 Good afternoon ladies and gentlemen and let me welcome you to Today's economics seminar which focuses on the topic of digitization and the quality and quantity of new music books and movies We are most fortunate to have professor Joel Waldvogel with us who is one of the leading researchers In the field of copyright economics. In fact, this is quite an interesting field where I would say digitization itself has been important inspiring research studies What is interesting that you know some maybe 15 years ago? Studies started to emerge and I think the first wave of studies looked at the effects of file sharing on Revenues and how the music industry is affected And I think you know the latest wave of studies that maybe started some five seven years ago and professor Waldvogel can describe that much better than I can Can has gone much beyond that and it has looked at you know, how creativity as such is affected by by digitization Today's seminar will not present a single paper Instead professor Waldvogel will present a synthesis of some of his latest research work that focuses precisely on the types of creative outputs that you see in the title of of the seminar and I don't want to say too much about that work in advance. We'll leave this to professor professor Waldvogel Just one or two words about his biographical background He's a professor at the University of Minnesota for a long time He used to be a professor at the University of Pennsylvania in the United States he's also research associate at the National Bureau of Economic Research and He has recently done a lot of work also in Europe working with the joint research center of the European Commission Where he has been able to get access to a lot of interesting data? I think especially on the music industry from which his research benefits and So, you know, he is in Europe quite often and the fact that I think he was in Zurich was quite advantageous We took advantage of this opportunity and invited him to Geneva to present today's seminar So I suggest as usual we have about 40 45 minutes of presentation Please feel free to ask any clarifying questions But maybe if we can leave the big substantial existential questions till till the end So professor Waldvogel, thank you for the very kind introduction. It's a great pleasure to be here It's it's unusual for me to be sitting down. I'm tethered to this microphone So you have to imagine me animated and excited, but I'm sitting down All right, so as Carsten said, I'm talking about a series of papers And so you might find that I'm not going into as much depth as you hope So we can if you want to go into more depth on any particular thing We can certainly talk about it at you know at the end in the discussion time All right, though the general topic though is as the title suggests It's about digitization and its impacts on the quantity and quality of new Products in these sectors its music books movies others as well talk a little bit about a few other copyright protected sectors When we think about digitization and media industries, it really it's a two-part story It started with bad news on the demand side and really the first the first affected industry was music It was affected by Napster and Napster came along and Napster and then subsequent technologies for consuming without payment They were bad news that reduced the ability for firms to generate revenue from any particular products that they might have created But that's just part of what happened, you know, and that's what you know as Carsten said That's what researchers were focused on for quite some time myself included I'll be a little critical of that research because I participated in it So I have a right to criticize it But you know eventually I think we've noticed that something else happened in addition Which is that there's cost reduction on the supply side, you know big Reductions in the cost of production promotion distribution and so these things together if you think about it You know would have different effects on what would happen in these industries Then if say only bad news on the demand side had happened now those two elements the Demand sort of the negative shock to demand and the negative shock to supply or negative shock to cost rather Along with something else that's very important about media products Which are typically the copyright protected sorts of products, which is that it's very difficult to predict which products Are going to be commercially appealing and successful and economically important at the time of investment So William Goldman the screenwriter has this very famous quote. Nobody knows anything I think nobody knows anything is a good description of the predictability of success in these in these industries and So these various effects together. I think have produced revolutionary effects on recorded music books movies television There are lots of new products and many of which are economically consequential So what I'd like to do today in addition to talking about findings about a variety of industries from a variety of papers is also to highlight Some rather nontraditional sources of data that I think I've had to resort to in part because I've decided you know once one accepts the notion that the right question is documenting the Quality and the the service flow of the new stream of products Then one has to become a bit Catholic about what sorts of data to look at anyway, so I'll get to all of that So I want to emphasize this again because I think this is a bit of a changed research priority relative to much of what's happened You know piracy remains an interesting and important question for many firms and you know actors in these industries But really when you think about the goal of intellectual property protection And I certainly don't have to lecture you about that I think the question, you know the public policy means to address is is the supply of new products robust and the quality of those products in the service flow and The question I guess is which evidence speaks to whether copyright is properly or adequately fulfilling its function So, you know, we could ask In the vein of my additional goals, you know, you could ask the question Are we experiencing a crisis and certainly there are some data that That says that we are some data that say that we are but there are others that really say very different stories And you know the last little kind of subtext is that better data are needed and more data Infrastructure so that it's possible to do things like what I've started to do but in in systematic ways there's a lot of Some data don't exist other data are expensive. So it's it's rather hard to to document what's going on Okay But to get to the substance of it I'm gonna start by talking about Napster and music sorry music rather and I'll talk about the quality of music since Napster Rising or falling. I'll just go ahead and hint now that I'm gonna claim that it's been rising And then we need to talk about why how on earth could that be in an environment in which there's been a collapse of revenue to recorded music So I'll talk about a theoretical explanation of why that might be so and think about its implications What does it imply that we should see if that were true? And look at those questions both for music and other and other industries Okay, and it's sort of uh in some ways I guess I'm telling you the story in the order in which it came to me But I think it's a relatively coherent order All right So every discussion of digitization especially in music starts with a picture of the bad news And so one of the kind of bad news pictures is the picture of the evolution of the recorded music revenue So this happens to be a picture of us recorded music revenue although pictures for uh for europe look very very similar It rises and had had risen for quite some time occasional fluctuations due to you know ups and downs of the economy But then it has fallen consistently and sharply since 1999 even if you add digital back in it's fallen The overall it's fallen by something like 60 percent in you in the us and europe So that's really that's big bad news. It really does look like the sky is falling Uh, oh, there we go. That's a way to think about it. I don't know if that's fair use. I hope that's fair use Okay, but the research response to this event the napster event Was what I would characterize as a kerfuffle, which is a you know a funny word But a kerfuffle about whether file sharing cannibalizes sales now in some ways that's a a seemingly Not stupid question But seemingly sort of trivial question in the sense that it's a the question is whether the ability of consumers to obtain products without paying for them Would reduce the ability of sellers to generate revenue from trying to sell them Some people didn't need research to know the answer But you know academics play by marquis of queensberry rules and so we do need evidence and actually to be fair It's fairly hard to garner evidence about this because of the problem that popular things are popular when and where they're popular So that means you'll see high volumes of Unauthorized or unpaid activity at the same time you see high volumes of paid activity Simply because those popular things are popular at the same time Moreover, you'll see people who do a lot of one doing a lot of the other doesn't mean that Doing the one that is stealing if we can may I use the word stealing? Anyway, some people object to it that that causes people to buy more anyways, I'd say that uh The dust has settled in that literature And most people believe that that indeed that unpaid consumption reduces the ability of sellers to generate revenue And so in fact a big chunk if not all of the reduction in revenue and recorded music is attributable to uh to piracy or file sharing Now that you know, I felt like and I think the the literature Participants in the literature felt like it was an accomplishment to be to answer that question even though in some sense Uh industry participants didn't need that question answered didn't think they needed it answered But about five six years ago. I had a little bit of an epiphany There's the road to Damascus and it occurred to me that that revenue reduction although again very important Especially for industry participants and incumbent participants is not really the question Relevant or not the only question relevant to thinking about whether copyright is working We should be worrying about the flow of new products. Okay, because I I shouldn't have to remind myself this But I'm putting it parenthetically there. We should worry about consumers as well as producers, okay now The the perspective of industry in the music context does have in it a notion of concern for consumers I mean the the uh, both the recording industry association of america and the ifpi do talk about how these are investment intensive products And so if there were a collapse of revenue, there could be a an end to investment and therefore Not their words. These are my words But an audio stone age a reduction in production of new things So it's not a crazy concern in the context of a 70 percent reduction in revenue But as I mentioned at the outset file sharing is not the only innovation really if you think we're living through a digitization experiment It's a compound experiment that includes both as I said before bad news on the demand side But good news on the cost side at least potentially I haven't shown you any any evidence of that yet But let me claim at least that it's possible that that's happened That that the costs of production and promotion and distribution have fallen in which case One wouldn't need as much ip protection whatever that means to generate the same level of output as before Okay, but that's just an hypothesis or you know a contingent guess Really the question that addresses whether any of this is so is what the question You know what has happened to the quote quality of new products in the case of music since Napster Now I put quality in these distancing quotes because usually once I say the word quality in the context of a product that has aesthetic connotations I start a fight that I don't mean to fight I can't I don't want to start and I can't finish because I'm an economist So for me quality doesn't mean anything aesthetic as a human it does but as an economist It does not it merely means whatever it is that causes people to go out and buy stuff Okay, or consume it whatever it is that shifts the demand curve around is an economist's notion of quality Okay, and I don't mean to sound vulgar but that that is I am an economist so I have no choice all right If one accepts that this this question assessing the quality or service flow Maybe that's a less contentious word But if one accepts that the question of assessing the service flow from new creative products is the right question Then one has to go after it even though I think at the outset you have to say it's a hard question It's a very hard question. So let me admit it's a hard question, but if you agree it's worth answering then I'll continue So I in the music context and more broadly I can think of two ways to go about studying this one is Critics best of lists by which I mean things like the you know the best of all time or you know Best of the last 10 years those kinds of lists not the best of this year since of course There are always 10 best every year. So, uh, you know a time series of the number of 10 best every year Wouldn't be very informative But if you have longer term lists Then you can learn something about the volume of work that critics believe Surpasses some threshold some constant threshold of quality over time. So that that's a kind of an unusual Even weird approach for an economist to take But then in addition to that I have a more conventional approach that I like to take when I have the data Which is usage usage based and so I'm going to talk about these two different approaches to to documenting quality Over time and our service flow over time One based on critics and one based on Usage information by time that is by calendar time as well as by vintage of the production of the products And I'll get to more about that as we go So when we think about the critic sort of information the underlying data are of this the following sort There's a famous, you know rolling stone volume the 500 best albums of all time that book came out in 2004 If you take the 500 albums in that volume and you bend them into the years in which they were originally released Then you get an idea of one of what what one of these sort of indices of the quality or sorry The volume of high quality music produced over time And you see that the the folks at the rolling stone think that 1970 was the You know the zenith and it's mostly been bad news since then now the rolling stone index Isn't really that useful for digitization because although one could mark the beginning of digitization at a number of different years Certainly 1999 is an important year and years since then are important and that index ends are out of four So I would need more than just this source of data to construct indices But this is an example and a famous example Of a data source like this now I spent a summer a few years back more than a few now finding a lot of A lot of critic lists like the rolling stone list and what this graph Shows you you can't read it. So it's just a graph that's meant to suggest how much work I did But vertically you can see different indices or different sources of data So pitchfork is in there and a whole bunch of different sources and across the horizontal axis you see the time coverage So the point is I have indices covering or I can construct indices covering the entire period since the beginning of time Which for me is 1960 Yes, I love coming to europe and talking about the beginning of time in music as 1960 all right, anyways If I make an index out of each of these data sources and then statistically spliced them together So it's sort of in geek terms all I'm doing is if y it Is the number of works from index i that were originally produced in year t I take the natural log of that and regress it on index dummy the index fixed effect And then a common set of time dummies. So what I'm doing when I plot the theta's Is I'm producing a spliced index of quality according to these critics Here's what it looks like now. I've put a little modesty panel on the years since 1999 just to maintain excitement Remember what's going on here is that the revenue to recorded music is falling by 70 in the 10 years following What's covered and so what happens to the the index of quality high quality stuff as perceived by critics Well, it goes flat. It doesn't fall. It goes flat Now if you are kind of a hair splitter, you might say that it had been falling And is no longer falling and so relative to falling. It's rising. That's a little bit too Cute I think an argument but certainly relative to the plummeting nature of recorded music revenue flat is an interesting contradiction Okay, so There we are index of vintage quality Now, let me talk about an entirely independent approach. I mentioned there were two one based on critics another based on usage data So let me talk about the usage data. This is a bit more conventional because instead of relying on the views of elites This is about uh actual economic decision making, you know behavior what people choose to buy or listen to Um, anyway, so let me talk about how I do this So suppose one were to have data on on some usage measure either consumption So sales, let's say or airplay something like that By both calendar year And year of original release, okay The idea is this if one Vintage's music is better than another's then its greater appeal should generate greater consumption or airplay After accounting for its age All right, so we know we should expect and i'll show you in data that it's true as well At any point in time older stuff is used less than newer stuff. That is to say that music depreciates But after accounting for the depreciation pattern If we have some you know more multiple years of data We can actually look and see the vintage impact on usage and try to infer something about the Quality and again quality just in the sense of service flow or usage of the different vintages Okay, so what do I have now there are there are good very good data out there at the time I started this research I I didn't have that much money. So uh, but what I could get I could get my hands on five years of airplay data aggregated So it's us airplay data Um based on millions of spins on the radio Uh, but all I have is the the uh the share of the spins on the radio From each year that are for music from each previous year, right? So it's kind of a lot of data, but it's it fits on a If it's in a spreadsheet I also had I I wished I had at the time like nielsen data on the sales of individual songs or albums But instead I couldn't afford that what I did have though is the recording industry association of america's gold and platinum certifications So what that is when an album hits half a million in sales It's certified gold in the us when it hits a million it's certified platinum when it hits two million double platinum and so forth Now there are something like 20 000 certifications in the data between 1960 and uh, you know 2010 It's a tiny fraction of music released, but it's the bulk of total music sales So it's it's not a crazy place to go looking and and by the way one since one knows wins things were released One can use those data to create shares of sales in a calendar year that are from each previous vintage Anyways, so what I do and again, it's another kind of geeky slide but what I do is I define stv as the share of Of this music used in year t that was originally released at vintage v Right, so in year t if you sum across the v's it would be one And so for a given year t the s is going to vary across vintages because of depreciation That's sort of the the noise here But then also because of differences in usefulness or quality. That's what we care about So to get at this I just regressed the natural log of stv on age dummies So and that's going to flexibly account for depreciation And then vintage dummies what's going to flexibly allow me to reconstruct an index of again quality and distancing quotes So yes the vintage dummies are the index of vintage quality So what do I get? This is the airplay. This is what I get from the airplay data. Now, what's interesting about this Notice how it has a peak around 1970 just like the critic stuff even though it's entirely independent It's not the same data set or something. It's entirely independent So we see that according to airplay That conditional on age the stuff from about 1970 is used a lot relative to other vintages And then it falls Then we get to 1999 and again, I've tried to build some excitement there with the with the blind, but any guesses what happens Well, I can't stand the suspense. It rises rather substantially after 1999 Which again, what that means in words is that given how old The music from the period since 1999 is it gets used more than music that age typically had gotten used That's what that means. That's what my my quality index means Now what about oh, sorry. Yes, maybe ask one clarifying question To what extent do you account for demographic shifts there? In a sense that I think You know, you have this baby boomer generation that got older and older I wonder if they boomed in the 1970s and to what you know that Is that something you may pick up? Yeah, so I mean the there are a couple of different things to think about here One is that the airplay data are all data based on airplay decisions made between 2004 and 2008 Okay, and so in some ways those are more vulnerable to this concern And in some ways the critic data is even more vulnerable to this concern because especially when you think about the Rolling stone editors, they are certainly they would seem to be nostalgic for 1970 the the The sales data here, however are actually data from the entire period in 1970 up to 2010 or so 2008 and so here I'm identifying patterns Not just from recent decisions, but from decisions made all the way through this time period So although there there could be some influence of that I don't think that that Given that I get this from all these different approaches that don't necessarily rely on the nostalgia of you know Of people at a point in time I don't think that's what's going on I mean I also since I did all this I actually have gotten the data I always had wished I had the nielson data for a number of well 22 countries And the patterns emerge all over the place even places with different demographic Baby boom patterns, so that's a good question and I get it. I've gotten a lot So here this is entirely independent now This is the sales data entirely independent of the airplay data. I mean in the sense that they're not literally You know connected in any way and entirely independent of the of the critic data And so when I do this process on these data Here's what I get I get an increase since 1999 All right, so again these data that are derived from actual economic decisions Show an increase in the usefulness if you will of music in this period when revenue has been collapsing Okay, so what's the bottom line from this first set of of Studies there's no evidence that vintage quality has declined in the digital era There's more compelling evidence that it's increased because I would argue that the stuff based on behavior as opposed to the view of elites Is more compelling Now you could make the It would be reasonable to register the concern that you know, I'm I'm showing the impact Or what has happened in a period when we had these this compound change both the bad news on the demand side and good news on the supply side And so I'm certainly not isolating Separately the impact of napster versus cost reduction. This is just on net what seems to have happened in this this industry So if one wanted separate answers to the question, what would happen if we just had a cost reduction? Or what what would happen if we just had revenue reduction? I can't do that here and I I do believe based on work I've done another context that if one just had a revenue reduction and not a cost reduction one wouldn't expect good news Okay, so like in india in the 80s When piracy was rampant in movies actually pretty strong evidence of reduction in the production of movies Big big reduction in the production of movies, right? So I'm not saying that the mix of good news and bad news would everywhere and always be good news for production Anyways, this is a big contrast to the the view That's implied by what happened to revenue and implied by what the industry talks about when they talk about the the catastrophe in revenue But if you believe me so far you might be puzzled. I was a bit puzzled. Well, why would this be happening? You know, why would there be a big increase in quality? Given a collapse of revenue. So let me try to talk about that So here we get to this nobody knows idea I think a fundamental feature of creative products not just music but movies books Probably others is this nobody knows anything idea that I've I mentioned before Here I'll credit it to both Richard caves, but it's really originally Goldman he wrote Princess bride he wrote the screenplay to the butch Cassidy and the Sundance kid So he's a he's not an economist, but he's a guy we should listen to Anyways, his notion is it's hard to predict what's going to be successful at the time of investment And the statistics that that people who study this tend to trot out our numbers like 5 to 10 percent Of products that get released end up covering their costs. So that's that's that supports the notion It's hard to predict now if you think about how digitization would change things It's useful to think for a minute about the world before digitization So traditionally it's been expensive to quote experiment So here I'm borrowing Marco tervio's term experiment But just meaning in order to figure out which products are going to be successful with consumers You need to bring them to market and then see who likes them in the music context It has traditionally cost about a million dollars To bring a new album to market a new album by a new artist Now that might sound like a ridiculous or astounding number But even in like last year in the last two years the IFPI's investing in music report Describes, you know in some detail what it costs to bring a new product to market according to the model that they By which they do it and that's the number that they that they cite it costs a million dollars So in a world like that The the strategy is to you know, do your best to choose which artists are worth investing in and then bet on those And then of course most of them fail even just because it's a hard problem Not because the traditional investors are bad at it Okay, so you bet on a few artists with x anti promise Many fail a few succeed and and then life goes on Well then bring digitization into the pictures We've already talked about some aspects of this on the demand side You've got some bad news but on the supply side as I've said a few times now the possibility of reductions in cost Now i'm going to show you shortly a lot of data about how there's been an increase in the number of products That come to market So that means that the cost reductions have to be important if they're offsetting The revenue reductions. Okay, so if we take the view if we take the view that That the data showing an increase in the number of products available aren't wrong and I don't think they are Then the one way to think about it Keeping in mind this is an unpredictable process is that Effectively it's become cheaper to experiment that is to bring a product to market and see whether people turn out to like it So we're taking more draws. Well, we society. We're taking more draws from the urn meaning Sticking some more in this case musical artists out there to see if if their work Is appealing to people and the question is if we can take more draws from the urn Do we end up discovering more? Artists who end up in the right tail as very successful and therefore important to consumers and valuable to consumers and sellers All right, so how would this work? Maybe I think in some sense what I've said in words is all I need to say But a tiny bit of formalism Might be helpful probably not but if you could think about you know Think about a little bit of a model here where uh, there's some true quality of of a let's say musical Artists recorded music product But the investor the label just forms a guess So the truth is q the guess is q prime, which is the truth plus a random error Okay, now as I said cost reductions obviously important even relative even given piracy because we do see an increase So let's just say that effectively what happens is That the cost threshold t Falls And so now more things can be brought to market and what would happen to the volume of quote good work Well, so just to pose for a second just for illustration's sake. Let's pretend there is no unpredictability All right, so this is a the world without any unpredictability What I have is a picture here with a threshold of this marketability threshold You can think about it as the the expected revenue from releasing a product and there's the old threshold t It used to be we release stuff that's better than that Then the threshold falls we release some new stuff Now let's just think for a second if this were the world Adding products would be beneficial to consumers because you'd get some new products that are here. They're not very good But they would add something This wouldn't get us very far resolving the puzzle of a big increase in the quality experienced So this this this story without unpredictability isn't going to get us very far But when we add unpredictability in there Now we're going to release all the products where our guess of how good it is is above the new cost threshold Well, indeed a lot of them will be terrible, but some of them will be very good and end up in the right tail of the distribution And that's what could account or explain an increase in this experienced quality Okay, so that's the idea the release of products with less x anti promise leads to a greater number of products with x post success and value Oops I just pressed the wrong button. There we go All right, so again, that's all just kind of blue sky thinking if it's true It should have some implications in data And so here are some empirical questions that I try to address with the in some cases unorthodox data that I've been collecting In some cases not so unorthodox are there more let me just articulate the questions on this slide Then talk for the next few slides about some data that are Uh, suggestive or indicative of answers. Are there more new products? I've already hinted about this But in including more low x anti Quality products for example in the case of music indie records that is records from Independent record labels the ones that are maybe wouldn't have seemed to have great promise at the time that they initially hit the market Second Do consumers have ways of learning about the proliferation of new products? As I said, there is a proliferation and here I'll show you some changing roles of traditional radio internet and other sources of information And then in some ways the most important direct test here Is the following two products with less x anti promise and again, I want to be clear about that since I'll come back to the notion You know, so imagine there are a whole bunch of would-be artists who say here invest in me invest in me And and the the the gatekeepers can order them from the most predictably successful to those who are less predictably successful understanding that we're not very good at predicting And the ones that are less predictably successful. Those are the less x anti promise Those are the ones that don't make it in the traditional world onto major record labels If they get released at all, it's on kind of in this independent fringe sector It's enormous, but it's a fringe kind of economically at least traditionally but Do they end up accounting for a growing and potentially high fraction of the realized exposed Successful artists that's going to be the the the test that I apply in these various industries Okay An anecdote is nice So let me talk about the the arcade fire album the suburbs So this is a arcade fire as a canadian band They're they're pretty popular with hipsters, but they're not very popular in a mainstream sense They don't get played on the radio in the u.s. Hardly at all So they they released their third album The suburbs in is it august 3rd 2011 it was critically acclaimed So the meta score. So if you go to meta critic, which is an aggregator of review information The meta score was 87. That's like an a pretty good grade They had very little conventional airplay if you look at the billboard airplay chart They never appeared on the on the billboard on the high airplay chart the one they released in billboard magazine On the other hand, they were pretty big on internet radio. Now here. I had to be creative I don't I was able to get data on last fm the top 420 songs per week at last fm By the way for a couple of years. I embarrassed myself by not understanding why the last fm ranking had 420 songs on it Anyways, enough said about that But what happened to this album? Well, first of all, so again, it's an x anti It's a low s x anti promise album even though they're They appealed to hipsters and you know the editors of pitchfork It's not a predictably commercially successful kind of a kind of an album But it turns out to be a winner on the sales side It's gold in the u.s. By now. Maybe it's platinum It also won the best album grammy in 2011 not the best alternative album No, it won actually the best album grammy. So it's kind of a an interesting story Oh, by the way, that's just a picture of their the the spins or streams of this at last fm Around the time that it came out okay Let me let me get back though to more systematic answers to the these questions about that I articulated. So is there a growth in releases? Yes, so I don't have the full neilson data But if you even look at billboard magazines accounts of the neilson data You see discussions of how the number of new products appearing in the data Rises from about 35,000 in the u.s. In 2000 to about 100,000 in about 2010 So more or less a tripling In the number of new products not the number of total the number of new products And I since then I've seen this in other data sources You can see this in the music brains data and the discogs data has clearly been an increase in the number of products released Okay That's all I'll say about that. I have more information but for today. That's all I'll say How about the changing Information environment? So this is kind of important because suppose you have a tripling in the number of new products How do consumers learn about them? You know if there are all these new products, but nobody can know about them It seems pretty hard for them to have any effect on consumption So what can I say here? I mean I should say at the outset I wish that I had you know super detailed airplay data And I wish I had the data on all the streams on pandora and spotify and easer Nobody has that except those private companies But I was able to find a few things So billboard as I mentioned releases an airplay chart the top 75 songs per week They call it the hot 100 or something, but it's only got 75 on it. I don't quite understand why And then there are so if you look at that you get like 3,900 listings per year Which is I think the product of 52 and 75 but only about 300 distinct artists because artists remain on there for multiple weeks All right Then if you compare that billboard list with the last fm list from the same time period What's interesting is the not that different total number of artists that show up there, but very little overlap And I'm saying that as evidence that the the new sources of information other than traditional terrestrial radio Are certainly providing avenues for artists who wouldn't be Promoted in the traditional way to be discovered by consumers Right, so there's more one would like to be able to say but certainly with what I could get my hands on one Can get that impression Here, you know if you're a music person you might find this chart kind of amusing Um, we have on one side the top billboard airplay artists that are not on last fm And these are sort of very popular well known, uh commercial artists On the right hand side We have the top artists in last fm in the same time period But who weren't in billboard and they're rather different a lot of it's sort of indie hipster stuff Some of it's not the Beatles obviously don't Correspond to that the left hand side one is so at the at the time I made this chart It was reminding me of there's a satellite radio station in the u.s. Called 20 on 20 Just the top 20 songs repeated every three three hours or something I I called it the soundtrack of hell My children enjoyed listening to it. I didn't anyways It's just a rather different just to give you a feel for what's being promoted through the different channels It's rather different All right second and still in the in the under the topic of Ways to learn about the new products. There's also been and this is also a digital phenomenon A big growth in both criticism and the dissemination of criticism So this picture is a little hard to explain. What is it? So if you go to metacritic and look at the underlying sources of critical information at metacritic So it's things like rolling stone pitchfork and a whole host of of sources of critical information and then you ask What year was that outlet founded? So then you sort of add up the number of reviews According to when the outlet that metacritic is drawing on was found on so you see there's a big chunk of reviews Attributable to rolling stone which was founded in 1967 or something But then you see that a lot of findings occur in this period And many of these are pure play digital products like pitchfork probably the most famous Which by the way was purchased by kandey nast yesterday or the day before So the uh, yeah the avant-garde becomes a commercial. I don't know. There's pitchfork So the point is all these are I think all of them certainly pitchfork and many of them Are uh, are free to access online. You don't have to go buy a magazine So there's lots of information out there that consumers can see and learn about these new products All right, so again, the big question is are there ways to learn about these new products? There at least are some new ways to learn about them Okay What's happening to the role of critics? I think that that question should have come before but let's talk about learning From critics versus radio. So now what I want to do with these pictures is think about the the albums or the artists who become commercially successful and the left hand side shows the share of the The commercially successful albums in the billboard 200 over this period of time 91 to 2010 That are heavily heavily played on the radio. See how it's falling Whereas on the right we see the share that are being reviewed in metacritic, which again is this aggregator Of critical information and we see you know a big increase So this is certainly consistent with people being able to learn about these new products via new mechanisms or new avenues Okay Let's talk about x anti promise and x post success in some ways As I said, this was the the most important direct test So to the artists with less x anti promise like who would not have made it to market at least not meaningfully Prior to digitization now achieve sales success And so here again what I'm looking at because at the time the only data I could get my hands on As sales data were the billboard 200 so just to be clear about that by week in the u.s It's the top 200 selling albums. Okay, so I can get that back to 1985 in a manner that I could afford And and then you can aggregate those up to the year level and and you can link them To to the to a separate billboard chart music on independent labels And so what I'm asking here to the indies account for a growing share of the commercially successful And the answer is yes if you look at the indie share among the billboard 200 It rises from about 15 to about 35 percent almost almost a tripling period between 2001 and 2010 Even if you look at the top 25 by week and aggregate those across weeks So the right-hand side picture shows that this phenomenon is occurring even at the top of the distribution Meaning that some of these x anti Just for shorthand sake call them x anti losers are x anti winners in a big sense They're at the top of the distribution that rises from about six to about I don't know 18 percent Right so that to me this is an interesting picture Because it says a lot of things that are starting out without a lot of promise In the new digitized world are ending up pretty successful Okay, yes, if you are a tom petty fan, then you'll like this quote even the losers get lucky sometimes So summing up the music evidence, I'd say a few things digital disintermediation If you wanted to use that word for this whole phenomenon, although we could fight about that word It provides a possible explanation for this increase in quality given the unpredictability more experimentation meaning bringing more things to market to see if people like them Leads to the discovery of additional good or successful music x anti losers become exposed winners and a lot of this stuff wouldn't have happened without Digitization is my argument Okay, but what about other cultural products? How are we doing on time? Okay, fine Okay, that's okay. That's perfect. So how about other cultural products books motion pictures television photography So of each, you know, I think you know the questions I'd like to ask I'd like to ask You know, are there more products? Are there ways to learn about the new products? Let's forget about sales concentration just for time Is there a growing success of the x anti losers? Are the new vintages good? All right, these are the questions now I should say that in these other media or these other Categories there isn't really a napster event It's not to say there isn't piracy. There is piracy But there isn't like a time when piracy began and became rampant So instead we might go into these industries thinking That we have mostly a cost shock And not mostly a revenue shock again, there is piracy plenty of it But I don't think there's like a huge change in piracy So you might even think of these contexts as contexts where we're seeing You know impacts of of cost reductions Distinct from a revenue reduction Okay, so let's talk a little bit about books So is there growth in new products? So it's it's actually a little bit hard to get measures of the number of new products There are a few sources like balker books in print There's a an almanac and if you get slightly different numbers from the different sources But they all show you pictures that look kind of like this. This is just one particular This is self published this one is increased enormously But all the indices that I can come up with That are measures of the number of new titles brought to market show pretty substantial increases At least in the u.s. I haven't had indices for other countries And you know at the same time there's a pretty big growth If you think about ebooks as an important phenomenon, and I think they are There's also a pretty big growth in the kind of ecosystem for consuming ebooks in the u.s It's largely kindle but also a few other reading products So there is this growth in the ability to consume as well as the number of new products that come into existence So yes, especially the self-published ebooks Supported by this diffusion of tablets and and readers Let me jump to what I think is the the most interesting test in the context of books So again, the big question is do do the new products that kind of wouldn't have made it past gatekeepers before End up being a big share of what it ends up being commercially successful So what these pictures are are pictures of the share of top Bestsellers that is the share of the top 150 by week in the u.s That are titles which were originally self-published Okay, so in the case of a book like 50 shades of gray It eventually is published by a conventional traditional publisher But it was originally self-published so it would stay in these data as a self-published or as an originally self-published book And so it's really kind of uh, I don't know I found this kind of astonishing This is a work with a former grad student professor at northeastern now We look at the share of these overall bestsellers And we find you know, it's a kind of it's typically it's kind of around zero until about 2011 Then it just starts to rise and it peaks above 15 percent. It's fluctuated a bit since then 15 percent of titles Are books that were originally self-published if you look at the romance category, which is the category where 50 shades lives It's 40 percent I mean, that's kind of astonishing These are people who you know, what does it take? These are largely people who who read a book rendered as pdf uploaded to the kindle direct platform And charge five dollars for it now most of these don't sell anything But a bunch of them sell a lot And it's it's it's kind of an interesting shift in that industry now again If i'm gonna use the word quality in the context of the word 50 shades Again, remember what I mean by quality is just stuff that people buy There's 50 shades, okay By the way, 50 shades is for separate entries on the list because there's it's a trilogy and a box set So there are four different ways for 50 shades to get in there. All right. So that's a quick quick jaunt into books Let's talk a little bit about motion pictures You know movies are are different and I mean my thought going into movies They're different in a couple of ways for one thing when you think about the movies that are economically most important Especially, you know from the us the movies from the motion picture association of america are are really big deal investments They you know, they they used to report a number every year the average cost of making one they were in last time They reported it 2003 it was a hundred million Per movie so that's a lot. That's a lot of money. And so you really would think that You might think things are somehow different here more over movies are kind of a An important export industry for the us and and you know people do talk about jobs even in this industry So I guess this slide is just meant to suggest there's even maybe more at stake in this industry All right. Well, let's talk about setting the stage here. So again, I don't think there's a big shock to revenue Of course, there are things that happen like dvd becomes a big deal than less of a big deal But there isn't and there's lots of piracy and much of the world But there isn't a Napster like watershed event after which revenue plummets So I think most of the interesting story here Is a story about what's happening on the cost side and here as in the others You could talk about production and distribution and promotion And so I'll talk a little bit about about those things on the production side Uh, we've seen a real growth in digital cameras, especially since 2005 You know, it used to be that the camera to make a movie you could release theatrically or for distribution It was a very expensive a pan of vision or something quarter million dollars You rented it and shot on film very expensive even, you know, just to make a movie of that sort very expensive Starting in 2005 you could buy a camera for about $2,000 that could produce Essentially cinema quality. Um, I mean people from the industry might fight with me about that But there are certainly many examples of movies distributed that are shot on, you know, the canon, uh, whatever It's 5d mark 3 $2,000 body Distribution well, you know, it used to be you needed to get into a lot of theaters make prints, you know, get Distribution deals. That's all very expensive But starting more recently a lot of movies are largely sold digitally So there are platforms like itunes netflix amazon There are actually quite a bit quite a few more platforms But there are lots of platforms that are involve, uh, less costly distribution amazon is essentially a passive player You take your movie and upload it there. If it's not porn, you can distribute it. They take 30 percent. You take 70 So again, it's it's not like having to buy your way into a theater Now promotion is really a challenge, right? So promotion is a big deal in in getting people interested in movies And so i'll show you some evidence in a few moments about about this Lots of movies that are being reviewed as a pretty rich information environment in this context Um Both professional reviews user-generated reviews. So all of this again, this is all just suggestive until I show you data But all this raises the possibility of first a bunch of new products and second Products that might actually be discovered by and useful to consumers. So is that is that going to turn out to be true? All right, so on the production side and I kind of mentioned this already but um digital cameras have been around For a little while and they've been widely adopted since Well in some ways since about 2005 but certainly since more recently than that And if you think about the price of the hardware needed to make a movie And I realize hardware is a small component of the hundred million dollar cost of the big deal movies but The price of sort of entry into being able to make a movie period has just fallen enormously And it creates big opportunities for especially for independent filmmakers, right? It doesn't really change things if you're thinking about spending a hundred million But if you're thinking about making a movie with three of your friends As uh, as some, you know, relatively famous artists now do it changes things quite a lot All right, this is uh, so imdb the internet movie database is quite a quite a neat Database has all kinds of information one of the things they have Not for every movie, but for some movies is the kind of camera used to make the movie Okay, and it's so it's pretty neat, right? It's pretty neat now It's missing a lot. So I don't know how representative this is but certainly this is just I've claimed that has been an Increase in digital the use of digital equipment to make movies and the digital The digital ones are the ones that are growing a lot towards the end of this bar chart. So just wherever you see things changing That's that's digital. Okay. All right Okay, now having said all that Let's think about the number of products. It's in sort of a nontrivial question Um, what is the number of products if you think about the number of products? Released by the motion picture association members. So these are the wide release movies produced by one of the member studios The answer is that it hasn't increased. It's kind of even going back to 30 years. It kind of fluctuates 150 200 Lately it's been around 150. So it hasn't gone up. That has not gone up okay But a very different notion of the number of products Is the number of movies that have an imdb page Okay, and so imdb again wonderful database and this gets owned by amazon by the way You can query the database and say show me how many movies with us origin were produced in uh in 1990 Or nine or 2003, you know just and you can ask it those questions And so what these pictures are we have us origin features us origin documentaries Now I have to concede it's possible since imdb is largely user generated and imdb has only existed since Early 90s It's certainly possible that their coverage of old movies is missing. There's some bias upward But they've existed since 19 again 95 and the period I really want to focus on is more recent And what you see is a very substantial increase in the number of movies that have an imdb page Especially since about 2005 So we're up around 3000 in the us for features another 4000 for documentaries And this is uh, I believe this is right. This is the whole world So the point is it's not just the us the us isn't doing all the work in these pictures It's not just a us phenomenon But you know having said this, you know, if you've spent some time at imdb, you know that a lot of those 3000 movies aren't really economically important products. All right, so I don't want to leave it at this Let's talk about some more stuff So let's think about now the the the gulf between the large scale wide release movies And and others so what this picture shows the bottom two lines one is the mpaa number again And then the dotted line is the number of movies released in the us On 500 or more screens. So that's kind of a threshold for wide release Again, the number of wide release movies isn't rising But the number of movies that show up in theaters in the us Up here is rising. So there's a lot of narrow release movies. Okay, they might be released for a very brief period of time In not very many theaters You know, you read the industry press, you know, a lot of this is about getting a review And then distributing it via itunes hbo or whatever. Okay, but certainly this is another notion of what's a product These are surely products people mean to try to sell because they're putting them out there and and trying to sell them What's another way to think about this? Uh If you think about the movies, this is actually a bad picture for what I want to try to say But let me let me focus on the year 2010 So in the year 2013, I asked the following question How many movies that had been released originally in 2010 Had been commercially marketed and via what via what sort of channels? And so again, the the number in theaters is about that 550 number We already saw that number. But then if you ask, well, how many 2010 vintage movies were there at netflix in 2013? and the answer was uh I don't know. What is that about a thousand and as of 2013 how many vintage 2010 movies were at amazon So amazon instant not prime prime is the subscription service instant is just the a la carte service And as of that time it was about 1200. I suspect it's more now. It's really growing. They have like 40,000 I don't know a really large number of movies in the a la carte service at this point okay So so I that's a lot of numbers my point is there's been an increase in production including And and an increase in the number of movies that I think are purposefully Marketed to consumers and available to consumers right, you know, even if you doubt the 3000 number At imdb. There's certainly 1200 is bigger than 150. Yeah, okay So now let's talk about can consumers learn about these products. So this is generally the product discovery problem So there's been significant growth in review provision and availability And it's really there's a range. I'd say professionals plus amateurs providing information to others So what do I do here? What I've done in this picture is to try to think about the top thousand movies by um By by by time And ask how how probable is it that the movie is? Has a review. Okay, so I'm not saying this very well. So if you look at the the top Over here, what this is saying is this is like the top movies that have and these movies have Always been ubiquitously reviewed in traditional outlets Okay, but if you go back now, let's jump down to here. These are movies that are way farther down the distribution In 1990, it's essentially none of them Are or yeah the 1990 movies there's very little information about them But as you come closer to the present You can see that even movies very far down the distribution And this is the distribution measured by the number of imdb users who rate the movie Right because that is available for almost every product Whereas box office revenue by construction is only available for a small number of products The point is these rising bars are just telling you that the depth The distance into the distribution That movies have information available about them. Sorry for that bad sentence is rising Okay What are some other ways to think about this if you ask Um, the number of movies which have an imdb user rating now to have an imdb user rating You need to have at least five people rate you otherwise. There's no rating. All right So you can just see across vintage as of 2013 The the number of movies about which there is a rating is rising All right And this is yeah, maybe this is overkill, but take a particular movie argo kind of a popular movie a few years ago And let's just think about how many how many reviews were there for argo. Well, there were 588 reviews And again this imdb has this very cool You know you kind of click down and you could find all the different sources that gave a review And so this these range from very big outlets to sort of obscure blogs The way i'm characterizing this is to take the alexa traffic rank of the outlet So the outlet might be the new york times Or it might be Joe's film blog i'm making that up and joe's film blog might be ranked 2 655 You take the natural log of that number you can kind of put the histogram in the same picture And i'm just saying there's an enormous range of outlets that are providing information about In this case a pretty prominent movie So just gives you a feel for for what's out there okay Oh, and i should also say Although these aren't all professionals These are a notch above simply people who are giving a star rating at imdb because these are either actual public editorial outlets or blogs that have enough of a following that imdb would link to them Okay, so this is not just the same as as total kind of wisdom of proud stuff All right, but let's get to the the important question. Do the independent movies succeed So again the question i keep asking in these industries is do the products that begin with low x anti-promise Become an important chunk of what's exposed successful Well here there are a lot of challenges in deciding what is it that constitutes a low x anti-promise Product you know we could just say the word independent, but even that word has a lot of meanings And so you know if you look through Different sources in the movie industry, you know you get this this word means different things So one definition like the independent spirit guys They have a definition that essentially says well you can be made by an independent or a major as long as you have independent vision So that's kind of the i know it when i see it definition Um then indy wire so i discovered indy wire Which is a very nice magazine or website that actually has a chart showing the box office revenue to indy movies But then i i realized their definition is released on not so many screens So that's kind of like movies predicted to be unsuccessful. So that didn't seem like a great definition So what i went with is movies not produced by one of the major studios as a definition of independent Understanding that this is a complex environment All right, so what does one see well indies are actually a growing share of both box office revenue As well as uh dvd revenue now this is kind of surprising to me. I don't have a deep database of dvd revenue I just have like i think it's the top 30 by week for a bunch of weeks I would have thought that was dominated by non indy stuff But it really that the share indy is is both kind of significant and growing especially recently even among Uh in that distribution channel Uh, yeah, if you look I apologize for all the data, but i kind of i'm enthusiastic about it This picture shows you the share of indy stuff Among what's being distributed through various channels across So it's kind of hard to even say that but if you look at amazon instant, let's say the red one and you ask as of 2013 when i made this picture and you ask of the stuff at amazon instant That was originally released in 1980 What share of it's indy and how does that compare to today and you see a rising share of these movies that are available for sale That are indy now in some sense that has to be true Right because there's a growing number of indy movies as long as they're As long as they're made available for sale given that the major stuff is pretty flat That has to be true. So that can't be surprising Okay, are the new movies good? All right, so there again two kinds of approaches to this based on critics and usage So let me start with the one that's that's sort of easy to understand rotten tomatoes is a pretty neat website Like it's an aggregator of review information that turns critics reviews of movies into two-digit numbers And so what I did here is I just took all the movies back to 1993 that have a score of 87 or above And you just see the absolute number of movies 87 or above is rising Okay, so the number of absolutely good movies according to the the the elites who who you know who make these decisions Is rising the absolute number of movies going up Now more over so that by itself that's kind of good news, but it's not clear that it's my mechanism It's only my mechanism if these are being driven by the indies And so if you ask the share of the rotten tomatoes movies that are getting great grades How many of them are indy? Oh, it turns out it's a it's a big and rising sharply rising share Okay, now again, you know, you could say that Yuri dreams of sushi is not an economically important movie Because it's we don't know it's box office revenue I don't think but if we knew how many times people streamed it at netflix we could say a little more, but we don't All right, one other point to make though you know, so I keep kind of dumping on elites, but In some data sets you have both the views of the lay people and elites So at metacritic you can get a user score and a critic score for every movie And what this picture shows I mean at one level it just shows a cloud But you know if you're a statistician, you know, it's a highly correlated cloud And so the movies that elites like more Regular humans like more too So that suggests some economic Consequentiality or kinds of that's a word consequence of of of the elites liking things more All right now Okay, yeah, so this is a perfect time to not talk about this next slide because it's a little geeky and It's all good news it turns out tv is under the going of renaissance if you'll forgive the expression and Yeah Yeah, and so is photography That shouldn't be too surprising, you know I mean photographers are pretty upset about how cheap their competition is But the number of say the number of photographs at getty by vintage is just absolutely exploded Because it turns out that even I can take a good picture if I shoot 5000 of them All right, so this is a good way to kind of slow down New digital technology has brought threats to creative industry piracy for sure. I mean and it's a realized threat in music I would be the first to agree that piracy is largely if not completely responsible for the The collapse of recorded music revenue But it also has brought opportunities through other aspects of digitization not the piracy per se That is not good news in and of itself There's been huge growth in the number of new products in a lot of industries as well as Though, you know, the number being distributed and a lot of the new products You know and when I say new I mean a lot of products that wouldn't have made it past the sort of traditional intermediaries Make up a large and growing share of what turns out to be important and successful So threats to revenue, you know, they're real and so I don't want to dismiss that that stuff's important But when we think about what How I think one should assess whether copyright protected industries are working one should assess it according to Whether there's a bunch of good new stuff being created. That's useful to people And here we see no sign of diminished output in these industries And we see that that a lot of the good in in many senses The new works are good if not better So I think things look pretty healthy Now what should we say about public policy and I don't want to over interpret anything But let's just talk about what things are potentially relevant So, you know, rights holders are concerned about revenue and that's fully understandable But I think, you know, copyright exists to provide incentives for creative activity And despite the revenue performance in music actually we didn't talk about newspapers But if you look at revenue pictures for newspapers, they're just as awful as the revenue pictures if not worse revenue pictures in Music So despite those really horrific looking pictures and I think well-founded fears in other industries There seems to be no crisis at all in creative activity So that's where I'll stop Thank you Joel and I really hated to to cut your shoulder Even I was tired of hearing me. It's good I think you've done an outstanding entrepreneurial effort to look for data. Um, you know that can say something intelligent about Um, you know how creativity has evolved in these various content industries I have one question if you allow me and then I'll open to the floor And maybe based also on your conclusion I think you've made a fairly convincing case that There is not fairly you made a convincing case that there's no crisis in creative activity um Now as an economist, you know, one of the things that you know has Recently recently received a lot of attention, you know, is that the the the evolution and the distribution of incomes and you know creativity and And sort of how much people earn from from from from creativity doesn't necessarily correlate with one another And and you know, there has always been the case in the creative industries that you know, a few people get very rich and most Don't But is there any evidence and I know this really goes beyond the research that you've presented today But is there any evidence on how the income of of you know, first of all the musicians But also those more broadly involved in the creative industries has evolved Yeah, that's a good question Not that I've done I mean, so there are people who've done research on the evolution of revenue sources in music, for example And I think they make a pretty convincing argument that uh revenue from live performance has become more important ticket prices have risen And but but more importantly revenue from live Performance has risen and that seems like a natural A natural consequence of sort of losing control of the of the of the recorded music product But you know, the there are a few folks who've studied like peter dakola at northwestern has a study of income sources. He surveyed musicians And trying to see the relevance or the importance of copyright as a as an actual or copyright as a direct revenue source for them So there is some work out there. I don't know much about it I will tell you that there are a lot of artists including artists I know and and you know respect and find interesting who are pretty unhappy Right, so there's a lot of grumbling out there Which stands I think an interesting contrast to what I see as a creative flowering But nevertheless, there's a lot of grumbling and there's I mean look at like spotify. So spotify In one sense is providing revenue in contexts that where there wasn't revenue before but a lot of artists look at it and say This is terrible. This is not enough compensation um So there's there's a lot of grumbling about income Which is not really a very deep answer to your question except to acknowledge that it's an important one Okay, please and maybe if you could identify yourself when you pose a question Thank you costum. My name is Benoit Miller. I'm an attorney in Geneva And I have several clients in the content industry and that for many years although I've also worked for the IT industry You know of interesting Presentation. Thank you. I have a lot of comments and questions and certainly I would also Equal the the comment by the chief economist On you know, I can report some anecdotal data From film directors whose uh, for instance from Ghana whose film series television series has been viewed more than a million time On youtube and the revenue she got from that is less than 2000 us dollars So, you know to your question as to whether copyright works for Creativity, I would put some question marks. It works for very large distributors Yes, but does it work For those who actually create I would also Question your definition of what is a product. I think that's a bit, you know, at least to me It's a bit confusing as to whether you are referring to the underlying work or whether you're referring to a physical or immaterial copy of that work and it's economic value And certainly point out that copyright Has an incentive to, you know, get more works Out there in the world But but I think it's it doesn't quite fit with The very heavy focus on the supply side of your research And I would You know compelled white bow. I I think attended all of your seminars Certainly in the field of copyright To think about You know in the future featuring research that looks more from the perspective of creators And not that from, you know, big companies We know are closer in Zurich or, you know around the world But the view from those who create and not just in america I work a lot with creators in africa and I think they have a different worldview. Thank you Should I um should I respond? Oh, yeah, so no, thank thank you for those, uh, those Oh, sorry, I was I was still I'm sorry. Um, so thank you for those comments I mean, there is a really interesting question You know, what does it mean for copyright to be working because I certainly again I hear from a lot of artists who are grumbling and saying it's not working, but um, you know There's this new thing on online education which threatens my income as a professor And yet people might get educated by it and I might say it's terrible, but people might get educated Is education working if people are getting educated, but I have a lower income Probably it is and so with sadness. I might have to say that if a creator says I'm I'm I'm upset. I'm unhappy, but nevertheless we as consumers Observe that he or she is creating a bunch of stuff By their behavior Copyright is working And so I don't know we can ask people how happy I mean certainly whether people are happy is important Because they give voice politically to their through, you know, to their happiness or unhappiness But I feel as though the evidence about production And stuff made available and when possible, I agree with you consumption is very important when one can get the data If we see stuff produced and consumed Then that would be a pretty good way to think that things are working Yes, please Thank you. Oh, sorry No, no, please. Thank you. Whichever Why don't you go ahead ladies first Microphone works. Thank you. My name is capri lebe. Um, I wanted to say that Of course, it's very interesting to see Who after all the statistics that we have seen in the presentation and the word revenue came up quite a number of times As the gentleman next to me said who is really Earning revenue in the context. Is it the creator the musician the The filmmaker or is it the producer and the industry? And my point about the copyright in the existing form in our traditional way we use copyright has to obviously change in this new digitalized air And I I'm sure that from our point of view at wypo that there needs to be a lot of adjustment made To find ways to protect the artists In a new revolutionary way not in the old existing traditional copyright law So may I say something about I actually I I agree that there's a lot of interesting questions about our kind of creators versus intermediaries And and let me answer the question not with an answer so much as a puzzle So a traditional author contract What is it 15 of cover price, you know, seven and a half percent of wholesale price And then the presumption is the publisher will do some promotion and maybe I'll get some money If I if one were to publish a book through amazon directly The structure of the contract you set whatever price you want you set a price above five dollars you the creator keep 70 percent Now Some and it's pretty similar, you know the itunes contract pretty similar other contracts are similar Now the question is Is this good news or bad news for creators if they can It's essentially disintermediated. I mean there is an intermediary, but it's a passive intermediary Is this good news for creators or bad news for creators? In some sense, it's great news But then one wonders how consumers learn about the product and to think about that is it was the guinean movie at youtube that Maybe amazon would have been a better place for that movie. I don't know. I don't know but yeah So i'm sympathetic to the concerns and I think those are good questions And one of the big problems is A lot of the intermediaries don't share any data You know, we don't know how many times things get consumed We don't you know, we know contractual revenue shares, but we don't know how many times things are being Purchased or downloaded so we don't really know how different entities are doing in the whole kind of ecosystem So I just I share the concerns, but I don't know it's hard to say much without more data Thank you. Thomas still on wipo And I confess that I used to work for the motion picture association. That's okay. I love motion pictures. Thank you very much It seems to me that we care a lot about The consumption of works Not so much about the production of works and I have a feeling that many academics are misled by the copyright clause of the american Constitution into thinking that the incentive it's a it's a sufficient explanation or even Satisfactory explanation of a copyright that it's an incentive for production I think that the self publishing on on amazon indicates very clearly That the real point of copyright is to permit distribution not production poets will write poems And 99 percent of them will never be published and to that extent art is advanced but As economy and most economies well, certainly all politicians are interested in the size of the economy not its quality so I think that when you you talk about production what i'm missing is some control for price and some some Assessment of of relative revenue I mean against what they used to be because you don't control for the shrinkage in the market not nor in your graphs Nor was there any place to that wasn't what you were trying to say But in fact if we if we shrank if we could control for revenue the actual size of the whole graph will be much smaller now Even though the relative pictures have changed So the second thing is I think that the danger with your conclusions Is that it encourages the idea that there's a quid pro quo that should be accepted by the The copyright industries they should say well, you know We'd have been twice as wealthy now if there were adequate enforcement provisions But hey, it's cost less to make the stuff So we're we're doing as well as we were 20 years ago We should be happy that is not how any normal industry operates You've got to compare against what it would have been or what it would be in a series of different states of nature It seems to me and finally a more sort of technical point Which probably betrays my lack of understanding in relation to your measurements of quality You're trying to tell us whether something is is in an objective sense good or not Sufficiently objective us to make policy decisions about it But isn't the problem that if you if you take critics Critics can only tell you their assessment of of 2015 critic tells you his view of a 1970 work Or modern work, but in order to have an objective Measure of Quality you would have to know what the 1970 critic Thoughts not only the 1970 work, but also the 2015 work. I don't think that data will ever be available So these are great points. I'd love to just uh to speak to these points. So thank you You're right. So two points about the size of the economy. So, um We what we care about Is just technically speaking is the sum of consumer surplus and producer surplus That's what we care about It is true and so and by that I mean, you know, how much would be people be willing to pay for stuff that they get? Less what they pay for it. That's the consumer side and on the producer side It's the revenue they get less what it costs to make stuff. So you could colloquially just think of it as profit It's clear in music that revenues weigh down. Okay, so the sellers are I mean revenues weigh down costs are down, but I would say that Revenues down more than cost. So I take the point about pain on the seller side But the point I mean We really we need to remember to care about the consumers and the producers And so it's not as simple as just saying that I neglected that revenue fell Right because consumer surplus ballooned So I think that's important a second related point is that On this quality issue really what I wish I could be measuring with quality is not Critical quality even though I end up relying on critics more than I in some sense I wish I did what I want to be measuring is whatever it is that animates people to wish to consume things And so when I when I have the right kind of data, you know, the consumption data by time and vintage I can use those data To to see which vintages are are more demanded given how old they are And I think that does get around the the real and you know genuine problem that you raise So I think it's a data problem that separates me from satisfying Of the concern that you raise, but I think I have the data in music And I had the someone on the slides I sped through with movies I'll just take a second to talk about was so in movies the way we consume Well the traditional way we observe movie consumption is at the box office So we never really observe much consumption of old movies. So we don't get vintage stuff However, if you think about the new world of streaming and digital, you know Distribution there's tons of data out there about consumption of old movies. I don't have any of those data I do however have a television listing data So I have like the all the listing or not all but I was able to get Uh, you know hundreds of thousands of television listings on premium channels And so with those data you can see which vintages movies are being used and it shows some promise in addressing And so I want to more show sympathy with with your your stated concern than assert something about the result Just to say yes, um with the right data one could study this if your friends in the motion picture industry Would be willing to share with me the streaming data by time and vintage We could do something together because I fully agree that that it's consumption You know, it's the the two surpluses the consumer surplus and the producer surplus that are important not the views of elites Those are really relevant only in as much as they can be shown to be correlated with the stuff that really makes the The the the gears go so that that's a thank you though. Thank you Hello, and my name is Mary. I'm Zehtoff tree. I'm a researcher in the economics and statistics division here in vital um Correct me if I missed it, but I have like two questions here the that you Touched upon quality and source of information Okay, so when it comes to quality you mentioned that Well, new works are generally good. How do you account for accessibility of big data? And what I have in mind. I'm looking at like Uh generation of original series when it comes to like house of cards and netflix and the behavior of the consumers and all those Statistics that they they had access to when it was not available. Maybe 10 years ago 15 years ago The second question is about source of information When you talked about the source of information were in music in film Was made mainly through reviewers And the distribution channels How do you account for the emergence of social media? And the word of mouse in the modern way sense of Yeah Yeah, yeah, so no, thank you. Let me just talk a little bit So the big data it is said that house that netflix developed house of cards You know making use of a lot of data analysis that helped Make it predictably successful or they knew that they should cast kevin spacey and so forth And and that may be true and I it's related to a phenomenon and I've I've heard uh venture capitalists tell me about projects their funding that supposedly can predict which music will be successful now to the extent that there are such technologies or methods for for Lessening the unpredictability problem. That would be great news I mean and maybe so it may well be I don't actually doubt that netflix was able to make use of data To predict I mean imagine if you could not have to produce the 90 percent of things that nobody wants That would be really useful to society. So I I'm I'm interested in that I don't I can't account for it But I just think it's a very interesting research question the other thing on a social sources of information and social media I entirely agree that It's quite likely and certainly possible That social media provides a way or provide a way of propagating information about about new products I haven't studied them, but I think it's very much worthy of study It's kind of true questions, but No, it's under the table. Yes, so being quick. Thank you very much for Your presentation in the fantastic work of data. So I feel bad that I'm asking for more But but one thing I missed was the discussion about concentration Or discussion about how many suppliers do we have in different moments of stages? Implicitly, we know because of your data that it seems that there are more producers Upstreams, right? There are more poets as they were mentioned there, right? We at least is indie I'm ready to believe that part, but I don't know how many intermediaries are now And I don't know how many distribution channels are now And some of the information that we get, I don't know how, you know, this might be Being ignorant, but the fact you have iTunes, Amazon, it seems that that has shrinked a lot Compared to how many shops you had before to buy DVDs or books and all that maybe maybe You know 2.0 fast and we get the impression of this is not real, but but maybe you know some literature I would like to know more about this. I think this is an important thing The other and very good question is I found really interesting that you went for music books and movies so very different pieces and Very different moments of times and I appreciate how you said that you had Napster in music And you don't have an equivalent for the others and I was thinking if there is anything Of that sort that provides an interesting maybe experiment Because at least from my perception not being an expert in the field It's like you had Napster and similar things paving the road For for iTunes, right? Well, and maybe for movies was the same. I don't know mega blow the this kind of things Maybe they paved the road for the Netflix There was a demand that wanted to be digital there Way before the more formal channels or let's say the legit channels were there However, for books, it's not the same right? We had amazon selling books physical books and then amazon went for the for the ebooks And and their own tablet system. So I was wondering if there's something special Yeah, no a lot of interesting stuff there the concentration of intermediaries is a really interesting question I've been looking at it. I haven't really done a lot of systematic on think about books You know amazon is really dominant in books in the u.s. And so It interestingly they remain a pretty passive intermediary in the sense that you just upload your book you get 70% But there are questions about how you get it promoted You know, you sort of have to agree to be in the kindle direct program in order to get some So there is an interesting question about about downstream power In video In the u.s. We we have right now we it looks like about 15 different platforms Although a few of them seem much more dominant than others. I mean netflix is by far dominant in the subscription space But in in the a la carte space, there are there are a bunch But I think but the point I wanted to agree with you though is that I think it's a very interesting issue to monitor because that's Even if you had democratization of production if you don't have democratization of distribution then issues could emerge So I I just want to agree that that there's an issue there The one thing, you know, you raise an interesting question about naps are prompting the development of itunes I mean music Was the canary in the coal mine, you know that went through this first and and suffered and and looked stupid in retrospect But i'm pretty sure they're not stupid. I'm pretty sure it was just a hard problem It's interesting that television, you know, uh, youtube appears february 2005 Within six months the major tv networks in the u.s. Are putting some of their programming online By the next fall season everything is online So I think there's you know, if you think about sort of prompting events and learning It's interesting Do you just to do case studies comparing these industries learning from each other? Because they're not yeah music had it harder It was first Okay, thank you very much apologies for not Being able to accommodate all questions, but Unfortunately, we don't have any more time So let me finally thank professor wildfogel for an extremely interesting provocative presentation I think we we we all learned a lot and We look forward to to keeping in touch and learning about your ongoing research Thank you And please if email me if either if I if you didn't get to ask a question or if you did and I answered it Badly, let's let's have a dialogue. Thank you. Thank you