Sprigman and Buccafusco on Behavioral Law and Economics and the Road from Lab to Law

Josh Wright —  7 December 2010

Christopher Sprigman is Professor of Law at the University of Virginia

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

In our second post, we want to discuss some of the implications of the study (the details of which we described in our first post). One of the consistent concerns about BL&E in this symposium is about the too-quick jump from data to policy. We should emphasize that we think more work needs to be done to support these potential policy suggestions, but, importantly,we think that the answers to the policy issues rest fundamentally on empirical questions.

As we noted in the previous post, our research supports the idea that creators of new works will value them substantially more than will either mere owners or would-be purchasers of the works. These valuation anomalies are likely to create sub-optimal transaction levels in IP markets. The higher transaction and negotiation costs associated with bridging a large bargaining gap are particularly troubling in the IP context where efficient transfer of rights proves crucial. In both the copyright and patent contexts, initial rights-holders (usually authors in the case of copyright and inventors in patent) often are not particularly well positioned to exploit their work. Given the gap between initial entitlement and effective commercial exploitation, an efficient IP law must provide a smooth transition between the initial rights-holder and the eventual transferee or licensee. In this post, we want to suggest two possible solutions to the transaction costs problems – one based on private contracting and another based on changes to legal entitlements.

One possibility is that private parties may already be dealing with some of these issues by using royalty contracts. The running royalty – i.e., an arrangement where periodic payments are made according to some percentage of sales or revenues – is a way of effectively “agreeing to disagree” over the value of a creative work. In cases where an author or inventor believes that the work is likely to produce substantially more revenue than the purchaser, use of a running royalty may allow both parties to structure a deal that matches their expectations and that reduces inefficiencies caused by optimism or regret aversion. We should emphasize that we cannot be sure how effective running royalties will be at mitigating endowment effects. Surprisingly little empirical research has been performed on the negotiation of royalties, so it is difficult for us to predict how endowment effects will affect royalty bargaining. It is possible, as we have noted, that royalties might reduce the effects of optimism by allowing the parties to move forward without resolving their differences regarding the likely return on the transaction. But it also seems plausible that the substantial differences in estimates of likely success will continue to hinder parties’ ability to agree on an acceptable split of the profits – the seller’s inaccurately high estimate of the likelihood of the work’s success may feed into a conviction that he deserves a more advantageous split of projected revenues. Indeed, these questions present another level of complexity: in many IP contexts the royalty rate will not be subject to bargaining, because it will be set by industry norms. Again, it is difficult to predict in settings where bargaining is impossible or unlikely whether the inability to bargain and the strength of norms will undermine endowment effects. Inability to bargain may result in an exercise of buy-side market power that partially or wholly offsets endowment effects. Or, it may simply result in a negotiation failure.

Even if the use of running royalties can play a role in mitigating endowment effects, it is very unlikely to be a complete answer to the problem of valuation. Running royalties are expensive to negotiate, implement, and administer. They involve the necessity of ongoing monitoring and periodic payments. As a result, running royalties are appropriate only for transactions that are valuable enough to bear the transaction costs of the running royalty arrangement.

Accordingly, it is worth looking to the possibility of restructuring legal entitlements to debias creators. Our results suggest two good reasons to consider debiasing. First, valuation anomalies proceeding from overoptimism present a better case for debiasing than instances in which the creators have idiosyncratically high valuations that are not driven by a mistake regarding likely market outcomes. Overoptimism can’t be dismissed as just an idiosyncratic preference – it is, rather, an information imperfection, or perhaps a failure to respond to information. In either case, it’s an error. Second – and importantly – our results suggest that the very old and continuing debate about the value of property rules vs. liability rules in IP has been incomplete. IP law is presently structured around property rules. But if the wide disparities between valuations that we found in our study characterize a range of IP transactions, then parties seeking to license or otherwise transfer ownership of creative works will face substantial negotiation costs arising from the need to bridge wide differences in valuation.

Thus far the law’s preference for property rules is based primarily on a presumption, driven mostly by theory and ideology rather than data, that markets and arms-length negotiations will allocate rights more efficiently than the alternative; that is, a legal regime based in liability rules, in which users are free to take, and in which the price of use is set not primarily via private negotiation but by a legislature, court, or government agency. (of course, there will still likely be considerable bargaining in the shadow of liability rules).

Our study undercuts that presumption. While liability rules require non-market price setting, which is beset by its own costs and is likely to lead to misallocation in some cases (although these concerns can easily be overblown, as parties often negotiate in the shadow of a liability rule), IP’s strong property rules may sometimes lead to significant pricing anomalies that hinder transactions and impose separate inefficiencies that liability rules may not create. The valuation spreads that we have identified add a significant and previously unrecognized premium to the transaction costs associated with IP bargaining. The gaps that we report may result in substantially higher costs of bargaining and, accordingly, fewer otherwise valuable transactions taking place.

When considering the respective costs and benefits of property and liability rules, it is important to remember that the inefficiencies created by property rules are neither different in kind nor necessarily less severe than those created by liability rules. Worse, property rules generate inefficiencies that are systematic in one direction—overvaluation and failed bargains—whereas the errors created in valuation under liability rules are more likely to by distributed symmetrically on both sides of the optimal price (that is, non-market pricing is as likely to produce under-valuation as over-valuation). If this is right, then symmetrical mispricing may not create substantial ex ante disincentives to engage in the creation of new works, for even if the creator understands that mispricing is likely under a liability rule, there is an equal chance of either over- or undercompensation.

Accordingly, if IP law’s preference for strong property rules is to be sustained, it must rest not on the basis of presumptions and ideologies but rather on evidence about the costs and associated inefficiencies of negotiation versus non-market pricing. These are empirical questions, and the answers may vary for different types of creativity and different markets. To make a start, we need more studies inquiring into whether pricing anomalies attend IP markets in a variety of circumstances, how large the valuation gaps are likely to be, and what can be done to shrink them.

5 responses to Sprigman and Buccafusco on Behavioral Law and Economics and the Road from Lab to Law

  1. 

    Geoff Manne has posted a long comment. We’ll respond to it (or at least to parts that ask for a response) by indenting portions of his post and then supplying our reactions.

    GEOFF: I remain skeptical of the relevance of this experiment. First, your experiment does not actually say anything relevant about the real world–does not produce relevant data–just because you say it does. We can discuss further the ways in which your experiment (or any experiment) may fail to mimic reality and the ways in which those differences matter, but I say tomato and you say tomato (you’ll have to imagine the different pronunciations), and you have the burden of proof, particularly if you want to draw policy implications from your single, flawed (they are all flawed–it’s the nature of the beast) experiment.

    RESPONSE: We don’t just assert that our experiment has real-world implications. We wrote a paper that explains in part why it does. The paper also explains the limitations of the experiment as a way of modeling the real world. In any event, the critique you launch against BL&E can be directed at any form of economics that involves abstraction and modeling — i.e., every form of economics.

    GEOFF: To add to the points I made yesterday: why do you think that even your evidence about creators of paintings “over-valuing” their works translates into the patent realm at all and the commercial realm (you know–the place where deals are actually most relevant to human welfare) for any IP? These are just assertions, and pretty simple logic suggests that the assumptions may be wildly inaccurate. I know you state in your paper that creation is much more of the non-commercial sort than economists assume, but does that hold for inventions as opposed to poems and paintings? And does that hold for the most-commonly traded works? And is it really true at all? You cite only work relating to copyright for this proposition, and it is a Habermasian (if that’s a word) criticism of economics, at that–hardly an empirically-relevant cite.

    RESPONSE: We used contests to model a market. Outcomes in markets, like in contests, often are based on a measure of quality, and prevailing in a contest, as in a market, produces a return. The basis for the model is clear. Like any model it’s imperfect. But we think it instructive. We don’t follow your remarks about what motivates creators. We have both written about that elsewhere, but that it not the subject of the Chicago paper (or the preceding Cornell paper). As for copyright vs. patent, whether our market is relevant to the latter depends on whether inventors are likely to overvalue their likelihood of success and be subject to the same overvaluation forces that we see in our protocol. We see at present no reason to doubt this — but of course this is something we can test.

    GEOFF: Moreover, you correctly note that, all else equal, over-optimism rather than idiosyncratic over-valuation (if there can be such a thing) should be easier to correct, but your own experiment pits painters/authors of their own works (as I understand it) in an emotional and fundamentally non-commercial setting and then asks them to engage in commerce. You really think their over-optimism is just that and not idiosyncratic valuation? Why? How do you know?

    RESPONSE: As we describe in the paper, we asked the painters to tell us how likely they were to prevail. And we also asked them follow-up questions designed to measure other possible bases for their high valuations. Their over-optimism re: likelihood of success was strongly correlated with valuation. No other factor was. Now, this does not mean that their overoptimism doesn’t have a deeper psychological basis, and our protocol is not designed to dig that deep. We’re not done, and we hope others will jump in on this question.

    GEOFF: And on that “over”-optimism–how did you establish your baseline? I do like how your assumptions go nicely with your preferred outcome (take the protections of IP down a peg) rather than the reverse, but on what basis do you claim this?

    RESPONSE: The contest is a zero-sum game. For that reason, the mean likelihood of prevailing is 10%. Our painters’ mean likelihood was over 50%. This is irrational, in the aggregate. Moreover, we can compare the optimism expressed by the painters and compare it with the other subjects. Painters think their paintings are much more likely to win than do mere owners, who themselves think their paintings are much more likely to win than buyers. There is no reason for the mere fact of creation or ownership to dramatically affect perceptions of likelihood of success.

    GEOFF: As the chance at winning the contest has monetary value, there could be an income effect. Could that explain everything? I don’t know–but do you?).

    RESPONSE: The bid-ask spreads are so large that income effect could not explain them. Also, our bidders are bidding well over the rational choice amount — they do this, we think, because they feel there is something special in being presented with a particular painting on which to bid. This effect, which tends to reduce the bid-ask spreads, counteracts any income effect.

    GEOFF: I notice in your paper that you run the experiment both as a lottery and a contest, and that there is more optimism bias in the latter–not surprisingly where there is a subjective determination to be made, and where my claim holds. I concede that in a lottery one could identify “over” optimism given that probabilities are objective, but a) this is where learning and commercial settings come in and make a huge difference and b) I notice that buyers as well as sellers are “over” optimistic in your experiment–and by not-dissimilar amounts. So where’s the problem? There is only a problem where there is a divergence in valuation, and if everyone over-values, so what?

    RESPONSE: Aggregate probabilities are also “objective” (i.e., predictable and stable) in a zero-sum contest. We discuss learning in the paper. IP is a highly differentiated product, and where products are not readily substitutable, learning has less powerful disciplining effect. As for the problems caused by over-valuation, (a) not everyone is overvaluing by the same amount (or nearly so), and (b) the wide bid-ask spread raises negotiation costs. See our comments supra about the relative amounts of over-valuation between roles.

    GEOFF: More important (in terms of an assumed baseline and the consequences of deviations from it), you seem to assume that rational choice requires some kind of objectivity in price.

    RESPONSE: We certainly do not. We are noting that according to the orthodox account, and with good information, we would expect to see prices *on average* converging around a relatively narrow bid-ask range. We do not see that.

    GEOFF: In actuality–in real markets–we don’t exactly know how prices are determined, and certainly they do not (how could anyone have enough information to know?) conform simply to “fundamentals.” Again this goes to the oddity of the experimental setting you have created–one where, arguably (and certainly in the lottery) there is some certainty about price.

    RESPONSE: Yes, and the relatively good information about price should create a very neat rational choice world. But it doesn’t (see our data). That’s precisely the point. We’ve tilted the playing field in the direction of a good rational choice outcome, and yet we don’t get that.

    GEOFF: Confronted with this setting in the real world, there is little or no commerce; in the real world, however, prices reflect a lot of different things, and it is an unlikely exercise to determine deviations from “correct.” In other words, I’m troubled by the claims that initial expectations about price (even in the lottery setting, allowing for idiosyncrasies and innumeracy) without any opportunity to bargain or repeat play indicate some sort of irrationality, where it is not at all clear that the market wouldn’t clear at some price, even if, to you, that price deviates from “correct.”

    RESPONSE: The market may well clear. And repeat play and learning may help. But if negotiation is expensive, it will clear at a lower level of output. That’s our point. For more on market clearing and learning, see our response to yesterday’s post.

    GEOFF: It also seems like a remarkable oversight that your paper doesn’t cite the Plott and Zeiler work on these sorts of endowment effect experiments, nor, as far as I can tell, take account of any of the problems they point out (although you may have done so inadvertently by avoiding the specific pitfalls they discuss). Perhaps you feel it is possible to do this (correct for biases in experiment design), and perhaps you feel you have indeed done this, but why don’t you discuss or even cite the literature to support your experiment design?

    RESPONSE: We cite a related paper by Zeiler in both the Cornell and Chicago articles. We are well aware of their work. We have designed our experiments to take account of their work.

    GEOFF: Finally, on remedy. I would first note that, in some sense, arguing for a “take now and pay later” regime in the face of what may be un-executed contracts because of idiosyncratic valuation or even merely transaction costs borders on advocating theft.

    RESPONSE: That can’t be right. You might as well say that 17 U.S.C. 115, which creates a liability rule via compulsory license for the re-recording of copyrighted musical compositions “borders on advocating theft.”

    GEOFF: As you point out, of course, there would be bargaining in the shadow of the law, but I do not agree as you suggest in your paper that this makes it all ok–bargaining with a gun to one’s head is not the same as freely entering into a contract. This is one of the great advantages of a pure property regime, and it is, at least to me, an important characteristic (freedom).

    RESPONSE: See above re: 17 U.S.C. 115. Interestingly, there is plenty of bargaining in the shadow of the compulsory license there, and most of it results in transfers at a price lower than the statutory rate. Whether your point about “guns to heads” holds is situational.

    GEOFF: At the same time, contrary to your assertion, there are systematic biases in the take now and pay later regime: The transaction cost of litigation, including especially delayed compensation, and the creation of poor incentives for inefficient infringers.

    RESPONSE: There’s plenty of litigation in property rule regimes, because the rule is rarely absolute (it’s never absolute in IP, and for good reason).

    GEOFF: I’m quite certain this is larger than the cost of bargaining.

    RESPONSE: How are you sure of this?

  2. 

    David: That’s only ex post inefficient, but surely the “fuel of interest” that IP adds to the “fire of creation” merits a great deal of ex post “inefficient” transactions for the sake of the great value of the “efficient” ones. Note–this is not at all the case for the inefficient transactions that occur because of a liability regime over-incentivizing inefficient infringers. In that case, innovation is harmed, not helped. Of course it is possible to have, overall, an inefficient level of innovation (but I’d prefer to err on the side of more rather than less–and definitely on the side of letting private actors decide rather than government), but it is hard to condemn the regime on the basis of the mere existence of inefficient transactions of the sort you mention. As you rightly point out, in effect, these biases are all over the place, the optimal result may be more or less invention than we currently have, and I’m back to the point that the mere identification of a single bias in a single experiment should be insufficient to even budge the policy needle on way or the other.

  3. 

    I remain skeptical of the relevance of this experiment. First, your experiment does not actually say anything relevant about the real world–does not produce relevant data–just because you say it does. We can discuss further the ways in which your experiment (or any experiment) may fail to mimic reality and the ways in which those differences matter, but I say tomato and you say tomato (you’ll have to imagine the different pronunciations), and you have the burden of proof, particularly if you want to draw policy implications from your single, flawed (they are all flawed–it’s the nature of the beast) experiment. To add to the points I made yesterday: why do you think that even your evidence about creators of paintings “over-valuing” their works translates into the patent realm at all and the commercial realm (you know–the place where deals are actually most relevant to human welfare) for any IP? These are just assertions, and pretty simple logic suggests that the assumptions may be wildly inaccurate. I know you state in your paper that creation is much more of the non-commercial sort than economists assume, but does that hold for inventions as opposed to poems and paintings? And does that hold for the most-commonly traded works? And is it really true at all? You cite only work relating to copyright for this proposition, and it is a Habermasian (if that’s a word) criticism of economics, at that–hardly an empirically-relevant cite.

    Moreover, you correctly note that, all else equal, over-optimism rather than idiosyncratic over-valuation (if there can be such a thing) should be easier to correct, but your own experiment pits painters/authors of their own works (as I understand it) in an emotional and fundamentally non-commercial setting and then asks them to engage in commerce. You really think their over-optimism is just that and not idiosyncratic valuation? Why? How do you know? I do appreciate the point that the experiment was set up not to require alienation of the work itself, which you would think would go some way toward mitigating this point, but certainly not all the way, and we have no way of knowing whether the deviations you observe could be fully accounted for by idiosyncrasies (the elimination of which would seem like a cruel and inefficient exercise).

    And on that “over”-optimism–how did you establish your baseline? I do like how your assumptions go nicely with your preferred outcome (take the protections of IP down a peg) rather than the reverse, but on what basis do you claim this? Frankly, it is a problem (as I suggested in my post yesterday) with all of BE: The assumption of a baseline. But without a theory (as David Friedman points out) to explain the baseline and the source of the deviations, it’s all just random noise. Perhaps the creators are exactly “appropriately” optimistic, and the prospective purchasers pessimistic (or skeptical, or resource constrained–did you consider an income effect here? As the chance at winning the contest has monetary value, there could be an income effect. Could that explain everything? I don’t know–but do you?). I notice in your paper that you run the experiment both as a lottery and a contest, and that there is more optimism bias in the latter–not surprisingly where there is a subjective determination to be made, and where my claim holds. I concede that in a lottery one could identify “over” optimism given that probabilities are objective, but a) this is where learning and commercial settings come in and make a huge difference and b) I notice that buyers as well as sellers are “over” optimistic in your experiment–and by not-dissimilar amounts. So where’s the problem? There is only a problem where there is a divergence in valuation, and if everyone over-values, so what?

    More important (in terms of an assumed baseline and the consequences of deviations from it), you seem to assume that rational choice requires some kind of objectivity in price. In actuality–in real markets–we don’t exactly know how prices are determined, and certainly they do not (how could anyone have enough information to know?) conform simply to “fundamentals.” Again this goes to the oddity of the experimental setting you have created–one where, arguably (and certainly in the lottery) there is some certainty about price. Confronted with this setting in the real world, there is little or no commerce; in the real world, however, prices reflect a lot of different things, and it is an unlikely exercise to determine deviations from “correct.” In other words, I’m troubled by the claims that initial expectations about price (even in the lottery setting, allowing for idiosyncrasies and innumeracy) without any opportunity to bargain or repeat play indicate some sort of irrationality, where it is not at all clear that the market wouldn’t clear at some price, even if, to you, that price deviates from “correct.”

    More on this experiment design point. As I say, your experiment creates a wholly unrealistic market and asks non-commercial actors to engage in some form of commerce around it–and then you want to draw policy conclusions? There are contests all over the world and, as far as I know, not a booming market in derivative contracts around them. Could there be a reason for this? Could it be that people do a remarkably good job of recognizing when exchange is not value-producing and avoid those situations? Could it be that an experiment run in a setting that doesn’t exist in any way (that I know of) in the real world is a poor indication of how real-world actors would act? At least the Cornell coffee mug experiment was a familiar, basic format: bartering for stuff among students.

    It also seems like a remarkable oversight that your paper doesn’t cite the Plott and Zeiler work on these sorts of endowment effect experiments, nor, as far as I can tell, take account of any of the problems they point out (although you may have done so inadvertently by avoiding the specific pitfalls they discuss). Perhaps you feel it is possible to do this (correct for biases in experiment design), and perhaps you feel you have indeed done this, but why don’t you discuss or even cite the literature to support your experiment design? Again, merely stating that your experiment is policy-relevant is not the same as it actually being policy-relevant, and you might want to go some way toward defending the experiment’s design, execution, and explanatory value (and the limits thereof). You may do this in the paper even without citing Zeiler and Plott (although you really should), but if you did I missed it in my perusal.

    Finally, on remedy. I would first note that, in some sense, arguing for a “take now and pay later” regime in the face of what may be un-executed contracts because of idiosyncratic valuation or even merely transaction costs borders on advocating theft. Do you really want to be in the position of advocating that people be forced to enter into contracts they prefer not to enter into? As you point out, of course, there would be bargaining in the shadow of the law, but I do not agree as you suggest in your paper that this makes it all ok–bargaining with a gun to one’s head is not the same as freely entering into a contract. This is one of the great advantages of a pure property regime, and it is, at least to me, an important characteristic (freedom). At the same time, contrary to your assertion, there are systematic biases in the take now and pay later regime: The transaction cost of litigation, including especially delayed compensation, and the creation of poor incentives for inefficient infringers. I’m quite certain this is larger than the cost of bargaining and results in either non-compensation or (as you point out but don’t identify the direction) bargaining in the shadow of the law–with considerable downward price pressure and the likelihood of inefficient transactions.

    Well, there is more, but I really must do some work. I appreciate the effort at bringing some empirical meat to these theoretical bones, but I fear that your effort has come up short.

  4. 

    It seems to me that your running royalty solution could produce an inefficiency in the opposite direction–not failure to produce IP worth producing but the production of IP not worth producing.

    I am a prospective author. My prospective publisher’s objective estimate is that the novel I plan to write will sell 1000 copies. On that basis, the publisher offers me a fixed fee of $2000. That is inadequate compensation for the cost to me of writing the novel, so I don’t–the correct outcome.

    Suppose, however, the publisher offers me instead a royalty of $2/copy sold. I, optimistically, expect to sell 100,000 copies, so gladly agree–and a novel not worth writing gets written.

    In this case, at least, the property regime produces the opposite of your asserted result–not the failure of a bargain that should occur but the occurrence of a bargain that should fail.

Trackbacks and Pingbacks:

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

    […] Christopher Sprigman & Christopher Buffafusco, Behavioral Economics and the Road from Law to Lab […]