 We are fortunate to have our next speaker, Professor Darren Asimoglu. He is the Institute Professor in the Department of Economics here at MIT. Now he's also associated with the National Bureau of Economic Research, the Center for Economic Performance and the Center for Policy Research, as well as the Microsoft Research Center. He's received numerous awards and fellowships, including the John Bates Clark Medal in 2005 for being and given every two years to the best economists in the United States under the age of 40. So maybe you're 41 right now, I guess. Just about, okay. Just the last thing I'll comment about is that he's a co-author on Why Nations Fail, the Origins of Power, Prosperity and Poverty and 2019 publication called The Narrow Corridor States Societies and the Fate of Liberty. Darren, it's really a pleasure to have you with us and I look forward to your comments and engaging with the audience. Thank you, Jim. Professor Asimoglu. Thank you. Thank you very much. It's a great pleasure to be here. It's, you know, this is the new rhythm for me. I hadn't given in-person talks for, you know, I don't know, 19 months or something. And then this is my third one in the last two weeks. So I think we're getting slowly back into the rhythm. So I'm going to talk about corporate responsibility in the age of automation, inequality and climate change. I was originally going to talk about automation, but then given COP26, I thought, broaden it a little bit. And I want to start by reiterating some discussions that people have been having. You know, the world is going through some tough times and there is increasing sort of recognition that perhaps corporations need to play more of a positive, active role in tackling these societal challenges. The business roundtable called for corporate, corporations to serve all of their stakeholders, especially worried about inequality and some of the social discontent. You know, if you listen to podcasts, the news or other online events, you'll see the Facebook ads that have been calling for new commitment to new internet regulations. Philip Morris is saying that they are committed to making smoking history. ExxonMobil has recently announced a five-year climate plan and on advertising pretty much every minute, so far as I can see. So is this the end of the shareholders' values and a new sort of approach to labor relations, technology, some of the challenges that are facing us? Well, I think hardly. I think current proposals are mostly not tackling the key issues and I think the two key issues really in my mind are inequality, especially tied in with the future of work and automation and that's what I'm going to talk about quite a bit and also climate change. So let me start with inequality and I think the first exhibit that I would like you to think about is this one which shows what has happened to how much private businesses in the U.S. are paying to labor. So this is what's sometimes called the wage bill, the total payments from private businesses to workers. And if you look on the left-hand side, what you will see is that the wage bill was growing very rapidly and very steadily for four decades after World War II. In fact, so rapidly that it exceeded the growth rate of population by almost two-and-a-half percentage points in real terms. And then look at what has happened over the last 30 years and you see a slowdown and then almost a complete flattening. So businesses are not paying much more to workers and I will argue that has really had major implications but it's also rooted in some first-order transformations of technology and the nature of work. Well, let me start with the consequences. Of course this has been associated with a sharp fall in labor share and national income but even more consequently it has been accompanied and caused, I would claim, with a very significant increase in inequality. So here I'm showing the real wages for 10 demographic groups, men and women, and going all the way down from the red curve which is for workers without a high school degree, all the way up to the dark blue which is for workers with a college degree, sorry, workers with a post-college degree, masters, specialized degrees, PhDs and so on. So if you look in the 1950s and 1960s and early 1970s you'll see that all 10 demographic groups were growing in tandem more than about 2% a year annual in real terms. So that was a really rapid growth of wages and it was the basis of American dream to share prosperity in the decades that followed World War II. But look at it, after the 1980s you see these curves going their own separate ways, a huge amount of inequality building up both among men and among women, but even more consequently if you look at the low education groups within those such as men with a high school degree, orange or less than a high school degree or women with the same educational achievements you'll see that their real wages are actually falling significantly. So this is not just lack of shared prosperity, it's really some groups losing out from the process of rapid growth that others are enjoying. Even men with just a college degree would find it surprising they have not experienced much growth. It's really the very well-educated specialized skills that have been valued in the labor market. Now inequality is not just a U.S. phenomenon. This is distinguished, a bad distinction, by being a leader in inequality. It was most unequal by measured by the Gini coefficient but it's the same if you use other measures at the beginning on the middle or early stages of the 1980s and then it has had one of the largest increases but pretty much every OECD country has had an increase in inequality as well. What is unique to the United States is this very sharp declines at the bottom of the wage distribution. There are some in Germany and the UK as well but the U.S. is really ruthless in the way that it has treated its low-skill workers. Now there is another element of commonality between countries and it's actually an important one to note because it's going to start giving us some clues about what's going on. If you look at how the occupation distribution has changed across countries there is a very striking similarity. So in all of these countries that I have plotted here and actually several more for which you can get data from other sources, you will see that these red bars are going negative. Those are the shares of middle-third, middle-class occupations. Those are things like clerical job, back-office job, sales jobs, blue-color operators, construction workers and so on. So actually construction workers wouldn't be here, sorry. Transport workers. So these are the basis of the so-called American dream, the pathways to middle-class well-off existence in the United States in the 50s and the 60s. Well, they've been shrinking in the United States but they've been shrinking pretty much in every country. And one reason why you might suspect there is this common pattern is exactly that those jobs are the ones, were the ones that were at the crosshairs of automation. The office-based middle-class jobs have been automated by office software and specialized software and production jobs have been automated by numerically controlled machinery, dedicated machinery and increasingly over the last 20 years by industrial robots. In fact, my work with Pasqual Restrepo goes quite a bit into this direction and tries to sort of understand what has technology done to inequality, what has technology done to wages and employment. And the most important part of this is that in contrast to many popular discussions and many accounts by economists is that you really don't want to think of technology like a monolith. There is some effect of technology, it increases productivity and then that productivity increases, it might trickle down to labor. I think that's completely the wrong way to think about technology. Technology is a very versatile set of approaches to production, but different approaches and different parts of technology have very different effects. And in particular, you want to distinguish the automation role of technology, which is the substitution of machines and algorithms for tasks previously performed by labor from other types of technologies that increase human productivity or even more importantly, create new tasks for humans. So on the basis of that, this is a macro picture and I'll show you a little bit more micro aspects of it in the next slide. What we do is we look at how much displacement due to automation there has been technologically and how much counterbalancing new tasks and other things that have put workers back into the production process and therefore helped workers also benefit from the productivity increases. So on the left, you have the developments in the four decades following World War II and it's actually a remarkable picture because what it shows is that there is very rapid displacement of workers from turkent tasks and rapid automation but it is almost perfectly counterbalanced by other technologies that create lots of opportunities for workers and in fact workers for different types of skills which I'm not showing in this graph but giving as a background. So much so that the sum of these two curves is the one in the middle, the thick blue one, it's hovering around zero. But now fast forward to the last 30 years you see a very different picture again in line with what I showed you in the first figure that there was a complete slowdown or stagnation of private business payments to labor. Now there is much faster automation, much more displacement and much slower other types of technologies and as a result this blue line now is going south. So this is to a first order approximation responsible for many of the other trends that I'm showing you. Now of course this is a bit abstract, it is the aggregation of many different technologies so let's look at one specific type of technology, industrial robotics. Industrial robots have been a god-send in many US businesses including in the automaking industry in the automobile industry or in metals because they have increased productivity quite significantly and have enabled them to compete against foreign competitors. So with such a technology if you had a view that somehow technology is going to help all of the stakeholders of the firm you should expect that it should be associated with higher wages, higher employment. So here we look at not just at the whole US level but across different local labor markets, different local areas approximated by what are called commuting zones and what you see is that places that have had more robotics technology, more adoption of robots, more exposure to robots have had lower employment and this is led by the industrial heartland of the United States such as Detroit, Lansing, Beaumont, Cleveland and so on but it's not driven by these so if you look at the rest of the areas within the United States you see exactly the same relationship. This is not confined just to employment you see the same thing with wages, wages have fallen significantly in those areas and inequality has increased significantly in those areas. This is of course about the past of automation in the United States, industrial robotics is about industrial blue-collar workers well they're not that many of them left in the United States anymore. The future many people think is going to be shaped by AI when potentially AI is different. In fact you would imagine that AI should be different because AI is a blow-technological platform you can use machine learning, unstructured data and other methods for performing a lot of different tasks and creating a lot of different platforms and different business models, different organizational capabilities. So in principle perhaps AI is going to be quite distinct and create a lot of opportunities. There are many many many I will say several many more enthusiasts of AI who are telling how we are all going to benefit from AI. Well what is the evidence? Well if you look at the data you'll see two very striking phenomena. One is that if you look at where AI technologies are spreading it is precisely in establishments that have a lot of jobs, a lot of occupations and tasks that can be replaced by AI. So here using a variety of different measures we rank establishments and the fourth quartile and the third quartile are the ones in the top occupations that can be automated. In fact they are the ones arriving almost all of the AI adoption. There is almost very little zero adoption in the rest of the establishments in the United States. But even more strikingly if you look at what's going on with hiring in those establishments the ones that are adopting AI have completely slowed down their hiring and the rest are continuing to grow. So the more detailed statistical analysis shows exactly the same thing that AI has been associated with a sharp slow down in hiring in vacancies in employment in these firms. And that I would say is not a necessity, it's actually a choice about how AI is being used at the moment which brings me to my next point. So why are we doing so much automation? Well I would say it's for a variety of reasons. Certainly automation became a focus businesses with increasing Japanese imports in the 1980s and 1990s. Many businesses felt that they had to cut costs and that meant they thought cutting labor costs in order to compete. It became strengthened with imports from China. But also I think particularly important has been the business models of and growing size of big tech. Today it would not be an exaggeration to say that a few companies set the tone of the future with technologies and approaches to technology among US businesses. This is not confined to AI, but it is particularly true for AI where about one out of every $3 spent on AI comes from these large big tech companies. Their model is very much based on algorithmic automations. Substitute algorithms for fallible troublesome, not so intelligent humans or so their managers and their researchers think. This is not just being confined to those companies themselves but has spread throughout almost all of the industries in the United States. Tellingly, if you look at Google today it's much more valuable and plays a much larger role in US GDP than say General Motors did, but its employment is less than one tenth of General Motors at the time. So these are extremely low employment corporations and the technologies that they export when they export it to the rest of the companies in the United States has a similar effect of excessively automating work. But US policy has also fostered this type of excessive automation in my opinion. One important element of this is the tax code. If you look at the tax code it has always treated labor worse than capital. If you install a machine you pay lower taxes than when you hire workers but that gap between labor which is taxed about 25% and especially software and equipment used for automation has actually widened quite significantly today it stands above 20%. So in other words if you lay off your workers and hire a machine to perform exactly the same tactic, exactly the same productivity you make 20% more profits because the government is subsidizing it. But some people will tell you this is the inevitable pre-ordained path of technology this is what technology wants. Everything has to be automated everything has to be algorithmically implemented. And we're actually benefiting from it because we're getting much higher productivity, much cheaper goods and much greater variety. Well I think the evidence for that is scan. It is true if you look at patents they have completely exploded. We have today more patents in the United States, four times more patents in the United States than in the early 1980s but productivity we are going through one of the slowest period of productivity growth in US history. So whatever it is we're not actually getting the fruits of this type of automation. And the post COVID world is set to accelerate this automation. Many companies say they have already automated especially in the hospitality sectors or in the process of automating work because of labor shortages, social distancing concerns and so on. Now I will say that if we go back to the issues of social responsibility, corporate responsibility, you know can we have ethical automation? Is it unethical to automate work? Well absolutely not. If you look at throughout history it has always been part automation has always been part of the broad suite of technologies that employers have used. British Industrial Revolution was started by automation mechanization in the textile industry. Mechanization of agriculture was a big milestone for the US industry. But what has gone wrong is that there has been an excessive focus on automation partly because of an excessive focus on cost cutting and automation has been the only game in town. So in other words it's become such that other technologies have been left aside. And this I would say is not because of the new opportunities or the new feasibilities that are presented by our knowledge. It's very much a choice or a direction of technology is a choice and it's a choice that businesses make but it's a choice made by businesses in a context shaped by governments. And we cannot ignore the regulation of technology in the heyday of rapid productivity growth and rapid share productivity prosperity in the United States in the 50s and the 60s for example this was made in the context of a regulatory framework set up by government institutions. So we have to go back to thinking about what it is the corporate responsibility and how that corporate responsibility actually has to show up itself and in what context it is situated. But before I do that let me come to the other crisis that we're dealing with. The climate crisis but perhaps businesses are working hard to solve climate crisis perhaps ExxonMobil and Chevron are really serious that they're trying to deal with this issue. Well, yes there have been big advances in some dimensions especially in renewable technology but not because of big oil and big energy companies but despite them. So here is the data. If you actually look at the cost of producing energy with solar or biotechnology or wind actually it was prohibitively cost expensive even as late as the late 2000s. And there has been a tremendous decline in the cost of renewable technology today such that in all about 90% of energy tasks we can use renewables cheaply than fossil fuel especially some of the implicit tax subsidies that go to fossil fuel which some people estimate are in the order of about 6 trillion dollars around the world are removed. But this was not thanks to big oil this was for a variety of other reasons that took place outside of the big oil in solar panel manufacturing in both the United States and abroad and because of subsidies that governments provided both in the United States and abroad and the pressure that consumers placed on companies. But in particular the innovation aspect is really important. So here what I'm plotting is the number of renewable patents and green patents that were awarded in the United States relative to either fossil fuel patents or total patents. So you see this huge takeoff none of this very little of this is driven by big oil but there's a problem it actually gets reversed after 2010. So if you want to think about what's the role of the energy sector or traditional energy sector in the United States well actually it's not in this run up but it's this run down. So this run down is very much caused by actually by it's actually international as well but let me since time is running short let me not dwell on the international but you'll see that in Germany, Canada and France the pattern is the same and this run down will slow down in what we are investing in renewables at a time where we really need to depend on renewables has one major simple cause gas prices, shale. So the US sector, energy sector has big time invested in shale and shale prices have fallen and with the fallen shale prices any investment that was going into renewables has halved or more than halved. So in particular the share of gas is increasing in our energy consumption precisely because its prices are falling and if you look here at the two curves that show the first one the natural gas price index and this one is the green fossil fuel electric patterns you see how closely they track each other. When gas became more expensive and that was the energy shortage period that was the real sort of impetus into renewables and as the shale gas boom reduced gas prices investments in renewables ceased. So again you will be hard pressed here to see any corporate responsibility especially from big oil. So what's going on here? Well I think we have to put this into a broader perspective. Jim was very kind to mention my book with James Robinson and in that book one of our second book actually Why Nations Fail we defined a period as a critical juncture if there are major threats and existing institutions prove to be inadequate for dealing with these threats and it is a juncture because it is not clear which path you are going to take. It depends on agency, it depends on choices it depends on our recognition of the challenges I had such as what I argued with inequality, automation and climate change. So indeed I think we are going through a critical juncture and I think we would have probably gone through a critical juncture even had it not been for the pandemic but the pandemic has accelerated and made those processes clear. Throughout history the major barrier to inclusivity has been the excessive power of companies and political elites especially when they are able to use this to control people or to control key assets or via coercion. But I think today we are dealing with somewhat different even though related concerns. Climate change is obviously a global one it's not just about distribution although it has major distributional implications but also automation creates a different type of challenge what we would call different kinds of extractive institutions. It is in particular the labor market has always been the place where most people have earned a livelihood and that's been the place where people have a chance to participate in shared prosperity. What we are seeing with automation is that these opportunities are disappearing for the vast majority of the American people including even those with just a college degree. And in fact if we destroy labor market opportunities I think this will lead to a failure of inclusive institutions failure of democratic institutions more broadly and of course we cannot confront the climate change crisis unless we have some democratic institutions that organize people and aggregate their preferences and transmit their voices to policymakers and to firms. So I think this is the sense in which the automation inequality crisis and the climate crisis are intimately linked. So therefore my view is that we have to really embrace this critical juncture and try to rebuild inclusive institutions. So how can we do that? Well it has to be done both at the political and economic level. That was again one of the main thesis of the why nations fail that economic institutions so for example corporate responsibility what firms do are shared cannot be separated from political arrangements. And this really is in my opinion reinforced in the current crisis because as I have indicated regulation of technology is key. So we have to regulate technology in so that for example we use technology not to improve shale gas and fossil fuel sector but to improve renewables and come up with new better ways of storage and allocation of renewable energy so as to take the first step towards a transition towards cleaner technology. Regulation of technology will play an even more important role in my opinion in dealing with the automation inequality challenges so that the only thing we do with our technological know-how and our collective knowledge is not to automate which comes at the expense of destroying opportunities for the vast majority of the people but we use technology to create new opportunities new tasks, new new feasibilities for American workers but you might say why not let firms themselves use that technology? Well indeed the firms need to be the ones to be at the forefront of the technological developments governments cannot invent technologies but what I have the reason why I have emphasized corporate responsibility or lack thereof is that I don't think you can rely on firms to do the right thing unless there is the right regulatory framework or institutional framework so in some sense corporate responsibility is not going to appear miraculously you have to force it and the institutions that will force that are the regulatory institutions that make sure that corporations have the right incentives and have the right behavior because they are regulated by the institutional structure in some sense what this requires in my mind is something you might call welfare state 3.0 I think the biggest institutional transformation that western world experienced over the last 300 years was the rise of the welfare state starting in the 1930s in Norway and Sweden and then in the UK in the 1940s and then finally in the rest of continental Europe and in the United States partially through Johnson's great society program in particular the welfare state tried to create an institutional structure for regulation both in the labour market and beyond so that some of these shares gains were shared fairly and some of the hardships were not unduly on the shoulders of one small group of people but today we are dealing with other challenges such as inequality, climate change and perhaps global pandemics and so on so that means a different type of welfare state which also as I have highlighted really has to take regulation of technology very seriously put it in the center stage because both the climate change and inequality challenges cannot be tackled without this regulation of technology framework but anytime you seriously want to think about design of new welfare institutions you would be a well advised to take this gentleman seriously this is Frederick von Hayek one of the most important social scientists of the 20th century and at the time when the beverage report the blueprint for the welfare state in the United Kingdom was being articulated by a government led by Lord William Beverage he had just arrived from Austria as an emigre fleeing Nazism and he was teaching in my alma mater the London School of Economics the British people were enthused when Lord Beverage's report was published many described it as a mana from heaven in the darkest hour of the population in the midst of the war and the provisions of the Beverage report started being implemented right in the middle of the war and some people credited even in turning the war around because it made the British public fight for Britain which they were not as keen on doing when it was so unequal and so unfair but Hayek was very worried he thought that when you build a welfare state of this sort you are actually going to empower the state and you are going to destroy liberty so he wrote a memo and then an essay and then a book which became a best author the road to serfdom and I want to end by one last slide from my more recent book with James Robinson the narrow corridor and in fact one interpretation of the theory that we develop in the narrow corridor is an interpretation of why Hayek turned out to be wrong in fact liberty did not get extinguished it flourished as never before in the decades that followed World War II in most of the western world so why is that well this figure sort of summarizes our conceptual framework you really have to think about the balance between the power of the state and elites and power of society for example as experienced as realized through democratic processes or civil society or media and liberty democratic institutions and more inclusive forms of prosperity flourish in what we call this narrow corridor when the state and society are in balance and in this context you can reinterpret Hayek's concern as saying if you make the state so strong with the welfare state and all these new responsibilities on his shoulders you're going to create an imbalance you're going to take societies out of the corridor but in fact what our theory shows and I think it's relevant and that's why I want to end with it is that that's not the only solution there's another solution and that solution is that as you make the power of the state greater by placing greater responsibilities on his shoulders society also becomes empowered to become a better monitor of states and bureaucracies and politics as the deepening of democracy that goes hand in hand with these much more complex and much more necessary institutions so in some sense Hayek turned out to be wrong because he did not give enough credit to democratic institutions and to regular people's ability to work within that political system so I think in this critical juncture it is an imperative for us to prove Hayek wrong again to deal with inequality and climate challenges but for that we cannot just bank on corporate responsibility I think we have to induce it to make regular people democratic process really becoming much stronger and much more attuned to these challenges and taking on its own responsibilities thank you thank you professor Asamoglu you've given us a lot to chew on and thrown out some pretty challenging questions for us my colleague Katie is going to manage this microphone and my colleague Ken is going to manage that first let me offer some questions in the audience we have John Polito up front here Mike down to John and when you speak just state your name and the company you're with so we all can know where you're coming from hi my name is John Polito I work with a company called Raise Holdings but I'm also an alumnus of the CTO program so glad to be back I had a question on your topic of regulation of technology and as it pertains I think to one of the main topics of this forum is around the idea of autonomous vehicles coming down the pipe in industry particularly on over-the-road tracking so in the context of regulating technology from a societal point of view do you have a perspective on how society should address the regulation of autonomous vehicles that will be coming down the road here shortly well thank you very much yes I think unfortunately because of the time shortage I was quite loose on what I mean by regulation of technology and I think it has several elements I think one of the most important elements is you need some sort of government leadership in defining what are priorities so if you look at many of the technologies that became defining including sensors the internet antibiotics you know the private sector played a very important role but so did the government in setting the agenda so I think that is the context in which the regulation of technology has to play a role but then the other aspect is exactly what you've pointed out which is you know how we actually use that technology I think autonomous vehicles are a reality they're going to be here I think some people are too optimistic about how much around the corner they are but it's undoubtedly they're going to be here but the question is how do you actually combine them with other technological adjustments so that we create opportunities you know in the United States today there are I think more than a million track drivers so if you roll out autonomous vehicles for example in tracking right away and lay off one million people first of all it's going to be completely disruptive but secondly it won't actually be very productive it won't be very productive because I think the current technology is certainly not up to the task of being able to do even the majority of the tasks that drivers can do but there are certain simple tasks that trucks can do like interstate driving autonomous trucks you know again we're not there yet but we probably will be there in about three four years but you would know that better can do the very simple form of interstate driving but then they would have to liaise and work together with human drivers so if you do that you're at the same time rearranging tasks in such a way that's actually not disruptive to humans it might actually lead to new tasks for humans that there might be more adjustments and more initiative and more autonomy requiring tasks depending on the current circumstances that drivers can play but you are actually automating some tasks that are some of the hardest ones so I think that is part of the difference between the more sort of healthy automation which goes hand in hand with other adjustment and other productivity improving technological investments versus the more disruptive ones that excessively automate and do so without creating any opportunities for affected workers or a broader industrial workforce of the country thank you for raising that question well can I do a follow up before we go to the next question can I do a follow up on that you said for companies they really need to be thinking about creating opportunities when they do that how do they do that? what does that look like? let me give you one example and I think that's where the complication comes in that may sometimes not be the same company that's doing the automation and providing the opportunities but let's look at the following issue let's think of all of the people that are in this room or actually that you interact with your professional daily life and think of the tasks that they perform during their daily working life my guess is that about 80 to 90% of them are actually performing tasks that did not exist 80 years ago so as a professor you know that occupation existed 80 years ago but almost all of the things that I do today are very different yes I also read books but I also lecture but the lecturing is very different uses very different technologies to a very different audience so that means that my job has changed often as a result of technologies and if it weren't those changes my job would not actually have the capacity to employ so many workers and not have the capacity to play to pay decent wages so those new tasks that we have created both in teaching and in management in supervisory position in design in programming are the lifeblood of the economy and they are part of these technologies that really help workers and the difficulty of course is that some companies are not going to be the ones at the forefront of it and some companies may be doing more automation than others so that is the difficulty and the corporate responsibility you cannot say every company has to do new tasks as well as automation but you have to create the environment where there is enough incentives and I think the onus is much more on researchers so a given company that employs 300 people they may be just so much into automation or some other things but when you look at Google or Facebook, Microsoft those companies they are shaping the future technological trajectory so the responsibility is much more on them that it is a balanced portfolio and do you think that they are creating new tasks? No not enough that is exactly what the data that I showed is that there has been a complete slowdown of new tasks and worker friendly innovations. Is anybody doing that well? Good question, I am not sure I know enough but I can tell you that there are many areas in which it can be done but nobody is doing it like education is one of those areas so almost all of the investment that goes into education in AI or AI education is going towards automated grading, automated teaching, automated homework help but nobody is investing in the low hanging fruit of developing AI technologies that create much more individualized teaching environments which of course would require more teachers to be hired and that is the point that we want to create more opportunities for workers and actually that would be a very valuable sort of employment because it is also quite clear I think to many commentators and people in this room the US education system is failing and is failing exactly for the low socioeconomic status kids and if you talk to teachers they will say well you know we need much more help for these students we just don't have the time and we don't know exactly what problems they are having that is exactly where AI and more teachers can help but nobody is doing that. Interesting, so a very different take on automation, can I think we have a question? Yeah, okay. Hi Dr. Ashmoly, first of all I just finished an error corridor and thought it was tremendous. Thank you, that's very nice of you to say. So I work at the MIT age lab and population aging is always on my mind so I wanted to ask you about the effects of automation in terms of displacing labor in a moment of labor glut and where prices are very low so I'm thinking about your chart and when perhaps Japanese imports were driving down prices and Detroit had to counter versus thinking about the future maybe when labor population is quite a bit older are those conditions going to make the automation process less socially disruptive? Wonderful, that great great question so absolutely you're 100% right if you look at the countries that are at the front of automation, South Korea, Germany and Japan, actually automation doesn't seem to have much of a negative effect in those countries why not? Because those are the three countries that have aged much faster than the rest and they have a real labor shortage so automation is making up for that layer of shortage so automation in those countries is leading to much less displacement because they don't have the workers to work in these tasks in the first place and so that again highlights another really important point you know automation is not bad by in and of itself if we use automation correctly just like it's in the long-haul driving, you know long-haul interstate driving is one of the hardest tasks people are away from their families they develop back aches it's completely debilitating mentally and physically so if we can automate these jobs but provide other jobs that's great so there are many useful things that we can do with automation but if it goes ahead of other things and it creates a problem of what I've called the excessive automation that's when we get the inequality and that's when we get the adverse labor market consequences and I think aging is going to change the landscape and in fact when we have a more aging more aged population in the United States I think we're going to have greater need for technology of automated technologies on the other hand I think it's also a more complex situation because at the same time we are also having the skills and the health conditions of the older cohorts improve as well so many more people are going to be able to do tasks that they weren't able to do or the older cohorts were not able to do so we need another broader adjustment and again balance rather than a focus excessive focus on just cost cutting would be very important in that context please Dave, go ahead Thank you very much for the presentation really enjoyed it one thing that stuck with me the most right away was I take your argument that we need to regulate automation because we need to protect the economic well-being of people who trade labor for wages and when I think about how we think about that issue in the US politically we say yes but the way we do that is we buy American we preference American purchases so if people take your same argument we need to protect these folks but we do it by sort of economic nationalism could you comment on are they both valuable tools to the same problem yes I think that's a great great great question it's a difficult question for economists you know because for economists including myself you know we grew up with the religion of free trade is good you know it was the thing that distinguished you know the most amazing thinkers like Adam Smith and Ricardo from people who didn't understand economics and I think there's a lot of validity to the wonderful ideas that Ricardo and Adam Smith developed but if you look at what happened in the 2000s in the 1990s also by the 2000s with cheap Chinese imports you know work by my colleague David Otter with his colleagues and work that I have done with David Otter and others myself estimates that three million jobs may have been lost in the United States because of Chinese imports in the course of about what 17 years it's a tremendous numbers some communities were completely destroyed because there were so specialized in things like toys and apparel that competed directly with very cheap Chinese imports businesses completely closed endemic unemployment and you know the theory is well perhaps we all benefited because we got cheaper toys that may well be true but that doesn't take into account all the costs that we incur and it also ignores the fact that the redistribution that would be necessary to turn this into a broadly beneficial gain never takes place that was very little redistribution towards these areas over these people who lost their job so I think in hindsight if Chinese imports were more slowly ramped up not completely stopped Chinese imports but it was a more gradual process give time to firms to make technological adjustments because many of them could have actually the more successful firms did upgrade their technologies and were able to compete but only a handful of them so firms were given that opportunity workers were given the opportunity to relocate I think the outcomes would have been much better but if you look at the Trump era economic nationalism I think it's completely useless why because it's like closing the barn door after the horses have left you're not going to be able to bring back the metal working jobs back because they have been automated or they're completely gone the factories are not here anymore so I think you know we have to have a more holistic approach to some of these issues as well but not like pining after what's already being lost thank you we have one minute left and I'll take we have a number of questions online I'll take one of them how do you recommend addressing new jobs from automation that are not typically suited to lower educated and they're particularly talking about the case where the folks who are replaced are not the ones who can do the new resulting jobs and they suggest the possibility of is it through a different approach to vocational programs absolutely 100% the answer to the question is excellent and the answer is in the question so I gave the example of Germany one of the things there are a couple of things different in Germany Germany has adopted three times as many robots as the United States has done per industrial worker so in at least some dimensions of automation they're far ahead of the United States but as I said they've done that because there was a shortage of workers and they did it in the context of an export boom so German machine tools cars were being exported and there was exports grew while they were automating so that's a very different condition again sort of underscoring the sort of the trade aspect but there's another very big difference so the German workforce is extremely well trained because first I think the middle school education is somewhat better than the United States but also because of two years in most industries two years vocational training that's very technical and very rigorous in nature so what German companies seem to have done is that when they automated these jobs and eliminated the blue collar workers they retrained these workers that already had a lot of industry specific skills and they were promoted to non-production jobs that has not happened in the United States and part of the reason is exactly because I think this workers did not have the basic skills to even be retrained but secondly employers were not interested in retraining so it's it takes a general ecosystem of training that I think is absent here so I think it's another area where policy might help but definitely I think the broader issue of automation is hitting the lower skill workers and some of these skills workers if they want to find jobs need to be retrained and that's not something that just you can do in community colleges I think community colleges are very important but it's also partly an employer based training system that is necessary as well. I like the idea this concept of an ecosystem it's not just one piece it's going to be a series of different activities and you outlined the number of them let's give a round of applause for Professor Asimovili.