Saturday, February 19, 2022

WIPIP 2022, Session 7 (internet law/antitrust)

Sari Mazzurco, The Law of Social Roles for the Platform Internet

Law’s expressive function: how law tells people what social roles various institutions are supposed to carry out. Policy discourse should explicitly consider the social roles legal reform would construct for platforms to help public understand whether platforms are doing what they should, guide platform behaviors, and guide further reform. Legislation should be attuned to multiple social roles platforms play. Social roles are fundamental: shorthand that help us understand what behavior is appropriate or inappropriate based on relationship context. OK to ask barista to make a latte and pay her for it, not so much for one’s PhD advisor. Law can shape roles: set exit and entry conditions (e.g. licensing) for roles. The roles of business and consumer have been applied flatly and wrongly. Businesses have been assumed to be asocial; consumers are satisfying self-interest. Businesses harm consumers only in a few ways—false advertising, monopoly prices, defective products. But these roles support only very thin social norms. It’s ok for businesses to do most profitable things. It’s ok for consumers to care about price and quality, and not much else. What did FB do wrong when sharing information with third parties when it told people it was going to do that? Business frame obscures broader privacy harms.

A woman who feels shame when her breastfeeding photos are taken down is not reacting as a consumer; she is reacting as a speaker whose culturally meaningful speech has been censored, and censorship is not a business-related concept.

Evaluates various proposals with attention to roles, e.g. digital due process—platforms as speech governors and users as democratic participants interested in collective and individual autonomy, interested in accessing a wide variety of speech that complies w/public morals. Platforms should moderate content in the public interest.  Lawmakers should enact regulations that create

RT: (1) When you say platform, do you mean FB? Conflating the two is common but misleading. Mike Masnick’s test suite is worth thinking about. (2) If FB should be democratic, why would we accept due process in place of voting? Being the authoritarian in authoritarian constitutionalism is a role, but is it a role you are happy with? (3) “Sir, this is a Wendy’s” meme: mistaking a place for having a different role is pretty common (and also correlated with racial and gender privilege). Sometimes we just tell people that they’ve mistaken the role, even if they feel real shame/outrage about how they’ve been treated. (4) Doug Kysar’s Preferences for Processes is an interesting exploration of the consumer frame.

A: Any platform could occupy multiple social roles, which help us connect harm to expectations to behaviors. Use FB, Google and Amazon b/c those are in the press and stir up controversy, but don’t mean to talk just about them except as monopolies.

The implications of the various bills frame platforms in various ways, and that highlights dimensions that may be problematic. Technological due process sounds good but may also accept the authoritarian constitutionalism frame. [Which I think deserves attention—if we convince enough people that the “role” of a democratic institution is carried out by providing individuals with individual due process on specific rulings but no ability to set the rules, that’s really bad for democracy.] Privilege: if a person needs to know what the role of the place is, then they do need to know what’s appropriate, and knowledge needs to be provided by normal social channels or by law. [Fair enough, but our current mechanisms of doing that—including law!—clearly do convey different messages to different groups about their social status/ability to speak back/dignity interests/etc.]

Role definitions can render certain interpretations “irrational.” If the law says that Google is a common carrier, expecting more speech governance is irrational. [I’m not sure anyone is unclear on that, though—different politicians/groups just want conflicting things from regulation and Republicans in particular are presently pushing common carrier while Democrats are presently not.]

Sharon Sandeen: Unfair competition is a different potential frame than either business or antitrust. Law of warranties is another example of an apparently obvious analogy of how we might regulate.

A: doesn’t see that as a path forward b/c that goes to price, etc. and harms that are physical or financial—a more holistic approach to dignity. [This is part of my “platforms aren’t FB” hobby horse. I just don’t see Ravelry as having the same relationship to dignitary interests.]

Jess Miers: Think about services’ countermoves to these regulations. Removing UGC aspects of services is often a reasonable response, not the expected “moderate more/moderate less.”

Gary Myers, Old Wine in New Bottles: Applying Antitrust Law’s Aspen Skiing and “Essential Facilities” Doctrines to Address Big Tech Challenges

As an antitrust lawyer, does antitrust have anything specific to say about the situation? Trinko case limits idea of access to essential facilities. Aspen Skiing/essential facilities were traditional antitrust doctrine that the SCt basically said were possible avenues for dealing w/large firm. Trinko, dealing w/highly regulated telecom industry, said that Verizon’s duty to deal w/competitors was governed by complex regulatory scheme and antitrust can’t be allowed to add anything to that. There are exceptions, which the Ct said were not overruled: Aspen Skiing/essential facilities.

Does Aspen Skiing indicate that a large tech firm might have duties to competitors/consumers? Possibly yes. If monopoly power is shown, which is possible, and barriers to entry, and significant network effects, FB might be monopoly; there are also attempted monopolization claims which don’t require as much market power, only a dangerous probability of success.

Likewise, essential facilities doctrine, edging toward common carrier—firms that control essential/important bottleneck for market participation can be required to give access. Similar to Aspen Skiing: access requirements can be imposed. Potential First Amendment issues, though. Business justifications are also a defense, but usually create jury questions. In Aspen Skiing, D claimed that P’s mountain was lower quality, which made it unworthy of cooperation; its facilities were older/not quite as nice. Jury rejected that business justification.

Would I take this case as an antitrust lawyer? It’s up in the air; an uphill battle partly b/c antitrust has become so narrow and business justifications get so much weight; having to prove each element is hard—tech cos will say they aren’t monopolies and the market is so dynamic that things can change tomorrow. The precedent is there despite today’s judges’ skepticism.

RT: What are the key differences between Aspen Skiing and essential facilities that make them different doctrines?

A: they are very close. Essential facilities is designed to deal w/a specific kind of bottleneck—telecom, bridge, road—a channel that’s needed. Aspen Skiing isn’t about a bottleneck but about changing a course of dealing, doing something harmful to consumers, etc. You could often assert both claims in parallel and Aspen did also feature an essential facilities claim. Their general monopoly claim was stronger.

Sandeen: can you find bad acts in terms of service especially as to businesses using the platforms to conduct business?

A: yes, you could definitely find problems there—restrictive terms, deplatforming, refusal to deal on nondiscriminatory basis. Amazon’s interactions w/ 3d party sellers are possible examples.

Tyler Ochoa: has essential facilities ever been applied to a telecom or other company where there could be a 1A compelled speech claim?

A: Not sure. Miami Herald v. Tornillo is not an antitrust case [but does mention the newspaper’s dominance in Miami, as emphasized by the recent Florida case striking down that state’s pro-spam law].

Rosenblatt: are all these situations (FB, Amazon) the same? This question arose w/net neutrality—access to the underlying pipes—versus kicking someone off of Twitter. When is this a useful tool?

A: there could be a myriad of possible situations, which is why this area isn’t amenable to broad brush statutory or regulatory rulemaking.  Does the D have enough power for us to be concerned? Not all tech companies have power that make their actions an antitrust issue. Power is a big deal; secondly, is it doing something that looks exclusionary/anticompetitive v. legit competition on the merits.  

Q: As a marketing professor, the word I say every day is Google. What are you going to do instead, advertise on Bing? [cue laughter; I did just see a TM case entirely about Bing, but that was probably dumb] Isn’t that an essential facility?

A: quite likely. Access to organic results for sure, though that shouldn’t disable them from putting sponsored ads up top.

Sandeen: a unique dynamic online: use consumers to make the decisions about whether or not somebody is going to be kicked off. If your ranking goes down to 1 star on FB, you get kicked off. FB would claim that wasn’t their decision.

A: they would! They have a case for that, too.

Jess Miers: Reddit is now a big search engine; Tik Tok has surpassed Google.

A: it is a dynamic market! One irony of antitrust law is its slowness.


Kristen Osenga, Can Antitrust Learn Something from IP? (working title)

Rep. Jayapal says that Amazon harms competition with its private labels. Collects data on sellers and produce competing goods. 60% of overall sales according to Amazon are 3d parties; they say only 1% is private label products. Europe has complained about this too. Also allegedly favors its own products in search results.

What nonpublic data are we concerned about? Sales, revenue, consumer claims/warranty claims, etc. Lets Amazon focus retail competition on best-selling products. EU has proposed prohibiting use of sellers’ data and prohibiting self-preferencing. The US is following suit in proposed bills. One would prohibit self-preferencing; one would prohibit offering both first- and third-party offerings. Amazon says this would hurt consumers.

As an IP person, has qualms. If something isn’t protected by IP, it’s free for copying. We like copying! Gives consumers more choice, lower prices. We have provisions to get generic drugs to market. Confusion is what we try to avoid, not copying. Why treat Amazon private label products differently?

What’s really bad here? Amazon’s ability to collect lots of data and use it to decide what products to make; Amazon preferences itself in ads. But the same activity happens in brick & mortar stores and no one is calling for CVS and Wal-Mart to stop making private label goods. We know it’s riding on coattails of national brand’s research and development and advertising. But we have recognized their benefits when not deceptive. CVS has lots of data on sales, when Vaseline is having sales, etc. CVS is not going to make unpopular generic products. They are going to use the data to ensure they’re underpricing third party sellers.

Is Amazon different b/c of scale/type of data? You can see how long something waited in a person’s cart or how long they spent on site. But you can get the same data if you’re CVS on your website. Putting them side by side decreases the likelihood of confusion.

Focusing only on these aspects—possible that Amazon is doing really bad things.

Justin Hughes: isn’t the answer that by all estimates Amazon controls 36-49% of ecommerce, and CVS etc. don’t have anything like that, not even Wal-Mart? [Wal-Mart is pretty big though in bricks & mortar.] Amazon is the bottleneck.

A: would accept the point, but maybe market share is ok.

Hughes: but don’t abuse your market power.

A: but what is abuse given the benefits of data.

RT: Counterargument: The price differential with the strong brand is empirically robust over decades because of the consumer preference for national brands. With the current subjects of complaint—small businesses—Amazon is the real brand driving sales for most of these products, and so the free riding is less sustainable for a business that doesn’t (yet) have a strong brand. If you think about misappropriation rationales, whether the product would still be produced w/the free riding is a consideration. Maybe we’re wrong about the empirics, but I don’t think we can exclude the possibility that more damage will be done to weak brands that can’t continue to command a price premium. For one thing, Amazon is not free riding on the advertising of these producers, because there isn’t any off Amazon; most of them aren’t like Head & Shoulders, which can advertise and sustain its brand despite the house brands. I’m not convinced that this is a distinction that makes a difference, but I’m not convinced it doesn’t.

A: two buckets—small and medium companies may be different. Batteries often comes up.

Sharon Sandeen: Trade secrecy background—these bills have no definition of what nonpublic information is. Small producers are mad at Amazon and want something done, and this is something, but it may not address what is really going on. Amazon may also be able to redefine information by contract.

Q: the complaint is that we don’t have patents or trade secrets, and we don’t yet have a brand [secondary meaning in the making!] and so Amazon swoops in. One question is whether that’s what Amazon is doing—appropriating innovative products that aren’t protected by IP—or whether they’re making batteries and towels, which I would care less about.

A: it’s popular goods—not clear about size. But Wal-Mart does the same thing!

Lunney: Generic house brand just creates price discrimination where otherwise there would be a [tiny little] monopoly; the parties can split the resulting surplus as they see fit.

Hughes: Wal-Mart has been accused of the exact same behavior [of getting into the OEM space and demanding supply at lower prices] [Lunney: that’s where Wal-Mart v. Samara came from!]

No comments: