Capitalism in an Age of Robots with Adair Turner



good afternoon I'm Vali Nasr the Dean of Johns Hopkins University School of Advanced International Studies and it is my great pleasure to welcome everyone to this event today it is also my pleasure to welcome Lord Adair Turner who is the chair of the Institute for new economic thinking to discuss capitalism in an age of robots as part of our Dean's forum evolution or revolution series which focuses on restructuring finance for a new global economy for over 70 years size has served as a platform for policy makers faculty and students to engage on the crucial issues shaping world events today's event continues the tradition of dialogue by bridging opinions of leaders and practitioners in international business trade and Finance with those of academics to share their experience with our students faculty and the broader community Laura Turner has been chair of the Institute for new economic thinking since 2015 and the chair of the energy transitions commission since January 2016 he's also currently the chair of the European subsidiary Board of chop and it's the new executive director of Prudential PLC luer Turner's career has spanned business public policy and academia he chaired the United Kingdom's Financial Services Authority from September 2008 to 2013 he also played a leading role in the post-crisis redesign of global banking and shadow banking regulation as chairman of the International Financial Stability boards major policy committee his expertise in the area led him to write the book between debt and the devil money credit and fixing global finance in 2015 Lord Turner is also a fellow of the United Kingdom's Royal Society and recently received an honorary doctorate from Cambridge University today's event will be moderated by dr. john lipsky the Peter G Peterson distinguished scholar at the Kissinger Center for global affairs and senior fellow at the Foreign Policy Institute here at size we begin with the presentation by luer Turner and then dr. Lipsky will join him for a conversation and question answer please well Valli thank you very much for that introduction it's a great pleasure to be here although I'm quite daunted because I've managed to leave all my notes back at my hotel so there's gonna be there's going to be a little bit of a just trying to remember things as I go along here but I hope it will still go okay my title is capitalism in the age of robots and what I want to consider is what will be the impact on the shape of the economy and the challenges which societies face in a world where I think we are going to automate more and more activities more and more work activities are going to be automated and throughout the lecture I am going to use the word robots as a shorthand for any machine or process that automates a work activities but I just be clear I do not necessarily mean a sort of humanoid looking thing with arms legs and a smiley face that is not how you've got to understand robots in this discussion um it's quite a lengthy lecture that I've got and within a week I will have a written version of this up on the i net website it's not quite ready yet but it will be there and because it's quite a lengthy lecture I thought I would begin by deliberately setting out some of the paradoxes and provocations that follow from what I've going to say so that in the more technical bits of it you still have an incentive to keep going till at the end I am going to assert and I don't know how easy it is at the back to see this I'm going to that the following propositions are true that the forced of the underlying pace of technological advance the lower the measured productivity growth rate will be the automation threatens income from activities which are really essential to human welfare but not income generated from purely zero-sum competitive activities that the more rapidly information and communication technology progresses the more that wealth and income will accrue to some of the most physical things of all such as land or sporting prowess or physical beauty that increasing productivity growth is not a priority for public policy in the rich developed world that skills are not the all-purpose answer to inequality and that aging and low fertility are no particular problem I wouldn't worry too much at all about the aging of Japan or China where Isaac would be worried stiff about the rapid demographic growth of Africa in particular so I hope I've stirred you up enough with some deliberate count as to what I think are many conventional wisdoms I'm going to set out my argument or my lecture sets out the argument in six elements first of all I'm going to say that everything is going to be eventually automated that as relation to automation the question is when not if secondly I'm going to take what might seem like a bit of a detour which says but how do we then deal with the solo paradox which is computers are everywhere except in the productivity statistics if we're automating more and more of the economy why do we see such low measured productivity growth I'm then going to talk about ok if you lived in a world which was totally automated what would GDP mean and my answer is pretty much nothing and where would welfare and wealth and income reside and then going to talk about the challenges for the advanced economies and I can say the fundamental challenge is probably literally rising inequality then I'm going to talk about challenges for emerging economies and the fundamental problem of what I told the old ladder of development destroyed and end with some implications for economic theory but depending how we're doing on time I might finish at the end of section 4 leaving you with an incentive therefore to get hold of the written copy and look at sections 5 and 6 but very happy to take questions on that when we get to the discussion with John so let's begin with the assertion that automation is going to eventually the information and communications technology will ultimately result in the automation of pretty much all work activities now whenever one talks about information and communication technology and says it's going to transform the world somebody pops up and says yes but there's been waves of technological change before there's nothing new about information communication technology which wasn't exactly the same for railways or electricity or motorcar but I think there is a reasonable case set out for instance by what I think is a very good book which is Eric brew Nelson's and Andrew McAfee is second Machine Age that information and communication technology is not just another technology but something quite unique in economic history and I think there are at least these five things that make it unique first is the extraordinary process of hardware productivity improvement in the guts of the hardware system according to the famous Moore's Law named after Gordon Moore of Intel every two years or so indeed he originally said or some people he originally said every two years and some people worked it out that it was every 18 months but let's say 18 months to two years we get twice as much computing power for the same price we get twice as much bandwidth for the same price we get twice as much memory for the same price and there's an extraordinary thing about exponential growth which is how big it gets if things double every two years in 60 years they increase a billion times if they double every 18 months in 50 years they double five billion times and then in the next 50 years they double they increase another five billion times so did in a century they increased by five billion squared which I've got it right is 25 times 10 to the power 18 which is certainly a very very big number and what this means is that we all have in our pockets mobile phones with massively more computing power than NASA used to put a man on the phone massively more than a supercomputer produced by crazy 30 years ago which would have taken this entire room but you ain't seen nothing yet because in another 50 years we're gonna have a billion times as much computing power or memory or bandwidth available to us in the next thing which is unique is software software is amazing in the sense of once you've got one copy of software the next 10 copies a hundred copies a million copies a billion copies don't cost you anything there is a zero cost of replication and that's not how the world of the electromechanical age work thirdly there is the development of artificial intelligence and here the book which I would recommend is a book by man called Nick Bostrom called super intelligence he basically argues that it is inevitable as computing power develops and as people try and help the development of artificial intelligence that we will eventually create artificial intelligence which is equally good as human to intelligence but also inevitable that the moment we get to that pace point we will accelerate the progress because then we'll have something as clever as human beings trying to develop it further but human beings that never get old never get tired that can work 24 hours a day improving the intelligence still further and so that a super intelligence is inevitable big data analysis is hugely powerful the bigger the ability to take huge amounts of data and work out what it means work out patterns that we don't know that we are even looking for and machine learning before I really understood how robotics worked I used to assume that if you were to automate somebody sewing a piece of apparel you'd have to go through the incredibly tedious process of observing exactly what they're doing and writing the code to do that you don't do that at all what you do is you take the fingers of a computer through that process and as it does that it writes its own code the code is written by the robot learning the physical process and I think if you take all of those I think it is a reasonable belief that we will eventually be able to automate a way pretty much everything that we now count as work but of course that still leaves the question as to how fast does this occur and what activities in work get automated earlier and later in the progress and here this I think a very good piece of work by the McKinsey Global Institute called the future of work which has tried to get to quite a close level of granularity in this analysis what they start with and you're really not going to be able to see this at the back of the room because I can't see it at the back of that room is a process of trying to take really granular capabilities of human beings and work out where we are on a sort of red orange green in terms of can we yet do that what is the capability of a computer so sensory perception is moderate recognizing known patterns is far advanced gross motor skills that means the ability of a machine to pick up a pallet and move it around in Amazon work a workshop and put it in a correct place that's far advanced fine motor skills you know sewing something on apparel that's less advanced if you take that as the snapshot of where we are at the moment you can then work out which type of activities are more automatable than others and what this chart sets out is McKinsey shot at how many of the hours performed in these different categories could be automated and what are the number of hours of this type of activity in the US economy so that reading from the top what you have is predictable physical repetitive of physical activities 50% of that we could automate tomorrow with the technology we've got and that accounts for about 18 percent of all hours worked in the American economy are essentially repetitive physical activities doing the same physical activity again and again processing data about sorry knows 81 percent of that could be replaced sorry 81 percent of that could be replaced and 18 percent of the time in u.s. occupations it deals with that processing data 69 percent could be automated and that accounts for 16% and then down at the other end you've got management activities whatever precisely those are but only nine percent of those magically complicated managing activities can be automated by computers today but they only account for 7% of all the hours and once you've got that pattern you can then move that through to actual occupations and what's interesting then is you get a hierarchy of the things which are highly automated and the things which are less automatable so essentially they're saying that pretty soon sewing machine operators graders and sorters of agricultural products will be highly automated all these are these are categories defined within the US Bureau of Labor Statistics definition of the Occupational balance of the US population whereas at the other end psychiatrists and perhaps unfortunately legislators cannot yet be automated apparently they're doing something of such unique human value that they cannot be replaced and then once you've got that pattern you can roll it through – what that means for scepters high ultimate ability of accommodation and food services manufacturing and then down at the other end so far less automate ability bit quite a lot of Health and Social Services professional certain law such as law etc etc so this I think is a very useful way to get us into the granularity of the debate now what of course this doesn't tell us is when this automation is going to occur but again McKinsey I think a useful shot at it and they're usefully distinguished to time scales which are relevant the time scale in blue is the time scale over which it would be physically possible to automate away a different proportion of all existing work activities and they reckon that today it would be physically possible to automate away 50% of work activities and that by about 20 50 or 60 we will definitely in physical possibility have reached a hundred percent or ninety five percent or so but they have a range of time periods over which that happen where the technological optimists say it will all be done by 2035 and the technological pessimists say well no it won't really occur till 2060 but the range of the people who've thought about this is is it 2030 or 2060 its when not if and then the green bit is well when will it occur when you allow for the fact that it always takes time for a possible technology to diffuse across the economy it always is determined by what is the relative cost of buying a robot versus employing a human being and the relative cost of capital and labor so the fact that something's physically possible doesn't mean that it's economically optimal and they're short of when that occurs is that the current work activities and here you start at zero because you're starting with things which haven't yet been automated that they could be fully automated under some scenarios by sixty years time but if not that they will definitely get automated by a hundred years time so the assertion I'm going to make is pretty much everything that we currently call work where we get up in the morning and do something for which we get paid could be replicated by some category of machine and the question is when not if and it feels like it's more likely to be between fifty and a hundred years time than three hundred years or five hundred years time and the question I want to ask is suppose that is true suppose that that's where we're heading fifty to a hundred years time and suppose that with steadily heading in that direction so that we get sort of a bit closer to that over time what follows for the nature of our economy for the degree of inequality for the policies that we need to deal with that changing environment however before we get to that I have to take my detour and my detour is to deal with this question of the solo paradox because of course you can always say after my sort of last ten minutes of sort of techno optimism yes but a dare if we if information and communications technology is progressing so rapidly why don't we see it in the productivity statistics and economic historians such as Robert Gordon have pointed out that the last what it is 40 years or so since we really had computers have not been a glory period of rapid productivity growth the great glory period of rapid productivity growth in the US economy and in the European economies was sort of between the 1940s and the 1960s and it was the great era of electromechanical machinery and the car factory and you know chemicals processing and things like that so he's argued that you know there's a slowdown and of course it's now 40 years ago that Robert Solow a great economist in 1987 30 years ago made this great comment you can see the computer age everywhere except in the productivity of Statistics he began to notice that everybody was babbling on about computers but you know a technology that the productivity rate wasn't going up and indeed over the last 10 years this paradox seems to cut almost more extreme because since the global financial crisis of 20 2008 we've had very slow productivity growth in many advanced economies we seem to have had an even bigger slowdown so what is going on and does this a fact of this a slow productivity growth suggest that a I'm wrong or at least that it isn't happening yet in terms of this automation well what I'm going to argue is that the Solo's paradox is inevitable because of three effects and these three effects I'm going to suggest are not just a detour in the sense of that to convince you of my argument I have to deal with the solo paradox but that once we understand why these three effects explain the solo paradox you will understand something quite deep about the way that economies evolve in the face of rapid productivity growth these three effects I call boom all effects and the high-tech high-touch paradox the zero-sum paradox and nil or low cost benefits missing from GDP now the boom all effect is named after William Bommel and a very famous article I think it was 1977 the economics of unbalanced growth and to understand that I think it's useful to do a little bit of basic economic thinking about what our mental models are and what they can tell us I'm going to suggest that most of us when we start thinking about productivity growth have a sort of standard paradigm in mind and the paradigm is the shift from agriculture to factory so you start with a hundred self-sufficient farmers on the prairies of America in their little log cabins and we work out how to have technological process they're all producing 100 units of food a know each one of them is producing one unit of food and they work out how to have only 50 farmers produce a hundred units of food so we double productivity but the 50 workers who are freed up from doing that they move off to factories where they produce a hundred units of cars and washing machines and televisions etc and so the product of the economy has now gone to two hundred four hundred people so measured productivity has doubled and in the standard model that we tend to have in our brains this is an endlessly repeatable process because in the next era we can have twenty five farmers are producing a hundred units of food we can have fifty factory workers but you 200 cars washing machines and televisions we can have 15 factory workers now producing 60 units of computers mobile phones and software applications that we didn't have before and then we can have 10 service workers producing 40 units of healthcare and then in the next period we can repeat it again and again and again and there's no just by to be clear there's no importance of a distinction between manufacturing or service here services are as much a automatable as manufacturing and it's an endlessly repeatable period in which in every period in this simple model we double the productivity growth and one thing I want to say is that despite what I'm going to say later there's an element of truth about this and as part of what's going on in our economy is like this but it's also important to realize that theoretically something quite different could happen suppose when we started with a hundred farmers producing a hundred units of food we ended up with 50 farmers producing a hundred units of food and we simply didn't know how to produce washing machines or cars or the farmers who now have the money from the productive farms have no interest in cars or washing machines what they might have simply done is employ the 15 non farmers as domestic servants and those domestic servants might have received a lower rate of pay they might have received no increase in real wage their product per person might be half that of the farmers agricultural productivity would double and total economy productivity would increase over 50% and indeed in this environment you don't have an endlessly repeatable progress because what actually happens is that it asymptotes to zero i if each time you get an improvement in the productivity of the farmers eventually you get one farmer producing 100 units of food with 99 servants looking after him and at that point unless you can fractionalize human beings productivity you know has come to a complete halt and it's slowly asymptotes mathematically heads towards 0 over 2 anything what's interesting is this may capture Oh know what one thing I want to add before I say this what I've shown is this with relatively low paid domestic servants but this effect can also work with some quite high paid non farmers as well the fifty farmers could produce a hundred units of food they can have forty five domestic servants paid half as much of them but you could have five artists and singers and priests at the temple and other high falutin are highly paid activities which however were themselves not subject to productivity improvement over time so the non productivity improving sectors of the economy could be high paid or low paid and you'd still get the productivity asymptotes to zero and the thing I want to suggest is that this may not be what a bad description of what happened in the first agricultural revolution which occurs in Assyria and Mesopotamia in the Fertile Crescent of what is now Iraq in about 8,000 years ago we achieved a breakthrough in agricultural productivity we as we move from hunter-gatherers to settled agriculturalists but a few centuries to millennium later most people were no better off because all of the surplus of that agricultural revolution went into the hands of an upper class who then spent a lot of money on domestic servants or on people building pyramids or on priests at the temple and artists and people who provided those sorts of services to them now we tend to believe that that sort of spurt of productivity and then it ceases won't occur in our modern economy and we're probably partly right as that because what we've created is a set of institutions of scientific R&D and innovation and research and entrepreneurship which means that the moment we've automated one thing or improve productivity and one thing we start searching for what we can do elsewhere and so it's highly likely that we won't just come to a halt in the way that we did in that first agricultural revolution but the point I want to make is that the balance between the standard model and this Bommel effect model does not must depend on the automation potential in the newly emerging economic activities it doesn't just depend on whether we're clever enough to work out how to automate new activities it also depends on two other things the impact of the productivity increase on income distribution which is in part determined by asset ownership and if a large part of the increased income from productivity accrues to already rich people who've already got as many cars and washing machines of the roof they want but who have a limitless desire to give parties at which people serve for them you may get a proliferation of low paid service jobs rather than a further increase in productivity so it depends on the income distribution and it depends on the consumption choices of the winners from the initial productivity spurt so basically what I'm arguing is that it's quite possible that you can have incredibly rapid productivity growth in one sector of the economy and what we see is a proliferation of jobs in low paid service activities and it is that which brings down the average observed productivity growth rate even though in the productivity improving sectors of the economy they're improving even more than before now are we seeing this effect well I think we are seeing this effect I would like you now to look at this extraordinary example of 21st century technology in London this is what is called a delivery driver and he is willing to or he or she is willing to delivery your pizza using that incredible state-of-the-art 21st century piece of technology called the bicycle and this person is paid very very little money based basically above the minimum wage and those sort of jobs are plural rating they're proliferating as we can see from the Bureau of Labor Statistics which I think is the best source of labor statistics that I know in the rich developed world we're anywhere in the world and one of the things the Bureau does is every two years it doesn't look back at where jobs have come from in the economy and it lists the ten biggest growth categories of jobs and it forecasts what will be the ten biggest growth categories looking forward over the next ten years and it's been pretty accurate in the past and I want you to notice what the growth categories are in this incredibly high-tech age in which everybody becoming a computer programmer they are personal care aides home health aides food preparation and servicing workers customer service reps cooks restaurants and construction laborers and every one of those is paid well less than the median annual wage and you have to go down to number 14 before you get people like software developers software developers and application developers I think we are seeing a proliferation of these relatively low paid and at least initially difficult to automate jobs and indeed we're seeing it even in environments where labor is incredibly cheap this caught my eye recently which was a article in The Economist magazine where they were talking to a manager in a tea packing Factory and they just got the arrival of a crate with a piece of automatic packing machinery and the manager pointed at the factory and he said well his job will go and his over there and that one's too but the manager then insists that as in the past he'll someone have found jobs for everybody as drivers or Watchmen or people who you know tidy up the grass outside so you get this amazing phenomenon of the automation of some activities and the proliferation of low-wage jobs what we also get oops sorry is there oh and that gives us what I call the high-tech high-touch paradox the more rapidly technological progress enables automation of existing activities the more high-touch jobs will grow in activities which either intrinsically cannot be automated because you literally need face-to-face service or at least for now cannot be automated or where wages just keep so low that there isn't a desire to automate them even though they could be automated but that's where jobs proliferate and more of them proliferate the faster is the growth rate of productivity in the two sectors of the economy the other thing that proliferate s– is zero-sum activities let me explain what I mean by that by going back to my simple model suppose we started with a hundred farmers who produced a hundred units of food and then we ended up with 50 farmers who produced a hundred food and 25 of the surplus went off to be criminals and 25 were employed by the farmers as police to defend them against the criminals now I think we can all agree that there is no increase in human welfare here I mean we need the police in order for not to be a negative increase in human welfare but you know this there's no clear increase in human welfare I mean I mean most people don't go out in a morning and say what I'd like to consume some criminal services today or indeed some police services and what happens here in productivity terms well it's quite interesting what happens here in measured productivity is entirely dependent on our conventions for measuring GDP and as it happens we have a convention that when we have things like the army or the police we put it into GDP at the cost at which we pay for it it is priced at a input price but we don't put criminal services in GDP so in this case GDP has increased by 25 percent rather than anything more but it's increased by 25 percent even though there is no increase in human welfare or consumption and what this tells us is that we have to come to grips with two separate questions about any proliferation of zero-sum activities the first is when you get technological progress and you get more zero-sum activities like criminals and police does total measured productivity increase in the answer is yes but not as so much as in the standard model because that depends on our accounting conventions and does that is there an increase in world with human welfare No now you might say yeah okay that there's an increase in these criminals and police but does this really happen in the real economy well what I want to suggest is that quite remarkably large amount of activities and a lot of pretty highly paid activities or any sense zero sum on their impact on human welfare at the moment there are a lot of cyber criminals out there and there are a hell of a lot and I deal with them in my business career have incredibly well-paid and devoted cyber experts who are defending all our companies against these cyber criminals we have a lot of bad selling practices out there and then we have a lot of financial regulators to stop those bad selling practices and then we have compliance officers and then we have compensation lawyers and then think about all legal services think about divorce lawyers I mean suppose over time divorce lawyers got cleverer and cleverer and cleverer would that produce an increase in human welfare well presumably not because there'd be a cleverer divorce lawyer on both sides of the case so they're extra cleverness would you know wipe itself out much of financial trading and complex financial intermediation I happen to believe is a zero-sum activity servicing the purchase and sale of already existing real estate necessary function but it doesn't increase the amount of real estate and a lot of people money get there from that indeed some educational services I think are a form of zero market signaling I you have to spend a lot of money to go to a university in order to signal that you went to a good university and that's as much as important as what you're learning there and then think about the whole think here in Washington about the whole of politics and elections and lobby groups and think tanks and even academic economists arguing in those think tanks for one side or another what they're doing is putting forward alternative points of view as to what our public policy should be how we should decide up the economic cake but they are not increasing the size of the economic cake and the cleverer they get and the more computers that they employ to do evermore sophisticated Facebook data based analysis does not increase human welfare it increases the intensity of competition for the distribution of human welfare and I would say that this is even true of some things which are undoubtedly a you know forms of human creativity if you think of the whole of fashion design and branding type activities I'd rather live in a world in which people are able to come up with new fashions new designs that sort of interesting area or halfway between art and commercialism it's an expression of human creativity which contributes to human welfare in the sense of I'd rather have an environment where there are fashions out there than everybody wearing sort of Mao type boiler suits and having to but think about it is that competition gonna make us in 2050 any happier than we are at the moment the answer is no there just be somebody different winning the fight to attract our attention in 2050 it will change each year but over time there is no improvement in human welfare so I suspect that there's really quite a lot of these zero-sum activities in the economy and whether they're not in measured GDP is arbitrary accounting conventions divorce lawyers more clever or higher paid divorce lawyers GDP goes up because divorce lawyers are paid for by consumers more highly paid corporate lawyers protecting intellectual property rights or attacking somebody else's intellectual property rights or they're not in GDP because they're an intermediate service and intermediate services are not in GDP so we have a whole load of stuff which doesn't increase human welfare with a completely arbitrary divide by whether it's in GDP or not we also have I think the interesting phenomenon that when we automate zero-sum services we don't automate them completely what I mean by that is on this chart if we go back to this chart that I showed earlier about the automation potential by occupation if we apply massive computer power to enable us to completely automate sewing machine operators there will be no sewing machine operators left because essentially we had all of our clothes will be produced by robots if we did give to lawyers ever more powerful data analytic tools so that they can before they go into court know every single res relevant precedent anywhere in the world they'll still turn up in court to argue against one another and I would suggest that this phenomenon will have no impact on their relative earnings I think that automation threatens the livelihood of the people who are truly producing the things that we absolutely need and not the livelihood of the people involved in a zero-sum activity which I could sum up by the following fact computers now for 20 years have been able very very easily to beat even the very best grandmaster in chess in the world but the world's top chess players now earn over 1 million dollars per year and if in 2050 robots can routinely beat Manchester United I don't think that's going to make any difference whatsoever to what we play top soccer players zero-sum activites competitive game activities are much less susceptible to a or tomato the things we really need in leaf so we'll have another paradox rapid technological progress could eventually automate away almost all of the activities which are truly essential for human welfare while supporting increased intensity of zero-sum competition for relative income status so that zero I some activities account for an increasing percentage of employment and measured output over time in the world where robots do everything will all still have jobs but they are all fundamentally be zero-sum competitive jobs for how we divide up the output of the robots rather than stuff which is actually producing welfare in itself and this suggests an interesting thing that one thing that keeps the rate of productivity down is that we find things to do 80 years ago almost a very clever man John Maynard Keynes in a beautiful short article economic possibilities for our grandchildren gave a rather similar account of the future to one I've giving now and said that in two or three generations time I about now it would be quite possible for us all to work only 15 hours a week but to enjoy a cornucopia of goods and services which were unimaginable in 1929 and of course it hasn't come about you know we haven't got that amount of leisure but what I wanted to suggest is that it could have come about because the productivity has been there that if people had a higher ledger preference and had decided to only work 15 hours a week and if the distribution of income was somewhat more even so that everybody could enjoy a good standard of living with 15 hours a week I think we would probably be now be producing most of the goods and services which we really need for human welfare on 15 hours a week and we would have a far faster rate of measured productivity growth but what we would be doing what we have done is we found things to do we have created all sorts of ways to compete with one another in elections and financial trading in legal services competing who's going to get what share of this cornucopia from productivity improvement we've got individuals who can't choose to work for 15 hours a week because they're not paid enough to do well enough out of 15 hours a week we've got the fact that work is a social activity and that we've got the fact that people invent creative activities per se we find things to do but we could have had if we wanted to or if society had been differently organized we could have had most of the things which amount to the quality of life of people today most of what's in your houses most of what goes on at a restaurant meal the food you eat we could have had that with a far lower labor input now so far I seem to been telling a pretty depressing story in the sense of you know we get this amazing technological advance but we find things to do we have lower measured activity we do lots of things which are of no particular value to human welfare they're just like sort of poker games between one another but the third effect you might be a little bit more up to about because the other thing that could be going on in the measured figures is that we simply under measure some of the benefits that productivity gives us because some of the men if it measures that productivity gives us come from the collapsing price of things and we probably don't well measure the collapse of pricing things in our GDP measures so imagine that sometime in the next hundred years a very small group of very clever scientists create a wonder drug and we all live to a hundred very healthily you'll die at a hundred that's fine but up to a hundred we have we will have very good jobs no more Alzheimer's no early deaths from children no middle-aged deaths from heart attacks you know I'll even say you know everybody's still playing tennis up to 98 undoubtedly a massive improvement in human welfare how would it show up in measures of GDP well how it shows up in measures of nominal GDP depends on how we organize the drug delivery process if we organize it in the standard way that America organizes it with private development and then a period of prate and protection you have a period of research and development where whether or not the expenditure there is in GDP depends upon rules on capitalization of research and capital development and they're a little bit fuzzy but that's not the key thing you then get the patent protection period where this stuff is being sold at a high price and quite a big chunk of revenue enters nominal GDP at that period but when you then get the end of the patent period the price collapses the contribution of this to nominal GDP collapses and in the world that I'm imagining of robots this manufacturing process is so automated that steadily over time the production of this wonder drug becomes a miniscule eventually an almost zero part of nominal GDP by 2100 this incredibly important thing for human welfare if you look at the nominal GDP figures it's absolutely my new because we're paying almost nothing for this drug how does it stay up in real GDP well what whether it turns up in real GDP depends exactly what the national income accountants do at the end of that patent protection period if they realize that there has been a massive fall in the price of a very valuable human welfare benefit and they put that in to the measures of the GDP deflator with which they convert trends in nominal GDP into real GDP then we will have by the end of the century quite a lot of real GDP growth but what if they get that wrong what if they just don't adequately allow for falling prices and the answer is that you then something which is hugely beneficial may not show up in GDP to the extent that it should and there is a very strong argument that we underestimate the productivity and real income growth which is coming from new drugs mobile phones and tablets streamed films and music computer games social network all sorts of things that we enjoy for free and indeed Marty Feldstein has spent some time looking at the way that the statisticians look allow for price falls in these categories are of activity and his conclusion is the result is that the increase in real incomes is underestimated and that the common concern about what appears to be the low growth of average household incomes is misplaced since these low growth estimates fail to reflect the innovations in everything from health care to Internet services to videotape attainment entertainment which have made life better in this these recent years now are they broadly speaking share Feldstein belief that we are probably under measuring productivity growth in some of these sectors of the economy where I'm a little bit less willing to accept Marty's point of view is that phrase at the element which have made life better during these years because I think we have to realize that there are two separate issues raised by Feldstein analysis two separate questions the first question is dunno productivity growth estimates fail to reflect super rapid productivity growth falling prices increasing quality and innovation and specific products and services and my belief is the answer is almost certainly yes and that's part of why we end up with the solar paradox but if you ask a second question does this mean that human welfare improvement has been understated I think it depends on different areas I think if we're understating the productivity improvement in the development of new life-saving drugs then the answer is definitely yes but is it the case that a world where children have endless evermore sophisticated computer games more and more sophistication for the same price each year mobile phones on all the time social networks on all the time is that a better standard of living for children and the one that they enjoyed twenty years ago it's not obvious to me that that is the case and I return to the fact that you have to ask two questions one is productivity measures reflecting productivity underlying technological productivity growth and two is improved GDP appropriately adjusted actually delivering greater happiness and utility and the answer is that could be slightly different answers to those questions so my overall belief is that when you add up the Bommel effects and the high-tech high-touch paradox the zero-sum paradox and the nil or low-cost missing benefits from GDP you can explain the whole of the solo paradox and in the formal version of the paper which I put on the web I have a little appendix which does a little piece of calculation which will set out you know little parameters from a model which illustrates that you could have an environment where underlying productivity growth is actually quite high and growing from two and a quarter to two and a half percent but where measured productivity growth will be 1.7 percent and going down so I don't think there is a solo paradox in the sense of that we're not we haven't really got productivity growth I think what we've got is dramatic productivity growth in some sectors of the economy either in some cases undervalued or by the proliferation of Bommel type activities or zero-sum activities now what does that mean for the shape of the economy in the long term what does the economy look like does an economy even exist let me make a few points about this the first is to return to the point I just made that we have this standard assumption inherent in our economic thinking the technological advance drives productivity improvement all the way across the economy which shows up efficiently in GDP measures of output per hour worked and per capita and that per capita then produces a good measure of improvements of human welfare and what I want to suggest is that up till 20 years ago both of those steps in the logic were imperfect but they were sort of good enough that when we were moving from factories farms to factories broadly speaking productivity growth was showing up in GDP and as we were moving from $1,000 per capita to $10,000 per capita as is happening in many parts of the developing world today and people are for the first time getting you know piped sewage systems and washing machines and dishwashers and cars and the ability to have a paid holiday at the beach that in that period of time increase in GDP per capita is a pretty definitive increase in human welfare but that increasingly over time this relationship becomes both of these relationships become more imperfect in the face of information technology and the proliferation of zero-sum activities and becomes more imperfect as incomes rise and basic needs get satiated what does that then mean for what the economy means in the far future let's imagine a sort of end point of what I've been suggesting this was in 2100 we have solar powered robots guided by artificial intelligence you know which are basically able to do all the world which we currently call work but they can do everything I mean they can mow your lawn they can clean your shoes they can empty your dishwasher they can you know all medical services are fundamentally better done by robots without these dangerous people by called doctors getting in the way of making medical mistakes but this work accounts for a very small proportion of measured GDP while almost all human work activity is devoted to zero-sum competitive activities or to artistic and design activities you know people in you know nice cheap locations you know making crafts gins and you know works of art and doing you know organic farming and in local urban farms and other nice things like that and then there are some low wage activities which are just uh neck anomic to automate because people have to live though so they offer themselves at very low wages of pay I suspect says yes that in this environment the growth or not of GDP per capita measured on current conventions will tell us next to nothing about what is happening to the dynamics of human welfare in that environment it will have just become steadily over the course of this century a imperfect but still sort of tangentially useful measure to something which is completely and utterly useless where then will the economy reside economics is about the allocation of scarce resources and consumption and production so if we haven't got scarcity because the robots are doing everything for us do we have an even an economy well I think we still do have an an economy in the sense of there are money incomes there is wealth denominated in money there are some people have more money than others and that determines who gets what share of the goods and services but what do I think that monetary economy will be dominated by I think it will do dominated by real property ownership and rents I think land and real property will be worth a lot because everybody who has money income however derived given that we've got a sort of limitless cornucopia of washing machines and food and clothes will will devote a very large percentage of our income to competing with one another to live in the nice parts of and we will drive up the prices of the nice parts of town I think there'll be a lot of income will be from intellectual property rents that one of course depend on the legal convention that we apply but a very small number of people in the course of the 21st century have been able to make those wonder drugs or to you know to program the robots to produce all of this stuff if they have managed in that process to a certain intellectual property right they'll be sitting on enormous dreams of IPR rents there's probably be a lot of value in subjective brand values and rents people just have the ability to make clothes look a particular way or to convince you that perfume X is just the one you want to buy those people will be paid a lot of money and there'll be a very high incomes of a very small number of people skilled or lucky in IT in subjective value creation or zero-sum competition and will also probably pay enormous amounts of money for inherently physical skills like the ability to kick a football around or to be a really beautiful supermodel and this is the irony of it this income distribution will determine the distribution of the consumption of scarce positional goods or status Goods and I think we're in a funny way heading to this already and let me in particular illustrate it in relation to this issue of land if you pick up UK national income accounts and look at the big about UK wealth national assets is rather interesting phenomenon being there for the last eighteen years which is that what the national income and accounts called non produced assets non-producing far faster than produced assets produced wealth what are these non produced wealth well they are essentially 99% land here is the increase in the value of UK housing which sits on land and the urban land on which the housing sits and I can attest to this because I am the very very lucky beneficiary of this effect in 1994 I bought a house a relatively small house that would sit on about a quarter or less of this room in the middle of London and it is now worth 6 times as much and I can tell you the fact it's worth 6 times as much has got nothing to do with me putting in some central heating or you know a new piece of double glazing it's entirely a dramatic increase in the value of the land and this isn't just the UK where this has happened if you look at thomas piketty's figures now i'm sure that all of you have heard of thomas piketty's capital in the 21st century i'm sure that some of you have a copy of thomas piketty's capital in the 21st century i think some of you but a suspect of ferret is his small number will have read all 800 pages of thomas piketty's capital in the 21st century but i have i've read it all and what is really interesting when you look at thomas piketty's figures his famous you charge for capital in France or or UK or anywhere else and he points out this amazing movement in the wealth to income ratio what Paquette II himself does not stress is that pretty much all of that increase in wealth to income resides in the blue area there which is land so we have this extraordinary paradox that in a world where we talk about things becoming distance less and you know it's all IT it's all a weightless economy the thing which is growing in value more than anything else is the most physical thing of all which is land in desired allocations so I have another paradox which is the more rapid the progress of information and communications technology the more that value is placed on inherently physical assets and attributes desirable land sporting capability physical beauty or created subjective values you know clothing brands you know crafts gin brands you know things where somebody just managed to convince you that there's something amazing about the provenance of that for which you ought to pay a hugely higher amount of money so if all of that is true and I will finish with that I won't deal with the last two what are the challenges in public policy for the advanced economies well I think the fundamental challenge is inequality I think if we go back to the fact that we're going to automate away activity we are going to automate away a lot of activities and we can automate away quite a lot of middle-income and low-income activities and that's going to depress the income of those people even more now some people faced with this say the problem is going to be there won't be enough jobs to go around I don't believe that because I think we can always create new jobs as there are some very bad arguments about how we can always create new jobs when you get a businessman involved in this debate they will say completely irrelevant things like well my company is in IT but I've created 100 new jobs so I'm creating more jobs well it's just a nonsense argument because it's not taking account of the fact that in the whole sector that they're involved in their IT intensive business is getting rid of other people that's a bad argument but the good argument is there's no limit to the number of jobs we can make if that we can have if we have flexible labor markets and if there's no other source of income because people will have to offer themselves at a low enough wage that you'll say ok mow my lawn you know delivering my pizza you know do something for me they'll find jobs somewhere there's no limit but the issue is the income distribution and as we know in the rich developed world two different degrees in every country most dramatically in the US but also elsewhere we have seen these very significant increases in inequality over the last thirty years and we've seen a phenomenon both at the top end of the income distribution and at the bottom end of the income distribution the top end enjoying extraordinary increases in income and real wages essentially stagnant at the lower end of the income distribution and I think this has a lot to do with the impact already seen of technology what is quite striking about our amazing companies of the information in context ology age the companies like Microsoft and Google and Facebook Instagram whatsapp Netflix etc is the extraordinary disconnect between their equity value and the unbelievably small number of people they employ now I keep on updating this slide I'm afraid I lost did it at the end of 2016 and Facebook's equity value has been a little bit volatile of late but I still think Mark Zuckerberg has got a penny or two so last time I looked at it was three hundred and seventy four billion I think it may be down to two hundred and fifty billion or so and last time I looked at it he employed precisely fifteen thousand employees what these are these phenomenon is a very vectors are drops in the ocean of the global capital market but they give very very high incomes to the people who are lucky enough to enjoy that a that position while at the other end as we see we get this proliferation of low wage low productivity jobs now a crucial thing to say then is how seriously then should we take Marty Feldstein argument that even though monetary income are low even low income earners are enjoying a huge flow of nil or low cost benefits from drugs development low cost computing and communications power and entertainment services private literally for free well I think there's an element of truth of that but the problems that make me think we've still got a problem of inequality are the following not all the free services necessarily deliver welfare stroke life satisfaction benefits I am not convinced that ever more sophisticated computer games for kids are increasing the welfare of kids welfare satisfaction may be heavily driven also by access to positional Goods and relative status where you are relatively may matter and finally and crucially the low income earner living in a poor location and facing a low crowded commute and high housing costs may not feel that their life is okay because the computer game that their kids apply has a billion times as much computing power for the same price as it had ten years ago so I still think we do have a problem of rising inequality what are we going to do about I do not think that more Skills is an adequate answer when you get to this stage of the argument in for instance a Financial Times article or an economist article the sort of articles that have written for nice rich liberals I mean like me probably like quite a few people in this room the default position at which you arrive as well there's this world of automation but provided we provide people with skills they will flourish in this new world and we will offset inequality now let me play make plane I am passionately in favor of good education and high skills but I don't think it's going to solve the problem for two reasons first I think we only need a very small number of very skilled people to deliver rapid productivity growth in the automatable sectors I think we need a fraction of this population to be great computer coders and software experts and we will drive a Productivity miracle we just don't need as many people and the incomes will be set by relative skills not by average skills however many people have learnt to code and could create a perfectly adequate computer game all the income from computer games will accrue each year to their very small number of people who have created the perceived best computer game in that period and all the way across the economy I think we have these winner-take-all and winner take most effects and I think the fact that somebody is a relatively good computer coder is not going to make a blind bit of difference to the wage that they are paid as a delivery driver driver if they are not one of the best computer coders I also think that increasing productivity growth is essentially an irrelevant answer I think the underlying rate of productivity growth is all very very fast and that the existing incentives to invention innovation and entrepreneurship are already sufficient to ensure continued rapid and possibly increasing progress and that the faster the underlying productivity grows the lower the measured productivity will be as low-wage Bo multi packed progress so I think if you think that the answer to this is more rapid productivity growth I'm suggesting that's just an irrelevant answer so if not that what should we do about it or should we do anything about it one argument for doing nothing about it has been put forward in an interesting book by a man called Tyler Cohen some of you may know him in a book called average is over and Cohen is deliberately put forward a sort of dystopia it's a thinker to a degree a deliberate provocation but I think not entirely what he fundamentally says is the world is going to get more and more unequal there's gonna be ten or fifteen percent of the American population who live in the elite cities go to elite schools learn high skills become lawyers investment bankers politicians lobbyists computer programmers doctors etc lawyer you know ten to fifteen percent and they can have a great lifestyle they're gonna work bloody hard and they're going to compete with one another like mad for you know high quality a property in Manhattan in San Francisco so they you know they may not feel all that rich because you know if you keep on having to buy two million-dollar houses you know you don't feel rich even if your incomes pretty high up cetera etc but that's what ten or fifteen percent they're going to do and then he says at the other end we can have a lot of poor people um but the good news is in America we've got lots of space some of it's quite warm some states have pretty easy zoning rules so poorer people are going to live in tiny houses in warm climates on low land loss locations so just about have adequate health care I'm not sure he's right on there but they'll certainly have lots and lots of zero cost entertainment which will be provided to them by Netflix and and they'll be okay and then he says and there's another group of people who are the talented Bohemians and these are the guys who were going off to Berlin or they're going off to Detroit and they're finding you know old industrial warehouses and they're doing you know organic urban farms and you know Kraft chin and and you know actually I think it's rather lovely lifestyle I don't you know and what Cohen then says is you know people say oh yeah but this won't work because there'll be a social revolution and he says there won't be a social revolution because you know people will be happy enough we really shouldn't expect he says a rising what are the exact words so to see here we shouldn't expect rising income and wealth inequality to lead to revolution and revolt the long-term picture will be fairly calm and indeed downright orderly now it strikes me that Cohen's arguments raises a normative ethical question which you can debate and but undoubtedly has some descriptive insights the normative ethical question is is that degree of inequality okay I don't think it is I happen to want to do something about it we're gonna argue that by the way that's a normative ethical question but I think even if you disagree with Cohen's suggestion that that's okay I think his analysis has got some important descriptive insights one is that in already rich world and even in much of the developing world increasing inequality no longer means absolute poverty in the sense of people are starving in the way that they were in Oklahoma in the 1930s they're not cold you know that they got adequate things and that may change their attitude to the balance when they politically revolt about whether they revolt about economic things or about identity things and the other thing is the centrality of housing and commuting costs that what you can afford to standard of living you can afford with a certain amount of monetary income is incredibly heavily determined by the cost of land and the cost of property and that issues to do with urban design access to property cost of property are absolutely fundamental to understanding the dynamics of our society what does that mean for policy responses that could make a difference well if my purpose of my lecture today is more to sort of stimulate the fact that we need this debate rather than to suggest a particular set of issues but suppose it might be this set of issues we might want to have there I think there's pretty little likely of this occurring in the US political system redistributive taxation on real property wealth capital gains and inheritance I think we should argue for a significantly less favorable treatment of intellectual property rights I think we're heading to a world in which more and more income is going to derive from intellectual property right rent and I'm not sure that that's a good idea I think intriguingly urban design and urban development to make multiple cities and as many parts of cities attractive places to live so that you have a good lifestyle even with the low monetary income are important I do think we have to debate universal basic income to provide adequate monetary income for people who won't get it from the labor market or alternatively the provision of high-quality public goods health care education shared public spaces and I think we've got to adequately pay for basic services what some of the things that we're seeing is this explosion of face-to-face service jobs home health care jobs social care jobs at least in the UK however what we've done is we've competitively outsourced those from the private public sector we have driven down the wages of those and we've made that part of the low paid gig economy I think we've got to stop doing that and pay people for those face-to-face service jobs adequately and there are some favourable developments which if they occur would be great we should have increased leisure we should have an increased focus on artistic design and craft skills I made fun of those Bohemians in Detroit and and Berlin but I think that's a rather good development oh and by the way I think low fertility rates and falling working age populations probably mean that Japan is going to find it easy at this to deal with them countries with rising populations all of which I think suggests that intriguingly although we've solved the problem of production and productivity is soaring already we've got some very very tricky challenges ahead and indeed these challenges may be more difficult than solving the problem of production which unfortunately is not an original thought because one of the really frustrating things I've discovered in life is that every time I think I've had an original thought john maynard keynes got there 80 years ago and 80 years ago john maynard keynes said in economic possibilities for our grandchildren thus for the first time since his creation I'm sorry that Maynard Keynes was a bit sort of gender-specific here but let's assume that that means his or her creation thus for the first time since his creation man will be faced with his real his permanent problem how to use his freedom from pressing economic cares how to occupy the leisure which science and compound interest will have won for him to live wisely agreeably and well I think that's quite a challenge and I think we're gonna find it quite difficult to deal with I won't cover the fact that the problems for the developing economy are even more challenging and that there are some challenging problems for economic theory but end at that point thank you very much [Applause] you well what a terrific presentation and the long one we certainly wouldn't want to have missed any and we could I'm sure that stimulated a lot of questions a lot of we could we could have a discussion easily as long as the presentation yeah but we have a time limit so I'm going to try to move this quickly okay and you'll also have the secure in the knowledge that in just a few days the the text will be a on that these you're not you're in light of your written your last comment you're not going to rename is the Institute for old economic lien are you interestingly at the very first meeting of the Institute for economic thinking in King's College Cambridge a point of pilgrimage for Keynesian in April 2010 we had a debate entirely devoted to the relative merits of Hayek and Keynes and people did point out the irony that we were starting our Institute for new economic thinking with a detailed look at two very fine and they are both fine both hiding and Keynes economic thinkers it's true it's often sobering to look and or disobey as you said yeah that they've thought of it they thought about a long time ago and I'm not going to claim not going to claim that you're thinking on this subject is overly influenced by the fact you grew up in in the UK yeah where if I remember every year The Times Sunday magazine had a publication of the hundred richest richest people they still do that they do the Sunday Times Rich List and when I lived in London it was remarkable one quarter of the hundred richest Britons listed there their occupation as Lando landed oh yeah whereas if you went to the United States and read the Forbes 400 yep they were inevitably either the founders of companies or the founders of companies so large you say they found it in yeah or their immediate heirs yeah so a slight effective no I think if you would look at Chinese figures today a hell of a lot is land as well I mean the aggregate figures show that wealth inland dominates I think even in societies which are different from you know old societies like the UK we are seeing these very significant accountable and a very non-trivial point that a lot of the change in income distribution does represent that that brings up a broader point and again we could go on but I'm going to no limit myself to just a couple questions and then a few from the audience and then we'll be out of time certainly the concern with income inequality has grown with the growth the obvious the apparent growth and income inequality in the developed countries not just here but pretty much everywhere and I don't know if you would agree though the lack of of economic theory ex-ante economic theory that predict that would have predicted it in other words we are looking backwards to say we didn't are we hadn't anticipated this yep and now how do we how do we explain this how do we do something as always as I like to say ex post explanations are never quite as satisfying as Exxon TX donations right but let's let's go straight to the issue of of Technology and the way you have portrayed it but one more quick comment you talked about the you reference Solo's quite clever comment yeah but of course he made it in 1987 87 yeah the response was paul davids computer in dynamo f published in 1992 that said basically professor solo Bob if you're very clever fellow but you don't know much about how productivity actually works use real world as opposed to engineer econometrics models yeah and Paul David said you're about to see a surge in productivity growth from some networked computers yep right yeah and that's what happened yep right so the was david wrong because no i think i think what is interesting is you know as often in economic several things are going on i think david correctly pointed out that the technologies always take time to diffuse it takes time before computing you know initially you had word processors and you know people just did you know fifteen drafts of a letter rather than they used to satisfy themselves as one draft but you know he still only had one letter of yeah you know it's a set of absorbing activities did to start with and and he from economic history had looked at the way that electricity had diffused in the economy and that only when it came out of streetlights and began to sort of automate the factory did now and I think therefore the Paul David thing correctly managed to forecast the spurt which we did see in productivity more in the US than elsewhere in the sort of to that nineteen ninety-three era and the era where Alan Greenspan was talking about it but I think what's interesting is when you look back at that that turned out to be relatively brief and it sort of lasted ten years and then it petered out again so if you if you were to sort of now come back and say okay let's compare the last 25 years with the previous 25 you you'd still have the solo paradox there the the David thing was a correct description and that's what I believe is going on that he correctly forecast that these technologies would start diffusing and start having a very very big impact on productivity in some sectors of the economy but that doesn't undermine the fact that the crucial thing is where do the people move to and the crucial thing I'm saying which is the Bommel analysis is in order to say understand productivity growth at the aggregate economy level you have to keep asking not simply what is happening to available productivity growth in the activities which people are currently doing but to which activities of those people going to shift and if they shift to relatively low wage and initially an automatic elective you may get in quick incredibly bad I mean we are seeing incredibly rapid productivity improvements you know I'm who was mentioned by Valley I'm on the board of Prudential in a major insurance company I can tell you the opportunities that we're looking at to take human beings out of back office processing uh absolutely enormous and we're gonna see lots and lots of people coming out I mean for instance how it gets to the issue of the developing world I think the whole of the call center is is a phase of human employment which is going right I don't think we're gonna have call centers in future because I think we're going to have automated bases which are better and more efficient and make less errors and a less susceptible to miss selling claims and etc etc so I think we are seeing in we are now seeing what Paul David said these network cluster effects but I think they're being ops offset by the delivery drivers and the Bommel type activities that's my hypothesis indeed the but also one of the I would say one of the David insights if you will for shorthand is to to perceive the productivity gains you have to install the investments you have to make the investments so if you look for example I was looking at the McKinsey report of early this year that you you mentioned and as they pointed out at present we're experiencing a period of job rich and productivity poor growth yeah they ascribe and I'm sure you're familiar with their their calculations 60% of all sectors have lower productivity growth yep than they did 10 years ago today 50% of sectors have declining capital relative to ours worked and only 5% of all sectors today are producing rapid acceleration accelerating productivity growth in other words I would my first response is of course we've gone through a period of very low business investment yeah so why would you expect to see productivity gains from low am i – my hypothesis is this isn't a sick the fact this is in Larry Summers terms of structural effect I don't think business investment is low because of you know short term is or more tax effects etc I think it's low partly because capital goods are very very cheap and if you go back to my chart on Facebook and you do a calculation which said how much investment was required to make Facebook so economists buy common investment economists mean the devotion of labor activity not to the production of current goods and services but to the building of the Machine which will produce goods and services in subsequent years my best calculation I did a little attempt at it is that Facebook was basically built out of 5,000 software engineer years so that even at the point of flotation that was thirty four million dollars of equity value per software engineer building the machine I think this is just a completely different economy from the economy of Henry Ford and how much investment he had to put into his factory equipment and how much investment somebody had to put into a steel forge to build that equipment before he could create equity value so my hypothesis is I look at those figures from McKinsey and I think that's the way it's going to be we are going to see sectors of the economy which achieve incredibly rapid productivity growth but with relatively little investment and we are going to see the proliferation of bomb old-type low-wage activities with very little capital equipment and that's the way it's going to be for structural reasons rather than something which we can fix or even should try to fix by different policies okay now let me ask about the following way there now this will be my last question is it would it be unfair to say you I'm gonna characterize your your argument as a very extreme version of skill-biased technological change namely it's going to be very small group of big beneficiaries and yeah and I would say well what remember that's an excellent that's an ex post explanation yep we didn't we didn't hear about that yep before that the inequality appeared when you step back you think we'll wait why hasn't that why wasn't that expected because in the past you would say technology made the stupid smart and the weak strong yeah so why wouldn't that why wouldn't the benefits of that be generalized including to the weak and the stupid who've been made strong and intelligent well I think I mean that's part of my argument that this technology is different because it's just more radical possibilities for automation I think it's also partly that to a degree we are reaching quasi Association in some categories of basic goods and services and that in that environment you know if if we wanted all to have ten mobile phones and eight cars and 22 washing machines we would be observing huge productivity improvement because in the production of those you know the productivity is being achieved but the areas where we are achieving productivity growth we're just reaching the limits of how many of these we can possibly want I mean and so I think that's that's what's going on and I think that's very different from the 1950s and 60s I think in the 1950s and 60s you know America is still taking people off the farms putting them into factories and those factories are producing washing machines dishwashers new kitchens new bigger houses cars for the first time that people have ever got those and there's just a huge demand for all of those things all of which can be automatable and I think you know partly what's going on here is a change in the nature of consumption patterns over time relative to our ability to produce these things you know really really fast I think the other thing which is going on is that I I think there's a period of time in which higher technology still empowers very large number of people you know the classic factory worker and the GM factory of of the 1950's they're still they still need their skills alongside some elements of factory automation whereas I think we are now heading to you know Amazon warehouses where almost nobody works and I didn't talk about my section on the developing world but actually I'm far more worried about the developing world than the developed world and let me sum up why I'm worried about that adidas is building a huge Factory in Bavaria now Bavaria I can tell you is not a low-wage part of the world indeed I think it's probably about the highest wage part of anywhere in the world so why are they doing it they're going to have 160 employees those 160 employees are gonna produce 500 thousand pairs of shoes a year and this is going to be low-cost and it's also going to be better customer service because when somebody in Paris or Berlin buys a de – shoe you know with the red flash with a blue flash buy the very next day they'll have reprogrammed the automated equipment to produce more of that and less of this whereas it would previously take six weeks for that message to go down to the supply chain to the people who sew it in Bangladesh etc and if you apply that added US productivity and the supply factory to the whole of the added that supply chain the million people that they employ across the world in their supply chain now will come down to below a hundred thousand and I'm terrified about what on earth do countries in Africa which think they're going to follow the low wage low cost manufacturing route as the first step towards prosperity I think that allowed is being destroyed so I I think this is different you know I mean these are very good challenges John and of course you're absolutely right this would be more compelling if I told you this 20 years ago but you know I think I'm gonna be right over the next 20 years so let's come back in 20 years time and see so but you know a bit but they're very the challenges okay the truth is we're out of time but so let's have a if it with your indulgence yeah I'm a couple questions from the audience yeah yes sir and please wait for the microphone introduce yourself and ask a question so Martin Smith actually went to school here a long time ago you said that people need his money the cost of capital is down and I think that's absolutely right cloud computing means I don't even have to buy my computers anymore right but isn't it a paradox to say you don't need capital and you don't need labor yeah I think it comes back to what you said I'm not sure you need land but it's the intellectual property that's just a golden ticket is that is that yeah look III think it's very interesting this you know and I think we're struggling with the theory as what should happen to returns to capital labor in which you know we have plenty of labor around because we keep on automating existing activities so there's lots of low wage labor but capital is also getting cheaper capital in the sense of investment capital and it's interesting there I'm effort I mean you might have said to me well Facebook that's an extreme but what about the generality there are figures from the IMF a few years ago that showed that the average cost of plants and equipment has fallen the price has fallen by 25 percent relative to the general price level so capital goods are getting cheaper and they're getting cheaper because they're ICT intensive and because software is cheap and cloud computing is cheap what does that then mean well I think it means that you just end up with an environment where you know we we don't we neither need a large amount of current capital to do production nor do we need a large amount of capital labor devoted to building the machines we don't need either I mean you could imagine another environment where we've envisioned these fantastic machines but first of all we're going to do enormous amounts of labor active to build those machines but that's not like this we we've got a vision of machines which will increasingly build themselves and you know I mean honestly I think we're just getting to terms with with what this means but I do think that ironically you know and this is the paradox that it increases the value it seems to increase the value of locationally specific land and it undoubtedly increases the value of intellectual property rights or of network externality effects I mean Facebook is a network externality effect everybody uses Facebook at least till about three days ago because everybody else uses Facebook you know and if everybody else didn't use Facebook you wouldn't use Facebook it's a sort of completely circular network externality and those clearly give huge returns but intellectual property gives huge returns I mean there is a very big issue as to whether America has over protected intellectual property I mean the tendency of the change over the last 25 years is to extend the possibility of patenting to things where you wouldn't have thought yeah I mean at one stage I think Apple tried to patent the swipe thing that switches the phone off and this is not something with it's not an intellectual property right you know you can't say something's an intellectual property right if I think of it 3 seconds before you think of it I mean the intellectual property right has to be something where you've devoted some effort and some scientific research to doing it so I think value a lot of value will reside in intellectual property right and we've got to decide how much we want but I think there's a very good argument for reducing intellectual property right protection rather than increasing it first of all thank you very very much very intriguing and compelling lecture and I look forward to of the 1.6 0.5 and 0.6 on your lecture as well hopefully it'll be as compelling as you know in person as well one of the things that I've been I found so compelling is your thoughts on winners and losers frankly and seemingly you win the gold medal and it's a linear effect you win the silver medal and okay you wouldn't bronze it's nothing and some and honorable mention means absolutely nothing whether it's in a brand whether it's shoes whether it's singing singing it right it's a sports type of event as well and they've been the people that are associated with that and I was wondering if you could follow up more with respect to that type of well look this is not an entirely new phenomenon there's a wonderful article in the economic literature by mine called Sherman Rosen about thirty years ago called the economics of stardom and he made this very interesting point that the radio and television and he was just talking at that stage about that let alone Spotify had dramatically reduced the number of people who made their living as entertainers while dramatically increasing the income of the small successful number of entertainers because 100 years ago in every town somebody was singing popular songs or singing Verdi or Puccini to you know the the local people of every small town throughout the world or doing a circus act or doing something and you know 100 years later you everybody could listen to Beyonce or Placido Domingo and you know all the others didn't get paid very much and Sherman Rosen pointed that out at that stage and that phenomenon has gone on the other phenomenon that goes on is that the higher people's income becomes the more that that the rents from celebrity fandom you know having fans great comparison I use in one of my books is between the income paid to Stanley Matthews you have never heard of Stanley Matthews Stanley Matthews was one of the most famous soccer players of 1950s Britain he paid for Stoke City he was in an absolute top of his career and he got a salary somewhat equivalent to you know reasonably well-paid doctor David Beckham is a multi-billionaire reason why David back there was a multi-billionaire or the reason why JK Rowland is a multi-billionaire and no previous a a writer of children's books has become that is that back in Stoke City the only people who knew about Stanley Matthews were in Stoke City and they had very little money and all they could possibly afford was to buy the ticket to go to the match now everybody throughout the world can go about decadent David Beckham and all of them can spend at least an element of money on buying a shirt which says Beckham on it and every time they buy a shirt with Beckham on it another penny ends up in David Beckham's a bank account and billions and billions of pennies month after month after month gets a very large amount of money now at one level does this matter because these are very small numbers of people on the other hand those wonderful phrase by the British economist Lionel Robbins who pointed out the existence of many rich people is there because other rich people exists I mean once you have small numbers of stars you then have they spend money on fashion designers and they spend money on party organizers and they spend money on lawyers for divorce cases and they spend money on you know you know all sorts of other things which proliferate other high-paid rich people so there's a whole set of complicated phenomena going on which are winner-take-all or or winner take most type effects and I I think you know they are quite pervasive you know all the way across the economy so final question does the Premier League enhance human welfare well the answer is I think yes but not in a way that we can anticipate that it will hunt enhance it more in 2050 and I do think there are two bits of my analysis that you have to deal with one that there are a lot of activities which we will automate radically and there are activities which won't get automated and there will be a breakdown between how fast is as it were inherent productivity potential and what's happened to measure GDP and the other breakdown is I think we we get to a stage where we should not expect the economic machine to deliver more improvements in human utility once we've got to a certain level of income and so we need a richer debate about what it is that creates good societies and human welfare but we resist that because we like the idea that we've got this measure called GDP that means something we we tightly hold the hand of nurse for fear of finding something worse that's very good you see Anthony but I'll say it anyway this thank you thank you very much thank you so I went on for so long oh no I did you could tell the a the the folks loved it we did and we're greatly honored that you took time out of your busy schedule to join us and to share this with us and thank you again people thank you very much we hope you will come back yeah I like it great thank you [Applause]




Leave a Reply

Your email address will not be published. Required fields are marked *