just this past year, a team of geeks said let's try a different approach and configure a system and show it a bunch of examples of games played at a very, very high level. we have a library of "go" games played at a high level. let's show the computer a bunch of examples of those games, that's it. we won't try to elicit the strategies or point out the strategies to the patterns most salient here. we will show it a bunch. they showed it high level games in midstream and they said, hey, what's the smart next move here? they're at the point right now where that system is able to come up with the exact same move as the human expert more than 50% of the time after six months or less work on this problem. i made a bet on twitter platform for all deep thoughts, i made a bet the world's best "go" player will no longer be a human being. with a couple of examples like

that it becomes more and more clear to me and erik that the future is coming at us more than the experts are predicting. that means the economic consequences are coming more quickly than a lot of us are expecting and makes the discussions for today all the more important. thanks very much. [ applause ] thank you all for joining us this morning. my name is melissa carney. i have the privilege of moderating our first panel discussion this morning. this panel is going to take the premise that andy and erik laid out for us, that there has been rapid technological advance particularly in the information center. we will ask the question what does that imply for the future of work, the future of workers and the nature of employment in this country in particular? as we tried to lay out in our hamilton project framing paper there are a lot of views on this

topic and in particular if this will be good or bad on that or how good or bad on that for society. fortunately this morning we have a really expert group to discuss these issues with us. truly, i would say some of the leading minds in the world on these very questions. you have their full bios in your program. i won't run through them in detail. i will just briefly introduce them. to my left is david otter, professor of economics at m.i.t. one of the nation's leading labor economists who has probably contributed more to the nation's literature on leading trends and labor market than any. we have larry summers, university professor, president emeritus at harvard university and served in a number of senior policy positions including secretary of the treasury of the united states and director of the national economic council. and aneesh chopra served as our first nation's technology officer appointed by president

barak obama and served as the virginia leader of technology and now in a technology firm. and erik has already been introduced and still a professor at m.i.t. [ laughter ] >> the way we will do this, i will pose an opening question to each of our panelists and we will move to a moderated free flowing discussion and leave the final 10 minutes for audience q & a. we will be collecting your questions on note cards which then will be brought up to the panel. david, i'm going to open it up with a question for you. you have written extensively about the nuanced relationship between technology and computers and workers, particularly noting that there are certain things that computers can do that substitute for tasks historically or traditionally performed by humans and other things computers do that complement tasks performed by humans.

so, in light of your research and the framework that erik and andy have laid out for us, how do you see this all shaking out for workers? >> that's a great question. i'm honored to be part of this discussion and really like the work they've written. i'm glad this topic is getting the thoughtful discussion it deserves. 15, 20 years ago, erik and i started talking about this when i was a graduate student and erik was assistant professor. at that time we felt people weren't taking this issue seriously, if anything, i thought people should not panic at this point. [ laughter ] >> i think there are a number of remarks i could make. i think there's reason for some skepticism about how fast things are actually moving and a lot of aggregate data that don't

support the idea that the labor market is changing or economy changing as rapidly as the story so dramatically the premium for higher education has plateaued over the last 10 years and we see evidence highly skilled workers are moving -- have less rapid career directories, are important part of the puzzle and productivity not moving rapidly and a lot of growth has been in relatively low education with a public service element to it. it's easy looking at these examples to see an inflection point but when you look at the aggregate data there's nothing to suggest there is an inflection point. it could be in the wrong place, but a reason for skepticism for things not changing that rapidly. the second point i want to make, when we think about how technology interacts with labor

markets we think of substitution of labor with machinery. that's a completely natural thing to do because technologies are made to substitute tasks we were doing. we've been substituting machinery for labor for as long as we've been able to think of ways to do that. that's a first order effect, a mechanical effect we can automate transportation, we can automate calculation and automate information stories or retrieval. in general, what is neglected is that complements us as well. many activities require a mixture of things. it requires a mixture of information processing and creativity, motor power and dexterity. if those things need to be done together if you make one cheaper and more productive, you increase the value of the other. doctors have not become less valuable as medical technology has advanced, right? they can do more, diagnose more

and that makes them more valuable. ultimately, there are three things that sort of contribute to how an aggregate results in and one whether it directly substitutes you or helps you do one thing so you can do something else. if you think about diagnosing medical testing obviously physicians can get a lot more information in the course of a day. the second is how elastic is the demand for those services? we are so much more productive in medicine, we could do all the medicine we did in 1950 in 10 minutes a week and people would probably be healthier given the state of medicine at that time. as people get better at it we get more of it partly because of the medical system and because the services are a much greater value and demand for them is quite elastic. third, from a labor perspective,

it matters how scarce the skill-set is that's complemented. it takes a lot of education and training to become a doctor. when doctors become more productive, we don't just get an infinite number of doctors at minimum wage because they have to have a lot of training and it complements slowly and tends to raise incomes. there are many examples productivity increases lead to making jobs more interesting and challenging. that's not the case. i don't want to take up too much time. that's on one side of the labor market. on the other we see a lot of growth of work that requires generic skills and hard to automate. let me make my final point. a lot of things that matters is how rapidly things change. if tomorrow amazon introduced

the $1,000 bezo spot that could cook for you and clean your house and comes on amazon prime and you could have it by monday, that would be a dramatic advance and we would all buy it. it would be extremely productive because a lot of people, that's their primary activity, driving and childcare and cooking and lawn manicuring. if amazon said we will have this in 2045 for $1,000, we would be well situated to adjust to that, because people would recognize that was not the place they wanted to be over the long term for a career. it matters how quickly we get there. i think -- i think a lot of the debate is not whether these things will occur but it's whether we're at the second half of the chessboard where the inflection pointing all of a sudden things are doubling from a small number to small number doubling again to a large number or whether it's a very incremental process. i would say the academic computer science technology

is very divided about this. if you go to silicon valley engineers believe everything will be accomplished immediately and you talk to the crowd that's skeptical. this is hard and making progress not true 20, 30 years ago, but we're a long way away. as andy said we live in very interesting times. >> i'm sure we will revisit those ideas as we have our discussion. as our nation's chief technology officer you were tasked with using technology and innovation to further our nation's goals of job creation, reduced health care cost, protecting the homeland, tall order. you've spoken very optimistically about the power of technology and innovation to improve our lives on a wide scale. i'm curious to hear how your view of what technology has done compares to that as andy and erik laid out, and in particular, how have you seen

technology impact a variety of sectors, including education and health care among others? >> thank you very much for the question. i have three general observations, all very bullish on this next decade. the first starts with my first trip to google, which was probably in '06. i was virginia secretary of technology and we were trying to open up government data to search engines. to make it more accessible to the american people. most people were getting information about government through search engines, not coming to the url of xyz.gov. i saw this globe when i walked in which had a light emitted for every search on google. the globe was spinning. as you get to north korea, it was dark, this stark observation. large swaths of africa and many parts of the world had darkness. you think about the american economy, what sectors are on that level of darkness as it

impacts the internet has had on the sector. health care, energy and education have not necessarily been plugged into the internet especially around data sets constrained by regulatory policy, medical records aren't flourishing on the internet and your energy usage data isn't flourishing on the internet. when you look at all this amazing capability and productivity gains in manufacturing and others you look to more than a quarter of the gdp, you're thinking these groups of sectors have been completely missing from this revolution. obviously, incentives start to change and data opens up at the same time, you might see an explosion of innovation. we're seeing that now in health care. we've made great strides opening up data, digitizing and eventually connecting medical records systems. more venture capital is flowing into this sector than you would have ever imagined, not necessarily because they're

trying to make the traditional system functioning incrementally better, now incentives are changing to reward a different type of health care delivery system which makes it a wide open terrain for entrepreneurs. that's very exciting because it's creating new types of jobs that never existed before in the health care sector. not all of which require a phd in physics. you can be a relatively low level employee whose utilizing the technologies to help on home health needs so forth. category number one is we're now opening up these big sectors to the internet age and i think that will bode well to ensure productivity gains hit them. second, again when i was virginia's technology secretary, the north carolina-virginia border used to be the world's hot spot for furniture manufacturing. that's it. we went through a policy of debates, those jobs aren't coming back, how do we build a safety net down there and

broadband is the answer and we did everything we could to improve that north carolina-virginia border. something interesting happened around this concept of automation. manufacturing is cheaper because you no longer have to have the same labor intensity and can in-source jobs back to the u.s. at a faster rate in response to china. so ikea opens up a manufacturing plant for furniture. where? right in the heart of the north carolina-virginia border, the same place that was written off for its capacity to build furniture and you're being told in the neighborhood you have to do different things because your life as a furniture person is over. all of a sudden, robots as co-workers, automation, you can actually compete on a more effective footing. we're seeing that in-sourcing trend now all across the country. manufacturing jobs are coming back. they're not the same labor intensity they were when they were previously here but that's still net positive. i would say the third

observation if i had any is this democratization of entrepreneurship is pretty much the most exciting thing i've seen. in that same north carolina-virginia border, there are people who used to have parents and grandparents work in textiles as well. now, they're building designs for clothing that can be 3d printed or their intellectual property can be transmitted over the internet to textile production all over the world and they're creating economic value in that same market because folks who didn't previously think of themselves as silicon valley entrepreneurs can plug in because of the democratization of capital innovation. i'm really fired up over the impact this has in the next decade acknowledging in certain sectors the challenge. too bullish? i don't know, but i'm very excited. >> larry, to you, you've been thinking and commenting on these issues a long time and you wrote a 2013 npr piece.

it raised a lot of the, i we're talking about this morning. and you sponsored the center for american progress on inclusive prosperity, the goal of the commission to address rising levels of income inequality and stagnant wages at the bottom of the distribution. in your thoughts and views on all of this, what do you see as the long run implications for the macro economy? >> thanks, melissa and thanks for the chance to be here. i'll leave the question of what we should do until later. let me focus on diagnosis and make a confession of ignorance and observation and express a worry. confession of ignorance is this. i think it should apply to everybody who speaks confidently in this area. on the one hand we have enormous antidotal evidence and visual evidence of the kind that erik

marshals, that points to technology having huge and pervasive effects. whether it is complementing workers and making them much more productive in a happy way, that's one possibility, whether it is substituting for them and leaving them unemployed is a possibility and can be debated. in either of those scenarios you would expect it to be producing a renaissance of higher productivity. so, we on the one hand are convinced of the pervasiveness and far greater pervasiveness of technology in the last few years, on the other hand, the productivity statistics on the last dozen years are dismal. any fully satisfactory synthetic view has to has to reconcile those two observations and i have not heard it satisfactorily

reconciled and something we have to figure out. it is a big problem to believe -- if you believe technology happens with a big lag and it's only going to happen in the future, that's fine. then, you can't believe it's already caused a large amount of inequality and disruption of employment today. so that is a major puzzle, which i think hangs over this subject which i just want to put out there for discussion. second observation i think it is a mistake to think of the economy as homogeneous producing something called output as we approach these issues. there's an aspect that doesn't get enough attention, which is sectors through progress working themselves into irrelevance. let me give an example. the illumination sector, providing light. it actually has had about a

ten-fold increase in productivity every decade for a century. and we now think of it as a trivial sector in the economy. no doubt we could continue to produce ten-fold increases in productivity but actually most of us want it to be dark at night. [ laughter ] >> so, in fact, there are more little league night games than there used to be, parking lots lit more brightly than they used to be. basically, what's happened is illumination has become quasi-free and whereas candle making was a major industry in the 1900s, illumination is a trivial industry today. we need to recognize that a sector that has rapid technological progress but the world can absorb only so much of it becomes ultimately unimportant in the economy. is that kind of thing relevant in thinking about the world?

here's a fact that continues to astonish me. i concede there are a million measurement problems around it. but it is a fact, what i'm going to say. in the way they compute the consumer price indices by definition, they were all set to be 100 for every good in 1983. consider two goods today. a television set and a year at a university. and instead of using a year at a university, i could use a day in a hospital. [ laughter ] >> the consumer price index for the latter two categories is in the neighborhood of 600. the consumer price index for the former category is 6. so there has been a hundred-fold

change in the relative price of tv sets and the provision of basic education and health care services. if nobody is wondering why governments can't afford to do the things they used to do, i just gave you a big hint. if nobody's wondering where most people are going to be working in the future, i just gave you a big hint. if everybody's -- if anybody's completely confident we will have rapid productivity growth in the future, they should be giving pause because no matter how fast in productivity we have in agricultural or illumination it doesn't really matter for the aggregate economy. increasingly, that's becoming true of a larger and larger fraction of what it is that we produce. third, i was -- when i was an

undergraduate at m.i.t., in the 1960s, there was a whole round of concern about this. will automation displace all the employment? and what i was taught as an undergraduate was that basically the people who thought it would were a bunch of idiot ledites and obviously there would be enough demand and work itself out and if people got more productive and they'd spend and maybe we needed some transition assistance, but it would all be okay. that's what i thought and bob solo thought and he was a hero and the other people were all a bunch of goof balls, kind of what i learned. i believed that for many years and actually repeated it often. it has occurred to me that when i was being taught that, about 6% of the men in the united states, between the age of 25 and 54, were not working. and that today, 16% of the men

in the united states, between the age of 25 and 54, are not working. and it won't be very different even when the economy is at full employment by any definition. so something very serious has happened with respect to the general availability of quality jobs in our society and we can debate whether it's due to technology or whether it is not due to technology, we cannot debate -- we can debate whether it's the cause of dependence or whether it is caused by policies that promote dependence. but i think it is very hard to believe that a society in which

the fraction of people in -- choose whatever your most prime demographic group is that should be working, whatever that group is, a society in which the fraction of them who are not working is doubling in a generation, and seems to be on an upwards trend, is going to be a society that is going to function well or at least function well without major social innovations. i would want to leave you with that concern, as there, whether you think it's due to technology or think it's due to globalization or due to the maldistribution of political power, something very serious is happening in our society. >> great. thank you. i definitely want to make sure we return explicitly to the questions that larry has raised about policy and where we need to push. but before i do, i want to pick up on the first observation

larry made, this is a great point, erik, for you to jump in on, given all these technological advances, really celebrated, why is it that gdp per capita isn't rising more rapidly? why is it that median wages are essentially flat and in particular what does that imply about the impact technology is having on our living standards? are we just -- we're not seeing it in the numbers. are we not measuring it appropriately? >> that's a great question. it's good for larry to bring up and what spurred andy and i in the beginning and others talked about a great stagnation and these amazing things, andy touched on a few of them. there's lots more and we could spend days talking about the wonders of technology we're seeing. it is a bit of a paradox there. there are a couple of parts there worth decomposing. the part about median income, i

don't see that being such a paradox. i think, as i suggested earlier, there's no economic law that says everybody is going to evenly benefit. it could be some small group is left behind. it could be unfortunately a big group. you can have biased technical change that grows the pie and some people are made worse off. i think that's a fair description, at least in my mind, i don't know if other people would disagree about a big part of the story of what's going on, that people with certain types of skills are in much less demand than they were in the past, in part because of technology and many in the median income and david is one of the people that has documented this and lots of people have touched on it. the question of overall gdp per capita is more puzzling, although as i showed you the chart, you don't see as much of a problem in that decoupling chart with the median.

the angst whether tea party occupying wall street is that median line, not the top line, even there it hasn't been quite as robust as maybe some of us would have expected. >> it should have been. technology has been super and more strong and more potent and more everything than it should have been before. the question isn't whether it's slowed down, the question is why didn't these new gale force of technology lead to a big acceleration. that's what you would have expected. >> let me address that. i've spent a lot of time visiting companies installing these technologies. some are quite complex. an ert system or customer relationship management system. we documented it takes five to seven years to roll out. during that process there's a huge amount of organizational disruption and you can do this on a case by case basis and case studies of disasters at m.i.t. and elsewhere trying to roll these things out. quite disruptive being rolled out and no productivity or

decrease while being rolled out and we have aggregate data from hundreds of these firms and i've written papers to show there's a long lag. if you roll that up to an entire supply chain or entire economy you can imagine these organizational disruptions, organizational complements often about 10 times larger than the technology investments themselves. they take much longer to roll out, can be part of these -- enormous disruption and until they're in place with the complementary pieces are in place you don't get the full benefit but with previous technology, it can take 20, 30 years. that may be part of the story. i think we're in a big organization of the economy and yes it's disruptive these people see these people have to be laid off, these other people

have to be reskilled. and in doing that you don't get the full productivity gain but get a lot of disruption. that can partly answer larry's question how you can have disruption without getting the full payoff. if i could take a moment to touch on some of the things that david brought up, i think those are also very interesting, partly about the leveling off of skill by technical change or the college premium, i should say, is actually very consistent with changes we see in technology as i have showed addressing, different parts of the labor market. more broadly, i think he raises the right question about complements and substitutes and what's happening. you look, oftentimes technologies initially are broadly complementary as many pieces of the system require humans or others to fill in. you look at horses, the number of horses increased all through the industrial revolution up to about 1901, that was peak horse, because, you know, whether saddles or carriages or other things made horses much more

valuable. but then the numbers plummeted once the remaining component that horses added wasn't so -- was no longer not automatable, if that's not too many double negatives. you could see similar things potentially, you know, are humans different than horses? of course we're different in many many ways. we have a much broader skill-set and can think a lot better mostly. >> and horses doesn't own capital. >> also, once labor starts disappearing, you can have humans own capital or at least some of them. humans can vote, humans can have guns and do other things if they're not happy with their income distribution. there are a lot of other things potentially different. as an economic fact i don't think there's any necessary inevitability, as larry was saying -- inevitability that it

take cares of itself and we should discuss the policies and in the first industrial revolution a lot of policy changes helped navigate that in a way we did create shared prosperity or inclusive prosperity. >> larry, you want to jump in? >> just on the productivity and disruption thing, i think it's a difficult argument. let's take retailing. you can imagine you can have all kinds of spiffy technology so you no longer have to have people behind cash registers all of that. the problem is you wouldn't expect the people behind the cash registers would get fired before the people working the systems got the new systems working. so the challenge about right now is people see that there's a lot of disememployment that's of disemployment that's already come from the technology but they don't see any productivity

increase. i understand why it might take years for it all to have an effect. what i have a harder time understanding is how there can be substantial disemployment ahead of the effect of the productivity. that is, if you thought that it just was impossible to put in these systems and so forth, then you might think that in the short run, it would be a big employment boom. you have to keep your old legacy system going and you have to have a million guys running around figuring out how to put the new computer system in. i understand low productivity. but i think it is hard to square, and it's not like i have the answer to this puzzle. but if you think about it hard, i don't think it's easy to square low productivity and substantial disemployment. i don't think the lags to reorganization story quite does it.

you shouldn't be getting the disemployment ahead of the productivity. >> it is a complicated story. i don't think i've totally nailed it yet but i think another part of the puzzle is that there are a lot of rents in the economy as well. if you get the types of people who do the reorganization being different than the type of people whose demand is falling, you can have big changes where the rents are happening way ahead of the changes in the overall output. >> let's deal with the fact there is disemployment. >> we all agree we don't want go the way of the horse. [ laughter ] >> what do we do about this? i want to talk about policy and i'm going to pose this to the panelists, a two-part question, so bear with me. first, it seems to me in large part the way this is going to play out for the american worker

is going to depend on how labor supply responds, in particular, in terms of skills. in other words, will -- is there a way to imagine that a sufficient number of people in our population will acquire the skills or the talents that are needed to economically prosper in the second machine age? and what would it take? is our -- is our education system broadly defined up to the task of delivering those skills and talents? the second part of my question is, what about those workers who simply can't acquire those skills or don't possesses those talents or even the ones who do but simply aren't enough high paying jobs for everyone. i will admit i am in part worried about a scenario where a small share of the population commands increasingly high wages and a larger share is relegated to low paying service jobs

presumably providing services to the high wage folks. it doesn't make me feel much better that robots are not going to be able to give a good manicure or clean houses any time soon. is that a reasonable thing to worry about? if so, don't we need to really rethink our social contract and dramatically expand our system of wage subsidies and income supports? >> may i -- i might want to take a stab at this, starting with the premise that if we applied the same capabilities we said may have a positive or negative effect but to unleash them in this particular question of how efficiently are the skills being communicated by employers, the training programs communicating what you get if you join and what the job seeker has or might wish to get. to me, we're like in the dark ages of the quality of that experience. you log onto amazon.com, there is feedback loops they've been analyzing, what's the probability i'm there to shop

for a video or lawn equipment or whatever. if you ask the same question of the workforce, the sad answer to that is drastically, no. we just did a study on the unemployed veterans skills gap. what we tried to do is we read every job posting in the economy and said what are the underlying skills associated with the job postings? we then looked as best we could through open government data the underlying skills of unemployed veterans. we took a spotlight on the commonwealth of virginia. you had hundreds of technology companies post jobs from employers who made a commitment to hire veterans. they're going out of their way to want to hire veterans. and they -- but they communicate the job in such a manner that feels like it's not really available or attainable to some set of the population. by doing this sort of skills assessment, what we figured out was every single entry level

technology job, every single one in april of 2014 from an employer who made a veteran hiring commitment could have been filled by a tech trained vet at that time unemployed in the commonwealth of virginia yet neither the employer knew to look for the tech trainable vet whose background might have made the initial screening nor did the vet know they could get that tech job because it wasn't in their suggested career path saying this is an attainable opportunity to get you to the next stage. these "new york times" story. how many young people have the ability to get into m.i.t. they didn't know there was financial help available.

if every person in the economy had a helper that said given where you are, here's the shortest path to awesomeness to land the best job available to you. is that anywhere near in our system today? how exciting would there be if there was a marketplace of tools to do that? would we solve this. before you get into the income subsidy question. just make the system work better. i think that's an initial place to start. >> and where is that going to get those skills? from the employer, local community college? >> this is the other fascinating question. we did a little panel, the center for american progress, highlighting this at&t partnership with udacity for six month chunks of learning. nano degrees. they're great for cybersecurity and other interesting areas of growth, so, i asked the question, are any of you regulated as learning programs that qualify for government subsidy, whether they be title

iv funding or qualify for the gi benefit or workforce investment board vouchers and the sad reality is these innovations are disconnected from any actual government support because there aren't thoughtful regulatory on ramps for these new entrants to be reimbursed in that manner, so, these are the areas where i think there's opportunity. >> sure. i agree with the direction you're going in in terms of skilling. there's a policy focus on sort of college for all and that has been healthy at some level, but it's, it's very incomplete. our education system is geared towards get people out of high school and college. if they don't go to college, well, it didn't work out. that's not productive when less than half of young adults are going to complete a four-year degree and that's not going to

be 75% within ten years or 30 years, although there has been an increase in both high school graduations and college completions over the last ten years. i think we need to think about the skill sets that allow people to do evolving jobs in health care professionals, in technical positions, many of which require real skill sets, but don't require four year liberal arts training, so, i think we push too many people towards expensive four-year degrees, which are either not as efficient or not as appealing as they could be. there are opportunities in kind of areas, written a lot on kind of this new middle skill occupations. there aren't things you can just get with a high school degree. it is foundational, credential, but for further training. i think there's a lot of productive room for investment there. hopefully, technology will allow us to be better at that. unclear. as with so many things, there's a great potential and

uncertainty about how fast and how well it will work. my biggest concern in this is the sort sort of inequality with which people have responded to these market signal, so you might have thought at a time when college had become more valuable and people going to college, the gradient between household income and college going would get shallower and that has not a occurred. the rate in college going has become much steeper and in college completion, much steeper still. and so, i think that works against economic mobility. it means that kids from low ses backgrounds are less likely to be going to school and be gainfully employed. when larry talks about the declining employment rate among u.s. workers, we're talking about young males, many of them minorities, many from poorer families, so it's a pretty

concentrated problem, which makes it worse, not better. if you sort of look from the median on up, u.s. society looks mobile, healthy, looks like it's making the right investments. if you look below the median, there are just that message and tools to correct that problem are somehow not coming together. >> let me say a few things and i'm actually more confident about these than i am about the technology stuff. given the productivity question. first, with great respect, i would engage in the experiments. i think the policies that she's talking about are largely whistling past the graveyard. the core problem is that there aren't enough jobs. and if you help some people, you could help them get the jobs, but then someone else won't get the jobs and unless you're doing things that are affecting the demand for jobs, you're helping

people win a race to get a fine night number of jobs and there are only so many. this was powerfully demonstrated by a study done in france where they looked at a variety of these kinds of job matching innovations and found that in a low unemployment areas of france, it worked and in the high unemployment areas of france, they only helped some people at the expense of others with no net impact. folks, wage inflation in the united states is 2%. it has not gone up in five years, there are not 3% of the economy where there's any evidence of hyper wage inflation of a kind that would go with worker shortages. the idea that you can just have better training and then there are all these jobs, all these places where there are shortages and we just need to train people

is an evasion of the problem. it is fundamentally an evasion. second, what we need is more demand and that goes to short run cyclical policy and the enormous importance of having tighter markets, so that firms have an incentive to reach for workers, rather than workers having to reach for firms. it's quite remarkable to look over the years at the harvard economics department. 30 professors and 40 graduate students, it's remarkable how badly graduate students get treated. when there's 30 professors and 20 graduate students it's remark able how well they get treated. people who have been to school in environments where there's 60% men or women are not unfamiliar with this. having the labor market run

tight is fundamentally important. it is important for generating investment and work. third thing i would say is that i -- and this is in the same direction as what david was saying. i agree with him and he knows much more about it than i. i think we can't think of education a human blob of capital where more is good. the idea used to be kind of, the way i would have sort of thought about this 30 years ago, was that part of what would be good about having more education is that people would be able to work in an office rather than being plumbers. that was good. that would upgrade people and give them new opportunity and plumber's children could work in offices rather than being plumbers. it's kind of the essence of the changes that are being described.

that they are more heavily bearing on people who work in offices. than they are on plumbers and so, the whole idea of working with a craft and a specialized skill rather than this generic, general manager with liberal arts competence, is i think central to thinking in a rational way about wages. if i could say one other thing, i think that the broad empowerment of labor in a world where an increasing share of, increasing part of the economy is generating income that has a kind of rent aspect to it and the question of who's going to share in it becomes very large.

one of the lesser puzzles, but very large puzzles of our economy today is that on the one hand, we have record low real interest rates that are expected to be record low for 30 years if you look at the index bond market and on the other hand, we have record high profits. you tend to think record high profits would think record high returns to capital would mean really high interest rates and what we actually have is really low interest rates. probably the right way to think about that is there's a lot of rents in what we're calling profits that don't really represent a return to investment but represent a rent. and the question of who's going to get those rents, which goes to the minimum wage, goes to the power of union. goes through the presence of profit sharing, goes to the length of patents and a variety of other government policies that confer rent and then when

those are received, goes to the question of how progressive the tax and transfer system is, that has got to be a very, very large part of the picture. and i am concerned that if we allow the idea to take hold, that all we need to do is there are all these jobs with skills and if we can just train people a bit, then they'll be able to get into them and the whole problem will go away. i think that is fundamentally an evasion of a profound social challenge. >> raise the issue of the minimum wage and unions and bargaining power of workers, but it strikes me we're faced with a conundrum in a sense these changes make the imperative of giving workers more sort of bargaining power and a higher minimum wage make that more compelling and important, but at

the same time, those same technologies make it easier for employers to replace workers who become too expensive with machines, so how do we thread that needle? >> well, i think that is a real challenge and one of the ways i think a lot of us have talked about, not just the minimum wage, but the earned income tax credit, which is a way of encouraging people to work and sharing some of the benefits from the economy to people who are are working and maybe not making very high wages. >> through the tax code, not the employer, employee relationship. >> one of the differences is that while it increases the incentive for people working and helps with income distribution, it's a broadly shared cost that lots of people bear as opposed to the employer who comes up with a way of employing that person and i think that you can make a good argument that those employers, those entrepreneurs,

who figure out how to put those people to work shouldn't be the only ones to bear the burden of having to raise the incomes of people who are having their skill demands fall. sharing that more broadly. and the net effect is not only encouraging more people to work but also there's a spillover. it could actually encourage more people to look for creating those kinds of jobs. >> let's just have some numbers here just to put this in perspective. roughly speaking, if we had the same income distribution in the united states that we did in 1979, the top 1% would have $1 trillion less today and the bottom 80% would have $1 trillion more. and that works out to about $700,000 a family for the top 1%. it works out to about $11,000 a year for a family in the bottom 80%.

but it's a trillion dollars. i don't know what the number is. my guess is that the total cost of the earned income tax credit is $50 billion. nobody's got on the policy agenda doubling the earned income tax credit and the big, aggressive agendas for the earned income tax credit are probably to increase it by a third or a half. so we're talking about -- so i'm for it. i'm all for it. but we are talking about 2.5% of the redistribution that has taken place. and so you have to be looking for things and there's no one thing that is going to do it. my reading of the evidence, i think it's a fairly general reading of the evidence, is that while there may be some elasticity, the elasticity around the current level of the minimum wage is very low.

perhaps a good way to make that point is to observe that the real minimum wage in the united states today is about 20% below where it was when ronald reagan was president. and even ronald reagan when he was president wasn't really complaining that the then existent minimum wage was doing a lot of damage to employment. and productivity has gone up since that time. it's tempting to think that everything's trainable, but if you ask not across international borders between the united states and other countries, if you take the boston smsa, say how much of the boston smsa's gdp is tradable, it's less than half. and so, there's a lot of scope for raising wages in areas where there isn't going to be some broad kind of competition. >> good. >> so, we have some questions

from the audience and i'm going to ask one that relates directly to that last point. so, the question is what is the role of trade on technology and vice versa? how does this relate to the relative skills of people in different countries? >> actually, this is what i wanted to bring up. a lot of the disruption that people attribute to technological change over the last 15, 20 years has had a great deal to do with change from international trade and the you know, especially succession of china to the wto in 2001. to everybody's surprise led to an enormous surge in imports in the u.s. and a very sharp decline in manufacturing and has had large spillovers to surrounding communities and work i've done with david dornan and gordon hansen and mogler and price. really documents -- i think we've all been surprised by how big a factor that is. and fortunately, we're much closer to eckquilibrium now.

policies aside i don't think the next 20 years are going to look anything like the present. but i think this kind of disruptive power is underappreciated and it's been quite significant. that doesn't have -- i am actually in favor of the president receiving fast track authority for negotiating trade agreements and so on but i do think that we often find an effect and we look for a cause and sometimes we get it wrong. waent to a we want to attribute it to the most obvious thing rather than something more subtle. we thought the internet economy was an amazing thing from about 1995 to 1999. i don't see how it all of a sudden became a disastrous thing around 2000. i don't think that's plausible. the other thing i would say is i do think that we're, you know -- that trade does circumscribe, or the possibility of trade, some of the things we do. we could not in my opinion restore unions to where they were 40 years ago without having very substantial and competitive

effects because we face -- we don't have the the kind of rents, the kind of market power. and you can point to countries like germany that have much stronger unionization but then german companies are outsourcing a lot of their employment to eastern europe. so i think that is a real constraint. on the other hand i also want to point out we bring a very western-centric perspective to globalization and technology jointly have brought more human beings out of poverty in the last 20 years. any time in world history. and that's not just because we have a larger population. as a percentage. we look at what's happening in china and india. and a lot of things we view as threatening are creating much broader prosperity. at the same time they create greater concern. when i think about machines substituting labor i worry much more about indian textile workers than i do about the general employment patterns in the u.s. if i'm thinking about this, i think about a world where you know, poor people in kenya have

solar cells but they don't have any -- there's no job for which their skills are really scarce. but i think we should be thinking about technology and globalization as working hand in hand at this point and the view in the 1980s, early '90s, that trade was sort of irrelevant to what was happening in the u.s. is i think quite out of date and still has not fully permeated the consciousness of how people thought about it the last 20 years. >> i want to agree in part and disagree. in part. i think first, it's right to say that trade and technology in a sense are strongly associated with each other. we wouldn't have much more trade but for the much greater ease of communicating and transporting across countries but for the technology that represented the container ship and a great deal else, so what

we call trade and the great increases in trade are very much tied up with technology. that's the first thing i'd say. second, i would agree, but be inclined respectfully to disagree with david on one aspect. i agree with david and certainly, my thinking would have evolved over the last 20 years, on the question of how much has changing trade patterns impacted the u.s. labor market? i think there is pretty clear evidence there have been significant impacts. i think some exaggerate them, but i think there have been significant impacts. i think it is a quite different statement to assert that all of that is due to trade agreements. and i think one has to look carefully for example, at the counterfactual. david asserted that since china's ak accession into the

wto. well, what is the counterfactual? i have some familiarity with the level of u.s. tariffs on china prior to china's accession into the wto. and they were not high. and so a very -- the main reason why china is exporting more to the united states is that china's producing six times as much as it was in 1999 and producing in much more technologically sophisticated ways. now, it's true that if they had not been in the wto, conceivably, we could have passed a whole new set of protectionist measures, but i think if you ask the question, if the united states had maintained its trade policies, vis-a-vis china as they stood before china was admitted into the wto, what fraction of the increase in chinese exports to the united states would we have observed?

i think the answer is the vast majority of that increase in exports. and i think that's very important. because i think there's a tendency to suppose that if trade developments impact the wage distribution importantly in the united states, then presumptively trade agreements are a bad idea and i think that in order to analyze any given trade agreement one has to ask the question, how much are barriers being changed in the united states and how much are barriers being changed in the affected country, and my reading of the evidence is that in many of the cases, because rightly or wrongly the united states market is already substantially open, if you look at the proposed trade agreements, the reduction in barriers and the consequent increase in exports to other countries looms quite large relative to any impact in the united states, so i just think that's an important qualification on the

globalization story. >> i'm going to ask a different one that takes things in a bit of a different direction. so someone asks from the end of the 19th century, technology let the workweek decline. why can't that process continue with the benefit of technology being a shorter workweek with no loss of income. and i'm going to add a bit of a maybe existential question or spin to, this which maybe is too existential for an economic policy group, but still. so all of this technology really changes our view of the good life and how we think about our time. and aneesh's story about the expert craft furniture makers who now put together ready-to-assemble products for ikea, i mean maybe this is too nostalgic for an economist but something seems lost in that to me. erik, i'm going to open this to you. this must be something you've thought a lot about. >> this is a great question and

for those who haven't already read keynes's great article "economic possibilities for our grandchildren," he talked-b made predictions about what would happen to our generation. he was more or less spot on in terms of gdp per capita. he extrapolated those functions exactly right and got that right. and he inferred that he looked around people who were that wealthy and inferred people wouldn't want to work a lot. they would go fox hunting occasionally or whatever, but there wasn't much else to do without wealth. of course that part he got wrong. people are not working 10 to 15 hours a week. mostly those who are working are working a lot more than that. there are a number of reasons for that. i mean part of it is we have a lot more we can spend our money on now than we did back then. there's lots of new goods that i think people enjoy. part of it also, and this gets into sociology. i think a lot of people, they enjoy working. there's a meaning that it gives to life for a lot of people. bob putnam described what happened when work leaves a community and it's really sad to

see how all sorts of other social indicators just plummet because of the way we're wrapped up in having a job and having work. going forward could there be shorter work weeks? i think certainly, we could. there's a trend in that direction, not as rapid at keynes imagined but we have to start thinking about new ways how we get meaning in life. i don't think that's an insurmountable problem. we'd also have to continue to have enough productivity growth to make that work but i don't think that's insurmountable. >> if you look at -- i'm uncertain what i think about this. if you look at an introductory economics textbooks from the 1960s or 1970s, in about chapter 5 there's always a discussion of something called the backward bending labor supply curve. and the idea is that as your wages go up at first you work more and more because it's attractive to work and then after a while you have enough in

income and then when you have enough income you take a bunch of it in leisure and so the labor supply curve looks like this. if you look at an introductory economics textbook today, that idea is largely not there. and the reason is that it used to kind of be true. that high-wage people work less hours than low-wage people. your image of the 1930s was that the ceo sort of went out to play golf at 4:00 and the workers worked 60 hours a week. and if you look today for the first time basically in economic history people who have higher wages on average very consistently are choosing to work more or are finding themselves working more hours than people who have low wages and it's in part a matter -- it's not all that the people who

have low wages are not able to get more work. there is choice that's people are working more hours. and that's why this idea of a more leisurely nirvana is less in fashion. that said, i must say, i have to be impressed that americans work about 50% more hours, maybe 45% more hours, in a year than northern europeans. i'm not sure that i would want to call that a great virtue of american society, but i think we have to think carefully about what the alternative to work is and how meaning and community are found in the absence of work.

classical economics has this simple view, which is working is bad, leisure is good. those who spent time in communities with 20% unemployment i don't think find that a riveting formulation of human motivation and desire, so i think it's something that needs to be thought about a great deal and i guess the thought i have without knowing where to go with it is it sure seems like in our society, whether it is taking care of the young or taking care of the old or repairing a lot that needs to be repaired, there is a huge amount of very valuable work that needs to be done. it's much less clear to use a modern phrase that there's a viable business model for

getting it done and i guess the reason why i think there's going to need to be a lot of reflection on the role of government going forward is that if i'm right that there's mightily important work to be done for which there's no standard capital business model that will get it done, that suggests important roles for public policy. >> there's actually some activist work that make it with a business return on investment the social good. so addressing climate change feels like a big priority for all of us to do something about and the fastest growing job in america, energy job, solar panel installation. solar panel installation. now, you made earlier comment about trade and technology. partially this is because the cost of importing solar panels is low. here's an interesting conundrum. if you benchmarked germany against the u.s. on the cost to install solar panel in a world where you've got globally competitive prices for importing these technologies, these fancy

panels, what explains the delta? it's a little over a billion-dollar delta if you kind of ran it out. what explains it is something called the soft costs of solar installation. the lack of automation in something as simple as permitting the ability for a home or business to put solar panels on the roofs. so if we took this powerful concept of information technology and innovation and if we had the ubiquitous same-day permitting processes and much more efficient financial markets for credit in order to finance these things and we had much more readily available matchmaking services so that folks who want the solar panels at a lower price can get them more rapidly with the local installers if all of those technologies could be put to work that way, we would reduce this billion-dollar hidden tax on the american solar panel economy. which is already the fastest

growing energy job in america, so you'd create more jobs on the backs of what is essentially a low cost trade import. you would address issues important to the world like climate change. and what's getting in the way? the lack of the adoption of these innovation -- these capabilities in of all places the mayor's office or the government. can you do same-day solar permitting in new york city? you cannot. but you can in certain parts of california where they're making an emphasis here. so my only comment about, there's too much work to do to have a leisure question, but if you're worried as larry said, there's work to be done, but no return, we can make it profitable. there should be many, many companies that are organizing labor to put solar panels in. if we could decost the process by applying these technologies that would grow the market create jobs, and not all of those installation jobs requires a ph.d. in physics. you can do them with different levels of skill in the economy. i think there's a role of government to like create market opportunity in places where we need it. >> we will be having another hamilton project event on march 11th focused on removing

frictions in the labor market with the goal of increasing jobs. so hopefully you'll all join us again for that. but mt meantime, please join me in thanking our panelists. [ applause ] we'll have a ten-minute break. moderating the second panel in this amazing conversation. i want to start by really congratulating the hamilton project for pulling a wonderful event together. and also in case you have not read the future work of the agents of machine which i think you put together in record time if you want a summary of the issues that we're trying to discuss here with a very distinguished set of speakers read this. it's really a great overview. and it links to everything you need to know. and besides that, they brought everybody here you need to know right here. so i recommend this. okay.

>> -- from reading entire books on the subject. >> i assumed that was the price of admission. that you actually had to read "the second machine age." in any event this panel is going to focus on business innovation. it's going focus on it in two ways. one is what technologies are we really talking about? so much of the conversation is about software coding, a little bit of robots thrown in there occasionally. but what technologies are really transforming, are they enhancing productivity, have we reached a technological plateau, which some people argue we have. but then even if the technology is rapidly changing, are our organizations changing rapidly enough? is our innovative capability to absorb all of this on the decline? and then finally, what are some of the business models that might work to actually help us absorb all of this and make the most of us -- make the most of it for all of us?

so that's what we're going to do and i will start with someone i think a number of you know well. john haltiwanger. he has done a lot of really important work documenting the fact that the pace of innovation in organizations in the united states may be declining, that entrepreneurship may be stagnating that labor market fluidity may be less than it has been. and believe me you need fluid labor markets to make the most of this. i thought i would start a question to john by making an observation that eric and andrew make, which is that in the face of rapid technological change, let's assume that is happening for a minute the benefit will not go to labor and it won't go to capital as we traditionally know it, it will go to entrepreneurial people and entrepreneurial organizations who can create new products, new services, and new business models. but i think john's worried about

whether we are capacity to do any or all of these things is declining. so john, is it declining? why? and what are the implications that if you got a stagnant sort of structure trying to absorb this rapid technological change. >> thank you. it's great to be here. i think all of us when we get a chance to listen to eric and andrew we're struck by the sort of gee whiz aspects of all the rapid technological changes. but then we also heard in the first panel i think nicely some of the skepticism that doesn't seem to be showing up in key numbers like productivity statistics. and i think actually the evidence on the declining entrepreneurship and the declining dynamism in the united states economy may be part of why we're not seeing it in those numbers and also raises questions about again about whether the u.s. is well poised to deal with this. so, if we were having this discussion -- in fact, a number of these people were having this discussion back in the 1990s. so we have lots of folks who

were playing important policy roles in the 1990s. and the u.s. was just surging in the 1990s in terms of both productivity growth and jobs growth and earns growth was doing okay. this was before the great departure. and i have quotes from some of the notable policy people in the room including folks on this panel. that indicate when they were giving speeches and saying why was the u.s. doing so well in the 1990s? and oftentimes the two words were used by these distinguished policy makers were the dynamism and the flexibility of the u.s. economy. so what have we sort of seen? one thing we know about innovation, at least from what we see in the data, i'm someone who looks a lot at the data, is innovation and productivity growth a very noisy and complex process. it's not just here's some new idea come along and we can talk about it. it takes maybe many years to figure out how to use it. and that noisiness, that

complexity at least historically from the united states has been when we have booming times and booming sectors we see a very high pace of entrepreneurship and lots of volatility. and also what's kind of stliking about the nature of that volatility, you know, it began in the 1990s. it's only a small fraction of businesses that make it, so again, a striking feature of the united states economy is we have the surge in good times, surge of entrepreneurship and only a small portion make it. and then talking about the dynamism is the flexibility part. well, what this means is that there's lots of restructuring going on in the united states economy as a result of this. and the u.s. has been quite flexible and fluid in terms of being able to move workers to other kinds of productivity uses. and the fluidity has helped, we've got to remember, it's not

just businesses experimenting. workers themselves experiment. a key way workers build their careers is job hopping. the way they find the right match, it's not just at the high end, it's everywhere. where earnings go up and where you could say they find lifetime careers is through lots of job hopping. so it should at least cause us concern that we see now in the data several indicators i'll say of dynamism and fluidity down, especially since 2000. some of these have been going on before that but it's especially in 2000 where things have accelerated. so what have we seen? we've seen a decline in entrepreneurship actually that predates 2000. but the tech sector was rising through 2000 and it's fallen. some might say was that the dotcom bubble? no, it's not just that. if you look at data entrepreneurship was rising through the 1990s in the high-tech sector. then it has this sort of mountain peak where everybody

was apparently starting a dotcom business. then it came down. and again if you look at entrepreneurship in 2005 it's lower than it was in 1995. >> john, can i jump in? is it on the upswing since 2005? >> no. it continued to decline. >> that seems -- how can that be true? we hear about crazy internet billionaires. i'm honestly puzzled. >> so -- yes. so when i talk about an entrepreneur, i'm talking about a new business that hires at least one worker. so what i've been doing and others have been doing is tracking the number of businesses that hire at least one worker. what we've seen is these entrepreneurs as so defined have been critical for job creation and productivity growth and innovation with this large pace of entrepreneurs coming in and a small fraction really taking off and disproportionately creating lots of jobs and being high in terms of innovation and

productivity. again, we've seen a decline in the post-2000 period that's continued. because the great recession was insult to injury. so young businesses were already on their way down prerecession and they just got hammered and they've been slow to recover. so we're obviously concerned about what this might mean for productivity, earnings, and wage growth. so do we fully understand this decline? and the answer is i'm happy during the questions and so on and the panel to talk about what are our ideas. but let me just talk briefly about the costs and consequences. obviously the consequences are going to depend upon the causes. one thing that we see -- one thing that we've done in work i did with steve davis presented at the jackson hole conference last august is look carefully at the decline of dine miz sxm labor market fluidity for

something that was a topic on the first panel, which is unemployment rates, and what we found is it looks like they're closely connected. this decline in employment rates we talked about a lot in the first panel has been especially for young and less educated workers and especially young less educated males. and what we found is it turns out there's lots of variation across states in where we've seen this decline in dynamism and fluidity. what we found is the states with the biggest declines in dine miz sxm fluidity are exactly the states that have had the biggest declines in the employment rights for young less educated males. and we push it even further, i'm not going to talk about ec econometrics but we try to generate causal effects and in fact our results were consistent with that. so then it looks like there's at least some adverse consequences. the second thing we've done is we've tried to look at what the productivity implications of all this are. so one possibility of this

sideline in dynamism and entrepreneurship you say maybe the business model in the united states has changed. maybe we don't need to do so much experimentation as we used to. maybe we've even gotten better. of it was waste. we had this massive set of entrepreneurs come in. many of them fail. very costly reallocation. maybe we got better. maybe information technology has made that -- we need less of that. so one way to look at that is what we've seen in the data the microdata is we've seen that there's lots of dispersion in productivity across firms. that's actually very much what's driving the dynamism. that's consistent with economic theory. again you get this wave of entrants in, lots of them are experimenting, some of them do very well, they take off, some of them are not doing so well, they shrink and contract. what we want to do is just look at the dispersion of productivity across businesses, particularly young businesses and particularly young businesses in high-tech. so one possibility is maybe that dispersion that's declined the

business model has changed so, we don't need to see smoch so much of that. what we see in the data is the opposite. in anything, dispersion and productivity of business has gone up instead of going down. so there's actually a need for more reallocation rather than less recalation. something's causing -- at the end of the day it's not that the shock's hitting businesses and you could say not that the technological changes have slowed down, so this is kind of consistent with the story we're hearing, but the responsiveness of businesses to this change has slowed down. and again that has by construction, that has adverse consequences for productivity. so it is in a mechanical sense part of the way we were getting productivity gains and we always get productivity gains is we're moving resources away from less productive to more productive businesses, we're just not doing that as much as we used to. >> a question, and i don't expect -- not an answer now but a question to put out there to try to relate to the previous discussion as to what extent might you think these amazing changes in technology that are

labor displacing or potentially labor displacing actually are discouraging the formation of a entrepreneurial venture with one person. what's the point? a lot of entrepreneurial ventures with one person are very labor-intensive local things and if you can do those things online you don't need the local entrepreneur and the one person that person hires. and by the way, i would link that to the way those one-person shops, they're more like mom and pop shops, a lot are founded by women. there's a whole issue of labor participation rates of women stagnating around 2,000, right around 2,000. technology and as a cause of this is something to think about but let's go from there to technology and to our team. we have here the head of darpa,

a very well-known source of technological innovation in the united states. and economists tend to, when they're looking at this technological change, tend to raise two issues. one is it's not showing up in the productivity numbers. we heard that very eloquently in the first panel. the other is just you know what? these technologies are all not that important we're not doing major innovations in health, we're not doing major innovations in transportation and energy nothing similar to electrification or telephonic communication. what's the big deal? so i don't think we have a better person on our team to talk about these pessimistic views of technology coming from the economics acommunity and your sense of where the big technological changes are coming from and is it all i.t. and software? >> let me just say how much i appreciate the chance to participate in this dialogue, the work that you and eric have done, andy and the hamilton

project. i think we're talking about unraveling this complicated set of linkages between teng and work and it's so core to what's important for our values and our fiech really think it is terrific. i want to take the conversation a very different direction and talk about technology itself. the word technology has almost become synonymous. many people use it synonymously with information technology. if you read "the new york times," their technology section is only about information teblg. everything else is relegated to the science section. but in fact, technology is much, much broader than that. and what i thought i'd do is take a few minutes and just share with you some perspectives about the bubbling pot of things that are happening in some very, very different areas. first of all let me say there's a lot more to say about the technology factors driving the many dimensions of technology. we can come back to that. let me set that aside. let me give you some different examples. one is something that is bubbling today in the maker

movement. part of that is new tools like 3-d printing which everyone has talked a lot. but part of it too is finding different ways to make those kinds of tools available to lots of people. one example is a tech shop here in alexandria, virginia that we at darpa helped get started a little while ago. we co-founded it working with the veterans administration in part to be able to provide a gym membership-like access to the tech shop here for veterans. so for about what it costs to belong to a gym they or anyone else actually can have access to advanced 3-d printers but also sewing machines and welding tools and every kind of production equipment, a wonderful machine shop. and i went by to visit there right before christmas. i wanted to just get a sense of the vibe there. and one of my most engaging conversations was with a young fellow who's in high school. he found out about tech shop somehow. he drives an hour and a half

each way many times a week, as often as he can break away, to come to the tech shop and build things. i said what do you do with these things? he said i put them up on pinterest and all my friends and family buy them. i got his card and i raced home. it was right before christmas. i thought maybe i'd buy something he built. it turns out i doentn't know anyone who wants accessories for guns to play paintball which is pretty much what he's building. but that's what's happening today. people are experimenting. this is a kid who's finding a way to build a business. i don't know what tomorrow holds. it might be something that broadens and taps the skills and the energies and the creativity of new sets of communities. that would be awesome. i don't really know yet if that's going to happen. but that's one thing that's bubbling. a major area of research that we're very excited about at darpa because we think it holds the seeds for technological surprise, that's our business is -- that's happening today in

research as biology is intersecting with the information sciences and technology but also the physical science and technology world. let me give you two examples of things that are 457g.happening. i'll tell you what's happening today but also try to take you out into what could be a very wild future. let's start with biology. this is the ability to engineer microorganisms to create chemistries and materials the world has never seen before doing it in dishes and trying to scale it up and doing it in factories. so what is happening today is we're able to build new specialty chemicals, new medicines, and very interesting beginning. but in fact it's only a beginning. we can see in these capabilities we can see a progression of new materials chemistries, but also functional systems. and self-repairing systems. we can imagine a future where you might be able in your built environment, you might be living in a building in which the walls

are able to sense the environment around them, adjust temperature and humidity and lighting conditions. these walls might support microbial communities that could disinfect the air, that can purify the air. they might be derlz that can self-repair so when your teenager puts a gouge in the wall it can fix itself. and when the time comes and when you want to the wall could biodegrade and not create this sort of perpetual waste that we live with today. that all sounds a little crazy but imagine a century ago if someone had told you about some magical new pvc material that was going to be super lightweight and incredibly easy to work with but would be so corrosion-resistant it would change the way you do plumbing that would have seemed a little crazy. and i think today some of these things that might seem crazy might become actually -- i think technically they look like they

might become possible. and of course how we turn those into businesses and products is another question. let me finish with an example from neurotechnologies, another area where biology and technology are coming together in some very exciting ways. and we're just at the very beginning of this adventure of understanding the brain and how to harness its amazing capabilities. today much of our work in this area is about the restoration of function. but in that work you can start imagining what might be possible out into the future. one of our areas of focus here has been revolutionizing prosthetics, moving beyond the hook that's been the standard for upper limb prosthetics. to pursue that vision our program manager developed a very sophisticated robotic hand with many degrees of freedom. and that was one branch of it. but because he's a neuroscientist he also did the research that helped us understand how neurosig naflg from the motor cortex actually controls that arm. that work culminated in some early human trials.

most notably a woman named jan who lives in pittsburgh who's a quadriplegic volunteered to have these two small implants surgically placeded on the surface of her brain on the motor cortex and from that neural signals are directly picked up and in real time as she thinks she's able to move this arm and she can shake hands, she can give you a fist bump, she can offer you a stack of cookies with this robotic arm just by thinking about it. first and foremost the health care implications in that dimension but many other dimensions of restoring function as we understand the brain, that's going to be amazing in itself. but as we do this work we also understand that we have opened the door that could free the brain from the limitation of the body. and as we start thinking about what else is going to be possible, beyond restoration of function, we actually open i think some amazing possibilities. some are going to be great and some are going to be terrifying.

i think the societal issues are going to make the work issues we're dealing with today look simple. technology has many quandaries that it raises. but i hope some of those ideas give you a sense of the very wide range of things that are happening today that could lead to alternative futures. >> so you mentioned the work issues and then you mentioned that there are other issues that we're not even thinking. there are obviously issues you think about too and those would be in the realm of national security. you didn't mention that as an area, but i did think that's one we might want to talk about. one of the things that comes to mind in thinking about technology and moving to business models is a concern that i have and i think many americans have about how we finance basic science and applied science and whether we are doing enough. and because we're pulling back on the support for basic

science, this gets to andrew in the following way. most research and development in the united states, most spending is done by the private sector, not by the public sector. and most is done by very large companies. 85% of the r&d spend in the united states private sector is done by the u.s. nullity national companies that are big. this gets to andrew. because andrew is concerned that the existing existing companies cannot cope with the new technologies and that we have to develop whole new organizational structures and business models to take advantage of. but remember, right now we have the situation where we've got the government doing the basic science and being very resistant to doing more and wanting to do less and wanting to know what the business model is right away. and then we have these large companies worried about they won't be able to make it.

talk about what kind of companies and how would they support their ongoing investment in r&d. unless you become huge overnight like google or apple. and then you're running your own venture capital firm, your own r&d facility your own everything. >> and can i just highlight first of all how interesting this panel is? i say this with no pride because i haven't started talking yet. but look what we're hearing about. we're hearing about this time of deep technological ferment which is unbelievably good news coupled with institutional sclerosis. this is weird. it's bad news in a lot of ways. john, as you point out we don't fully understand this phenomenon at all. i really don't understand why at a time when the tools of entrepreneurship are really good more widely available than ever and getting better all the time. the maker space you described is a really good example of that. it's on the decline in this country. this is a deep deep puzzle and we'd better spend some time figuring out. a lot of it is completely opaque

to me. some of this is clear. some of it is self-inflict damage. some of our ways of approaching the situation and our reactions to it are making the sclerosis worse. i live in cambridge, massachusetts. pray for the people of cambridge rights now by the way. i live in cambridge massachusetts. and the city council in cambridge made a very sincere effort last year to specifically ban uber from operating in the city of cambridge. they wanted the taxis to continue. they wanted nothing but the taxis to be able to pick up anybody in the city of cambridge. i think my head came close to exploding when i heard about this. i have trouble imagining a worse response to this situation that we're in. because uber's a very controversial company. i think their management has done themselves no favors in some ways. but i want to go on record as saying i love uber and its business model. not just as a person who likes to get from point a to point b. but the more i understand about the opportunities they're providing to put labor back in

the economy and to provide a decent living for people, the bigger fan i back of the company. laura, you probably saw the study that alan krueger just worked on and published where he says that the average uber driver, the comparison is a little difficult to make, pretty clear they get paid at least as much per hour as a cabbie does. >> per hour. >> per hour they get paid at least as much. they have great flexibility about how many hours they want to work. in contrast to almost other part-time workers, they appear not to suffer a penalty on a per-hour basis from working fewer hours. you made 15 bucks if you drive three hours. you made 15 bucks an hour if you drive 30 hours a week. these are all really good things to have. especially to the point that larry was making earlier, if we want to bring jobs and demand back, here say platform that's bringing jobs and demand back. the harshly negative reactions to it honestly don't make any sense to me. and i get the idea that some people kind of want to legislate

secure jobs in the middle class back into existence. i think that's a fundamentally misguided approach. the question you brought up was are today's great big successful companies going to be able to navigate this transition that a lot of us feel is beginning? and the pattern of history is not an encouraging one. when we look back at these big technology trends steam to electricity being the biggest one most recently that we went through, the pattern is fairly clear that the companies that are on top at the beginning are usually not the companies that are on top at the end. and there appear to be two main reasons for that. one is financial. when you've got a factory totally set up for steam it's really hard to get out your sharp pence sxl justify the retrofit forget this weird new thing called electricity. that's part of it. the deeper pron erer problem is a mindset problem. if you're used to thinking about a factory as this thing with a big thing in the basement, belts and pulleys that drive your machines, when some weirdo shows up with an electric motor, you

say that thing have less powerful, costs more per horsepower and whatever, why do i do that, you don't see eventually replacing them with overhead cranes and assembly lines and things like that. the mindset challenge is a really severe challenge. as i look around at a lot of very successful well-managed companies today i see that mindset challenge coming up over and over. one example i give, one of my messages to large established enterprises is your management needs to become a lot geekier. and by that i mean a lot more driven by the numbers a lot more rigorous, a lot more evidence based, in things like human resources where the dominant mode right now is you interview me and you look deep into my eyes and judge my character and fit for the job and make a recommendation based on that. we've got ample evidence that's a terrible way to make human capital decisions. being a lot more analytical and a lot more geeky is the right way to do it. today's managers didn't grow one that tool kit. they didn't get to where they

are by virtue of their geekiness and their familiarity with quantitative stuff. so that particular transition i think is going to be difficult. there are a lot of ones that feel similarly challenging to me. i think a huge open question is whether today's successful enterprises are going to navigate into this technologically very different future. i think some of them certainly will. i think a lot of them are really going to struggle. >> can i sort of connect your question, which is this confusion, and i agree with it, about sort of the technology not only changing rapidly but creating all kinds of enabling technologies. and i think the makers was a wonderful example. and yet this decline in entrepreneurship. might is that be -- john how have you looked at the issue just of the overall demand level in the economy? let's go to larry's point. larry's point is a very important point. the overall macroeconomic climate here is in fact possibly the biggest determinant?

productivity doesn't change that fast, and it's -- but the macro conditions in terms of the excess demand or excess supply in the labor market can actually change. and then i think about 2000 and 2007. a lot of mom-and-pop entrepreneurs particularly in 2007 there was no capital for these people. absolutely none. you could not keep your establishment open from one day to the next because your demand fell off the cliff and your ability to finance fell off the cliff at exactly the same time. so i wonder in terms of thinking -- maybe it hasn't -- the technology may be enabling but if you don't have the demand for your product and you don't have the financing to set up your little enterprise which may someday be google you can't do it. >> i think that's right. all of those factors are right. but i think that's mostly a post-2007 story. right? and i think that's really important. i mean the great recession

clobbered the whole u.s. economy, especially clobbered young businesses for exactly the reasons you're talking about. this was going on before that. we've seen the decline in dine miz sxm fluidity and the decline in employment rates and the decline in productivity predates this. >> [ inaudible ]. -- for why that is? >> this is going to be the two-handed economist kind of thing. >> laundry list. >> on the other hand of what might be going on. one or more more benign interpretations but may actually have adverse implications for the united states. we talked about this in the first panel. i think david autor in particular talked about this. we have seen this big shift away from young small businesses toward more large mature businesses. particularly the multinationals. their share's going up. and one can make the case that actually what i.t. has been

especially good is it's enabled the multinationals to be able -- to be abling to communicate with all their activities around the world instantaneously. so as a result the question is the big guys are saying maybe the u.s. isn't the best place to do this there's other places, maybe that isn't good for the world economy but the u.s. is going through some disruption kind of factory. the second thing that does seem to be going on that again may be associated with the change in the business model not so much an explanation is again it's good to remember that most entrepreneurs fail and only a small fraction really grew rapidly. what was really striking in the data were these fast -- these high-growth young firms that played such a vital role in the '80s and the '90s. we've seen i don't want to say a disappearance of high-growth

young firms but a tremendous decline in high-growth young firms. >> so google and facebook are distracting us from the real story? they're exceptions? >> one question is whether the business model now is -- maybe it used to be you wanted to be google and now it's i wanted to be bought up by google p. you could make the case maybe that's not such a bad thing, that's a change in the business model. but then again, if this is all good news, if this is all entirely benign then why are the productivity statistics so bad? then let me go on the other hand of things -- where are we looking? the obvious concern is the u.s., has it changed its business climate, its regulations its labor market regulations in some fashion so that we've seen this gradual decline? and we're work on that and i think we find some evidence of this. trying to think about that question, it's useful to think

about all the cross-country studies that have been done on this. on the one hand we have now increasing evidence that countries that are successful are precisely those countries which have this successful productivity-enhancing reallocation. so that countries that are not doing well are countries in which they have lots of dispersion in productivity and they're just not able to get resources to the most productive businesses. it's actually been harder. and that all suggests -- we know countries different quite a bit in terms of their business climate and labor market and all the rest pf financial markets and so on. we're pretty convinced that this matters a lot in terms of whether a country's successful or not. we struggled a lot more actually finding, even in cross-country evidence, of finding exactly what are the causes associated with why this country seems to have a bad environment relative to the other. one view that's increasingly

been taken in the macroeconomic community is to think of this by death by a thousand cuts. lots of little things. don't look for one big thing. and in fact i don't think there's any -- there's no evidence there's a big smoking gun that says the u.s. suddenly did something around 2000 and so we're seeing this decline. so i think we are beginning to look at some smaller things. what are two smaller things that look like they actually might matter? and at least the suggestion we need to push harder on this. this is not a complete explanation. one builds on david autor's work. david had done work like this and we took it and applied it particularly to our more recent data and also applied it to this more comprehensive data. one part of david's very nice -- he's got a lot of nice research contributions is he found there has been an erosion of employment will doctrine in the united states through precedence in the u.s. court system that made its way gradually through the nature of how judicial

precedents are set in? states before other states. it's precisely that variation that allows to identify these effects. in his work, and we followed up that actually looks like it's associated with this decline and dynamism. that's working in the right direction. that's not saying this is the whole stair. it's just at least saying oh, yeah, you actually can see. here's one of the thousand cuts we're looking at. the second one we just started to look at is inspired by very recent work by krueger and kleiner. and that's this work that occupational licensing requirements have risen dramatically in the united states over the last couple decades. and those are the kind of regulations where you could easily imagine wait a second, that's exactly the kind of thing that again could sort of stifle the regulation. we even heard a little bit about this on the first panel. it wasn't on ok's-payingal license occupational licensing but the permitting process. has the u.s. become more sclerotic because of

accumulation in problems in the way things are working? i'm not going to argue that i think we know this for sure now but i am going to come back and say, one, we've seen this decline in dynamism and two, when we look at the productivity statistics at the micro level we ought to be seeing an increase in dynamism, not a decrease. that almost can't be good news. >> one follow-up here. you suggested that you're looking cross-country. historically speaking for a very long time the view was that by far the u.s. was the most entrepreneurial, most innovative, most fluid. so who's surpassing us,000? or is maybe we're just less but our gap to -- so our gap to excellence has declined but we're still number one. >> i think it's more the latter by the way. i don't think that the rest of the world is necessarily has

become that much more dynamic and entrepreneurial. and erik tried to push this before. you've got to remember things are not that dire in the united states, at least on the aggregate productivity statistics. they slowed down, but we're not in a crisis. we are side say in a crisis in terms of unemployment rates. the point that came up that larry summers -- and i think that is connected to this. again, my concern about this decline in dine miz sxm flexibility, i'll say it, both at the top and the bottom. at the top are we poised to take advantage of all these technological changes or is that going to happen elsewhere and are we not going to be as successful, for example, as we why in the 1990s when the economy was rocking? i'm probably even more concerned about what's going on at the bottom end because the lack of fluidity and dynamism means we know there's disruption that goes on in all of this and it doesn't seem like we accommodate that so well.

and the workers that get caught up in this, what used to happen is because we were such a dynamic and flexible economy other opportunities were arising. they're not. so i think they're just not participating in the labor market. >> i guess what i would like to say is a number of the many pieces of david autor's work, which actually suggests that the issue is not -- there's the issue of demand and the fact that we were running at a very low level of demand so we had an employment problem across the skill spectrum, it's increasing an employment problem that existed before which is the employment problem of high school dropouts. but in fact, in fact, that issue -- and you run a sort of high-demand, high-intensity economy shows up not in employment numbers, it shows up in poor job numbers. it's about the quality of the job, not the percentage of people who are employed. and you know if you think about it if you think about someone

said on the last panel we have to worry about the fact the technology itself may be leading to disruption of jobs that are quality jobs and then you have care-giving, education janitorial services, all of the things which actually in a number of the periods david was looking at you saw employment growth there. it's actually pretty strong. it was one of the reason that's unemployment rates in the u.s. fell so far, is because low-wage jobs, low-quality jobs not even part-time, increased and people took them. that's a demand phenomenon. but then you relate it to the technology thing. the technology's changinging out the middle. it may anybody david's more recent work taking out some of the top in order to run a high employment economy if the technology's taking out those kind of jobs more people are going to be employed in low-end jobs. which takes me back to talk a little bit about -- when you think about some of these

technological breakthroughs they are very very exciting but think about them in terms of i would say distributional issues. maybe it's a little bit unfair. but every time i hear about these prosthetics i say how do we as a society decides who gets this stuff? this is not inexpensive stuff. does everybody get it? and if so how do we generate the revenue stream for the societal promise that everybody gets to live in their brain outside of their body or everybody gets an arm when their original arm no longer can hit -- can be a baseball player-level hitter? >> those are huge issues. but before we boil that entire ocean i think that one we do have a little bit of time before it hits us. we need to be thinking those things through. but i want to come back to what some of these new areas i think will do that i hope -- my hope is -- i don't think we know, but my hope is they will create a

plethora of different kinds of opportunities because in the information technology world if the only answer to the challenge is, well, everyone needs to go back to school and learn how to code, that's going to be really world if the only answer to the challenge is well, everyone needs to go back to school, that will be really good for some people but it won't get everyone in a 300 million person society. so think about tying it back to some of the things we talked about a minute ago. you think about synthetic biology. i was talking to a small company startup that wants to tackle synthetic biology pathways to new specialty chemicals and to enhancing the production. they think about scaling up, they are going to look more like a traditional chemicals or manufacturing company. lots of smart people with lots of education can get employed but when they scale up they'll also need technicians and people at all different skill levels

with different kinds of skills. coding but different skills as well. i think we're going to really need that kind of diversity of different technological opportunities. i don't know how this will play out. one of the interesting things coming back to your point about where r & d investment happens, it very much -- as you said, twice as much of our nation's r & d is made in the private sector. that ratio was flipped when we were all small children. it's a trend that actually is overall healthy. right? it's good if we have a more innovation driven economy. but that investment, of course, companies make not thinking about the jobs. they make it in order to pursue their business plans and their profits.

the government part of that investment, i make my share of that for national security objectives, national science foundation and national institutes of health are seeding basic research but with no particular focus, drive or ability to shape how that turns into jobs. in a market economy, there are no formal drivers to shape how this comes out. that's the richness of our approach and also, it maybes it hard to predict where any of this is to go. >> i think one statistic that is misleading -- maybe all of these are misleading. i will say especially if you go look in the more high-tech sectors, you still don't find that the young guys are reporting much r & d. and that's because the questionnaires are gauged and sort of specified. then you'll be able to report these statistics. think of all the tech companies

we're about here today. they're tiny. they don't have r & d labs. they're doing everything. so i think actually if those companies aren't spending basically 100% of their time -- indeed, that was the division, of course, that these are businesses -- they have no revenue but they're spending an enormous amount of resources, it's all r & d. i think actually the decline in entrepreneurship particularly in the r & d sector, is troubling. we're seeing less innovation and less r & d and we're not measuring it in the statistics. >> issues aside -- in fact, i think we are capturing a lot of that but measurement issues aside, when private companies invest in r & d, it is almost exclusive product development for known markets and a small

fraction of that will be the kind of next generation, more exploratory work. >> not to be too critical, it's not that difficult to pick up applied innovation. it's more core to pick up -- >> but at the top level, i think we can agree that this growing corporate share is much more product development driven. >> sure, i agree with that. >> and it still remains government's function to fund the basic research, the university core. >> and i think that's part of -- there are many messages that if we go to the policy side of this discussion, both in the first panel, we're all thinking about a world in which the pace of technological change has picked

up, we need a policy debate. including even the fact that so much of this has been driven at the beginning from support for basic science and universities. and one thing that the large companies do do -- and then we train people who have phds. if there's not a lot of research support for them to do basic research on, they go and do applied research for the companies. that's where the jobs are going to be for them. we need to have a policy debate which focuses on if we think these are going to improve our lives dramatically how we finance this appropriately. i do worry a lot about that. so we have some questions here. one is a question, i think we probably should address. this is regarding the decline and fluidity. focused on increasing shareholderer wealth and investing in risky adventures?

i wonder, you're sort of thinking about what kind of business models out there. is shareholder value particularly driven by activist investors a good environment for promoting the kind of technological change that you think we should have? is it discouraging entrepreneurship? what do you think? >> i gave a version of the talk to kick off today and you can start to see these great decoupling happen. i was giving this presentation to the open societies foundation and he said you're misattributing the root cause here. he said it's the rise of what we call market fundamentalism, which he associates with the idea that the job of a company to return money to its shareholders, not to think more broadly about stakeholders. it's that turbo charged selfish version of capitalism that's causing a lot of what you're

seeing here instead of any surge in technology. i think that's a really intriguing idea. but my career, as someone who kind of tries to understand what's going on in the business world is on the order of 25 years old, throughout that entire career, i had been reading about the excessive short-term of american business, and over reliance, in keeping wall street happy. the names have changed. that critique has really not changed as i have been reading it for two-plus decades. so i don't know -- honestly, laura, i don't know how much weight to attach to that. i don't -- i don't see that as a major factor here. and the most -- among the tech companies that i know, the most ruthless, growth hungry, vicious competitors are also investing the most in the basic fundamental technologies. >> yeah. so, these are a hard set of questions like all the ones

we're about today. but again i think there's evidence that historically the major innovations have come not so much from the incumbents but from the newer businesses. so that's kind of one concern. the second concern here is whether -- what the incumbents are trying to do, obviously -- they haven't installed basic products out there. certainly in the popular press, are they just trying to protect their -- and grow their market share? back again i said maybe the goal now is not to be the next google, but to be bought by google. the question is, what's google's goal in that? is it actually to take advantage of new technology or actually to shut down competition? those are the kind of concerns that fit into the kind of questions that you're about. i don't know that we've got

overwhelming evidence that that's what's going on. but it's not inconsistent with the evidence we've seen decline in entrepreneurship and we've seen this increased share at the top end of businesses. >> i spent half of my professional life here in washington, the other half in the private sector, 15 years, 10 of which was in venture capital and the other portion in a couple of different companies. and i think you're right, andrew, that perhaps that's the core market drivers, shareholder value, nothing has changed. on the other hand i think those are actually huge drivers. they are fundamental core drivers of every business decision i ever participated in. i just think the fact that this unchanging market-driven decision-making process -- how

it's grappling with this new set of changes in technology, maybe it's that nexus and that's what's different. i do think -- the premise of laura's question is really a core issue. we do rely on the market to solve -- right. >> whoever asks, we do rely on the market to solve a whole host of problems. and i think we are in a regime where these problems will not get solved that way. >> every venture capitalist i know encourages their companies to hit home runs, swing for the fences, not to do an incremental -- >> because that's how you return for the lps, that's a business model for the venture capitalist. >> absolutely. that's not part of the problem we're talking about. i think that's a good thing instead of a bad thing.