to say. I think so. This we innovate with sort of atypical persons, that's the first point. The second point, we had a very good start this morning. I have in my mind many questions as you. So we have to keep time for the audience to ask these questions. Things which didn't come very much in the discussions. For example, in the labor market is skills versus talent. And I would like at some point also to ask questions about this. Now I also was struck by David this morning. When you made this distinction between history versus science fiction. And that gave me history versus science fiction. So I think that gave me a good idea for the sequence of this panel. So we start with economic history, with the economic historian, Joel. You're science fiction. Then Marianna, you will follow with Reinhilde. And then we finish, I don't know if you take it badly, or positively with Hall and more science fiction sort of person. So we get two extremes, and more, I guess, reasonable persons in between. That's the two ladies here. Very welcome and very pleased to see you here. So 10 minutes, and yes, we'll be challenging. And there is a clock there, so you can, and I will discipline you. I'm sorry for that. So Joel, and please. Thank you. So as Peter said, I'm an economic historian. I'm going to take a very long run view, so I'm not going to talk about investment and things like that. In fact, I want to talk about only one topic. And that actually goes to the subject of history versus science fiction. So I want to address a literature which I sort of call technopessimism. And technopessimism comes in sort of two main varieties. One is what I call the technostagnationist, people who believe that all the sort of low-hanging fruits have been picked, that we've invented antibiotics and air conditioning and smartphones and so on, but we can't even get them again. And so, obviously, things are going to slow down now, and eventually we'll sink in some kind of stationary state and life will get very boring. And then there are the technodistopians of people who actually think that technological change will be so fast that artificial intelligence and machines will basically take over the planet. They will all end up in some kind of Kurtzweilian singularity or maybe some of you may remember a Kurt Vonnegut's player piano written in the 1950s in which nobody's working and life is terribly depressing and awful and so on. So there's these two different technopessimists and so the good news is they can't both be right. The better news is they can both be wrong and they are. And so let me sort of give you a bit of an economic history view on that. And what I want you to think about is some kind of an imaginary regression. I'm not going to run this regression because I don't know what data to put in it, but we can still do this mentally. And you think about this kind of regression as you put something on the left-hand side that you may call technological progress, innovation, productivity growth, whatever and then you stick in a bunch of right-hand variables that you think may have mattered to technological change in the past and you go, how about if we took the beta hats and plugged in today's values, what would that predict? Now I want to warn you that that is probably not the best way to do this simply because if you run regression you assume a fixed structure and if we've learned anything it is that the industrial revolution and the subsequent technological changes constitute what physicists would like to do as a phase transition, as to say the whole structure of the model changes. The way technological progress occurs now is very different than it did in the 17th century even during the period of the industrial revolution. That said, I'm going to do it anyway and then if the model is wrong, so be it. But at least we'll give you some kind of indication of the things that have mattered in the past. And the central argument I'm going to make is something like that. So technology is knowledge and it's based on knowledge. But what that really means is that when we invent something, when we build some kind of new gizmo and then patent and so on, it's based on some knowledge base which I've called an epistemic base. And that basically means that do you actually know why the damn thing that you're building is working? And we can think of many inventions that were made over through history in which people got something to work but they had no clue why it worked. And what's happening of course in modern times is that we understand more and more about why things work. We never know absolutely everything but we know a hell of a lot more about mechanics or biology, molecular biology. You name it, that explains why the world is and we can use technology that's based on that. Of course the more you know about something, the more likely it is that the technology will work properly. So the question is what kind of knowledge exists between the science that underlies that technology, what science is a bit ambiguous here but leave that alone. And here's what I'd like to propose. It's not only true that you need science to get technology to work. Technology feeds back into science and that's something that's often neglected also the late Nathan Rosenberg, my dear friend, wrote a very famous paper about this. And basically the point is that if scientists have better tools to work with, science will move faster. In fact we can explain most scientific revolutions by the tools that scientists work with. The classic case of course is the quote scientific revolution of the 17th century which was to a large extent driven by the invention of lens grinders who came up with microscope, came up with telescope, barometers, vacuum pumps. I mean I can give you a longer list but you get the idea. And that is what produced people like Galileo and Hooke and so on and so forth. And you know making science work better and then a century later you get the industrial revolution. In our 20th century maybe the best example that I can give out of hundreds is X-ray crystallography which is a tool that scientists use. It's helped gain 28 Nobel prizes in the science including the development of the DNA of the person who used it. Rosalind Franklin didn't get the prize. That's a different story. But these are tools that scientists have. The technology makes available to science. Now you ask yourself what kind of tools have modern technology made available to science and it boggles the mind ladies and gentlemen. These are things that, some of these things we had before. So we had telescopes before but surely Galileo would never have dreamed about something like a new space telescope that's about to launch next year or the Hubble or any of these things. You look at microscopes, yes. Robert Hooke had a microscope but compared to a basic hell microscope this is like a toy. But we have tools that nobody in the past even dreamed about. You start with lasers. I mean the lasers, we use them in daily life but people don't fully realize how much of a revolution laser has made in scores of investigative fields. Or you look at what computing has done to science. I mean forget what computer has done to productivity. We've talked about that. But think about what computers are doing to our ability to discover science. I mean we have entirely new field that have computational ex-biology, physics, chemistry. You name it. These fields would have been unimaginable. We can simulate equations that people have known about for centuries and could never solve. And now we simulate them. And as quantum computers are going to come online hopefully after 30 years of waiting it will finally come online we can do this faster and faster and faster. So what I'm suggesting basically is that modern technology is going to pull itself up by the bootstraps using its impact on science. And that's an indirect effect on technologies and productivity. It's very difficult to calculate because actually it may cross fields in ways that we can't even imagine. But this is what's going to happen. All right, so that's independent variable number one. Here's another variable, incentives. We're all economists. We understand about incentives. Now what kind of incentive structure do we have that drives science and technology? We basically know the story, right? Technology is largely driven by people trying to make money whether they make money through the patent system which doesn't work too well but we can imagine worse. But there are other things that incentivize technology. First movers, advantage, secrecy, blah, blah, blah, the whole thing. Now, did this exist 300 years ago? The answer is yes, the patent system did exist in some countries. It was not very well working very well but it was working to some extent, worked okay. Second question, how do we incentivize the people who produce the science? And there the problem is difficult because they're actually, you know, very rarely does a scientist capture anything of the social surplus that the science produces. So what do we do? We have devised a system which we call reputation mechanism. Every academic in the world plays that game. You don't do it for the money. You do it for the reputation. Now reputation is kind of nice because it's correlated with the 17th century patronage. I've written a book about that. We don't call it patronage. We call it tenure. But hey, you know, it's the same thing. But it is about reputation. You sit in a tenure decision, you realize what's going on. People look at what people think about your work. Now, as I wrote the book about these reputation mechanisms, the 17th century, you know, it started the 17th century. But what they had was nothing compared to what we have today. We have a huge system of incentives. It's not just tenure. It's named professorships. It's prizes of membership in learned academies. I mean, you name it. We have perfected this system. So the scientists today work harder than ever. And they work basically all to maximize the reputation. And the nice thing, of course, the globalization, much more now than in the 17th century, means that you have a global reputation. And that, of course, does yield certain phenomena like winner-take-all and so on and so forth. But basically, everybody understands that. And so what I foresee is that science is going to continue to thrive. Don't ask me where it's going because that takes me to science fiction, not to, oh, I can give you a list of things where I think it might be going. Basically, if I had to make any prediction, I would follow Freeman Dyson's prediction that if the 20th century was the century of physics, the 21st century will be the century of biology. I think our ability to manipulate life on this planet by C-S-C, or maybe more advanced gene-editing technology is going to truly be an epochal invention that we will be able to design the kind of lifeforms that we need to cope with a changing environment. That is, I think, the great future that he has lying before us. That, I don't think, is going to be science fiction. Thank you. So, no, thank you very much. Now, my conclusion is then it's a problem of diffusion, you know, of science in the society, not so much innovation per se. I mean, because you stopped there. You give me another 10 minutes? No, no, I don't. I don't because Mariana, Mariana, I'm sure, knowing what you write, she's going to pick up, you know, the role in particular of the state because you've written a lot on this, but maybe not only on this, but you may make this point. And then we'll go back to, you know, what's the problem then if, you know, things are so rosy. And we have some, as you know, some colleagues of you also here in the room, they argue in the other way. I can't imagine who you're talking about. You can't imagine, no. Thanks. It's the problem with passionate people, you know, in panels, you know, it's very difficult because you are so much interested in learning so that, okay, thanks for the discipline. Mariana, your turn. Great. So you address this issue now. What's the point of attack, you know, of that? So I think this session is incredibly important precisely because of the context that was laid out last night, which is that lots of the growth that's been happening over the last decade has actually been consumption-led. So if we want, well, consumption-led and private debt-led, so I'm sure you all know that the levels of private debt over disposable incomes are back-to-record levels and that's what caused the financial crisis in the first place. So if we want investment-led growth and if we want that investment to be directed and I think that's where this concept of smart growth, innovation-led growth comes from, I think what this panel then is, you know, does if you want to give us some insights on how that might occur. And what's extraordinary with the Economist is on the one hand this one sort of area where I really think we all more or less agree in both micro and macro that innovation is one of the key factors that determines long-run growth, then a key area of disagreement is how that innovation actually comes about and in particular what the role of public policy is. And I have six quick points that I'll get through in these nine minutes I have left, which I want to sort of unpick some of the myths and some of the problems, I think, that exist which are not going to help us get that investment-led, innovation-led growth unless we get over those problems. And the first one is that we have a very useful framework in economics to justify public policy, which is you have to first identify some sort of market failure where it might be the public good problem, so basic research, you need government to finance that, it might be asymmetric information, so we need SME financing, it might be negative externalities, so a carbon tax on pollution, but actually if you look at history, and Joel is an expert here in history, it would be incredibly hard to explain anything that has happened in innovation in terms of big innovation, so general-purpose technologies that have really affected productivity across many different sectors, also across production distribution and eventually consumption if we just had that one lens, that the role of policy was fixing some sort of problem, one of those different market failures. Again, very useful framework, but very limited also in terms of guiding future innovation policy. So that first point is, the sort of one-liner is actually what's happened is that the few places in the world that achieved smart innovation-led growth, few regions even within countries had policy that was much more about shaping or co-shaping and co-creating markets, not just fixing them. And just the quick sort of way to also describe this is if you look at the type of organizations and their level, the sort of breadth and depth of intervention, for example in Silicon Valley, whether it's the DARPAs, the ARPA-E, the SBIR, the NSF, the procurement programs, the NIHs, their breadth and depth was across the entire innovation chain, basic research, applied research, the linkages between them, which Joel was basically highlighting why we need the linkages between them because innovation is quite serendipitous. The search for one thing often leads to the discovery of another like Viagra, they were looking for one thing and something else popped up, very interesting history. Then early stage, you're supposed to laugh, early stage finance for the few companies, there's actually very few that are often even willing to engage with some big innovation programs, procurement programs, and then also bold demand side policies. Don't forget that mass production, which was one of the biggest innovations, organizational and technological, would not have had the effect that it did without, for example, suburbanization programs, which were from policy, which actually allowed this new innovation to get fully diffused and deploy throughout the economy. And second, if you actually look at how those innovations occurred and the way that policy interacted, it wasn't saying, oh, let's go after that technology, let's do the Internet or let's do GPS. Often the big innovations, all the ones that are in your smart products like the Internet, GPS, touchscreen display, Siri, voice activated systems, were spillovers over time, and you need that sort of long run perspective of mission oriented investments. So going to the moon is the most classic one, but some of the more interesting ones today are, for example, those directed at the climate change problem, but going to the moon required more than ten different sectors, including clothing. You couldn't just go there with jeans and a t-shirt, and our ability actually to also capture and assess and evaluate, because we often have to evaluate these public investments, really must adapt to really capture that full sort of dynamic spillover process, including which everyone thinks was a failure, but I've never seen one proper article that has captured the full spillovers that came from the Concorde investment, right? Anyway, these mission oriented investments are quite important also in terms of capturing the multiplier effect. You know, if you're just doing innovation policy, kind of patchy here and there, the effect that it ends up having on the overall growth of the economy is probably less than if you have this more sort of directional push, which also really creates linkages between not just different sectors, but also between manufacturing and services. Third, the actual structure of these investments is incredibly important. This is not just public money thrown out into, you know, two different companies or different areas. The organizational capacity of, you know, organizations like DARPA, ARPA-E, YOSMA, this public venture capital fund providing patient long-term committed strategic finance in Israel, key to the whole start-up nation dynamic in Israel. What actually happened in those organizations is incredibly important to learn from, right? So they were public but unpolitical. You know, you didn't have a political process getting their hands inside. Same thing, by the way, and Mario probably knows the history of EDE. EDE in Italy had three different phases, very important. It began not very well with Mussolini, but then it had three phases, public unpolitical, public ultra-politicized and then privatized. And the second and the third phase in Italian industrial development were quite problematic. That first phase was incredibly interesting also in terms of their ability to attract the top managers wanted to work in this public entity called EDE in Italy. Same thing in DARPA and ARPA-E, their ability to attract real experts, real scientific experts who think it's an honor to work in these agencies is very much related to their mission-orientedness. If they said we are out there simply to de-risk the private sector, to level the playing field, to fix some sort of market failure, it's not that exciting. You know, someone like Steve Chu, a Nobel Prize-winning physicist who ended up directing the DOE under Obama, you can bet that that was very much because Obama, at least initially, was thinking about directing this 800 billion stimulus program very much in this green direction. And that ability to attract that level of expertise in Europe today where we like to blame the public sector for everything without admitting just how many problems there are also in the private sector with this incredible inertia I think is really hurting our ability to reform the public institutions. So again, coming back to Italy, the reform of the public administration is not very much about rethinking how do we develop this dynamic, decentralized network of different types of public organizations able to think in the strategic way. Fourth, this whole issue of, well, what kind of investment? Direct, indirect, well, of course we need both, but one thing I think internationally is we have too much emphasis on these indirect kind of tax incentives, whether it's the R&D tax credits, which are actually the more smart kind of tax incentives we need or again the patent box, which I think is a very bad idea, but anyway, in general this assumption that business wants to invest and you just have to facilitate and de-risk that process is simply wrong. Often you know, and Keynes actually talked about this in a wonderful letter he wrote to Roosevelt in 1932 where he said, we have a problem here, you know, these businesses out there, they're actually not wolves and tigers and lions, they're actually domesticated animals and so this notion that the animal sphere, it's actually have to be created endogenously through public investments to get business to even want to invest, so through direct investments in nanotech and biotech in the life sciences, which by the way preceded the venture capital entry into biotech by 30 years, and then on top of that ad as icing on the cake but not the cake itself, some sort of tax incentive, that balance is incredibly important because otherwise all we do through these incentives is increase profits and as we just heard from Thomas, whose work I love, you know, there's no profits problem, profits are already at record levels, the real problem we have is the investment problem, so how to create quality, how to get business to invest when it wouldn't have invested anyway is key and this kind of mix of direct and indirect but the structure of that direct investment is really key to understanding that process and I wrote a book called the entrepreneurial state trying to really understand how that's happened over history. Lastly because I have one minute left and because this is, you know, a conference organized by the ECB, the kind of summary in terms of finance, I think it's very important first, there is a finance problem, there's a demand for finance problem, in fact SMEs by the way, and I always get killed by politicians when I say this, are overly financed, you know, in the UK we spend more on SMEs than on the universities, right, like something like 9 billion, there's few SMEs about 4% that are even interested in innovating, so either we should just do it for folkloric reasons to finance SMEs but if we're interested in productivity increases and innovation, we really need to think of mechanisms to pick the willing and this is independent of size. Second, it's not the quantity of finance but the quality, patient, long term strategic finance is what is required and unfortunately in many countries we don't have that, some countries might do it through a public bank like the KFW in Germany, some as I've mentioned already through YASMA in Israel, some like the US through procurement programs, the SVIR program which was fundamental to providing early stage finance before VC came, and lastly and I'm five seconds over so I'll go ahead and finish this. I think the financialization of the real economy is a massive problem, the proportion of share buybacks to R&D spending has escalated and it's getting worse. Apple under Steve Jobs reinvested the profits, Apple under Tim Cook, not, what's the future of innovation and Apple that corporate governance issue is key to that. Thanks, it doesn't tell us really how you pick up the winners I mean, in what you say. We'll give you another 10 minutes. I mean, it's a very credible play with that provision, I mean how do you do that, but that's correct. So you worked a lot on R&D, so what is your view on this? So you raised a very specific question that we needed to address so my question was do we measure productivity well enough to draw really conclusions of do we have a productivity slow down? I actually took this question to really come from a concern that the innovation machine as a driver of growth was in trouble in Europe and so rather than talking about productivity growth and its measurement through the residual TFP, what we typically see, I would like to immediately turn to innovation and its measurement directly of that to see whether we see the same product, the same downturn, slow turn in innovation when we look at the process directly. But before I do I want to make two specific points about what's so specific about innovation. One is innovation investment, different from capital investment, and one first important point is that innovation of course creates private welfare for the innovator, that's where the incentives to invest come from, but the really full creative potential of innovation comes from not only the innovator but also indirectly from the diffusion and the spillovers, that's not just imitators but it's others that apply the same idea to completely new applications and that's really where the power of innovation as a growth machine comes from, it's from this diffusion and where you have the social rate of return rather than just the private and we saw this with the work of David Otto very well. Unfortunately it's much more difficult to measure this diffusion process and particularly to which channels this diffusion actually runs. What we know is that of course diffusion requires some kind of communication but the transfer of know-how is very often embodied in people in R&D and innovation personnel and what we know from micro evidence is that lots of these spillovers actually come from transfer of people that move within firms to different jobs but also across firms and across sectors so that movement of people is really very important to get these spillover effects here and also important to note is that it will typically take a much longer time before you see actually all these private returns so the returns on investment for investment really take a much longer time window than other capital investments and then find a lot of important very specific point of innovation is yes it's a creative power but it's also a destructive power, it destructs all technologies, products and skills and particularly the creative destruction power comes from the more radical innovations and they often come from firms which are challenging the incumbents here so we also need to identify the type of frontier firms here and making distinction between new versus incumbent firms here so what do we now know about measurements of innovation directly do we see the same kind of downturn taking all these effects into account so unfortunately we don't have that good measures about the effects of innovation and the social rates of return, the fusion is very hard to measure but we do have better data on investment for innovation and of course that's an input variable so it doesn't reflect the effects but it's nevertheless a very good indicator because input variables already reflect the capacities of firms to invest in innovation and it's just only the creation of no ideas but also the adoption and it's also a reflection of do they have the right incentives to invest so capacities and incentives should be reflected in these input variables as well so if you take a look at innovation expenditures and that's really the most broader definition of investments because that includes investment not just for creating own ideas but also for adopting and adapting other innovations what we know we have community innovation survey evidence which is run across countries here and across time unfortunately it's usually it's a repeated cross-section it's not a panel and that's why it's sometimes very hard to really get good time trends on that but for Germany it is really a panel that can be traced and there we can look at really both the extensive and the intensive margins so do we see more or less companies over time engaged in these innovation expenditures and what we see is actually very interesting is that overall the number of firms that are active in innovation investments has actually decreased over time and that's actually a trend that holds particularly for SMEs that are less likely to adopt new technologies here so we see actually a drop in extensive margin of innovation active firms here and also a drop in the intensive margin as well nevertheless overall the expenditures for R&D and particularly so for the creative part here is increasing in Germany even the share of GDP here but that comes mostly from the large companies that are increasingly investing in R&D so overall we see a larger inequality in the innovation landscape here so number of smaller firms particularly dropping out and a more concentration in the leading firms in Germany so that's why we also look at corporate R&D so that's the R&D investment for creation of new ideas by the way most of these expenditures actually don't go to infrastructure but go to R&D personnel here so it's more than three quarters of that is investing in people rather than in equipment here. These aggregate R&D numbers is something we constantly trace with the OECD birth numbers and they show actually there is no drop in birth here it's pretty it goes slightly up and in Europe it's pretty stable with a slight increase here of course the China is a stellar growth performance on this dimension here but at least the good news is that in Europe we're not going down on these corporate investments here but if we take a look at the firms that are behind these corporate R&D investment trends here and we look at the scoreboard firms which are traced by the European Commission since a couple of years but these are only the largest R&D firms the 2,500 largest firms traced over time we see that overall that total investment of these firms also has continued to go up so we don't see the same kind of downward trend there again but in that recent work we've been doing at Bruegel on these corporate data we do see that there is a difference in terms of concentration that's taking place and that world of large R&D spenders here so these R&D spenders are highly concentrated in a few firms so that corresponds to also with David's analysis here it's even more concentrated than their sales and employment here so for instance the top 10% of these 2,500 firms so the 250 largest ones they account for 3 quarters of all the R&D expenditures here and the top 1% so the top 25 almost one third so it's very highly concentrated in few firms if we look at the trends over time in this concentration we did see downward trend in concentration of these R&D expenditures so the concentration went down but this downward trend has stopped more recently and is now starting to increase again and particularly the share of the top one here so it becomes increasingly more that concentration is particularly high in what we call innovation based growth sectors so these are the higher R&D growth sectors where there are more new R&D leading firms so these are the sectors where there's really much more dynamics going on that's typically the pharma biotech nexus and then the ICT nexus both hardware and software in pharma and biotech we see it's very highly concentrated in terms of R&D expenditures but at the same time we also see it's a very stable set of top firms here despite the biotech revolution here, classic pharma is still very much dominant in that sector here and they have really identified this biotech challenge by taking over a lot of these small biotech firms here but it's still the same kind of leading firms that are dominating this sector if we look at ICT there we actually see it's also very highly concentrated but we see here we see this upward trend in concentration of R&D and that holds particularly in the US ICT actually basically Europe is not present in ICT in terms of the landscape here so we see that in ICT in the US there is recently an increase in this concentration of R&D not an employment only in R&D expenditures here the question is should we worry about this this higher concentration of corporate R&D in fewer firms so it could be that indeed these leading firms are simply the more efficient ones here in terms of R&D they're exploiting better R&D economies of scale and scope and in that respect then the increase in concentration is a higher efficiency of the R&D landscape this could hold particularly in pharma and biotech because there are indeed very large economies of scale and scope in R&D however in ICT what we see is that this increasing concentration comes also with a higher turbulence in leadership here there are also much more new leaders within this landscape here so it's a concentration but with a higher turbulence so you could argue that these new leaders bring in the more newer R&Ds here so just to give you some illustration of these numbers so in pharma 93% of the R&D expenditures by the top 10 in 2015 came from firms that were already top 10 in 2005 here so you see how much it's concentrated in ICT that's only but only 70% so it's still highly concentrated but it's less and at least we have two new firms that belong to this top 5 that's two way in Facebook here so there is a bit more turbulence going on here I'd just like to make the point that we really need to look beyond the sheer size of investments in R&D here we also really need to look at who is making these investments here the distribution of that investment into leading firms here and the diffusion aspect here but overall I think although the total investments is not such an issue here it's an increasing concentration that we need to be more careful about understanding the process Thanks the discussion will be on policy of course and we will concentrate on that and you will tell us probably that science fiction is today already yesterday well when I thought about what I talked about today I asked Peter for some guidance he sent me some questions and they were such good questions I'm going to structure my talk around those questions first one was were innovations of tomorrow be driven by new firms or established incumbents so that's very tempting to give an economist answer well on the one hand on the other hand but actually both is the answer and let me explain that so here in Europe we hear a lot about GAFM now Americans never see this acronym stands for Google, Apple Facebook, Amazon, Microsoft these big American incumbents in the IT area but far from being some kind of sleepy cartel with highly concentrated industries what we've heard about earlier they actually compete very intensely among themselves very intensely compete in operating systems office productivity applications cloud computing devices advertising services and many others if you sit down for a few minutes you can create a table that shows you this effect and it's because of that competition that the prices are low often free and innovation is very delivered as promised, lots of innovation low prices lots of activity but more importantly it's not just those incumbents that are benefiting from this environment but even the startups the new firms because now they can go by cloud computing and networking from Amazon, Google, Microsoft they use open source tools like Python, TensorFlow, Hadoop and that's virtually eliminated the fixed cost of entering a business I like to call this the return of constant returns because you can go out and purchase the business services on an as necessary basis as you grow you can add more in a seamless way and this has made a huge difference in the startup world Susan Woodward at San Telecom Metrics says that since 2010 there have been 4,000 new companies founded in Europe raised over $27 billion and a large part of this is due to the fact that the business services and the technological services necessary to enter have changed dramatically when Google started, when Microsoft started you had to build your own data centers you had to create your own software you had to go out and build all of those fixed costs before you could produce anything but now those services are readily available for computing, networking, tools in the cloud and it's easy to get started and easy to grow so take an example, Netflix Netflix is based on the web, it's based on serving data but it has no data servers of its own it has no network of its own it just uses Amazon Web Services to provide that Spotify, Evermore, Airbnb Costco, Snapchat they all run in the Google Cloud they're not providing their own infrastructure they can go out and buy the infrastructure because you have this very competitive supply from the firms I alluded to a minute ago and it's not just on the computing side you can hire people through LinkedIn you can find customers using Salesforce you can do your accounting using QuickBooks, your outsource user support using Zendesk on and on and on if you want these services you can go find them and you can go implement them on an as needed basis there are literally hundreds of these specialized support services can be purchased at whatever scale is necessary so Adam Smith once said the division of labor is limited by the extent of the market famous quote but nowadays would say the division of business services is limited by the extent of the market but the market is a global market so these business services can be provided in a very very fine degree of granularity and specialization second question was what are the latest technologies not being adopted by sufficient large share firms and what can be done about it well I think the answer there is very easy it's machine learning predictive analytics all of these stuff we've been hearing about in the news there's a computer in the middle of every economic transaction now and because that computer is in the middle of the transactions you can capture data from the transaction you can store that data manipulate it analyze it and so on everybody has a data warehouse every major firm the software is readily available as I mentioned earlier cloud computing provides the platform for the analytics what's left over the expertise what you were talking about the expertise is a problem but even there not for long because the universities have responded very quickly to start offering programs in data science and analytics not a week goes by that I don't get an email that says we're putting together a new program in data science what should the curriculum be can you come talk to us can you help us and so on just in the last few months it was at MIT at Stanford at Berkeley Michigan all of these places are getting into this area and turning out the students in fact the choke point now called area is actually on the firms hiring because if you don't have the expertise internal to the firm it's very hard to understand who to hire what skills you need what are the differences and so on what are the differences a statistician an econometrician a data scientist a machine learning specialist and so on without that internal expertise it's hard to build up but of course you bring in consultants you bring in outside experts you develop that in house expertise and then you can utilize that technology and finally the last question what's the most important factors ensure cross fertilization between fundamental university research what we've been talking about here industry research and development and then business investment so I wanted to talk specifically about Google because I know what Google is doing but this is widespread phenomenon we provide a lot of open source software for example recently we've released this system called tensor flow tensor flow is a system for doing machine learning at scale at huge scale the software makes it completely transparent if you want to run your analysis on a single CPU on a GPU graphical processing units on TPU the tensor processing units you want to run your analysis on a thousand machines at once that's just a simple instruction to the system we also provide the training data because to do this kind of deep learning you need to have data that's trained so we donated nine million labeled images with six thousand labels to the open image project so if you want to do something in image recognition you can go out and get these nine million actually there already was nine million so nearly twenty million labeled images to train your system on we also provided eight million YouTube videos that's five hundred thousand hours worth of video along with the video labels from a diverse set of four thousand eight hundred different labels to not intimately connected with the machine learning image processing image processing by machines now beats humans so you can get a better set of once you've trained it on this labeled data the computers beat humans but humans are still way ahead in understanding YouTube videos because all those cats you know no what happens is it's very hard to understand actions you see people are they dancing are they fighting are they doing kung fu are they exercising computers are having a quite a hard time with this so now the YouTube videos are out there to help make progress in this area we've also donated one thousand cloud tensor processing units to universities each one offers a hundred and eighty teraflops floating point performance 64 gigabytes of memory and they're available at the university so why do we do this is this great altruism on our part well there's a little bit of altruism there I have to say a little bit but on the other hand we want those students to know how to use those systems we want them to know how to use tensor flow we want them to know how to deal with these labeled images we want that to be a skill that they come to the job with and in fact we hire thousands of interns a year for exactly the same reason because the big competition now is competition for talent talent competition for talent that's the single most significant definition of competition and of course from the viewpoints of the job applicants who are looking for jobs what are they considering they're saying I want to go where they can learn the most I want to go to the place where I can develop my skills make myself highly marketable and succeed in my chosen line of work so that's what's happening now it's a very vibrant time I think we'll see more of that in the future thanks thank you it's interesting because to prepare this session I googled actually I googled now to see the difference between skill and talent you know if you came exactly and then I asked myself about what is a low skilled job I mean certainly look at the next 20 years and I thought basically what cannot be replicated you know it's a lot of emotional interactions which is one of the points you made you know when described you know the development of technology and Roberts actually maybe in the end they will be able to do that and then I thought about many what we call low skilled jobs are not very well paid in spite of the fact that very often they have this comparative advantage compared to machines and learning machines in the future and then you think about the low skilled companies you know and these things which are really make a difference in many jobs now and they may be asking myself the relative price so it's going to evolve in the future because now it's usually low paid jobs even for those who have sort of talent of dealing with interacting with other people in general so I mean that's because you provide a sort of very positive picture of the environment now when you aggregate that at a society level I mean there you don't have to worry about it well yes because now we enter the reaction of the panel phase which takes a couple of minutes so that's one I mean just to kick off and then you react to each other and then I open the floor but very quickly just I give my sense when the job demands this level of skill and the available workers have this level of skill there's two ways to solve that you can bring the worker to the necessary point or you can lower the job to the skills that are appropriate for the worker so it used to be to be a cashier you had to know how to make change irrelevant it used to be to be a taxi driver you had to know how to get around a city now irrelevant it used to be to be a physicist you had to know 100 different integrals now no longer necessary we have symbolic integration systems that can calculate those integrals in many many cases the skills that were prerequisites for a job at one point have now become provided to the user via the information technology and I asked needed basis that's why the distinction skills and talent you want to react well you know it's more policy that's because I'm tickled by what Hal just said because I think there's this has been around for quite a while this didn't start with digitalization and in fact I think there's a paper written in the 1970s in which one economic historian basically said that if things are going the way they are going in the asymptote we will have machines designed by geniuses to be operated by idiots and I don't think it's going to be quite that bad but that is precisely the kind of thing that you were talking about I want to make if I may respond to Mariana's talk I love the entrepreneurial state and I think she's absolutely right but I think behind that actually that we see technological change occurring at very rapid rate when there is something called a focusing device when society puts in front of it a problem that it feels it needs to solve okay and so what the government does basically it aggregates society's preferences and it channels it to other reasons so this has always been true in the 18th century British society had a number of problems that it knew it had to solve okay one of them is how to pump water out of mines another is how to prevent smallpox and the third one is how to measure longitude at sea in 1800 all of the three were solved we have exactly the same problem in the late 19th century we had to learn how to fix nitrogen from the atmosphere people worked on it they solved it they had to make cheap steel they solved it what is the situation today that's a very well clear cut defined issue that we have to solve it's called climate change and that is where a lot of and I believe we'll tackle it I don't know how exactly but we will tackle it and by the way climate change is evil twin is ocean acidification good luck solving that one so very good point and this whole focusing device is what I basically meant by the mission orientedness and why we need that to focus broadly defined directions and then allowing this bottom up experimentation there was 500 homework problems that got us to the moon again cross sector however if you look at what's happening currently in Europe there's huge differences in how the climate change problem is being tackled in some countries I would argue in the UK very patchily very unstably you know feed and tariff here then we take it out a bit of funding here then we change the name of the agency then we cut the agency then we restart the policy versus Germany which is not perfect the energy van policy is not perfect but what's very interesting about that policy it's very mission oriented it's required every sector in Germany to transform itself so instead of having a sectoral based policy say just renewable energy or some other sector the energy van policy has created a challenge a mission and has rewarded those companies in those sectors that have been willing to respond so steel in some countries remains quite boring and has to be bought up by the Indians Tata Steel having to buy up the UK steel industry in Germany and in Belgium steel has massively lowered its material content through repurpose reuse and recycle not just because they woke up one day and wanted to do it but it was very much again challenged to do it that way but can I just ask a really quick question to Hal about Google so Google what did it get from government it got its algorithm right publicly financed by the NSF it got GPS which what would Google maps be without it my question to Hal is should and I'm really saying it sort of you know openly should government sometimes not for the basic R&D that's obvious retain a share in these investments you know Tesla got 465 million guaranteed loan success cylinder of 500 million taxpayers asked to pick up the loss it doesn't necessarily get a share of the success should we be thinking about sharing not just risks also rewards and what would that mean for Google in terms of the benefits and the profits that you currently have just one slight correction actually Stanford owns the original patent on PageRank I like to say Stanford it's like a university but with money and part of the reason it has all that money is because they've really funded scientific research and they've contributed dramatically to the economy not just the Silicon Valley but of the entire world so that's the system we have I like to see universities get money from intellectual property maybe there should be some sharing arrangement with the federal government I don't know it was a good polling question but we don't have that in our list we will have one of the polling questions just after I need so that will be question 2 in our list wait until I need this finish just wait a minute I'd like to go back to this issue of talent and skills because I think that's really very crucial for innovation everything draws around having these talented people being able to create and diffuse new ideas but of course the talent needs to be nurtured needs to be trained and who is the incentive to train the talent here so certainly because this training will actually be not just only firm specific but should be a general type of training here so all the examples that you gave of universities do they have the incentives to train we're all from the US in Europe I still think we missing a bit that single market for education and for training here which reduces the incentives for universities to train very specific skills here and also one other question I still have on the training here is if training happens in corporate companies here to which extent is that specific and to which extent do these companies still try to appropriate the rents from that investment here and that may actually reduce the social returns from that training and the mobility of the people to move outside here so that's definitely also so we have now the question the polling question we are asking not yet is not yet in the iPad no it is you know how to handle the iPad I hope so do you think that the productivity gains derived from the ICT revolution alternatives A have slowed down given that they have been largely by now B will accelerate even that the full potential of the ICT revolution has not emerged yet C are neither slowing down nor accelerating I think I should pay attention so of course you can guess by answer I think that's not for you that's for the audience for the audience then they're taking up our time no no I hope I need a pen actually an old technology we have 10 more seconds so you're going to I need a pen take a vote can we answer it the audience from the audience so I will we have a referendum yeah well 70 given that potential I think it has not emerged yet it has not emerged yet I guess would you answer that what you have said I see 72% I agree with me that's good I agree with you yes no no no I pick up questions 3-4 questions and then the first round of questions yes you give your name affiliation as usually I know you from Amsterdam one of the currents that I saw through all of the three is the incentives for doing the innovation R&D for academics its reputation I'm just wondering though whether the it was mentioned a little bit on the winner take all you think that's actually going to start reducing the desire of people in Silicon Valley to work day and night and end up at their 30th birthday going back to their mother's couch and only one person gets 50 billion is that going to undermine the incentives to have the winner take all that gets so strong as it is how to do about it I pick up three questions first so we take note of this one second and then I have Richard Portas London visit school and CPR the first session policy first session policy conclusion was focus on the quality of jobs and in a sense my question relates to the exchange between Peter how a minute ago how do you raise the quality of the job in picking fruit in care of old people in a whole range of services which are low paid services for which we in the UK cannot attract young people duty jobs we therefore import them but that's going to stop what happens after that is not clear nurses in the healthcare system it's going to be very difficult innovation led growth how does that help raise the quality of jobs for often poorly trained young people who don't want to move they don't want to leave their communities where they've grown up and that's a problem we see in the UK many communities Brexit land and a problem of course that you see in the US from plan and I'm interested in how things somehow innovation led growth it's a relative price which may adjust involving emotional interactions versus the others it's maybe too early but at some point we're going to see what you say but that will be for later my question is actually for Peter Pratt and Embracing Power as we discussed also in the last round so we've seen that there is a shift of investment towards intangible and generally certain type of skill of human capital beneficiary and the net result seems to be a declining corporate demand for external finance credit markets are simply not the way in which the news forms of investment are financed and that means that there is also truly a smaller role to play for policy intervention in credit markets so provokedly but what is the residual role for monetary policy in promoting investment in this context that's my colleagues on this maybe let me take a fourth question and then we close the first round thanks, I have a quick question for the chief economist of IBRG my question is for Mariana I think your idea of government taking a stake and a success to what extent you think that's already been done through profit tax government is a shareholder in every company because it has a share in every company's profit and in that sense when Google makes another billion in corporate profits government does take a gap to what extent you think your idea has already been addressed your idea has been implemented so there's different ways that the government could take a share it used to be in fact tax I mean do people just yell at a number what was the top marginal taxation rate when NASA and DARPA were founded and the president at the time was a republican over 90 so obviously what's happened to tax including also different types of tax loopholes that if we wanted to we could fix that has changed massively we also by the way when we go to war we don't worry about tax we go to war we create money through different means but let's ignore those two possibilities of create money to do innovation like we do with war or raise marginal taxation right back to what it was there's different ways one is I think downstream this is different from the upstream stuff where you really can assume that there will just be spillovers, knowledge spillovers etc downstream when you're giving particular companies 500 million in a guaranteed loan Elon Musk, one man 5 billion from the US government for his three companies right SpaceX, SolarCity and Tesla so these downstream investments to particular individuals in particular companies I really think it's completely naive not to think about sharing the risk and rewards as by the way Yasma does through royalties as Citra used to do through equity as any state investment bank through its investment fund would do is not allowed to make a penny on any of its investments it can only get money through the budget appropriations committee that would be fine if NASA was just doing basic research but when it allows Novartis to work for free in the international space station I think that's a problem or the space tourism which costs 30 million whatever it is to go up there Justin Bieber goes up unfortunately comes back down that money not a penny goes to NASA why? we can rethink these and by the way we think in terms of admitting that things like patents were a deal between government and the company now what we've been allowing to be patented hurts that deal to the public return we're increasingly patenting upstream so the tools for research are being patented so we're kind of going back to the world of secrecy that's a bad deal it's not about should we or should we not have patents but the public return falls if then the diffusion of that innovation once the patents up has been hurt and of course it will be hurt from that last example the Beidol Act which allowed publicly financed research to be patented from 1980 onwards in the act it's actually there in the act it says well we better make sure the taxpayer doesn't pay twice 32 billion a year with the taxpayer in the US pays for the national institutes of health 75% of new molecular entities with priority rating come from that kind of finance and the act says we better make sure we pay twice so let's allow the prices of those drugs that have been massively publicly financed to be capped sort of marching rights the US government has never exercised its right so this is also a power play when the narrative is so strongly skewed you know that somehow innovation all occurs in companies and at best what government is doing is leveling the playing field fixing market failures that also confidence for government to say okay well yeah of course we should cap the prices of drugs that we helped to finance not because we're just regulating the process but we co-created that system itself that you know I think is a power issue and it's been fed by ideology and so it's not just about equity stakes even though I do think you know sorry just really quickly with Tesla Obama did the opposite he said if you don't pay back the loan 465 million we get 3 million shares in your company it's like why would you want 3 million shares in a company that doesn't pay back its loan probably because it's not very good had he done the opposite 3 million shares in 2013 when the loan was paid back the price per share had gone from 9 to 90 that would have more than paid back the salindra loss which you can bet would have made the taxpayer and this is about the consensus building and not getting taxpayers pissed off that you know they're always just picking up the bill whether it's of the banks or of the salindras that would have helped but somehow this isn't you know a question that we're asking you want to actually was reflecting what Mariana was saying she's saying it's absolutely true on the other hand it's worth keeping in mind that you know there's a famous paper by Nordhaus who actually calculated that when firms innovate 97% of the social surplus filters down to consumers and so eventually the taxpayer is also the consumer so I'm not losing all that much sleep over this I do want to answer Eric Bartelsman really brilliant question about is there a winner-take-all system in some sense destructive for the incentives I thought about this a very long time because it is an important issue and here's something anyway I wanted you to think about this people buy tickets in lotteries and ex ante we all know the expected value is negative but they still buy the tickets in lotteries actually innovating where any kind of entrepreneurial activity is like buying a ticket in a lottery and people buy ticket in a lottery but there's a difference between a ticket in a lottery and engaging in innovation and entrepreneurial activity and that's actually pointed out by Adam Smith in a very lovely passage which actually that people are overly optimistic about their own capabilities and so we have the odd phenomenon which is that the unconditional probability obviously had to have to add up to one but the conditional probability are larger than one because everybody thinks she's smarter than the average right this is like like Wobygun we're all above average and of course that is what that over optimistic illusionary is actually what's driving innovation and so yes in the end there will be a lot of disappointed people who will not win the Nobel Prize and whose companies do not get the Google's and the Apple's of huge future and there's probably some negative social surplus but that is way way way smaller than the total social surplus that accrues to society from the fact that most of us are overly optimistic which is a really good thing I want to say a word to about the oh I'm sorry did I take you settle, you settle it's fine go on so I you want to I want to say one word that hasn't been mentioned this morning and should have been and that's demography because if we look at the US right now the labor force is growing at half the rate of the population next decade the 2020's you're going to see the lowest labor force growth ever in post war data if you look at Japan China, Germany, Korea Italy they all have a terrible demographic problem in terms of reduction in the labor force and what does that mean for the wage well we would think that you would see the wage rise for unskilled workers especially because computerization can provide the skills on an needed basis in many situations so I think all of our intuitions are about this period when we've had a surplus of labor because of the baby boom on the one hand and women entering the labor force and the other both of those have maxed out and in the future we're seeing the baby bust and we're seeing a low participation in the workforce all the discussions we've had suggest that and we're going to see a higher wage larger share going towards the labor bill I would also like to come back to that question of Eric on the extent to which concentration might be reducing the incentives to innovate and particularly the incentives for these new creative ideas that come from new entrepreneurs would that be indeed reduced if we have more concentrated markets and I guess indeed the answer could be yes as well as no yes would mean that indeed if these incumbents would indeed form barriers to entry making it less attractive for new players to come in but it doesn't have to be it could also be just increasing the incentives if these entrepreneurs with these new creative ideas can then sell their ideas at the right price to these incumbents which would be more efficient to use these new ideas here so it all depends on the extent to which we would have a well functioning market for technology here and definitely also requires that we have both competition policy and regulatory authorities here to make sure that innovation is high on their radar screen too to make sure that you have a fluid market there as well so that it could be positive Can I pick up the concentration? Another view of this concentration problem is you know there's different reasons why a market might be concentrated and sometimes it's also government permitted so when AT&T was a monopoly very concentrated market it led to innovation also coming back to my previous point government was a bit more confident they forced AT&T to create Bell Labs that was part of the condition to retain your monopoly status you had to reinvest those profits and this is very interesting in a context today where we have record level hoarding of cash in both the US and in Europe and record levels of financialization Google I think is very interesting because it does actually reinvest its profits in big areas sort of as Bell Labs did in terms of energy and no one asked them to but I think the other side of that is also this whole corporate governance issue that when you have large companies what they're again doing with their profits also then affects the small companies that they buy up so Cisco has undergone massive change in terms of its corporate governance and its level of financialization and speaking to companies that have been brought up by Cisco the experience of being brought up in its history when it was reinvesting the profits back into R&D and human capital is very different from its experience today because an extremely financialized company didn't used to be that way when it buys up a small biotech company it's the ability of that unit to continue to be innovating to think differently and to continue to be disruptive is very different if you're in a monster that is ultra-financialized versus one that's quite functional in terms of what it's doing with its profit so corporate governance issues I think are key more than just the static view of how concentrated a sector is and how big a company is and that's the question in all this is number one so so question number one is a lot of new technologies are being invented electric and self-driving cars linear engines, new materials what should be the role of the government? the options are A, the government should only provide the legal frameworks B, the government should actively support innovation with subsidies C, the government should provide both provide legal frameworks and subsidies D, the government should not intervene so here I have about a minute cast your board why subsidies it's very interesting why subsidies the reason Italy is so problematic is the way the state intervenes is through subsidies don't influence it I've been talking about investing co-investing let's see but you're missing a choice co-invest finally above these are the results and D is assumed in A it's not well-fraised can we rephrase for the next one well yes both the first one is very high only the legal frameworks so let me pick up a number of questions a second one of questions looking around oh yes Sylvester I have one question which is relating to the fine Google receives the 2.8 million with reserves of 86 don't you think that one of the reasons why all these firms get these enormous fines have to do with the tax arbitrage these firms do with the rents you know relatively they are maybe under taxed and then from a normative point of view maybe society on the European, on the global level things well we have to do something about that and of course they find of course all kinds of limitations competition, monopolies etc but maybe one of the things behind that is the general feeling from a normative point of view that these firms are having enormous rents and don't pay relative property taxes other questions well yes there is another question there can I identify yourself please I know always Robert Gordon from Northwestern I wanted to link together Joel's emphasis on the interaction between science and technology and how's examples of cloud computing as a new tool which has allowed many firms to essentially scale up with variable costs instead of massive fixed costs of developing their own IT departments one would have thought there would have been massive unemployment of people who used to work in the IT firm IT departments of Netflix and Snapchat and all the other firms that you gave examples of and we don't see that we don't see an effect on productivity of the cloud computing technological revolution analogous to the big productivity growth that we saw in the 1990s as a result of the invention of the personal computer personal software and the internet I wonder if you could compare those two types of IT revolutions that one of the 90s which showed up in productivity and this one that's currently going on that does not seem to be showing up in productivity just one hint as to where I think the answer lies is a lot of the economy which is simply not being impacted by cloud computing and all you have to do is look at bricks and mortar retailing construction utilities, education and medical care which taken together is a substantial part of the economy and one in which we don't see a big impact of these current waves of innovation Thanks if not you answer on this first I've had this conversation I haven't had this conversation before but you know, for one thing I truly believe that we can get too much so we're engaged in productivity fetishism of some form I mean, productivity depends on the way we measure the numerator there's to say output and a great deal of the impact that we've had that ICT in particular but much more technology has had really doesn't show up in the GDP statistics for a whole variety of reasons part of the reason is that they really affect the sort of transactions cost of life so to speak so for instance if you are now able to buy a pair of shoes online rather than spend half an hour driving to a shopping center and then waiting in line and searching for help that is again in consumer surplus in social welfare that doesn't show up anywhere in the transactions cost there's a whole array of areas in which this is taking place even in something like Netflix if you calculate the real impact of Netflix that Netflix engages because it's a heavily competitive industry it glades in something like close to marginal cost pricing marginal cost is very low because in digital technology marginal cost by definition are low so that means that these prices are extremely low often zero GPS is costless Wikipedia, Spotify go on all these things that have changed our lives don't show up or show up in very weird ways I think that really matters but the other thing Bob you're right, large part of the economy have not been fully affected by the ICT which is a reason to be optimistic because it means that there is still a great deal of unexploited opportunities that will be exploited when people write better software and the use of software gets more diffuse so I actually see this as a very optimistic point you are telling us that we are richer than what the statistics are we are not only richer but we are better off the national accounts the national accounts they've always been understating what's really going on but I think the rate of understatement is getting worse as the digital sector increases because of the nature of the technology I'd like to come back on this question I've addressed several times on the incentives of some of these low skilled service sectors like education and health which are certainly in Europe public sectors here so do they have enough incentives to adopt the latest technologies often digital technologies so why is it still such an issue to get e-health and e-education really up and running here I think in Europe there is still a very important low hanging fruit which is the single market for public procurement here so we could win a lot by having that public procurement market way more effective here so that we could increase the adoption of new technologies that are much more efficient and lower priced deal for these public service sectors here would be way more faster in adopting these new technologies Paul? Yes, well to answer Bob I think you're exactly right the system is a great example there used to be lots of them that was one of your biggest expenses with your information technology now that you have cloud computing that has dramatically reduced the demand in that area but we don't see people standing on the street selling apples at least not skilled system administrators and I will say on the sectors you mentioned just as Joel said those will be will adopt the technology in my view on this taxation question well remember corporate tax is supposed to be in profits which is revenue minus cost to the first order for a lot of these businesses they've got global revenue but the costs are located in the U.S. to a large degree Google actually pays a 21% corporate tax rate the debate is not so much about corporate tax but who gets it so there's some squabbling among countries on that point nothing new there Mariana you want to come back to which question? So Bob Gordon's question I think it was your work also that showed just how long it took electricity from when it came out to get fully diffused and deployed throughout the economy it was something like 50 years so if you look at personal computers only now we would expect their impact to really be to be fully diffused and deployed throughout the economy and even really affecting how households behave, how different types of businesses work in terms of both their production and distribution but electricity just like the mass production example I gave also did have these directional pushes I gave the example of suburbanization having been absolutely essential for the ability of mass production one way that we could think about this today is could we actually and I ask it as a question have bold green strategies that allow IT to get fully diffused and fully deployed so green not just in terms of renewable energy but really allowing how we work, how we produce, how we distribute using the power of IT in that sort of green direction Richard's question on care I think this is also an issue care seen widely not just social care or hospital care or nurses but a caring economy there's people who have been writing about care as well as a direction but that would require all sorts of synergies and really seeing the system holistically and also doing something we do in the defense sector for some reason we haven't done in health which is to line up the innovation system with the access, the issue of accessibility and the price system and how when the defense department funds through the DOD or DARPA technological changes relevant to the defense industry the prices that are then charged take that into account so that just shows the sort of lining up holistically of that innovation model where the downstream market accessibility is seen in that context I think there's ways to rethink health innovation systems that see the welfare state not just as a redistribution of wealth but also as intimately interlinked and this also I think comes to some I mean you didn't say it's not just about basic research and applied research but how we purchase, government purchasing programs have been intimately linked with how we have innovations come about there's one quick thing on tactics though it's not just about the Irish schemes or the Dutch schemes within the US it's scandalous what happens so Apple moving one of its subsidiaries from Cupertino California which is where the knowledge is more or less created to Reno Nevada which is more or less a third world country where they're not sorry someone here no no it's scandalous because they don't pay a penny of tax in Reno Nevada that's how they try to track capital and 3 billion was lost to the state of California through that scheme right so you don't have to look internationally at these really problematic sorry Dutch house can I come back to maybe on the was a little bit in the speech of Mario this morning about the price discovery process the price process in general in the context of the supply shock that you were describing the fact that it's much more transparency access to pricing of competition you have a view on this I mean what is the price today actually price in general we as central bankers we have this exactly price stability and how do we measure prices today the price level and the evolution of price level in the context of technological changes that you described and what is the price the time dimension the segmentation the quality of products and all these things I mean because you have been working on big data also and trying to reflect on that aspect of the economic system well I like to say GDP is a hard time with free and take a simple example take GPS systems which you already alluded to one point there were over a thousand dollars that were only used in trucks they definitely improved the efficiency of trucking and logistics systems they become lower a standalone device that consumers can use fantastic as the price gets lower and lower and lower the real GDP goes up and up until it hit zero at that point it's out of GDP and now what happens is if you take the smartphone smartphone has reduced the sales of cameras of GPS machines of music players of video players of alarm clocks on and on and on so overall in terms of measurement it's probably reduced GDP because of the reduction in sales of all those substitute products and for now at least there's no hedonic adjustment on smartphones so that missing link of incorporating other devices into this general purpose device it's problematic I think that through across the board a whole range of products that are digital Other interventions yes Haraguli University of Chicago so much of the theme in the panel was the interplay between government innovation and much of the theme was how governments maybe can start up innovation but let me bring up another theme of how government innovation interplays maybe that's a question to Joel but also to others and that's the safety by which innovators can finally get the rewards for the innovation so I've done research for example with Kojen and Thomas Philippan Flipson where we investigated the rewards to innovation in the health sector and we have shown that the risk of expropriation afterwards drives up the necessary returns that you have to have in the R&D sector in medical innovation so the risk that Hillary Clinton would have succeeded with the healthcare reform the discussions that we now have about profits in the medical sector being taken away from entrepreneurs means that in order to want to innovate in that sector you have to have the rewards of a high profit there and the effects of that are dramatic right so medical innovation much higher if it hadn't been for these risks of future government interventions the state of the medical knowledge would be better without that innovation what I don't know is how much that mattered overall how much that mattered in the historical perspective how much that mattered in other areas I would imagine matters in other areas as well right I think this reason that Uber was created in the United States not say in Germany because if I was an entrepreneur in Germany you know thinking about creating Uber I would immediately think about the association protesting not allowing me to provide that services to people right so it's an environment where you wouldn't want to innovate or if you think about companies that are becoming very rich and then the EU imposes large fines on you because they are seeing huge profits being made these kinds of things can really throttle innovation in certain areas so I'm wondering how much political stability the political safe environment matters who wants to take that one you want to go first I just have no idea what you're talking about so expropriation like what I mean just give an example because otherwise I'm not sure what you're talking about so Hillary Clinton would have expropriated whose profits so Hillary Clinton for example or socialized so the threat is that the United States introduces let's say European based health system we know that in medical R&D the key the key profits are earned on the US market there are papers by us in Mogul and so forth there's a lot of the profits in some sectors are earned in the United States because that's where the profits can be earned and then other countries can then use these innovations and provide cheap medicine to their people there's often an argument that medical innovations that have been made should then be shipped freely to Africa there are arguments being made that expensive medication should be made available freely to all so there's constantly talk in the United States for example that has maybe the most profit oriented medical system to go and introduce essentially Canadian or European based system and because of this threat the profits that you need to earn on R&D have to be higher and you can show you can calculate that this effect is pretty dramatic there for example if you invest in healthcare there's a lot of evidence that points to this healthcare expropriation risk in the healthcare industry as something that drives up necessary returns and that drives down the speed of R&D in the medical area so the question to the panel is this is the safety that entrepreneurs can count on in collecting profits from the innovations is that critical for innovations or is that not important So I think that's basically also what the patent system tries to address here is to find a good balance between on the one hand making sure that you get enough incentives during the patent protection time for recouping these investments here but of course with the price of after that patent protection to have the disclosure and the diffusion here so I think with the patent protection if we carefully think about what the optimal tradeoff would be in terms of giving enough incentives while at the same time also making sure you have diffusion that's what a good patent system would actually do here now what I would like to say on the patent system is that this tradeoff between rewards and costs here is something that's been done very general across different technologies here so the patent duration that would solve this here and the renewal fees is the same for all different types of technologies here and it's very clear that's completely different in pharma biotech versus in IT here so I would actually argue that we should also with the patent system with the duration and particularly with the renewal fee take more into account the differences between technologies to find a better balance between that appropriation that incentives versus the cost of patenting I see what you're saying but I would almost have put it in reverse which is at least how I see it and you know I might be seeing it in a narrow way from how I see the case of health and pharmaceutical companies they are very attracted to being in the US in particular parts of the US because they benefit massively they also talk about this from for example the National Institutes of Health program which is 32 billion a year it's quite stable it even went up under Reagan by the way the first time in US history where some of these organizations from ARPA E to NIH are facing potentially very very large cuts through the trumpet administration it wasn't true this is they have had actually cross-partisan support up until now so this is a new potential phase in US competitiveness we'll see what happens but then they move to Europe due to these ridiculous tax schemes and by the way the patent box was first introduced in Europe and the reason why the patent box makes no sense is patents are already monopolies for 20 years you don't have to reduce the tax on monopoly profits you have to try to incentivize the hiring if you want to research laborers who actually create that patent but also the corporate income tax as we know is actually much higher in the US than in some European countries so if anything it's sort of gaming the system not this kind of entrepreneurial depiction that you mentioned but gaming the system of where do I get the sort of patient public income tax through things like the NHS where should I then move my activities to benefit from different types of tax schemes and I think lining up you know having proper sort of international laws as I can't remember who was talking about it through the BEPS was at Dietmar is one way to deal with that but the entrepreneurship of the US system it's not about their DNA it's not about sort of entrepreneurs somehow being more risk taking in the US they have also through these different types of policies including as I mentioned VC in the US is very short term right they're exit driven they want their returns to happen in three or five years through a buyout or an IPO what happens before the VCs enter is fascinating they're often freeriding which is fine I don't say that negatively but they come in about 20 years after the biotech investments the nanotech investments the internet investments today the green investments they come in and perhaps the problem is also that their share this 20% maybe it's too high given the actual risks they took that we might kind of look at but this ability of firms in the US to take bigger risks it's not because of their sort of risk lovingness they're unable to do that through these different types of taxpayer guaranteed schemes which we don't necessarily have in Europe in the very very early stage almost at the end I'm not worried about the fact that innovation in the US or anywhere else is under incentivized and therefore that we're providing to a little of it in fact I think the main constraints about on innovation are not on that side they are actually on some of the problems that we're trying to solve are really really difficult one of the main issues that everybody is worrying about is how to improve energy storage right so battery technology is really a big deal and you know there are scores of firms working on this extremely hard and the reason they're not making as much progress as we should is not because I think once they break through the government is going to expropriate their profit that's not going to have but it's a hard problem and you know we need the actual physical issues and chemical issues are really difficult that's what's constraining progress now you do put your finger on an important point and that is there is resistance to technological change because they're always losing right it's never you know a Pareto optimal system and since they're losers and the winners and the losers don't always get compensated that I think is when you ask me where is the role of government it is to worry about the losers and because the losers will be able as or they will at least try to block technological change you know we've seen this in England with the Luddites we now see it with European Uber drivers I mean taxi drivers we see it all over the place government should it's not that they should squash this resistance but they should make it palatable for people to live in an age of progress in which unfortunately not everybody gains so we have a last polling question so can you do you think that in the next 10 years technological innovation will A replace more tasks than create new ones so the net impact on employment will be negative B create more tasks than replace ones so that the net impact on employment will be positive C have an insignificant net impact on employment it's difficult one minute while they're answering that the biggest battery storage invention was through ARPA-E ARPA-E just came out with the biggest battery storage innovation six months ago 10 more seconds let's see let's see the answers we still have one minute one minute and then we close on time we should I will submit it to you the next time nobody ignores the supply yes no no what you mentioned David Ricardo was asking this already and I have a critique of this question as well it should have wage in there there's no supply side here let's wait what about wages that's maybe the conclusion so you will comment what comes out of it last answers okay that's the results you see so no that's so so let's conclude this panel you want to comment on this very quickly 10-20 seconds maximum for each on this maybe you show the results again I have the same critique of the question that I raised before namely what about wages that's the real issue because of course if supply of labor goods down wages go up people may consume some of that in having additional leisure all sorts of things could happen so it's not employment per se but what the interaction of employment or wages it matters good so Mike David Ricardo chapter 31 on machinery he asked that question what actually happened for the next 200 years was that the profits being earned being reinvested in other parts of the economy the big change has not been robots just think of the effect that mass production again had on jobs, wages etc the big change has been financialization from the 80s what is happening to profits not necessarily being reinvested creating that renewal process that's what today is hurting both skills which is an endogenous function of investment and new job creation thank you if you ask somebody in 1914 are you worried about that and then you told them that a century later their great-grandchildren would be cyber security analysts and computer game designers they wouldn't have had the foggiest idea what you're talking about we are going to create new occupations that we don't even dream about anymore than say people at the time of Ricardo could have known that the sons of the handloom weavers who are being thrown out of work would become railroad engineers and telegraph operators I agree there is lots of potential and lots of these ideas which we may not at this time already be able to anticipate but we also need to realize that all these potential effects will take a long term before they will actually be realized so the long term window that we need to take into account for evaluating these investments thank you very welcome very welcome oh thank you