 Good afternoon, everyone. And let me welcome you to the last seminar of 2012. Today's seminar will focus on copyright. It was great because I think most of the other seminars that we had this year were on patents. We're very fortunate to have Professor Petra Moser with us from Stanford University. And she's going to present on the effects of copyright extensions on the quantity and quality of literary output. Petra is known for her work that she has done in the area of economic history. And in particular, she has used policy experiments that have happened in history to analyze the effects of intellectual property rights protection on economic performance. I think we generally know that there's a lot we can learn from history that applies even more so in the economics profession, unlike in physics or chemistry where researchers can undertake experiments. This is not something that generally we can do in economics, although they are also lab experiments. But we have history. And history sometimes does offer very useful experiences or very useful experiments that one can use for research purposes. Just one note of one biographical note. As I said, Petra comes from our teachers at Stanford University. She has written her PhD dissertation on the determinants of innovation evidence from 19th century world fairs. She has done a lot of work in the area of patents. So we are quite pleased to see that she has branched out to copyright, which is of a lot of interest to the World Intellectual Property Organization. She also had a stint at the Massachusetts Institute of Technology. And then she went to Stanford, where she is an assistant professor. And she really has published in some of the best economics journal. And as I mentioned, it's really well known for applying the methods of economic history to questions of intellectual property protection. So with that, I turn to you. Thank you, sir. Thanks for this very nice introduction. When somebody introduces you like that, you always feel like, who are they talking about? So thank you. So today we'll talk about copyright. I think it's funny that you think I'm branching out because it didn't feel like so much of a branching out at first because I thought I would still talk about intellectual property rights, which I thought was basically all the same. But copyright really is very, very different. So we're just starting this work. So any comments you have will be really helpful. And so we're taking it in small steps. So today, what I will show you here is just really a very limited study, very targeted study of an extension in the length of copyright on the price of books. And this is the price of books in the romantic period, which is from 1719 to 1840, as I had to figure out myself. So the big policy question that we're trying to address here is whether copyright should be more short-lived as they are at this point. So copyright does, by patterns, provide a temporary monopoly right to a specific work. In the US, this would be under code Title 17. It's for literature, music, and software. And so today it's life of Arthur plus 70 years. So it's quite long. And the argument for long-lived copyright is that it encourages creativity by increasing the incentives to invent and to create these products. And so I will not get to this specific part today, but so one part of the argument I think that we want to explore in our future work is whether copyright actually also makes it possible for a new group of people to enter these creative professions. And so one thing that's actually very relevant for the topic here is that the 19th century was a period in which the status of a writer changed. So writing was novel writing, was a scribbling profession that was pursued mostly by women. And then in the 19th century, you have a switch where more men move into this profession of writing. And so my intuition is that although I will go nowhere in actually showing this in this paper today, but one of the things that we want to look at is whether actually increasing copyrights, providing economic incentives for writing, makes it actually possible for different people to enter this profession. But for that to actually happen, we need to see some effect of longer copyright protection on price as the economic variable that would actually drive an increase in the profitability. So in the literature, an increase in price obviously is related to both an increase in creativity by increasing the incentives, so increasing the profitability of writing. But price is also important because it means that it restricts access. So if a book is very, very expensive, and I'll actually show you that books in this period were extremely expensive, then whatever knowledge is in that book is going to be restricted. Only the people who can afford to buy the books are going to be able to access this. So those are basically the two mechanisms by which price is really important. That's what we'll look at in this paper. But as I say, I will only show you an effect on price. So I will not show you anything about the effects on quantity or the effects on new people entering the profession or even the effects on diffusion. But what we're really just trying to do in this first step is there an effect on the price of literature or more generally on, in this case, it's literature, but it could also be other copyrighted products. So the causal effect is very difficult to identify with modern data because there's lobbying for copyright. So it could be, for example, that we observe an increase in the price of a copyrighted book after we see an extension in the price of the, in copyright. But that increase in the price is actually what drives the change in the copyright rather than the other way around. So it could be because the price of the product is high so that it's actually really valuable for a firm to maintain exclusivity. And because of that high value, then that firm invests in having copyright changed. So then we observe a change in the copyright and we observe a change in the price. And the two are correlated, but we really can't say anything about what causes what. And so in the modern period, so we have the US Copyright Act of 1998. It's called the Mickey Mouse Act, whether this is for right or wrong, but in public debates it goes by that name because it was very strongly associated with lobbying by Disney. And so it's really hard to establish a causal effect for this. The strongest empirical evidence that we have is on piracy. And when I'm talking about piracy, it's really Napster. So we actually have a really nice series of papers in this literature that look at the effects of violations and copyright protection on sales and on counts or the production of new songs. They find there's really not much of an effect. So we look at modern music sales. They don't seem to be very much affected by the violation of copyright. And in some sense, you can think about violations in the length of copyright as just a really extreme change in the expected length of copyright. Because basically if your copyright is violated with some probability, that means that the expected length of copyright is going to be much shorter. So what we are looking at in this paper, we're looking at the effects of copyright on price. And so there's really not much empirical evidence there. And the empirical evidence that there is actually goes, is very counterintuitive. It goes in the wrong direction. So one set of results are about copy machines. And so what we see with the introduction of copy machines, which made it easier for libraries to essentially violate the copyrights of journals, we see is that the journals started charging libraries a higher price after they introduced copy machines because they thought, well, now they get more value out of that. And our intellectual property is violated, so we just charge them a higher price. Then the other data that we have on this is that pirated books by European authors sold actually for higher prices in the US than copyrighted books by US authors. So again, that effect goes in the wrong direction. And so in this case, it could actually argument as well, it's just the quality of the European books was higher. European authors just wrote better books than US authors, which is why these books sold for a higher price. So to be quite honest, I'm not super satisfied with either of these studies. I don't think they tell us too much because there's a lot of other stuff going on. And we need to separate that. And so that's precisely what we're trying to do in this paper. And so what we're doing here is in this paper, which is called Dead Poets Property for a reason. So we're trying to investigate the causal effect of longer copyrights on the price of books. So what we're taking advantage of is a differential increase in the length of copyrights under the UK Copyright Act of 1814. And then what we'll do here is we'll compare changes. Sorry, I should back up. So the UK Copyright Act of 1814 introduced a non-obvious change in the expected length of copyright protection that favored books by dead authors. And so what we'll do in this paper, we'll compare changes in the price of books by dead and living authors before and after 1814. Because only one of these type, the only books by dead authors benefited from the increase in the length of copyright protection. But all books, all books by dead and living authors were affected by a bunch of other things that may have also affected the price of books. So that's why, so by this way, we're trying to isolate the causal effect of longer copyrights. And so to do this, we've collected data on more than 1,000 book editions between 1719 and 1840. So what we find here, just to preview the results, is that unlike the existing literature, we really find there's a very substantial increase in price in response to longer copyrights. So we've seen 8% increase in price in the price of books for each additional year of copyright. And that is controlling for the age of books, controlling for authors, controlling for changes over time in production technologies through time-fixed effects, controlling for literary quality. I'll show you our controls for literary quality. They're quite bad, but we control for it in some way. And then controlling for genres. And then so one thing that you're always worried about with these type of studies, well, could this be by changes that were independent of the copyright changes, but they happened actually before. So we're controlling for that too, and the results don't go away. And then we also look at, so in our placebo analysis, we also look at books by dead authors after 1814 that were not affected by stronger copyrights. And for those books, we don't see an effect. And then the other thing, the very last test that I'll show you is that if we actually believe that stronger copyright or longer copyrights increase the price of books, what we should also see is we should observe books to become cheaper as they approach the end of their copyright terms. And we do see this in the data. So now I've basically told you all I'm going to tell you, and we'll go in some detail down the road. So the UK copyright law of very short overview from 1710 to 1814, and I think my understanding is that a lot of you actually know a lot more about copyright than I do. But I will just tell you about basically very short review of UK copyright law. So copyright was first formalized in 1710 in the Statute of Anne. And originally, copyright was modeled after the 1623 Statute of Monopoly, which had a length of protection for 14 years. And that was essentially just intended to allow the training of two, well, it's an I can't remember the English word for it, two apprentices. You want to train two apprentices to produce a new technology. So they say one takes about seven years to train, so we want to train two, so let's make it 14 years. Then between 1710 and 1814, publishers were required to provide library copies to 10 university libraries and the British library. In 1798, there was a decision in Beckford v. Hood which said that the copyright is enforced even if the publisher has not provided a library copy to these 10 lending libraries. That called into question the deposit requirement and then a little bit more of a history, grand history tour. So in 1814, 1815, there's this period of lots of wars. And so you had restrictions on the price of paper. Paper was really expensive. The library copies were supposed to be of highest quality paper. So the requirement to deposit these books was actually really expensive for publishers. So they argued that this requirement should be abolished. And in English Parliament, the debate in part as a result of lobbying by universities counted this argument saying that deposit of these books was very important to promote learning. So you have these two opposing views. On the one hand, the university libraries wanted to create a record of all existing knowledge and make it available to scholars. And on the other hand, you have the publishers who say, hey, this is just too costly, especially with the high prices of books. And that's really the background for this act, the 1814 act. The 1814 act was mainly intended to reaffirm the requirement to deposit books. But the publishers got is they got a relaxation of the requirement to print books on best paper, which was this very expensive paper. So they only had to deposit one book of best paper to the British Library. And then were allowed to publish all the other books, submit all the other books on regular paper. But what they also got is they got this change in the length of copyright. And at the time, when you look at the records in parliament, nobody really cared, or very few people cared. So the publishers did not say, hey, this is really useful to us. This change was introduced in the very last minute of the debates. And so in some way, this is good for us as sort of an experiment, because that means that it wasn't. It's very unlikely we have been driven by the interest of the publishers. And so now let me actually just explain to you how the Copyright Act influenced the length of copyright. We'll just get some water. So for authors who were alive, the act of 1814 increased the length of copyright protection from 28 years to 28 years or the life of authors. But when you actually look at life expectancies in the 19th century, they were quite low. And they were even low for authors. And so what that mean, that the life of authors actually did not get you a real extension. Because the average author in our data set was 42 years old at the time of the initial publication. And they had a remaining life expectancy of 28 years. So for life of authors, this really didn't mean a real extension. For books by dead authors, however, the act extended the length of copyright from 14 years to 28 years. So that now, rather than having 14 years before the act, so they get a full 28 years regardless of whether the author dies or lives in this period. So that means that we have a differential increase in the length of copyright protection for books by dead authors. And that is really the key of the analysis. So anything that we do here is based on this. So what it means is that we have two types of books here that are differentially affected by an extension in the length of copyright. And what we'll argue and what a lot of the rest of this analysis with the rest of this presentation, the paper is about, is looking at alternative facts that could drive the effects that we see, which is this increase in price. So I'll show you the data set. So the data that we've collected is the price of 1,072 editions printed between 1719, 1840. And we got these data from Sunclear, that is a literature, it's a literary history of reading in the romantic period. And then we also got additional data from catalogs, from basically book catalogs, price catalogs of books in the 18th century. And so the first thing that we see here is that books were actually super expensive. So in the mean price of a book in our data set is about 17.7 shillings, which is very high compared to a weekly wage of nine, and very exceptionally, 40 shillings per week of a worker. So basically what this means is that the books in this data set or any book in this period, it's not something like you go to the bookstore and you buy something to read. So the books the re-reading was done in this period is that one person of the household would buy a book and then everybody would sit together and read the book together in a group, or a library would buy a book. So books at this time, a lot of it, they actually also used the status symbols. So if you have, say, you are a lord of something, then you have a nice library that you can show your guests. So, oops, OK, so now these are sort of the control variables in the baseline. So one of the funny things why I like seeing this picture because one of our co-authors actually thought this was a literary critic. So it just shows you how well-protected we keep our graduate students. So this is a picture of Boris Karlov in the Bride of Frankenstein. And it's not under copyright, so I haven't violated any laws now because it was published without a copyright notice. So there you go. So what we're going to control for, we'll control for the age of the book. And so why is that actually important? So it's important to control for the age of the book because you may think that books at different ages price for different, they were sold for a different price. So say a book, a first edition of a book might be much more expensive. Then we are also controlling for author fixed effects. So say a really popular author, say Sir Walter Scott, his books may sell for a much higher price. So we are controlling for that as well. And then we're controlling for time fixed effects. And that's really just a way to control for technical changes. So if you know a little bit, sort of, in 19th century history, there are a lot of advances in the printing press, which made it actually much cheaper to print books. And so we're controlling for that by time fixed effects. So we have year dummies in there for every single year. But then we also control for it again through our different analysis. So we have this control group of books by living authors. Then we also control for literary quality. So this is my, I don't know much about literature, especially not about American literature or English literature. And so what we're doing is we're using the Western canon, which I think my literary, more literally inclined, friends say it's a terrible, despicable list of books, of Western books, of canonical books. And they say, hey, first of all, you shouldn't have a canon of books. And then it shouldn't be Western. But that's what we use. So we're using books that were defined by literary critic at Yale as the canonical books, so important influential books. And so if anybody has ideas of how to better control for literary quality, we're open for that. But that's what we're doing. And then we're also controlling for genre. And so that's actually important for many reasons because, say, poetry in the 19th century is a very different sport from novel writing. So people write poetry because they think they're intelligent and they would like to show it off to their friends and are very often self-published. So Lord Dudley may write poetry because he's a lord and then he's going to pay a publisher to publish it. So that obviously going to have a completely different incentive structure from, say, a novelist who writes because they want to make some money. So are we controlling for that? The other thing, obviously, it's size, right? Novels are usually thicker, so they cost more to produce and you may also be more willing to pay for them, pay more for them. And then we are also controlling for page counts and page size. So page size is important because it influences production costs. And it also, page size almost functions like a paperback of a version. So for example, we have a quote there by Lord Dudley. He was waiting for a book by Sir Walter Scott and said, OK, the new version is just too expensive for me. I will just wait till it comes out in the cheaper Octavo edition, which was a very standard size and it was a little bit cheaper. So we're controlling for that as well. But all of these things, they all restrict our data sets. So the main tests that I'll show you, they'll be just based on data without page counts and page size. So then let me just start by showing you basically a difference in mean. So what we see is that the books by living authors become cheaper after 1814. And that could just be because the price of paper goes down or because printing actually becomes cheaper because of the advent of Stanhope's printing press. But then the other thing that you see is that books by dead authors go up by a lot. So the average price of a book by dead author before 1814 was 17.7 shillings. And after 1814, the average price was 33.4. So that means we have about a 15 shilling increase, which means it's almost double, so it's about 80%, 90%. And then the difference here is 16%. So you can really already see it in the means that there is a differential price increase in books by dead authors. So books by dead authors after 1814 just did become more expensive. So then I always like to just show you the raw data, though this is the raw data that we have. So on the x-axis, you have years. The vertical line is 1814, which is the advent of the act. And then the crosses are observations that we have for the price of books by dead authors. So these are at the addition years. So that means that some of our observations here are for first editions and others are for later editions. And we're controlling for that, but we're including both types. And so essentially what we'll do is we'll look at, so in some of the tests, we'll look at the changes in prices of books by the same authors before and after 1814 and also before and after the author had died. So essentially we're looking at books by the same authors in some specifications with shorter and longer copyrights. So in the regression, what that tells us is that after 1814, books by dead authors become about 112% more expensive. And so that is controlling for the book age fixed effects and controlling for the age of the books and controlling for the identity of the author. And then when we control for quality, which is really important, is that the estimated effect of copyright actually doesn't really change. So remember what I told you before is that in previous studies, people have said, we have this price difference, but it really is probably just driven by quality. So we throw in our actually imperfect, but we throw in our measure of quality here. It does not affect the results. And then what we also hear in this specification we put in the drawn effects effects. So we are controlling for whether the book is a novel or works of poetry or hymns or other types of nonfiction. And so that actually does lower the estimated effect a little bit, but it's still very large. So it still implies a very large significant increase in the price of books with stronger copyright. And then we're taking out the controls for authors. Because essentially, so what we do in the regressions before in one, two, and three, as I described a little bit earlier, we are comparing books by the same author before and after that author had died or before and after that book had gotten an increase in the length of copyright. So that actually takes out a lot of our observations because some of the authors only have one book. So we're taking that out so that we get more, basically more power, so we get more observations to estimate the effects. And so what you do here, then the effect actually becomes larger. So when we take out, when we don't control for the authors, the identity of the author, the results becomes larger. And then so whenever somebody shows you what we call the difference in difference test, you want to be really careful when that change happens because a difference in difference test, which is what I do here, basically tells you before and after. So it's very nice and appealing test for a number of reasons, most importantly, because it gives you a fairly clean control for what would have happened if a policy change had not happened because you kind of control by very similar, with very similar books in this case that have been affected by a change in the copyright, in the length of copyright versus not. But it's also, it's a very simple test because it basically says like anytime after versus anytime before. So what you should really always be worried about what you should be careful about is when you see these types of tests, is when that change actually occurs. So ideally you wanna see nothing before and then you wanna see something afterwards but with some delay because whatever the policy is, it should take some time to kick in. And so that's here what we see. So we see basically nothing until 1814. And then it takes really, it takes some time. It takes more than 10 years for publishers to actually realize this. To realize, hey, we actually have much longer rights on the books of that authors and they put up the price. So then another thing that you might be worried about is that books by really popular authors sell for a higher price. And then some of these authors died after 1814. So that then essentially the effect that we were observing even though it's there in the data is actually not driven by what I'm telling you it's driven by, but it's actually driven by something else which is completely unrelated. Which is the fact that there are some guys that were just really famous. Their books were expensive and they died after 1814. So then their books are even more expensive and that's what we captured. So that's what we were trying to do here. So we were looking at books, the most famous writer of this period was Sir Walter Scott. This is not him, but this is Waverly. So Waverly is like basically some of these romantic books. Like another one is Ivanhoe. I don't, I tried reading them, I can't. So when we take them out, the effect actually becomes stronger. And actually the reason, oops, what does this have? So the reason for this is that books by Sir Walter Scott were almost more, they're always more expensive. Doesn't matter whether they were off or on copyright. It's just, and that means that now when we take them out, we're comparing the estimate to a lower priced average book which increases the size of the estimate. So that means that now if you take out Sir Walter Scott, so as one of the famous poets who died after 1814, the estimated increase in copyright actually goes up to 145%. And then that's the other, this handsome fellow here is Lord Byron. So Lord Byron was like the rock star of poetry in the 19th century. And his books also were always very expensive and he died after 1814. So we take them out. And essentially we're getting the same effect. So we're getting, if you're excluding books by both by Sir Scott and by Byron, you get a larger estimated increase because both of these books, books by both of these authors were always expensive. So that's really not driving our results. So then another thing you might be worried about is this effect which is very well observed in art as well which is that when somebody dies, their works become more valuable. So that could also drive our effect but it could only drive our observed effect under two specific conditions. And the first one is that books by recently deceased authors have to be more expensive which I'll show you in a bit is actually true. And then the second one is that they have to account for a larger share of our variation after 1814. And that's not gonna be true. So books by recently deceased authors, yes they are in fact more expensive. So they sell for 36.2 shillings per book and that's within one year of the author's death. And other books by other dead authors, they sell for 28 shillings. And then books by living authors, they sell for 17 shillings on average. So we take those out. But then again when we take them out, it doesn't really affect our results. So we still see this very, very large increase in the price of books. So it's roughly even controlling for throwing out all the recently deceased guys. It still implies that the price increase after 1814 is about 140%. And then, so now I'm controlling for page numbers. So one thing that I think is really important is that bigger books may sell for higher prices because they're more costly to produce. And it's also important because you may actually pay more. So you may be more willing to pay for like 20 shillings for a 400 page book, then you are willing to pay for a very short book. And actually authors, publishers realize this. So when you look at a lot of these books in the 19th century, what you see is that they were very creative in spacing. And nowadays, obviously they don't do this anymore, but they drag out the words. So you look at these eight 19th century books and they have something written and then there's a big space and then there's the other line and there's a big space and there's the other line. And they also pull out things in volumes. So Frankenstein, for example, was published in three volumes. And lots of these other books, they basically drag them out. So we're controlling for that. And when you control for that by collecting data on page numbers, it still implies that the books by that authors after 1814 become 127% more expensive. Okay. And then you also wanna control for the size of the book, right? For two reasons. So it could actually be that really large books are more expensive. So some of these books are this big. So when you see a quarter, for example, that is quarter means four times. So the book had to be the page, had to be folded four times to produce this book. And so there's a standard size, which is the octavo, which is slightly cheaper. And as I said before, it's like the paperback version of the book. So we're controlling for those page size effects. And then again, now we're getting the samples gonna get very small. So now we only have like about 660 observations. And so you still see that there's a really substantial increase in the price of books by dead authors in response to stronger copyrights. So now I'll show you sort of the placebo test. Because so another thing that you may be worried about is that there might be something funny going on after 1814 when for some morbid reason, we just like books by dead authors more. But for that to drive our results, it would actually have to be that there's a differential change in taste that favors books by dead authors after 1814. Or there could be something else that we can't observe. There's could be something else going on here that we don't know how it affects the price of books, but that affects them differentially. So making books by dead authors more expensive after 1814. And so what we use here is that the extension only applied to young books. So it only applied to books between the first edition and the first 14 years. Because if you had survived the first 14 years of your book being published, then the book automatically even before 1814 was protected until for 28 years. So what the 1814 act does essentially is just harmonizes the copyright protection for dead authors so that anybody who was dead within the first 28 years gets 28 years. But so that with the books where the author had died sort of after 14 years, they already had 28 years before. And so those guys, so the books that are within 14 and 28 years of the initial publication do not get the extension. The only change that they're affected by is that they are written by dead guys after 1814. So now what we're trying to do is essentially find a way to control for anything that could be correlated with being dead after 1814 versus alive before 1814. So essentially we're trying to see whether that change in copyright was correlated with something else that we can't explain and we don't know what it is, but now we're trying to test for it by essentially seeing whether we can see a price increase also for those books. And there is none. So when you look at our very love interest now which is like post-1814s times the dead guys, there's no effect. So it is not statistically significant but it's also just really small. So that kind of tells me that it wasn't some kind of random change in taste or something else or it's very unlikely that there was a random change in taste or something else that could have driven that change in copyright. So that really in some way, for me that tells me that there has to be something. The only thing that I can think of now is this extension in the length of copyright. Yeah, go ahead, yes. I'm maybe not precisely understanding of policy experiment or what I'm wondering is what price increase from what I understand is you have these two types of books, books by life authors and books by dead authors. And what this legislation did is essentially extended the copyright term of the books by dead authors. No, but it extended them, no. It extended the copyright terms for both, books by dead and living authors, but it extended it by more for books by dead authors. Right, but in your analysis, you're essentially looking at books that are all protected by copyright. No, no, so the book can be un- or off-copyright. So we take any book that was published before or after 1814 and we take the price of that book at any edition. So we basically take the price of any new book that hits the market. It doesn't matter whether that book is un- or off-copyright. But okay, but what I'm wondering is what drives from an economic theory point of view, what drives the price increase in the sense that you have books that are protected by copyright and you would expect them to have a certain equilibrium price depending on the demand for them. And then this act occurs, and why would there be a price increase? I can understand why there wouldn't be a price decrease, but what is the argument that there is a price increase? So that's a very helpful question because so you can think of this as a durable good if that helps. Or actually I think what helps me more is to think about it in terms of what customers at this time expressed. So for example, the quote that I mentioned earlier by Lord Dudley when he said, I would really like to read this book by the new book by Sir Walter Scott. However, it is outrageously expensive. I will not buy it at the present time. I will wait for this book to drop in price. So he says, I will wait for the cheaper October edition to come out. The price goes down when an alternative publisher can come in and offer the book at a cheaper price. So what you actually see in this period is that you see many more new publishers entering the publishing industry who are just specializing in cheap books. So they are specializing in books of copyright and they get them and then publish large quantities of them. If the copyright is extended by a longer time, if I have a copyright on this for a longer period of time, I can keep out these competitors. And so that basically means that I can keep the price for the book by a dead author at a higher level for a longer period of time. Right, they don't understand. And I think this is what comes out of the difference in difference approach, but I think what is a bit confusing is to talk about price increases. Because it seems to me what you're capturing is that in some sense the prices don't decline for those books. Yeah, so yeah, that's exactly right. So I guess, yeah, so it's just a higher price. So yeah, in summary it's just semantics. So I just call it a price. It's a higher price. You could just call it that. So what you're saying is that the price basically, the price takes a longer time to fall. That's exactly what we're trying to say. So because the coefficient that we showed before is just relative to the other, yes. But again, when you look at the means, you actually do see that the books by dead authors, on average, the prices do go up. And that could just be driven because some of the prices don't go down as quickly. Because what I'll show you is a mean. The mean price, the only argument that we're making here is about the average price of the book. So it's not an argument about the entire distribution. It's just about the average price of the book. And there are some books that just, yeah, it could just very well, the argument, the mechanism is that the price doesn't go down as quickly, yeah. So, but here essentially what we see is that it's not something driven that could be just differentially affecting books by dead authors after 1814. And then the other thing I think, which is related to your question like, does the price decline as the copyright ends or after the copyright ends? And so the argument here is that, so in some way a lot of what I've shown you to this point is trying to exclude alternative explanations, try to investigate alternative explanation for what we see in the data, which is that books by, the mean price of books by dead authors goes up, controlling for a bunch of stuff. But then, so you wanna think about the mechanism. So what's the mechanism that underlies this? And so the mechanism that we believe underlies this is that now we have a more long-lived copyright in this book. And so if that's however the mechanism, if having a longer copyright increases the price of the book, what I should also see is that as the same book approaches the end of its copyright, that book should become cheaper. And so what we're using here is that I think this is also important to sort of clarify again is that all of these books get an extension because say everybody dies, right? So you're always becoming a dead author at some point. So what we're exploiting in the main test is that it's just as dead authors get more after 1814. But by that sort of fact, even books by living authors, or even living authors benefit from this because now if I sell my rights to a book to a publisher, I can actually sell that publisher a longer revenue stream. So even if I'm dead afterwards, if I sell the entire right to the publisher, I can offer that publisher a longer revenue stream. And actually what we do observe is that there's an increase in the payment to authors after 1814. So they get double the amount of money after 1814. So, but what you also, what you wanna see for this argument to hold is that the books that are, so if a book that's protected for a longer period of time should be priced at a higher level. And so this is the test here. And then what we see here is that additions of book titles become cheaper as they approach the end of copyright. So at the X axis here, you have the remaining years of copyright protection and it starts at 25 years. And then so as we approach the remaining years of copyright is basically it's almost gone. And that actually, so you see there's a decline, right? But the true decline may actually be even larger because of sample selection. So in the last remaining years, there are very few books that are still in print and those books are very popular. So you may actually expect them to sell all else equal for a higher price. And so what you still see is that as book approach the end of copyright, they become cheaper. So now basically what do we conclude from this? So yes, there's a substantial increase in the price to longer in response to longer copyrights. It's 8% increase in the mean price of a book for each additional year of copyright, which implies economists like elasticity is a surprise and elasticity of price with respect to the length of copyright of about 27%, which is huge. Then that's controlling for book age, author and time fixed effects, which really and controlling for all these other things. And I think that's really our main argument. It's like when you actually do this right, when you do it sort of in a kind of a clean way with a control group and like controlling for other facts, you do see an increase in the length of copyright. So you do see an increase in price in response to longer copyrights. So when we're looking at the mechanism by this, it can be the mechanism for the price increase. So we know that there's a price increase for books by that authors after 1814. But what do we know about the mechanism that drives this? So the placebo test is supposed to help address this. Essentially we can pretty much exclude that it's anything else about these books that drives the observed change because when we look at the price of books by dead authors that didn't get longer copyrights, they don't actually see that increase in price. What we do see however, is that books become cheaper as they approach the end of copyright. So then when you think about this, can you actually take these results and put them and apply them to another industry? I think it's really, really important, especially I think with historical research, but also with any other type of historical research, whether any other type of economic research, which is that you wanna think about differences in the qualities and the characteristics of goods across industries. So I think the closest parallel here is one to pharmaceuticals, but it's actually a very different good. So in pharmaceuticals, what you see is as a pharmaceutical goes off patents, the price of the patented product very often goes up. And the reason for that is that people who are not respond, yes, it's very unintuitive, it's a very counterintuitive result, but so the reason for this is that, say if I don't pay for the difference between a generic and a patented drug, say if my insurance pays for it, I have every incentive to just go for the patented drug because everybody tells me that's the original and that's, so if I don't pay, it doesn't matter, right? To me, the difference doesn't matter. So what you actually see when you see a generic drug entering is that actually allows the pharmaceutical firms to charge a higher price for the patented drug because now you get the customers who don't care about money, mostly because they don't have to pay for it. So then that actually allows the pharmaceutical companies to price differentiate, right? So you can do this for pharmaceuticals, you can't do that with books because the difference, the key difference here is that when I need a drug, I need it now, right? So before that, you basically see the price of the, so we can't wait, we can't delay our purchase of a pharmaceutical. With the price of a book, you can delay, right? If the book is priced too high, you just wait, you just wait until the book becomes cheaper. So I think that's sort of a key difference between our results here and sort of the, I think the most closely related industry which would be pharma, where we have sort of similar, where we have related evidence on this. And so now what we're trying to do here, and I think this is something also where I would really value your questions and sort of suggestions, is that we're taking this as the first step, right? So as an economist, like the first thing I wanna know, like is there an effect on price? Because if there's no effect on price, I have nothing to go by, right? There's nothing I can, if I don't see an effect on price, then what's the mechanism by which I see any increase in incentive or any increase in a decline in diffusion? So what happens if when the price goes up? And so that's what we're working on now. And so do we actually see more entry into this field of writing? And so what we're doing here is we're looking at biographical data on authors. So we're looking at, before and after 1814, whether it actually becomes easier for people who are not from wealthy backgrounds to enter the profession of writing. And that actually does seem to be the case. But very cautiously, so the reason why I'm not showing you these results today is because it'll take us another, at least a year to really have enough data and clean it up enough and do the analysis carefully enough that we could actually say like, yes, it seems to increase entry, it seems to increase entry into the profession of writing. So that's sort of the next step, like what does it actually do? When you create stronger poppy write, does that allow more suitable people to enter, say, a creative profession because it actually makes viable that profession by creating, making it profitable? And then the other issue is that having this higher price may also really discourage the diffusion of knowledge because, say, if there are books by dead authors, say the works of Walter Scott that were more, that were much more expensive after 1814, did that keep people from reading Sir Walter Scott? And then we can see how having less access to the works of really good writers, how that influences the quality of writing later on. Thank you. Thank you very much. This was the most interesting seminar and thank you also for patiently answering all the questions. That just leaves me to wish you all happy holidays and I hope to see all of you next year. We are in the process of finalizing the seminar program for 2013. Our first seminar will be at the end of January, it will be on trademarks. We have Professor Georg von Grevenitz from the United Kingdom who will be here. So I look forward to seeing all of you next year. Happy holidays and thank you again, Petra. To the formula of originator with some range of variations. So I may be in uncertainty because maybe that genetic is the... Question. Karen Carter on the other issue to the WTO. Thank you very much for your presentation and for insisting in the paper.