Sprigman and Buccafusco on Valuing Intellectual Property

Josh Wright —  6 December 2010

Christopher Sprigman is Professor or Law at the University of Virginia

Christopher J. Buccafusco is Assistant Professor of Law at Chicago-Kent College of Law

We would like to start by thanking Josh for inviting us to participate in what promises to be a fascinating discussion on an important subject.  We’re looking forward to engaging with the other members of the symposium.

To begin with, we would like to talk about some of our own experimental research on the valuation anomaly widely known as the “endowment effect.”  Over the past quarter century, laboratory and field research in the social sciences has provided considerable evidence for the existence of a significant gap between the valuations that people attach to goods that they own and the valuations they attach to goods they are considering purchasing.  Thus, in one classic and well-replicated study, subjects to whom a university coffee mug was given indicated substantially higher willingness-to-accept values than subjects who indicated their willingness-to-pay for the mug.  This and similar studies suggest that aspects of goods that should be irrelevant from the perspective of neoclassical economics – such as the fact of prior ownership – can systematically bias valuations of those goods and lead to sub-optimal exchanges and inefficiencies.

In a series of recent studies, we sought to extend these findings into the realm of intellectual property law.  We hypothesized that the valuations that creators attach to their works will be even higher than those of mere owners and would-be purchasers.We conducted the experiment with painting students at a major art school acting as our creators and law students acting as our owners and buyers.  In an attempt to model the nature of IP transactions, in which the value of the rights is based on the opportunity for rent-seeking and not on the goods themselves, we created a quality-based contest for the paintings.  The Painters submitted their works to the contest and were told that they would be competing with nine other paintings for a $100 prize to be determined by an expert.  They were further told that they would have a chance to sell their right to win the contest to another subject who would make them a cash offer for their chance to win.  They were asked to indicate the least amount of money they would be willing to accept to sell their chance to win.

In addition, we told another group of subjects, the Owners, that they had been assigned to be the owner of a painting’s chance to win a contest of the same value.  Like the Painters, they were asked to indicate the least amount they would be willing to accept to sell their chance to win the prize.  Finally, we told the third group of subjects, the Buyers, that they would have a chance to purchase one of the paintings’ chances to win a prize and were asked to indicate the most they would be willing to pay to buy the chance.  After indicating their values, all subjects were asked to complete a series of follow-up questions.

As we predicted, Painters’ WTA was significantly greater than Owners’ WTA, and both were significantly greater than Buyers’ WTP.  The mean probabilistic value of any painting should have been around $10 weighted by its relative quality.  The Buyers were fairly close to this number, indicating a willingness to pay $17 to acquire the chance to win the contest.  Owners, as in a typical endowment effect study, wanted considerably more, about $40, to part with their chance to win the prize.  The Painters, however, were unwilling to sell their ~1-in-10 chance to win the prize for anything less than $74.

We asked follow-up questions designed to help us sort out what was going on here.  First, it is important to note that subjects’ predicted probability that their painting would win the prize was strongly correlated with their valuation.  This suggests that subjects understood the nature of the contest and were making logical assessments of value.  Although their assessments were logical, they demonstrated significant systematic biases.  Just as with mean value, subjects’ role significantly predicted their predicted probability of winning with Painters > Owners > Buyers.  Interestingly, subjects’ emotional attachment to the work and time spent creating it were not significant predictors of value, while their anticipated regret at learning they had sold the winning painting was very close to being significant.  While any of these last three might create rational reasons for overvaluing one’s work, valuations based primarily on over-optimism about quality and likelihood of success appears more problematic from the perspective of rational choice theory.  Painters’ and Owners’ relationships with the works generates excessively optimistic assessments of value that lead them to over-price their works.  This over-pricing creates substantial WTA-WTP gaps that may create sub-optimal transfers of works and market inefficiencies.

These findings are particularly significant for IP law, because transactions between creators and other parties (e.g., publishers, producers, investors) are often essential to the ultimate production of works.  In our next post, we will discuss ways in which IP markets may already deal with some of these problems (e.g., via royalty contracts and collective rights agencies) and possible changes to the law that might mitigate them, including the use of liability rules, formalities, and the fair use doctrine.

8 responses to Sprigman and Buccafusco on Valuing Intellectual Property

    Larry Rothfield 7 December 2010 at 11:21 pm

    What you are describing as bias based on emotional attachment on the part of artists is much better described as a rational interest in augmenting the cultural capital of their artistic name. Winning the contest would not just be worth $100 but $100 plus whatever the prestige attached to winning amounts to (in this case nicely measured). Part of the $64 surplus value, though, is probably a function not of what winning the contest would be worth reputation-wise, but of what being in the game at all is worth. Phrased negatively, it would be very bad for one’s art-student cred (and therefore to career) to be seen to have so little confidence in one’s work that one is willing to fold.


    Thanks for your comments on our work. I’d like to respond to your methodological concerns here, and we will address your policy concerns in our follow-up post tomorrow. It’s certainly true that our protocol entails a substantial abstraction from many real-world IP transactions. Nonetheless, I want to make two responses. First, all experiments require some abstractions, and for purposes of our first foray into this area, we were willing to exclude the effects of negotiation and learning on valuation. In a future experiment, we plan to examine precisely these questions.

    Second, we believe that the effects we’ve shown are likely to be meaningful even if bargaining and learning reduce some of the WTA-WTP gap. There is well-established empirical research showing that, all else equal, the farther apart two parties start in a negotiation the less likely they are to make a deal. Thus, if the creativity effect that we demonstrate causes buyers and sellers to come to the bargaining table with more distant expectations than they would in the absence of the effect, we can expect to see fewer completed transactions. We are certainly not claiming that transactions in IP won’t take place – they obviously do all the time. We only want to suggest that there might be previously unrecognized costs to the current system.

    Additionally, we think there is reason to believe that market learning may play a smaller corrective role in IP transactions than it could in markets for tangible goods. Although authors, for example, may continue to publish new works, each new work will only be an imperfect substitute for the previous ones. Accordingly, given the over-confidence that we see, it seems likely that previous failures may exert only a weak signal about the likelihood of future success. One might imagine a law professor who, despite that fact that none of his previous papers have been accepted by the Yale Law Journal, fervently believes that his new, better one will be. Again, this is exactly the sort of question that future research could clarify.

    The Ghost of W.D. Hamilton 6 December 2010 at 1:36 pm

    Prof. Manne
    I’m inclined to agree with you. That said, I think there is a compelling argument (often stressed by Paul Rubin) that we are programmed for loss aversion. That is, thousands of years of evolution during times when trade was zero-sum has created an inherent and “sticky” loss aversion in some humans. For cavemen, trade represented a real risk that the counterparty would steal from us and we’d die from lack of food. So, man (perhaps) avoided trade for all but the last few generations. In the IP transaction world, this might mean overvaluing our work product.

    I’m skeptical of most behavioral arguments and think the research is too heavily based on experiments which happen in an unrealistic void. (i.e. no advice from trusted ones, no input from the market, no real value at risk, etc.). But the notion that we’ve been “bred” to overvalue our created products strikes me as possible… and seems to explain why we as humans may be unable “learn” to stop overvaluing our work product even in repetitive scnarios.

    P.S. As an alumnus of Chicago-Kent, its nice to see a prof on this blog… even though, you know, I’m firmly in the Chicago School (Non-thaler)on behavioral economics.


    I look forward to seeing your follow up post, and maybe this is covered there, but I see a real problem here of a disconnect between experimental and real-world conditions–sufficient to make your results suspect (but then you would have expected nothing else from me). The problem I see is the absence of repetition and learning, as well as (as far as I can tell from your example) the absence of bidding or even conversation between buyers and sellers. It is one thing to identify a minimum WTA; it is another thing to hold tight to that identified WTA in the face of the reality that no one is WTP at that price. Moreover, I have no doubt that asking someone to identify a minimum WTA or WTP in advance creates what the behavioralists would call a framing effect (or maybe it’s a focal point–who can keep all those biases straight?), and participants may hew to these numbers far more closely than they would in the real world where usually negotiation precedes any determination of price. Thus, in the real world, the previous failed negotiation is an important determinant in the next negotiation (or, in your example, decision about a WTP or WTA); negotiation often precedes identification of reservation prices; and often prices are determined by bidding and through conversations rather than stylized iterated processes. I don’t know whether these (and there are surely others) deviations from reality would affect your results, but I am strongly suspicious that they would and curious if you have considered this.

    Moreover, your proposed correctives (especially the increased use of liability rules) strikes me as perfectly backward. Where conversation and negotiation are essential to reaching agreement, liability rules frustrate rather than help that process. An inventor has a hard time negotiating with a potential investor or even manufacturer of his technology without strong property right protection. And how can an increase in uncertainty around price (which is what happens if courts are more involved in setting damages) help facilitate agreement over price? I fear that your point is that creators overvalue relative to licensees, so we should just “nudge” the determinants of IP value in a direction that will put downward pressure on creators’ valuations. But why not the opposite–why not apply upward pressure to potential licensee’s valuations? More important, why not conclude from your experiment (if it is sufficiently realistic to offer much of prescriptive value) that the key is to define prices more clearly, ensuring that any biases are readily overcome? This, again, seems to me to counsel in favor of property rules. Yes, I know, from your point of view this would just exacerbate the difference between WTP and WTA by bolstering WTA estimations. But how do you know that that effect is greater — or even significant relative to — the problems of uncertainty around price and thus the difficulty of reaching agreement via negotiation? Because your experiment doesn’t explore (as far as I know) the process of negotiation, learning and the like, you may be glossing over the most problematic aspect of this–and advocating solutions that actually exacerbate rather than ameliorate the problem.

Trackbacks and Pingbacks:

  1. Behavioral Economics, the Law, and the Regulators - NYTimes.com - December 8, 2010

    […] the “Free to Choose?” symposium. So far, people like David Levine, Ronald Mann and Christopher Sprigman have taken their turns. “Behavioral economics itself has made a significant contribution to […]

  2. Free to Choose Wrapup « Truth on the Market - December 7, 2010

    […] Christopher Sprigman & Christopher Buccafusco, Valuing Intellectual Property […]

  3. Sprigman and Buccafusco on Behavioral Law and Economics and the Road from Lab to Law « Truth on the Market - December 7, 2010

    […] Behavioral Economics on AntitrustFree to Choose: Day 1 Wrapup « Truth on the Market on Sprigman and Buccafusco on Valuing Intellectual PropertyFree to Choose: Day 1 Wrapup « Truth on the Market on Thom Lambert on Behavioral Law and […]

  4. Free to Choose: Day 1 Wrapup « Truth on the Market - December 6, 2010

    […] Comments Chris Buccafusco on Sprigman and Buccafusco on Valuing Intellectual PropertyPropsective Economist on Richard Epstein on The Dangerous Allure of Behavioral Economics: The […]