The Nature of Sequential Innovation
Christopher Sprigman, Christopher Buccafusco & Stefan Bechtold
How to pick between innovating or borrowing. “Cinderella Man” is harder to develop than another movie about Rocky. Risky, as is question about whether to develop another erectile dysfunction drug. Differently risky. In our framework, innovation is not always optimal, either privately or socially. Innovation doesn’t produce novelty by necessary; borrowing isn’t necessarily a change from stasis.
Literature on innovation and the firm; literature on IP and sequential creativity. There’s a rich literature on innovative firms, but what about borrowing firms? Factors that influence sequential innovation. Literature has looked at legal factors, esp. IP law. We want to broaden that focus to consider not just legal factors but others—market, behavioral, tech/artistic.
IP laws affect the scope of available innovation space. For example, blocking patents allow more room for sequential innovation than ©’s derivative works rights. But also: administrative law, tort law, tax law—differentially treats innovation v. borrowing. Nonlegal factors: maturity of the field. Borrowing expands until a certain point and then leapfrog innovations redefine/expand the innovation space. In smartphone, innovation is highly dependent on borrowing, as opposed to painting. Tacit knowledge: difficult to convey; can reduce borrowing. Market factors: consumers’ taste for innovation in a particular field v. borrowing. Low tolerance for borrowing in paintings; high tolerance for borrowing in pharmaceuticals. Market participant/intermediaries: PROs lead to a lot of borrowing of musical compositions.
Behavioral factors: risk/uncertainty tolerance (innovation has greater risk profile); optimism bias (sanguine about ability to create/invent around); creativity effects (upstream creators may overprice inputs to downstream). Much more complicated than changing the law to create a clear fix. Policy levers: if you try to shift pharma from borrowing to innovation, will there be factors in market/tech that push against this or are legal factors from other areas, such as tax, much cheaper way to shift mix of innovation and borrowing in a particular field?
Lemley: innovation often goes along w/having to take a license b/c of patent threats. How does that affect your space? The story for innovating v. borrowing assumes that I put in the uncertain, risky work but may get a valuable reward, one part of which is insulation from control by other people. That might be more true in ©, but not true in patent. May drive people in a curious, ironic way towards borrowing.
Sprigman: in patent, b/c independent invention isn’t a defense, part of risk of innovation is info risk, and that might not be fixed by search.
Q: overlap in rights?
Sprigman: location/negotiation of lots of rightsholders can be a problem. Smartphones = tons of transaction costs.
Rosenblatt: I could imagine exactly the opposite story—if creators are risk and uncertainty averse, they’re less likely to borrow b/c they don’t know what the law will do (uncertain about legal effects of their actions), so they’ll go away as far as possible. Optimism bias could drive them to think they won’t be sued or that they won’t be found to be infringing.
Buccafusco: there are probabilities on both sides of the equation, so it’s really hard. You can’t be uncertainty averse to both; what matters is relative salience; whether people treat those as losses v. gains, etc.
Q: what about people who just decide not to take either risk? [People v. firms? One might decide to be an employee, but can a firm decide this?]
Buccafusco: theory of the firm literature does assume that you plan to act. What decisions will you make and how? You’re right that you could ask why people decide to act in the market at all, but we are taking a different temporal slice.
Copyright and Distributive Justice
Justin Hughes & Robert Merges
Distributive justice: where does the money go? Claims in IP scholarship are that copyright has mostly enriched big corporations. Copyright could improve distribution of wealth in terms of money and property. © offers significant benefits of wealth to individuals at all levels, which should not be overlooked. © is particularly important in allowing African-Americans to convert labor and talent into money and wealth. Rawls: arguably the most important part of his framework is the difference principle. Inequalities are permissible if they have the greatest benefit to the least advantaged. But Rawls’ actual difference principle isn’t just concerned w/the bottom—should be concerned with society, from the bottom up. It’s fine to improve the middle and the top as long as you don’t worsen conditions for those at the bottom.
© improves wealth distribution to middle class. Peter DiCola’s money for music study: pro musicians derive 10% of income on average from ©; we think that undercounts what’s happening. Collecting societies distribute enormous amounts of money to creative professionals. Five years, two PROs collecting only for compositions only for public performance distributed $4.1 billion to individuals. Also should include amounts distributed under Hollywood’s collective bargaining system.
Procedural protections for individual authors to make their rights sticky. The most powerful but glorious and beautiful mystery—the statutory termination of transfer right. If we wanted to strengthen © as a distributive tool for creative professionals, we’d look more at these bells and whistles.
Second principle: inequalities are permissible only if attached to offices/positions open to all under conditions of equal opportunity. True meritocracy. American society has failed to provide equal opportunity, and no group has suffered more than African-Americans. But this is a bright spot for © distribution. ©, warts and all, arguably provides the most robust mechanism for the most disadvantaged group in American society. We say this acknowledging tremendous problems w/actually ensuring that African-Americans receive the full benefits of ©. Fumi Arewa, K.J. Greene, others, have written about this. But their typical diagnosis is not weaker rights but broader or better enforcement. And despite all those problems, list of wealthiest Af-Ams almost all derive from ©-based industries, principally music and broadcasting rights. In an era when tech seems to be weakening middle class incomes, we should pay att’n to ability for © to protect individuals.
Does © exacerbate or ameliorate the skewed distribution of wealth in our society? The latter; we should focus on strengthening income from ©.
Q: © can channel individual rewards, but does favor superstar imbalances b/t individuals. Winner-take-all: worse overall?
Merges: long tail distribution issues; very unfair and averages hide the unfairness. But when you look at distribution from BMI/ASCAP, have to compare that to research talking about income of average musicians.
Dan Burk: If I don’t buy Rawls and instead like Nash, do I have to buy your paper?
Merges: It’s a good way to analyze copyright in a rigorous way. You don’t have to ask if the least well off are rigorously compensated, but you can ask if © contributes to the wellbeing of non-superstars in a meaningful way. We think it’s even easier to meet that standard.
Buccafusco: you tell a story about a few people, which is odd in a paper about distribution. There’s no counterfactual showing that distribution would be different in another © world. But more importantly, your story is about the people at the bottom and middle but your evidence is about people at the top. Do you have evidence that in making the top better off you’re not making the bottom worse off?
Merges: nor is there any evidence they are worse off. No reason to believe that © has made them worse off w/o rewarding/compensating them w/ entertainment value. That’s just our assumption and we’re going with it. [OK then. It is a condition of Rawlsian justice and therefore of the premise of your paper, but ok.]
Buccafusco: Term of transfer rewards currently wealthy at the expense of the currently poor. People who are selling rights now = less valuable contracts because they can be terminated if they’re successful, and those people starting out are poor now.
Hughes: most people who know the industry think that’s dumb b/c no one goes into a contract thinking that any record will have value in 29 years. Termination benefits 10,000 songwriters you’ve never heard of and they’re just as important as Bruce Springsteen. [Is that how many terminations there have been? I’m pretty sure there haven’t been.]
Rosenblatt: is this exclusivity based? Mix tapes, sampling—a good deal of innovative copying. Worth taking into account.
Merges: Voluntary decisions to waive rights to build market share can be very effective. But once you cross a threshold © is valuable to protect your rights. [I don’t think that was the point, but ok.]
Anupam Chander: Income makes people better off, which is great. The other side is monopoly rents, and those rents are being paid by someone. What’s the distributional effect of the monopoly rents? You said this was good for the middle class, but your example was 25 people for whom © is very good. What is happening w/people at the bottom? Are you relying implicitly on violations of © by people at the bottom?
Hughes: We were certain that people would think we were talking about distribution of wealth to Af-Am community as a whole. [Possibly because that’s what distributional discussions usually entail and what Rawls seeks with his first principle, which you discussed at the beginning? Or possibly because of these sentences from the opening paragraph of the paper: "Is our copyright system basically fair? Does it exacerbate or ameliorate the skewed distribution of wealth in our society? Does it do anything at all for disempowered people, people at the bottom of the socio-economic hierarchy? In this Article we engage these questions." Stupid readers!] Rawls = equal access to stations and offices. For Af-Am community, access to highest offices of wealth is through ©. We aren’t talking about distribution to other groups or within the Af-Am community. [Not clear what distributive justice has to do with the claims, then; you might be making a claim about openness to talent, but that actually is only one part the overall Rawlsian framework--the way that inequality is justified within a society that has adhered to the minimax principle.] Access to Madonna isn’t the same thing as access to medicine, anyway.
[There is a plausible narrative here that has appeal: because of lower capital costs of entry, entertainment has been one of the easier ways for some extremely talented African-Americans to make lots of money, whereas other methods of discovering and exploiting talent often require capital—cf. Bill Gates and Steve Jobs’ access to significant social and physical capital, including their ability to trespass and break things, which could have proved lethal to African-American boys. However, the implication of putting it that way would not seem to be “© should be stronger” but rather “hey, that’s terrible; other means of exercising great talent ought to be equally open to African-Americans too.” If anything, the relative disparity feeds into a critique that American culture too readily channels African-American talents into entertainment and sports fields; the solution to inequality is not to close off those opportunities, but neither is it to double the number of NFL teams and have the government fund a lot of music purchases, nor to mandate Content ID for all websites.]
IP, Privacy Harms and other Fundamental Values
Misalignment of law with values of creators. Many described liability problems; problems w/trustworthy help such as studio assistants, business managers. These aren’t IP problems most of the time. Sometimes they’d be fit into IP, but not a good fit.
Equality, privacy, and distributive justice are the values that creators seek, but typically not a lot of place in our conversations about IP levers, efficiencies, markets, and entitlements. We need a theory of what Progress is. Not necessarily about wealth aggregation/more stuff. Some may consider some of the problems to be IP overreach, but that begs the question of what IP is for and what fundamental values ordinary creators want to use IP to protect.
Rough project: equality, privacy, fairer uses, abuses. Here, focus on privacy. Five clusters that explain different privacy interests and harms. Goal: understand in broader cultural way. Constitutional concerns: privacy as condition of spaces & things (houses etc.); bodily privacy; mind and relationships (religion, speech). Nonconstitutional: info privacy: public disclosure of private facts; misappropriation; false light; intrustion on seclusion. Three things we care about: independent thought (ability to formulate one’s own ideas); fortifying relationships (bodies and privacy among communities); flourishing culture and science (allows public realm to succeed).
Cases brought by heirs: copyright war over Duchamp chess set. JD Salinger, Ted Hughes, James Joyce. These cases are really about nostalgia for family relationships. They aren’t literary reputation cases; they’re about family memories trying to preserve them w/o interference of others. Owning memories is futile, though.
Authors and owners enjoining publication of previously unpublished works. Intrusion into seclusion; spatial privacy as well as intellectual privacy. Willa Cather didn’t want anything published after her death—free from constraint of oversight of others.
Limited publication: breach of confidential relations/trust. Snapchat, FB, email. Digitally, these limited publications don’t really exist any more. IP claims are tempting to punish a breach of confidentiality.
Fair use cases: recontextualization of works that have already been published. NYT sued David Shields for thumbnail images in a book criticizing NYT photos; graffit artists objecting to being used as backdrop for ads. His lawyer: if they didn’t want to work with him, they could fuck themselves and find someone else—he wasn’t interested in licensing.
Last case: suit is brought by subject of work. Copyright suit through assignment.
Except for unpublished work cases, process/purpose of creation is largely irrelevant in these cases. These are claims about identity and affiliation being inseparable. Privacy claims are bilateral—their contours are always defined by the interests of others. And so is IP as a balance b/t ownership and access. Privacy requires intrusion; authorship requires public domain.
Most of these cases involve privacy losing; IP claims are much more disputed. If the interests are similar and the claims are similar, culturally and values-wise, they probably should rise and fall in a similar way.
Payoff: (1) rethink progress; (2) rethink value of rule of law, given blending of different types of claims; if we care about that, we should think about alignment; (3) discursive shift asks us to think about moral narratives; economic claims are moral narratives but so are these. Language wars are policy wars, says George Lakoff.
Linford: privacy claims can be stalking horse for much less plausible claim. Is there a proper way to police against this? Monge v. Maya. P wins, suppressing wedding photos. Argument: Hurts them as celebrities.
A: that’s an unpublished works case; it’s a strong privacy claim and for copyright. Should be in the control of author/owner, just as intrusion into seclusion is strong interest. My views have also changed about heirs’ restricting access to unpublished (and unarchived) works.
Heymann: Law as communicated by judges v. law as understood on the ground. What does the feedback look like here in terms of internalizing how the law should be used?
A: trying to figure out recursive nature of law is sociological project. Giddens’ work on structuration has levels of feedback loops. Just beginning to figure out how meaning is shaped in different ways by different voices. Methodologically: what’s the data here? Anxious about that—need to be clear about where the claims about domestication are coming from.
Q: have we seen anything like this before w/r/t complicated legal framework w/ strong moral narratives among general public and strong distributive justice components, that is, w/tax? Is there an account of tax history that you could draw lessons from?
A: Not so many studies; there are studies of bankruptcy practice.
What’s the Harm of Trademark Infringement?
Have to say something outrageous to justify your hanging around for the very last speaker. I thought about saying that I’d developed a new and coherent justification for trademark dilution, but then I thought that even in the craziness 2016 has brought us that still wasn’t credible. So instead I will challenge the basic function of trademark and the crucial rule that I learned at my managing partner’s knee: the remedy for trademark infringement is an injunction.
My challenge is both casually empirical and more seriously normative: eBay asks us to rearticulate why the remedy for infringement ought to be an injunction, and it turns out that the reasons for granting injunctions in TM don’t make much sense outside the core purchase substitution situation that so many trademark cases no longer resemble. And it turns out that when plaintiffs are asked to particularize their harm stories, explaining why this confusion is specifically likely to cause the plaintiff harm, they often aren’t very good at it.
Old perspective, articulated by Jeffrey Sanchez: “The basis for the presumption of irreparable harm in trademark law is the known or proven fact that monetary relief from trademark infringement is ‘inherently “inadequate” and injury is “irreparable.”’” But by whom is this known and how was it proven?
eBay rejected near-absolute presumptions in favor of a patent owner, or plaintiffs generally, requiring a plaintiff to show irreparable harm to get an injunction. Okay then: What is irreparable harm? Easy case: the defendant has no money to pay damages. Might be reparable under other circumstances, but not on these particular facts. Counterfeiting cases may fall into this category. Medium cases: we’re convinced that there is harm, but we don’t think it will be possible to measure the amount. This is a relatively common justification for irreparability in TM cases, but it has a real weakness: we need to distinguish between lack of certainty about whether there is harm and certainty that there is harm plus uncertainty about its amount, and that’s pretty hard to do with current techniques in TM cases.
How have TM owners traditionally gotten around this problem? By claiming harm to the intangible value of their goodwill. Related question: What is goodwill? TM owners tend to treat it as a word to conjure with. But it faces the same problem as lost sales: if it’s a business asset, it can usually be measured, because accountants and investors like that. If it can be measured, harm to it ought to be measurable, which means it can at least sometimes be cashed out in damages.
Even if infringement leads to lost customers and not just lost sales, it is possible to calculate the present discounted value of a customer, not just a lost sale. And of course lost customers are a big if, in many cases—we say that disappointment in an infringing product may turn customers away forever, but as Mark McKenna has painstakingly documented and I’ve also written about, that’s really not likely to be true in most cases. Strong brands are extremely resistant to change, and we know this even in other TM contexts. My favorite example: courts and the
Trademark Trial and Appeal Board have found that university mascots and names have retained
trademark significance despite uncontrolled use by others for decades and, in one case, for nearly two centuries. People just have terrible incident memories and usually substitute general impressions to form their opinions about brands, which also leads them to make mistakes about, say, who’s sponsoring the Olympics. People think about prominent brands when you cue them with the product area pretty much no matter what, which, first, causes noise in confusion determinations, but second and more importantly for my topic, throws doubt on the basic theory of harm from non-counterfeiting infringement: people are walking around confused about the relationships between famous brands and others all the time, and the brands stay famous and profitable.
Often, what might be lost by infringement, instead of “reputation,” is licensing revenue, which we know from patent and copyright cases is usually reparable by damages.
Intermediate conclusion: Goodwill is intangible but not generally unmeasurable outside of TM cases. Perhaps notably, the one intangible harm that the SCt has been really clear is irreparable is suppression of First Amendment rights, which is not an irrelevant consideration in TM cases.
But suppose we accepted that these negative effects on reputation could really occur from bad infringing products. Would that be irreparable harm under eBay? Even if we required the plaintiff to prove a quality difference, the risk of harm to any particular trademark owner would still be low. When we wait for evidence of such harm it may fail to appear. At most, a poor quality brand extension makes consumers less likely to be interested in a different, related brand extension in the future. This is a market preclusion argument, not an argument for ongoing harm, and it’s particularly unlikely to reflect an immediate risk to a trademark plaintiff, which is what eBay supposedly looks for. And the weakness of the reputation argument also threatens another traditional argument in TM, which is that lost control over one’s own reputation is inherently irreparable. That’s just a misperception about risk, or a sub rosa lowering of the standard from likely to theoretically possible.
The paper discusses a few cases that demand more than just cursory statements about goodwill, reputation, and control, and I’m largely in agreement with them. There are still plenty of the traditional cases too, but their rationales are increasingly creaky and come down to “we’ll let the district judge decide there’s irreparable harm because lost control can be the basis of a finding of irreparable harm even if it doesn’t have to be,” which I don’t find very persuasive.
One interesting example: Uber Promotions, Inc. v. Uber Technologies, Inc. In this case, the well-known national brand Uber Technologies was found to have caused actual confusion with the transportation business of local senior user Uber Promotions in Gainesville, Florida. In finding irreparable harm to Uber Promotions, the court noted Uber Technologies’ extremely controversial and often downright bad reputation. For example, the court pointed out that, as of the time it wrote its opinion, top news stories for “uber florida” included numerous stories about Uber Technologies’ exposure of a driver’s personal information, including her social security number. It concluded: “With all due respect to Tech, Promotions has every reason not to want potential customers and other members of the public to associate it with a company that has inspired protests in cities around the world.” Bad product extensions are one thing” confusion about an association with poorly performing products isn’t likely to be harmful. Confusion with a brand that triggers riots and boycotts could reasonably be predicted to be substantially more harmful.
What next? I think greater attention to the harm stories of particular kinds of infringements could help courts understand what academics have been saying about the overexpansion of infringement liability to situations where there’s no real benefit to consumers and potentially severe harm to competition or free expression. My usual hobby horse: materiality is also useful in figuring out which cases might involve irreparable harm.
Q: What about harm to the consumer?
A: (1) Doctrinally, separate factor. (2) Turns out that w/o harm to TM owner, it’s hard to explain how consumers would be harmed either.
Lemley: Not doctrinally separate for sure except in 9th Circuit, and they’re weird. [True.] Why shouldn’t it be a balancing test? Why shouldn’t we enjoin when there is harm to public even if $ would redress the TM owner’s injury?
A: I am persuadable on whether there should be a balancing test, esp. with factors (1) and (2)—if there is no adequate remedy at law I would often want to call that irreparable. However, I’m not convinced that TM owners are good proxies for harm to consumers. If consumers are harmed, they should (or consumer protection authorities should) act on their own behalf; we generally ask plaintiffs to have standing by showing harm to themselves, and so too here.
McKenna: Throwing consumers into balance weakens argument for injunctive relief: evidence suggests that consumers can adapt to new marketplace if they learn they can’t rely on this as a signal—they just create sub-brands, excepting cases where the products are really close.
A: thanks; should also consider people benefited by D’s conduct.
Ramsey: when likely confusion is found, court shifts into anti free riding mentality, lets TM owner control uses. Makes sense that they’d instinctively grant injunctions.
A: Interestingly, eBay says you’re not supposed to do that even for “property” rights—as we’ve seen in other papers, property/liability isn’t that simple a split and it’s a lot more complicated than enjoining “property” violations.
Rosenblatt: Isn’t part of the concern durational—the longer confusion persists, the more likely irreparable harm is?
A: Except I think the evidence for that is poor. Confusion may be harmless and stay harmless even if it doesn’t dissipate. Suppose the D’s product isn’t yet crappy but might theoretically become so—that’s not likely irreparable harm, just possible harm; it doesn’t meet the standard set out.