Critiquing Buccafusco & Sprigman’s The Creativity Effect (2011)

Critiquing C.J. Buccafusco & C.J. Sprigman (2011) The Creativity Effect, University of Chicago Law Review 78

Critique by Prof. Paul Heald, Professor of Law, University of Illinois

Daniel Zizzo set this up very nicely by giving us the theoretical framework to think about experimental economics and pointing out some of the possibilities and pitfalls. Now we move to a specific study that’s actually an experiment actually carried out in the context of copyright law and I think, just by taking a hard look at it, we can illustrate some of the more general points that Daniel was trying to make.

So first I’d like to summarise the paper and the methodology for those who haven’t read it, do a brief critique and Daniel’s actually given me nice terminology to sort of summarise my critique, and that would be high internal validity and low external validity. If we have time, I’d like to note some other cognitive psychology experiments that have already been conducted that are relative to IP in the patent and trademark context and maybe have time at the end to point out some resources for your own experiments.

Some warnings. In general experimental economics is super promising and you have to be so careful if you want to generalise to the real world. Very hot topic, right? I have two colleagues, both who have PhDs in psychology and JDs on the law faculty with me, this is considered to be such an important new area of study and to preview a conclusion, experiments are actually most convincing for me and I think most valuable for us when they test assumptions that are made explicitly or implicitly by the law about human psychology. There’s a lot of legal rules and also economic justifications or explanations for legal rules that are based on specific notions of human cognition and testing those is what I find to be most valuable.

All right. We have seen this actually isn’t really new. If you think of it, if you’re a trademark lawyer, human subject experiments have been used in litigation for decades now to test likelihood of confusion, genericness and also in false advertising litigation. So on that scale, there’s been a lot of experiments already related to IP but just for litigation. Now we see more generalised experiments taking place not for litigation and seeing policy prescriptions based on the results of those experiments, which is of course what Chris (Buccafusco, a former student of mine) and Chris Sprigman do in their paper.

Buccafusco and Sprigman describe their first experiment in trying to find some sort of creativity effect, this suspicion that they have, this intuition they have that creators probably over value their work more than the market does simply because they are the creators. Their first experiment, though, only showed what is described in the literature as endowment effects. This has been shown in many, many studies, that owners of things tend to simply, by the fact of ownership, feel endowed in such a way that they value that thing more than a rational market actor would. If you’ve ever bought or sold real estate and run into sellers of real estate you know instinctively this is true. You meet crazy people who have crazy valuations of their own real estate, probably due to this endowment effect. And I’ll just throw something out right now. What de-biases that effect in the real estate market and allows a functioning real estate market to exist? Buccafusco and Sprigman talk about de-biasing the creativity effect, what might de-bias that naturally in the IP market is something we should ask ourselves at the end of their paper.

So first the endowment effect that they showed had been shown many, many times. The most famous experiment is probably the wild over-valuation of coffee mugs. You give people a $5 coffee mug in the experimental context and they won’t sell it for anything less than $8, right? Even though they’ve given it, it’s only worth $5, the fact that they own it, it’s mine, mine, mine, they over-value it. This has been shown in dozens, dozens of experiments. But query, you know check out the prices of actual coffee mugs in garage sales and boot sales. You get them for a quarter, 50 cents. In the real world, people are clearly willing to part with coffee mugs for much less than they are in the experimental situation, leading to a prejudice I have to prefer observation of what happens in real markets as opposed to experimentation when possible. So I view experimentation most valuable in situations where you actually cannot observe the behaviour in the actual market. Not surprisingly, the best experiments mimic that as closely as possible, what happens then in the real market that we care about.

Now the second experiment is the one that’s the subject of this paper, which I think is very original and very illuminating. They’re trying to find some sort of effect beyond an endowment effect, an extra value that the actual creators of works of art might have more than not only those who buy the works of art but those who own works of art. So they have three people involved in the experiment. There’s going to be $100 award for the best of ten paintings and one player is the painter, the actual painter, and the painter has the chance to sell his or her chance of winning to a buyer. Rationally, if you have an average painting, the painting’s worth $10. The chance of winning is worth $10. Each painting is assigned an owner, again, without looking at the painting, you would think your chance of winning would be one out of ten. It’s a $100 award. The value of your chance of winning is $10. And then of course the third player is a buyer and the buyers are given the opportunity to purchase a painting and take part in the potential $100 award. So for all of these people, the average value of a chance of winning is $10.

All right, results. Do not hire painters, owners or buyers to run your investment portfolio. We may have suspected that already. No exchanges occurred. In no circumstance did the buyer value the chance of winning more than either any of the owners did or any of the painters did. We’re not going to go through all of the matrices here but you can see that the painters valued their chance of winning at $74 and this is not explained, it correlates with their confidence in how good their painting is, but that’s not enough of an effect to completely explain the high valuation. Owners value their chance of winning at $40 and buyers at $17. So you see the endowment effect in the difference between the buyer valuation and the owner valuation, right. Just because you own it you value the chance more highly. What they call the creativity effect is the difference between the valuation of the painter and the owner, right, because you’ve painted it, because you’re the creator, there appears to be an extra marginal value that you place on your chance of winning.

As I noted, the chance of winning does correlate with valuation because it doesn’t explain it. Interestingly, painters’ emotional attachments to their painting doesn’t explain the difference and painters don’t have what’s called a regret aversion to any higher degree than owners. So they have, I think convincingly to me, isolated a particular extra perhaps irrational valuation that creators have over owners or potential buyers.

Now, what sort of criticisms can we make of the study? First of all, it’s quite a small end. They run this study twice, so they really have 20 participants. To me, I’m not willing to jump to some of the policy prescriptions that they make from such a small N. And more importantly, and this goes back to something Daniel said, if you want your study to have high external validity, you want to run it in such a way that it sort of mimics the external market that you want to say something about. And in this situation, notice there’s no round of negotiation whatsoever, and there’s no communication between the buyers and the painters or the buyers and the owners, they just write down a figure. There’s no negotiation, which is of course wildly different than what happens in the real world when a painter goes to a gallery and says my painting’s worth $10,000. I would like you to put it on display and ask $10,000. And the gallery owner laughs and the negotiation begins and the price drops radically.

Another problem I have with the paper is related to this and the genesis of the paper is their knowledge about the endowment effect and their suspicion that there’s an extra creativity effect, that there’s a cognitive bias that they want to show exists in artists and creators. It’s decoupled from any kind of legal question that we might ask. It’s decoupled from any sort of presumption that the law makes about human cognition in this situation. They’re giving us pricing data and valuation data. If that’s what we care about, if we want to know something about how artists’ value works, why not observe painters as they negotiate with art galleries or study Etsy online markets for creative works. Quite possibly you’ll gain more information about the way artists actually conduct valuation than in this sort of artificial, experimental context, I don’t know if you watch Pawn Stars at all, but it’s highly illuminating to this point of external validity of this study, and that is they always ask the people who go into pawn their great grandfather’s civil war sword – they go this sword is worth $20,000, I’ll take $12,000 for it. You go in and talk to the Pawn Star guys, they call in an expert who provides you information on the general value of swords, a negotiation occurs and they sell the sword for $500. That’s the way the real world works and my main criticism of the experiment is it’s simply too far removed from real world valuation that artists engage in.

Okay. So for that reason you know, we might want to think about how real markets in general de-bias buyers and sellers through exchanges of information during the transaction and you might think about this in the real estate market, too. You think your house is worth $200,000. Not after the agent tells you some comparables that have been found in your neighbourhood in the last couple of years.

That said, I think that the study has really high internal validity. I think they’ve taken tremendous care in considering exogenous factors like confidence in their valuation, confidence in their chance of winning and regret aversion which have to be considered. I would be more modest, I think, and not offer any sort of normative [unclear 13:28] descriptions, but instead use a study to maybe explain some existing intellectual property rules or at least the study might help explain why we have resale royalty statutes or inventor compensation statutes as ways to sort of dislodge works from owners, who might worry that they’re not being paid enough and that they’ll lose all possible income streams in the future. These statutes assure artists and may assure them enough so that this creativity effect is overcome and they’re willing to sell their works, work for higher doctrines, compulsory licensing, orphan works legislation, all actually force transfers from artists and it may be we have to do that because they have this irrational creativity bias which leads them to over value the work more than the market would. A little bit of modesty is always helpful both in experimental economics and observational economics.

Do I have enough time to go beyond that, Martin? Okay. There are some other studies out there that are worth looking at. Greg Mandel tries to take a look at hindsight bias in patent litigation, so this is experiments, really interesting. He tells half of his mock jurors, gives them here’s what the invention is, here’s the history of the invention. The other half of the mock jurors are told the exactly the same thing but they get one added fact, and that is the patent owner is the one who solved the problem, who actually came up with the invention. That’s the only change in the information given to the mock jurors. It radically changes the result of their findings as to the obviousness of the invention or the non-obviousness of the invention. The people who are told that this guy actually, that the patent owner actually is the inventor and solved the problem find this invention to be much more obvious and much less protectable because of what psychologists call hindsight bias. So again, and that’s I think an interesting finding in a way that experimental economics can help illuminate at least one, actually it’s a real litigation problem that you run into. You litigate patent cases in the US, at least, where jurors have to look backwards in time at the state of the art and then come up with a conclusion as to whether the invention was obvious at the time it was invented, maybe fifteen years ago. So hindsight bias is a serious problem.

It’s been pointed out that there’s also something called attribution bias, right, which has also been measured and proved in a bunch of psychology experiments. And guess what? When you identify somebody as the creator of a work, this attribution bias is going to kick in. Attribution bias is what makes us credit somebody with more talent or more contribution than they actually may have made because they’ve been solely attributed in the media, for example.

So does experiment prove that we need to change the jury instruction system in patent cases or de-bias jurors somehow? Maybe, maybe not. It’s illuminating, I think, and interesting again, policy prescriptions may not flow from a study that starts with knowledge of a cognitive bias and then tries to find it somewhere in the legal system. I much prefer studies, for example, that look at trademark dilution, where the law makes a specific assumption about human cognition, and that is when consumers see a famous trademark being used by a misappropriator, somehow their connection in their mind between that famous mark is weakened. That is the whole point of dilution law. The dilution happens in consumers’ minds, that’s the basic assumption of the law. Well, why not test human subjects? Expose them to famous trademarks in various contexts and see whether in fact recall is slowed or whether there does seem to be any sort of change in the way the consumer experiences the trademark. In the future, people have done that, what’s quite interesting, and I don’t want to go too much into the study, is that famous brands seem to be immune from any slowing in recall. So the brands that are most strongly protected by dilution doctrine, the most famous marks in the world, are actually completely immune, it appears, under their study to dilution and when consumers see those marks on different goods and different context, recall isn’t slowed but there’s a lot of problems with the study, mostly because it’s very a-contextual. The students are law students who know something about trademark law and there’s not much of an attempt at all to replicate what actually is found in the market. Other criticisms, I dislike the idea of the study. The study actually has really low internal validity although it might have had tremendous external validity if it had been done right.

Finally, if I’ve got just two minutes to rehabilitate Chris a little bit, we did a study, again testing an assumption that at least US law makes about why we want to extend copyright term. Congress believes that bad things happen to works when they fall into the public domain, that they become subject to abuse and economists specifically predict that, for example, audio books made from public domain books will be worse than audio books made from protected works because the public domain works lack an owner, there’s no shepherd to choose the right person to create the audio book. There’s no single owner with a monopoly incentive to do a great job in creating an audio book, so those audio books will be worse. Well, that’s a specific prediction and it’s one that can be tested and it’s one that can be tested in an experimental context which is what we did. We pretended to be on the Amazon Mechanical Turk, a company judging various readers for audio books and asking Turk workers to tell us which readers were better than other readers. So we had excerpts that Turk workers could listen to online. They rated the excerpts and we come up with an answer, not surprisingly to you maybe, that the public domain, that audio books made from public domain works are just as highly rated, the recordings are just as good quality, the readers are just as strong as those that professionally made copyrighted books. So maybe it’s my own preference for my own sort of work, but I really like the starting point of an assumption in law economics that drives legislation or drives regulatory regime and then testing whether in fact the assumptions about human cognition upon which the law is based turn out to be true or not.

Critique by Dr Piers Fleming, Lecturer in Psychology, University of East Anglia

I’m a psychologist so I’m going to try and consider this paper and some of the points that were raised from a psychological perspective. I just want to pick up on some of the related phenomena that I think are relevant to this creativity effect and that I hope you can find an application for within your own research.

The first phenomenon is effort justification. So this is a well established model in that the more effort you put into something, the more you value that thing that you obtain. For example, if you earn £10, you probably think quite carefully about how you spend it, but if you’re given £10 or if you win £10, you might be more willing to just blow it on something that’s less important to you in a more impulsive decision. And this is relevant to creators, because of they put more effort into the things that they create. So it’s a related phenomenon that might partly explain this creativity effect. And of course it’s not really rational because the £10 is, whether you win them or earn them, you should really think about them equivalently.

The Ikea effect is a very interesting, relatively new finding that’s actually mentioned in the study and in the Ikea effect you find that people who successfully build something value that thing more than those who haven’t built that thing, and it comes from the actual Ikea object itself. So they did some experiments and they found that somebody who had built an item from Ikea valued it more than the pre-built item. Which doesn’t really make sense, because isn’t it easier just to have the item pre-built and probably it’s done better than you would have done? But nevertheless, people did value it more and indeed, some people were willing to spend more to get a non-built item that they would build themselves than the pre-built item. What’s interesting about that is that the point at which that extra value was obtained was when it was successfully completed. So one of the questions that Buccafusco and Sprigman ask is when does this creativity occur? Well, it seems like it might occur once you successfully complete your task.

Finally, loss aversion is a very well established phenomenon and it relates to the endowment effect that Paul mentioned. So this is the idea that the pain of losing something is greater than the pleasure you might gain from that same thing. So you can imagine that, if I gave you £10 you’d be quite happy. If I then took it away, the same £10, you probably would be more upset than when you started, when you had £0 from me. And even if I did it the other way around, so if I took £10 off you and then gave it back to you, you’d probably still be having a less good day than if I’d just left you in peace. And you can explain things like most people wouldn’t take a bet if I, heads you win £105, tails I take £100 off you. So even if the bet’s biased in your favour, most people wouldn’t take that bet. There are other reasons, but loss aversion is a big part of that. So it’s not really, it’s not entirely rational again, and it relates to the creativity effect because a creator might think about it from the perspective of losing their creation when they sell it. So the painters in the study, although actually it wasn’t the case in this particular study, but creators in general, when they sell something, they lose possession of it. So that loss is worse for them than the gain of somebody else buying that same item.

Okay, now there are a number of things that affect these kinds of effects, particularly the endowment phenomenon and these are things that we might want to think about because they’ll either enhance or detract from this excessive subjective valuation. So the longer you own something, the more you tend to value it. There’s a nice study demonstrating that as time goes on, the longer you’ve owned something, your excessive subjective valuation that comes from ownership increases. So a creator is likely to have spent more time with something than a non creator. Also if you touch something, that again enhances this excessive subjective valuation. So some items might be more susceptible, or you might be more likely to touch some items than others, so it might differ between different types of medium as to the extent of this effect.

Uniqueness: so the more unique the item, the stronger these kind of effects are. So of course, if you’re creating something, the chances are you’re creating a unique item. So again, it’s likely that the effect is going to be stronger in the case of someone who’s creating something. And self-association: the more something is associated to you, again the stronger the effect is, okay. The demonstration is, so Paul mentioned the mug study, of which there are many. So if the mug has the logo of your university if you’re a student, or something that’s personal to you, you find that the effects are enhanced compared to a blank mug.

The last two are quite important. The value of the object has to be uncertain for the effect to occur, so if I own or I somehow fashioned a £10 note, which I imagine is illegal, then it’s not the case that I would then value that £10 note more than another £10 note, probably. Or at least if I did, it would be a small effect. And also, with experience, you can overcome this kind of effect. So as Paul noticed, it’s not the case that shop owners struggle to meet a valuation of a buyer coming in to their shop for things that they own and want to sell to them. So with experience, you find that people’s susceptibility to overestimating the value of their own goods disappears.
Okay, so there is a psychological explanation for this which might be useful in terms of framing how you think about this kind of effect. Now there are other explanations and I imagine economists probably wouldn’t favour this particular explanation but I quite like it as it gives a sense of what might be going on. So it’s called implicit egotism and the idea is that we like ourselves. People tend to like themselves more than they like other people, and that’s a good thing. It’s not necessarily entirely rational in the sense that probably everyone’s fairly evenly talented and kind and so on, but people tend to think that they’re better and prefer themselves. And actually, if you don’t feel that way, then you’re probably depressed. And that is one of the signs. Okay, and that preference for your own self is extended or transferred to objects that you own. This is where things like touching extends that sense of self to other objects. The longer duration gives longer for it to be associated with you. And again, in the act of creation, creation is something where something is much more likely to be associated with you. You put something of yourself into your creation, so it links nicely to why the creativity effect might be stronger than the endowment effect, as was found in the paper.

Okay, so what are the implications of these kind of phenomena? We’re going to get differences in subjective and objective valuations. That’s what they found in this study. So the painters and the owners valued it more than the buyers and the painters valued more than the owners. But we’re also going to find that the more you think, the more effort you think you have put into creating something, the stronger this effect will be, okay, because you will have spent more time and there will be the effort justification effect. So that’s something to bear in mind, that the more effort people have put in the more inflated their subjective value is likely to be. And also there are likely to be differences in people’s perceived contributions. So if you’re co-creating something with somebody else, you’re likely to value and rate your own contribution as being greater. Again, if you don’t believe this, firstly think about how much you think you do of the housework in your house, then ask somebody else who lives in the house with you how much they think of the housework you do and you’ll find that you have inflated your own valuation. And unfortunately, we’re not actually aware of these biases. So if you ask people, well how much did you think other people value this, I know you said that it’s a really brilliant creation, how much do other people value it, you tend to believe that your own valuation is the same for everyone. So we don’t compensate for this kind of inflation.

Also, the valuations occur retrospectively. So if you ask people in advance how much they think the Ikea table that they build is going to be worth, or the mug that they’re going to be given, they don’t realise that they’re going to inflate their value. But after they’ve got it, then that valuation, they still have that increased subjective valuation. So one way round this kind of problem might be to ask people to set prices prior to creation, but of course that has its own practical difficulties. Also, so as I mentioned earlier, there’s likely to be differences in different types of media, so if you are producing something physical there’s going to be a stronger effect probably but if it’s something virtual or a service then it’s likely to be a weaker effect.

The good news is that people really will want to innovate. So people want to build Ikea products, people want to create things and we know this just anecdotally, that there’s something intrinsically enjoyable about creating something. However, that doesn’t mean that we shouldn’t encourage people to create and support them in their creation by protecting them legally because if you lose something that you’ve created, perhaps because it’s stolen because it isn’t legally protected, you will feel a disproportionate loss to what other people think you should be losing, think your loss should be. So you lose your painting and people say oh, that’s a shame, but it wasn’t that great a painting. But of course you’re thinking it was a brilliant painting and I put lots of effort into it and I really liked it. So compensation is unlikely, is likely to protect an objective value and not the subjective value that the person has set so they’re likely to be excessively discouraged and might even be put off further creation endeavours.

These are mentioned in the paper, I’ve mentioned a little already, so one way to get round this is to have objective valuations because it’s the uncertainty of the value. So if it’s a £10 note, there’s no issue about the endowment effect, there’s my £10 note versus your £10 note. So if you could get objective valuations that ought to reduce the size of this kind of effect. Experience reduces the effect, so if you have experienced intermediaries that should also help and as I said earlier, if you agree the value before you start the creative process, if that is possible, then you ought not to have this kind of problem.

Also, picking up a couple of points that Paul made. I completely agree that observation is as useful as experimentation and if the two don’t marry up then you’ve got to think about what exactly is going on. So, well firstly, I suppose, in terms of the negotiation, that it wasn’t taken account of in the experiment, I think it’s true that negotiation introduces a whole range of factors, so it might be the case that some people need money more than others and so on and so there are a lot of individual circumstances and a lot of noise so it would be hard to analyse that. But at least you get a starting point for people’s subjective valuation. So you do need to take these additional factors into consideration.

The other point in terms of observation is, I think it was Tom who said yesterday that it was singer songwriters who were most interested in copyright and IP legislation. Well, could it be because they are the people who’ve put the most effort and are most involved in the creative process? So there is some sort of observable evidence to support these kind of factors.

Comments from the floor.

Prof. Martin Kretschmer: Thank you. Okay, I think then since we are at the end of the experimental kind of cluster, we swap the papers around so we could have the experimental perspectives together, it’s probably a good time to take a few questions if they are relating to that field, before we move on.

Commentator: Yes, I found this paper very interesting as well, but I wanted to pick up on a couple of points where we might expect or not expect endowment effects. So we’ve got the example of the mugs being for sale and the garage sale. What I was thinking was, well the mugs that are still valued are probably still on the shelf in the kitchen. And when we’re thinking of the role of experience, and experienced intermediaries don’t necessarily have the same endowment effect, I think there’s another aspect here in terms of why do we buy objects, and it’s because we either value or we think others will value them. If you are a trader you are not necessarily buying things to show to your friends or because you expect to get any social appeal or to get any aesthetic appeal from that object. You’re selling it purely as an item of trade and therefore that object is not something that you have any attachment to yourself, it is a proxy for money. And therefore if it’s a proxy for money, then its value is going to be far more close to what the market dictates. So I just they were a couple of additional points that I found quite interesting and I thought there might be one other additional possible confound in this study, which wasn’t mentioned, which was that the students who were creating were art students and therefore perhaps they value art more than most students do. And so I don’t think that’s a critical confound, I don’t think it negates the effect that we see, but it might perhaps exaggerate it. So if this was to go on, I would like to see that confound perhaps controlled. That’s all, I think.

Next Commentator: I was just wondering, are we missing, I’m just looking at the list that Piers has got up there and all the conversations we’re having about owners, buyers and creators, what about agents? What about the people whose job is to set, to get the people the money, to establish how much something is worth in a marketplace, to give that idea of setting up agreements, creating valuations? Aren’t they the experienced intermediaries that is on the slide that Piers has got there and are we looking at those people in CREATe? When we talk about creators are we looking at the people who legislate and create agreements for creators or is that outside the scope of what we do?

MK: There’s no answer to it. Does anybody want to?

Commentator (Prof. Ruth Towse): I could say something, if you wouldn’t mind, at this point. I work in cultural economics and you won’t be surprised to know that there’s a considerable amount of literature on art prices in cultural economics because economists love fiddling about with data and doing econometric studies and there are masses of data on art prices, so it’s a subject that has been very thoroughly studied, including the question of whether art is an investment or is a consumption good, and those kind of things, which comes into that. But there is a study, well there are two things that I can think of that are relevant. One is that a colleague of mine at Erasmus did a study, Olav Velthuis did a study of dealers, art dealers and gallerists and found that they were just about as random as the artists were in prices, partly because people who go into that profession are sort of, as it were, wannabe or would have been or something artists. So they have in a sense, the same kind of endowment sort of things except they have to make money.

The other thing is that there’s an enormous entry and exit into the dealer or the art gallery market. The other thing is, which relates to this experimental thing, is that you need a lot of information and that’s not come up in this discussion, how well informed people are about art markets. I used, Hans Abbing is an artist and an economist who claims that the problem with modern art markets is that so many people are subsidised to go into them that nobody has any sense of what’s good art and what isn’t, and therefore there’s an enormous information, not even asymmetry, which economists are used to dealing with, but simply lack of information. And so the whole question of valuation is very difficult. Sometimes people simply use price as an index of value, which makes life even more complicated because if you set a high price, people think oh that’s good, you know. Actually my mother in law’s a bit like that, she’ll say what’s that painting, so you say well it’s something, how much did it cost, oh well, in that case – and that seems to be quite a feature of these markets. So the market in itself doesn’t work in the sort of usual way.

MK: Any other questions?

Commentator: Yes, I’ll keep it brief. So I agree with a lot of what has been stated and indeed I did think like Steven that one weakness specific to this paper did relate to the sampling because since you’re talking about different sample treatment, the painters in one case, non-painters in the other case, I think you have got the sample selection bias type of problem that may confirm [unclear 39:39] effect. About the point about experiments vis a vis sort of natural data, I guess just to add in the specifics of this context. I guess what you can have in an experiment sometimes is precise control potentially about what is happening. So if you were just to observe data in the art galleries or on Etsy, you wouldn’t have. If you look at the bargain sales, you’d actually not know what the valuations are specifically. One thing that you, I mean there are points made by Steven on that, but one specific point there that strikes me by which I don’t think that it’s necessarily going against what has been stated here is the fact that there must be billions of mugs around, and I know how many mugs you have got in your kitchen. I certainly have many more mugs than I’m ever going to need. So the fact that as most action goes up on garage sale, consider to be the distribution valuations, needs etc, doesn’t actually mean that they’re wouldn’t be a bias in the other direction. For that we would need to have data on the amount of mugs that are not actually being traded at all. I suspect there are millions of them, frankly. And probably I will shut up.

Oh, just one point. Negotiation, I agree with Piers, I think there is very interesting things there to say. I also expect an additional question, because there are a number of other considerations that come in when you talk about negotiation. And of course the very nature of bargains is that you wouldn’t reveal your preference actually. If there are bargains you wouldn’t have revealed your preference as you begin as part of the bargaining process, so in a first study or an early study I think you want to start as simple as possible and that’s part of what you do. Although of course one thing that follows from there and that you do need to be careful about the conclusions and I agree with what Paul said on that.

MK: Okay, that’s last one now…

Commentator: Just commenting, now that I was confirmed that maybe there is a bias on the use of the data, that was my impression. In this paper, the impression that the conclusion ran the factual evidences, in the sense that the conclusion wanted to express the viewpoint that we should switch from a property regime to a liability regime, which is actually one of the hottest debates in the US right now, particularly was promoted by Professor Reichmann, Professor Lang [name 42:41]. Reichmann was for liability, Lang for apportionment of profits. So my impression is that the bias there, it is because the authors were looking for supporting evidence to motivate why we should go for liability rules. Now that you think that the bias is confirmed as the economists may say this impression is even strengthened.

MK: Paul, do you think that’s true, no? Sorry, even if there’s a presumption or an assumption, Buccafusco and Sprigman were looking to support a liability rule when the set up the experiment?

PH: I think they were asked to tack on these policy prescriptions at the end of the paper to fit the law review format, so I don’t know if that was what their point was. The paper is pretty transparent about their disappointment of not finding a creativity effect with the haiku experiment but they admit that and I think they have a good explanation why painters are different than haiku writers.