In 2015, Seattle passed an ordinance that provided for collective bargaining between for-hire vehicle drivers and their firms, including Uber and Lyft. The ordinance was challenged by the Chamber of Commerce, and the case quickly made its way to the Ninth Circuit Court of Appeals. After the Circuit ruled that Seattle’s ordinance was likely not “clearly authorized” by the State of Washington, Seattle recently entered into a compromise in which it removed payments to drivers from the subjects of collective bargaining.
As long as there is federal jurisdiction, federal law has “supremacy” over state law or over municipal law, which is a delegation from state law. However, antitrust’s state-action exemption generally immunizes state laws that authorize conduct that would otherwise violate federal antitrust law. In the Seattle case, the principally litigated issue has been the scope of the state-action exemption and whether it extends to the municipal ordinance in question. In a controversial amicus brief, the Federal Trade Commission (FTC) and Department of Justice (DOJ) sided with Uber, arguing that the state-action exemption did not immunize the Seattle ordinance and assuming that the collective bargaining activity contemplated by the ordinance would otherwise violate Section 1 of the Sherman Act. Given the uncertain borders of the labor exemption to antitrust, which protects collective bargaining activity as well as the unilateral collective action of workers as to the terms and conditions of their work, independent contractors live in a state of legal uncertainty when it comes to exercising their coordination rights. The competition authorities generally seem to assume that they are unprotected. Some commentators such as Marshall Steinbaum highlighted the dissonance between the FTC’s and DOJ’s anti-labor views and the agency’s original mission, as captured in the FTC Act, to deconcentrate private economic power from trusts to smaller economic agents. Uber itself has been the target of a recent antitrust suit. An Uber passenger brought an antitrust case, on behalf of himself and other Uber passengers, alleging a conspiracy against Uber, for fixing prices among all of its “independent” drivers. In March 2018, the district court granted Uber’s motion to compel arbitration. In a new article titled “Antitrust As Allocator of Coordination Rights,” Sanjukta Paul explains that antitrust law allocates the right to coordinate decisions such as pricing or output across economic agents, and does so favorably for large powerful firms but unfavorably for workers’ organizations and small businesses or “micro-enterprises.” The ostensible basis to prefer coordination by large firms is promoting competition through the pursuit of efficiency. But even that basis, Paul argues, fails to explain many antitrust decisions that yield significant coordination rights to large firms while undermining competition via concentrating power. To reach parity of treatment between these varieties of coordination, Paul calls for liberalizing horizontal coordination rights beyond firm boundaries while providing mechanisms for public oversight. Has antitrust steered too far in favor of large firms? And if so, what’s the best way to redirect the antitrust agencies back to a mission of dispersing private economic power? To answer those questions, we invited Wayne State University Law School’s Sanjukta Paul, International Center for Law and Economics’ Geoffrey Manne, Marshall Steinbaum of the Roosevelt Institute, and fellow Bytes contributor Jodi Beggs to a Bytes Chat. The Chat was moderated by Hal Singer, editor of Washington Bytes and Senior Fellow of the George Washington Institute of Public Policy. The transcript was edited lightly for readability, and each contributor had the opportunity to refine answers and add hyperlinks after the Chat. * * * Hal Singer: Greetings, y’all. Looks like some folks have jumped the gun on our chat. But this is the official version right here. Geoff Manne: Such a control freak, Hal. Marshall Steinbaum: Kind of like Uber. Jodi Beggs: If Uber were actually a control freak, it would make drivers employees so it could tell them when and where to work. Steinbaum: But it does exactly that without re-classifying them! In fact that’s the entire point. Manne: Amen, Jodi. In any case, here’s the key point: Why should employee/contractor status under employment laws be relevant to the assessment of injury under antitrust law? Steinbaum: Chat over. Singer: Chat beginning. Let’s start with the FTC’s decision to back Uber—contra the interests of Uber’s drivers—in the Seattle case. Why did the Ohlhausen/McSweeny-led FTC and the Delrahim-led Department of Justice feel the need to weigh in here? How does limiting the coordination rights of Uber drivers or any gig-economy workers fit into the FTC’s larger mission of dispersing economic power? Steinbaum:It’s part of their long-running campaign to roll back the state action exemption, which Sandeep Vaheesan has shown is part of a larger mission to “accommodate capital and police labor.” Manne: Hal’s question is the wrong question if you’re actually trying to understand the FTC’s interest here. The right question is: “Is the FTC doing the most important thing it can do, or just oneof the most important, in taking on state actions that prop up cartels and otherwise use the power of the state to violate antitrust laws for the benefit of a favored constituency?” Singer: Kinda liked my question tbh. How about we let our special guest speak? Sanjukta Paul: To answer Hal’s question whether the FTC’s decision to weigh in on the Seattle case squares with its mission, the short answer is, it doesn’t. Limiting the coordination rights of small players, including gig economy workers, runs counter to that mission, because looser coordination among small players helps to maintain the dispersal of power. Conversely, if we are permissive toward corporate consolidation and rigid toward the loose coordination of small players, the predictable result is the further concentration of decision-making and power. To be fair to the FTC and DOJ, they were acting very much in the spirit of the overall antitrust paradigm that has governed our world since around the late 1970’s, which favors economic coordination centered in large, powerful firms but looks askance on any other form: workers’ organizations, looser coordination beyond firm boundaries, and public coordination of markets. As I argue in the paper, that paradigm favors consolidation precisely because it concentrates decision-making and power—not despite it. Steinbaum: So, it doesn’t fit their larger mission, or at least fit the intent of the FTC Act, which is, as Hal says, to disperse power. Instead, that brief, among others, shows that they actually want to concentrate it. It’s also part of a long-running campaign on the FTC’s part to carve out regulatory space for Uber’s business model and prevent any challenge to that business model. See, for example, the letter that the heads of all the bureaus sent to a Chicago alderman in 2014 (I think) telling him that the “Free Market” means that Uber should be allowed to use surge pricing. Paul: And yes, rolling back the state-action exemption is another aspect of that same project. Manne: A theme of Sanjukta’s paper (and apparently Sandeep’s writings) is that the FTC is a witting tool of capitalist oppression. This is… uh, how to put this nicely… unsupported. The FTC letter Marshall cites isn’t about doing Uber’s bidding or cynically carving out some sort of antitrust exception for its business model. As the FTC letter states, “absent some specific evidence that a particular pricing model will harm consumers, [Chicago] should clearly allow for greater flexibility and experimentation in structuring fees in order to facilitate innovative forms of pricing that may benefit consumers.” Flexibility, experimentation and innovation for the benefit of consumers. Doesn’t really sound like protecting Uber from challenge. Do we have to/want to get into this now or at all? Beggs: As a side effect, the FTC’s policy drives a bigger wedge between employees and independent contractors as far as rights are concerned, which probably won’t help dispel the arguments that workers are having with their employers over their status. Manne: The FTC has litigated cases for years that on the sovereign immunity issue (like NC Dental or Phoebe Putney) because of its concern that state regulation can shield anticompetitive conduct—like dentists excluding competition from non-dentists to whiten teeth—for the benefit of a small group at the expense of consumers. Also known as rent seeking. The problem, of course, is that government-created cartels are protected by law. This is why the Supreme Court has repeatedly sided with the FTC in these cases and held that state-action immunity is “disfavored.” There are tons of examples of locally powerful industries capturing their regulators (like dental or medical boards, taxi commissions, or city councils) and implementing exclusionary regulations—like caps on the number of licenses—that bar new entry. The FTC amicus brief is wholly consistent with this long-standing interest in making sure that municipalities and licensing entities don’t facilitate or engage in anticompetitive conduct, except when expressly authorized by the state legislature and closely supervised by the state. Singer: Yes, Geoff, perfectly consistent with Supreme Court precedent that favors the powerful. Manne: Well, you asked if it fit the FTC’s mission of “dispersing power.” It clearly does. And what’s with the conspiracy theories? Steinbaum: It’s not a conspiracy theory. Uber’s antitrust lawyer is now the Deputy Director of the Bureau of Competition. And yes, as Sanjukta says, Uber is itself an illegal price- and wage-fixing conspiracy. Paul: Uber is the conspiracy. Antitrust is fine with conspiracies when they are controlled by concentrated capital, and engage in coordination through hierarchy. Manne: So the fact that a lawyer who worked at the FTC for 6 years, Verizon for 10 years, and Uber for a mere 2 years just returned to the FTC — this is proof that the FTC is on a cynical mission to prop up Uber?… Wait, Sanjukta: Uber is a “conspiracy”?!?!? Steinbaum: What Sanjukta said. Yes, it fixes prices for tens of thousands of independent businesses. Uber’s conduct is per seillegal, according to all the cases you cited above (barring state action, needless to say, which Uber has gone to great pains to avoid). Although now, of course, they want state legislatures to retroactively legalize their business model. See that awful Handy bill in New York State. Beggs: Apparently, there is a fine line between “price fixing” and “a business having a list price,” but I guess this just comes back to the semantics of Uber’s business model definition. Manne: So, the claim is that Uber engages in price fixing with its drivers? But it doesn’t coordinate with them. It offers them a service, at a take-it-or-leave-it price. Steinbaum: Oh, Uber offers a service to drivers? Then the brief you lauded above is exactly, precisely incorrect. Manne: The FTC/DOJ brief? How? Steinbaum: The premise of the FTC/DOJ brief is that Uber is a two-sided platform and the state-action exemption is limited to the customer-facing side of the platform, not the driver side. Ergo Seattle lacks immunity. But according to you, Uber doesn’t deal with customers at all. Therefore, the drivers’ collective bargaining is in fact covered by the state-action exemption. Manne: Uh, no, Marshall, I did not say “Uber doesn’t deal with customers at all.”Regardless, that’s not what the brief says at all, either. It doesn’t rest in any way on Uber’s two-sidedness, and it certainly doesn’t imply in any way that if Uber doesn’tdeal with customers, Seattle is free to authorize drivers to collude. What the brief actually says is “given the state statute at issue, the Supreme Court says the city can regulate transportation services; it doesn’t say it can regulate driver negotiations over Uber’s fees.” Steinbaum: That is in fact the FTC’s position. Read the brief. Manne: I have read the brief. And I’ve read your blog post on it. Your characterization is wrong. But don’t take my word for it; the Ninth Circuit said the same thing in ruling against the city: “although the statute addresses the provision of transportation services, it is silent on the issue of compensation contracts between for-hire drivers and driver coordinators.” Paul: Geoff, how do you justify Uber’s price-setting for rides under the status quo antitrust rules that, as I understand it, you generally think are doing fine? Manne: How do you justify any business setting a price for anything? Paul: Geoff, is it your position that Uber sells rides? Manne: Uber does not sell rides. Uber sells logistical services. The service it offers to drivers is different than the one it offers to riders. Paul: Does Uber sell rides to consumers? Manne: No, Uber does not provide rides. Singer: I want to get Jodi in here for the econ take. Let’s assume the FTC is motivated by a desire to promote competition. Let’s imagine a parallel universe in which Uber’s drivers can coordinate on their wages or “wage shares” (the portion of the fare captured by the driver). What is the harm to competition, assuming no increase in fares, if the effect of the drivers’ coordination is merely to increase the wage share from the current 61 percent to (say) 75 percent? Or is it not possible to shift the wage shares without also increasing fares for passengers? Beggs: Well one important distinction I would like to make for starters is between “impossible” and “not profit maximizing.” Impossible? Probably not, given that Uber’s business model generally seems to be to eat through endless VC money anyway. But yeah, if it was maximizing profit before a pay increase to drivers, then continuing to maximize profit would most likely result in higher fares to riders and fewer rides sold. One interesting thing is that the drivers could be less likely to feel the “fewer rides” part of this, since they are very small compared to the overall rideshare ecosystem. (In other words, the change might manifest itself as fewer drives rather than drivers each giving fewer rides.) This creates different calculations regarding higher fares for drivers versus for the company. Steinbaum: Driver collective bargaining would in fact be pro-competitive, because Uber is a monopsonist. Another example where the FTC and DOJ don’t understand economics. Manne: The set-up to Hal’s question is all wrong. Singer: Thanks. Framing is everything. Manne: It’s just not a sensible way to look at this issue. Ride-share drivers don’t earn “wages”; they’re not employed by Uber or Lyft or other transportation network company (TNC). They earn fares from riders (at a rate set by the TNC), and the TNC keeps a portion of that in exchange for the service it provides in offering the platform. Like a franchisor. Or… anyone else selling a service. So Uber drivers are users, in this set-up, somewhat akin to input providers (or consumers). To answer your question, Hal, cartelization by drivers to drive down the price they pay Uber would likely lead to higher rider fees, or most likely it would. Singer: Sanjukta, could this be the policy basis for the FTC’s intervening in the way it did? To keep fares low? I’m just trying to help them out. Paul: First of all, if you believe Uber’s own account of what it does, you shouldn’t think of the coordination as a wage share. According to Uber, drivers sell rides to customers, and Uber sells the use of software to both drivers and riders. Well, then the use of the software is an input into drivers’ production process, and allowing drivers to bargain in their bargains with Uber lowers the cost of an input, not vice versa. So prices drivers can charge should go down (on this logic). Steinbaum: Ok so we’re in nonsense Amex world and now Uber is in fact a two-sided platform? I agree with Sanjukta’s point about the direction of the pricing pressure. Manne:I’ll just do the sensible thing and ignore Marshall trying to get a rise. Paul: Second, to be perfectly honest I think it’s really difficult to think in terms of drivers’ share having an “effect” on prices at all, because I think Uber’s pricing decisions are both more opaque than they suggest, and are relatively unconstrained by the things we might think they’re constrained by. I most certainly do not think the algorithm passively implements the “market” price, as Uber’s CEO Kalanick famously said a few years ago. Third, frankly I think Uber’s business decisions, including pricing decisions, are much more geared toward their upcoming IPO and toward the financial markets than anything else. I think Uber has lots of reasons to set prices in particular ways that have little to do with exactly what percentage is going to drivers, so it’s far from evident that that percentage would affect prices in any particular way. Steinbaum: Sanjukta is right. It’s incoherent to claim that increasing the wage share that Uber pays to drivers would cause rider fares to go up. It seems like Uber’s pricing policy is more price discrimination now than it is raising fares across the board. Paul: It’s certainly clear that their pricing decisions interact with (a) their investor-facing decisions and (b) market dominance, far more than with driver share. Uber wants to prove future dominance to investors. Beggs: I think agency pricing (i.e., where one party’s “price” is a percentage of another price–the price that the consumer pays–as opposed to a fixed number) is a new thing that has not been well thought through from an economic perspective yet. Steinbaum: I think it’s pretty clear that Uber is engaged in some version of predatory pricing to monopolize the market anyway. Manne: [Mutters to self: “Ignore Marshall. Ignore Marshall. Ignore Marshall.”] Beggs: I wouldn’t put it as strongly as Marshall did, but I do think there is a difference in incentives between Uber and its drivers. Specifically, Uber might want to capture market share, push others out of business, etc., but drivers likely see Uber as a short-term gig and thus don’t really care if Uber takes over in the long run. This would make them less likely to go with lower prices in order to capture market share, etc. Manne: Of course there is a difference in incentives between Uber and each individual driver. It’s less clear if the interests of drivers in toto are so far away from Uber’s, but presumably they’re not identical. Steinbaum: Yes, I agree with Jodi. I think a key dimension of the mismatched incentives, given the non-employee status, is that Uber wants drivers activated as much as possible, but also empty, so it has spare capacity it doesn’t have to compensate them for. Whereas of course drivers want to be full when they are activated. Beggs: That’s somewhat different from what I was saying but plausible, yeah. Singer: In her new paper, Sanjukta argues that Uber’s own coordination on its pricing of ride-hailing services would constitute impermissible price-fixing under the FTC’s construction of Section 1 of the Sherman Act that barred drivers’ collective bargaining. And yet the FTC is not pursuing Uber on price-fixing grounds. That sounds like hypocrisy. Sanjukta, can you break this down for our readers? What allows the FTC to embrace these seemingly contradictory positions? Paul: Sure. The construction of Section 1 operative in the Seattle case says that price coordination between drivers is impermissible price-fixing, per se. For that reason, the Chamber of Commerce contends that the local ordinance that previously authorized that coordination is preempted by federal law. Meanwhile, Uber unilaterally sets prices for rides charged by drivers, the same drivers antitrust supposedly regards as competitors barred from coordination themselves. I’m sure we’re going to get further into details of the law here, but on a big-picture level this just doesn’t make sense. The FTC presumably ultimately cares about effects on riders (according to its description of its consumer-welfare mission) and cares about drivers’ bargains with Uber to the extent that affects consumer prices, as we just discussed. And they are subscribed to the idea that coordination of prices is bad for consumers. OK, in that case, how is it that price coordination is fine when laundered through a well-financed corporation, but not when the little guys do it? And if you want to distinguish between the types of coordination here (on consumer prices vs. bargains between drivers and Uber), shouldn’t direct coordination on prices to consumers by Uber be much worsefrom the standpoint of ultimate consumer welfare, on this logic? Manne: What coordination by Uber? This is where the fact that there’s no agreement comes in. Drivers don’t “agree” with Uber. Uber unilaterally offers a service at a price. In a labor market in which Uber has virtually no share, drivers decide whether to buy the service or not. It sets the time and distance rates (and other components that are part of the fare calculation) for rides booked through its platform. But Uber is a single entity. The millions of drivers who use TNCs do so in order to earn money by driving people around in their cars. They’re also independent actors. Just like someone trying to earn money by selling something on eBay, say. (For which eBay takes a cut and sets the terms, etc.). The platforms set the rates; drivers take it or leave it. Where’s the coordination? Paul: The coordination by Uber is, plainly, coordinating prices for ride services offered by Uber drivers, among other things. I think that the FTC’s stance here is due to a deeper tendency to defer to coordination when done within and by large, powerful firms. What is interesting about the Uber case is that here, that deeper tendency has actually outstripped the reach of the positive lawitself (though the tendency is certainly in the positive law as well) so it really lays bare some of the first principles that we don’t normally see or discuss. Steinbaum: Uber allocates and divides the market. It’s a cartel manager, same as Standard Oil was for the 19th century railroad cartel. Manne: So you may be surprised to hear I have some sympathy with the argument in Sanjukta’s paper that antitrust may prohibit coordination too much in various ways. Paul: Aww love this theme. Manne: But I don’t think that what Uber does is “coordination” of the antitrust / price fixing sort. Beggs: So usually the way that competition brings prices down is by increasing supply. In this case, increasing driver supply doesn’t necessarily get prices down unless Uber chooses to lower them. That said, I don’t agree that it’s coordination per se. Steinbaum: I thought Sanjukta put this well at a conference we both attended a few months back: If the church-organists professional organization the FTC was so enraged by had instead programmed an app to match organists to churches, and prevented members of the organization from matching with churches other than through the app, the FTC would have been sending letters about how #innovative the church organists were, instead of suing them. Paul: I’m trying to understand the straightforward appeal to “firms set prices” here. I hear you and Geoff saying that Uber is setting a list price for a good or service like any business. But also that it doesn’t sell that good. Geoff, do you think I as a firm have the right to set prices in a market I’m not myself (as a firm) involved in? At a minimum, we are way beyond any consensus understanding of what firms do at this point. Beggs: Call it what you want, but drivers definitely don’t get to set prices, nor do riders, so by process of elimination… Manne: So let’s do market definition. Singer: No, Geoff. Not in a per se, price-fixing case! Paul: Actually, before we get to market definition we have to do firm definition. Manne: It’s not a per se case, Hal. And there are a couple relevant markets here. But let’s go full Amex. The main issue here is the market in which Uber’s platform operates, which connects riders and drivers. Uber offers platform services — it sells logistics. To drivers. Uber sets the price for that. Paul: Ok. Do they set prices in anything else? Manne: Yes. But so what? They set various other terms of use on the platform including rider prices, as well as about a billion other things. Beggs: Price-setting rights are logistically independent from who actually sells the good (minimum advertised price rules, for example), so I would argue that Uber saying it doesn’t sell the good doesn’t mean a whole lot in practical terms. Manne: Right. Paul: Logistically independent? We are talking about the law here, and the understanding of the firm assumed by the law, I thought. Manne: Uber has the right to set prices for the thing (rides) sold by drivers via its platform. It doesn’t have to sell rides to do that. Steinbaum: It’s fixing the price of the transaction that’s taking place, regardless of whether it is or isn’t party to the transaction. Manne: Now you’re taking about the other transaction, right? Steinbaum: This is the sort of meaningless debate that Elzinga made fun of in his critique of the Chicken Delight case (Siegel v. Chicken Delight 1971), but I guess now that’s…. a matter of substance we all have principled positions on? Manne: There are several transactions here. One between Uber and drivers. One between Uber and riders. One between drivers and riders. Beggs: Use the Kindle store as an example. Over the last eight or so years, we’ve seen pricing rights allocated both to the publishers and to Amazon regardless of who is viewed as selling a Kindle book, the pricing authority can be held by a different party. Paul: This is the whole point. This sort of “pricing right” is novel. And Amazon has the same problem! Beggs: I think Amazon and iTunes are pioneers but it’s still a fairly new thing yeah. Manne: I don’t think it’s so novel. For example, construction firms set prices on large projects they run, even though generally what they are doing is project management, not providing the underlying construction services. Beggs: *whispers* Didn’t the Apple version of the Kindle store get sued for… price fixing? Steinbaum: Yes! Paul: Yes on a very similar theory! Manne: Also, I think the Amazon parallel is a good one. Or eBay. Paul: I agree the Amazon parallel is a good one, but draw a different inference from that. Manne: Not identical, of course, because eBay doesn’t set the prices of products, but it does set many other hugely relevant terms of trade, even though it doesn’t sell products itself. Beggs: The settlement in the Apple lawsuit lead to Amazon doing the “price set by publisher” thing on their site, so they did change pricing rights in at least some cases. Singer: Let me pick up something you guys mentioned earlier. Maybe an efficiency defense to its “price coordination,” a characterization that I know Geoff rejects. Ride-hailing firms like Uber and Lyft might argue that various efficiencies justify their price coordination when it comes to setting fares. Is there a reason why Uber the mothership would want lower rates than would a single driver (where pricing would be set absent the alleged coordination)? Could it be the commission-structure by which Uber is compensated? For example, might Uber the mothership be insensitive to some costs that are uniquely born by the drivers, such as the driver’s opportunity cost, which militates in favor of Uber setting lower fares? Also, the mothership might have access to better information than would a single driver—e.g., number of calls, number of other drivers, location of calls/other drivers/pricing by taxis—which uniquely permits Uber to exploit a price-discrimination strategy, thereby increasing profits without necessarily reducing output. Are these legitimate efficiencies? Steinbaum: Hal is exactly right—drivers face a cost for activation, but Uber does not, and Uber decides if they get a fare or not. Paul: Do we really think spending billions of dollars perfecting this pricing strategy is somehow efficient? Is it socially useful? Singer: You’re answering my question with a question? That’s not allowed here. Paul: Even crediting all of this, none of it affects the comparison at issue here: between Uber’s coordination of consumer prices and drivers’ hypothetical coordination in their bargains with Uber. The considerations you’ve identified all regard the consumer price, so these “efficiencies” could all be realized by Uber’s coordination while drivers coordinate as to their bargains with Uber. Beggs: Well… perfecting the pricing strategy for which party? The incentives are definitely not aligned to lead to a universally optimal solution. Paul: Ok, it’s not socially useful at all then! Happy to agree with that. Singer: But does Uber’s coordination perhaps lead to lower fares relative to a world in which drivers set fares? Paul: As I’ve said, I don’t think antitrust should hold out the “no-coordination case” as the normative benchmark for appropriate prices—or anything else. So to be clear, I do not think price coordination is a bad thing. However, I think selectively applying the no-coordination norm—to drivers but not to Uber—is wrong and harmful as well as inconsistent under the positive law. That said, if we are consistently applying the law, you don’t have to show that consumers paid higher prices. That is not a legal element of a price-fixing case. That’s what makes it per se! And that would apply to a hub-and-spoke arrangement like this one, as indeed Judge Rakoff held before Uber got the Second Circuit to kick the case to arbitration. Finally, if you get into rule of reason territory and you’re able to convince a judge that the consumer welfare standard justifies Uber’s coordination but not drivers … well, at that point I don’t see a clearer indictment of the consumer-welfare standard. Manne: Uber, et al. don’t even need to rely on efficiencies. First, they could argue they not coordinating anything, so go blow. Second, the alleged restraint is an ancillary restraint — i.e., it’s not the primary object of an agreement, but it’s necessary for the proper functioning of the objectives envisaged by an agreement. And third, the courts don’t really do per se much anymore, so efficiencies are more like the a priori debate over whether we condemn the conduct under per se or not (see, e.g., the Apple eBooks case (especially the dissent)). But, of course, over and above all that, there areefficiencies. Steinbaum: I’ll throw in the fact that not only is Uber in a unique position to price-discriminate, but they themselves also boast of increasing efficiency by monitoring drivers to reduce moral hazard. Paul: Hal, this is my least favorite question. Singer: But we need some econ Sanjukta! Paul: Do we really?? Singer: Econ Uber Law. Manne: Uniform pricing is important to Uber’s business. Hence, this is an ancillary restraint (if it’s a restraint at all). Paul: I agree with Geoff 100% on the first part—uniform pricing is important to Uber’s business. (Hal, did you see that?) Steinbaum: The efficiencies that Uber has claimed for itself were a public admission that it violates labor law. Which was funny to see happen in real time at the NBER Summer Institute when that paper was presented. Beggs: What seems to be happening in part is that Uber is willing to let the short term be a loss leader for the long term, and I doubt that most drivers are willing to make the same tradeoff, so what is good for Uber is probably not what is good for the drivers, even though the “but we only get a percent of what you get so everyone wants to maximize the pie” sounds good if you don’t think about it too much. Steinbaum: Exactly. Paul: Beyond that, it is impossible to think about Uber’s pricing strategy in abstraction from their investor-facing strategy. And investor decisions do not track operational profits, as both Uber and private equity show in converse ways. Beggs: This is true, but Uber doesn’t seem to have any trouble attracting capital regardless of their specific decisions. Paul: I think Uber cares more about selling itself to the (financial) market than anything else. It’s not totally obvious what the long-term strategy on operational revenue (or profit) actually is. But does it matter if share values keep going up? Long enough for the people currently making decisions at least? Singer: Let’s get back to Sanjukta’s paper. She argues that “if we intend to dismantle the monopolistic and oligopolistic market coordination that has proven so pernicious, we will require other forms of coordination to replace them.” She asserts that relaxing enforcement of Section 1 claims against coordination across small firms is necessary to attacking our monopoly problem. How far should we go in legalizing horizontal coordination beyond firm boundaries? Sanjukta, are you really suggesting legalizing all cartels? Paul: I do think we should liberalize a lot of horizontal coordination that doesn’t take place within firms, including some that might currently be referred to as a “cartel,” along with addressing monopoly, because I think the two are self-reinforcing. For example, failure to address monopoly contributes to differential access to capital among less and more powerful people; investors are interested in growth through increased market share, etc. Beggs: This conversation makes me think that it’s worth having a legal distinction between “1099 worker” and “small business.” Manne: Sure. Uber drivers are small businesses. Beggs: Except they aren’t really, at least not as it pertains to this context, if for no other reason than small business generally have pricing authority. Manne: I’m not sure why it matters for this 1099/small business distinction if they are price takers. Steinbaum: Here’s an idea: Coordination is legal if the parties do not have market power; illegal if they do. Beggs: That’s kind of my point. Uber drivers don’t have market power in any meaningful way above and beyond that of employees, so it’s absurd to give them different collective action rights. Steinbaum: Well Jodi, that’s why they should either be classified as employees under both the National Labor Relations Act and the Fair Labor Standards Act, or Uber should lose the right to set prices, the way traditional taxis have. Beggs: My concern about Uber drivers being classified as employees is that they could potentially lose a lot of autonomy. Paul: What I agree with in Marshall’s statement about when coordination should be legal is that we need independent criteria—independent of firm status—to make that determination, though I am not totally sure about “market power.” Manne: Totally agree with the first half of Marshall’s “coordination is legal if the parties do not have market power; illegal if they do.” Paul: And the current system is the opposite of what Marshall suggests we ought to have: you get coordination rights from antitrust precisely if you already have concentrated power! Manne: Hal, please don’t ever quote me agreeing with anything Marshall says. Steinbaum: Geoff is concerned he won’t be invited to the next FedSoc gala, but I’m sure he has some social capital stashed away as far as that’s concerned. Singer: If you would suck up to Geoff like I do, maybe you would have received an invite to FedSoc event in Malibu next week. (Seriously, Geoff got me an invite.) Manne: But why can’t a small business work for a larger business? Again, large-scale construction is done this way. Also movie production. There’s a great paper by Mitu Gulati and Bill Klein — a really fascinating discussion of the organizational dynamics of large construction projects. Paul: Can’t believe I’m not hearing Marshall scream: “Richfield Oil”! Singer: Geoff’s solution to most competition problems involves the little guy selling out to the big guy. Steinbaum: I actually was about to scream that and then I decided it didn’t quite apply (for once). Paul: The idea that a small business should not be subordinated in a “work for” way to a large business, under antitrust, is the whole point of Richfield Oil. Steinbaum: Exactly. Manne: We also see these small businesses working for large businesses in franchise too. Why are franchise agreements evil? Steinbaum: Not evil Geoff, just anticompetitive. Paul: Well, in franchising we have a similar situation where the much more powerful firm (franchisor) is permitted to engage in all kinds of economic coordination–even setting franchisees’ prices–while disclaiming firm status for every other purpose, such as labor law and corporate law. Steinbaum: Also, franchisees eligible for subsidized SBA loans! Paul: Meanwhile, franchisees are entirely denied coordination rights. They can’t get together to set consumer prices or negotiate their (oppressive) contracts with franchisors. Manne: “More powerful” implies something like an involuntary relationship in which the only choice the franchisee has is to be come the franchisor’s bitch. But that isn’t the case. So if the franchisee enters voluntarily into a franchising agreement, is the alleged power disparity really relevant? Steinbaum:Ah yes, the bilateral efficiency argument. In my econ PhD they referred to this as the “Barro Critique.” One of many empirically irrelevant ideas knocking around economics. Bork has a hilarious disquisition on how power is an “uneconomic concept.” Paul: Really the key that I want to communicate is that antitrust itself is creating and allocating these power differentials. You may agree or disagree, and we can have that conversation, but in allocating pricing rights to franchisors over franchisees’ prices, it’s making a specific policy choice. It doesn’t have to make that choice. Just like it doesn’t have to make the choice that franchisees don’t get to coordinate. That choice is also inconsistent with the treatment of “franchising families” under corporate or labor law. Singer: Good chance to get back to your paper, where you speak of the “firm exemption” to antitrust. Love that phrase, by the way. So long as things happen within the firm, they are outside the scope of antitrust enforcement. It’s like what happens in Vegas, but for firms. An example I like to use is exclusive dealing: An exclusive dealing arrangement between a large input provider and a standalone distributor is subject to antitrust scrutiny, but bring the exclusive distributor in-house and the exposure goes away. Briefly, how did we get here? Paul: Yeah, so we’ve had two parallel developments: 1) the firm exemption has been broadened, and 2) other coordination rights have been contracted. The legislators who drafted the Sherman Act did not obsess over the firm-market boundary. They cared about whether economic coordination was socially helpful or harmful. Steinbaum: They had a more sophisticated understanding of economics than the charlatans who’ve had so much influence on the last 40 years of nonsense antitrust rulings. Manne:[Ignore. Ignore. Ignore…] Paul: Over time antitrust has moved away from the original legislative vision, and has reified the firm as well as an idea of the competitive order that legislators had no concept of or concern with. In that competitive order all coordination outside the firm is inherently suspect, and the firm becomes a black box whose internal coordination becomes more invisible to the official way of thinking, even as it becomes more powerful. Can we step back for a second? I was going to say something about balancing power but I want to say something maybe we can agree on. Manne: As if I haven’t agreed with you enough! Paul: At the end of the day we are having a debate here about differing visions of a social and economic order. But what I want us to get to first, analytically, is admitting that. Singer: In Geoff’s economic order, we all work for the man. Sorry, go on. Beggs: And should like it, if I’m reading correctly. Manne: Admitting what, Sanjukta? That we’re debating different visions of a social and economic order? Too abstract for me. Can you unpack that? (And also, let’s do without the inaccurate caricatures of classical liberalism, Hal; you do enough of that on Twitter!). Paul: Sure. That’s the reason I’m so focused on unearthing the firm exemption: because I want to show that antitrust is making basic choices about who gets power in society. Manne: Ok, but I still need help with what you think we’re agreeing on. Paul: The current antitrust framework elides that fact, and it does so through neutral sounding talk of firms and cartels. So that is the deeper point. I am actually not a lobbyist for would-be cartels. The fact is that antitrust used to be far moretolerant to economic coordination that happens outside firms, and far more scrutinizing of coordination that happened within them. If I have one point, it is that we should come up with independent criteria for whether economic coordination is good or bad,not the formalistic distinction about whether that coordination takes place in a firm or cartel or something else. Singer: Shifting gears again. Assuming the FTC could stop sucking up to the platforms (as it did in the 1-800 CONTACTS case) and somehow be motivated to bring a Section 1 claim against Uber for price fixing of passenger fares, what are the chances that such a claim would prevail under current antitrust standards? I’m trying to figure out if this is serious case, Sanjukta, or if instead you are merely pointing out the double standard. Steinbaum: If Uber is a hub-and-spoke conspiracy, like Apple e-books, then the antitrust case against it has a very good chance. If Uber is successful at claiming that unlike the organist professional association (or its own drivers), Uber operates a vertical price-fixing conspiracy, then that case is much harder to make. And given the mandatory arbitration clause, any antitrust case relies now on either the FTC or a state AG. If a state brought it, I bet you’d see another spurious brief from the Feds like the one they submitted against Seattle. Beggs: I think there’s an important difference with the Apple case though, mainly that if I recall correctly, Apple was pressuring publishers to set the Apple prices on the Kindle store as well or something similar. The Uber case doesn’t have a parallel, since it’s not like drivers are setting prices on Lyft or whatever. Manne: The chance of success of a Section 1 claim areclose to zero. Just saying; not claiming that’s good or bad. Singer: Why zero, Geoff? Manne: In the market for rides, in which you hypothesize Uber is price fixing, it isn’t even dominant. But even if it were, it isn’t price fixing. Paul: I’m pointing out a conflict both at the level of policy and principle, and at the level of positive law. However, I do want to note that I wouldn’t actually advocatebringing such a case, because, as is clear from the paper, I don’t espouse the general principles that it rests on. I do think those general principles are being inconsistently applied. I think that under the positive law there are two possibilities. Either the case against Uber shows there’s a horizontal restraint and the per serule applies, in which I think the plaintiffs prevail, period. Or for some reason they get outside the per serule—I think this is hard, because the setting of ride prices has nothing to do with the contract about software use that is supposedly the subject of the relationship between Uber and drivers, according to Uber itself, and therefore it can’t be a vertical restraint—and then I think it’s hard for them to argue that they get outside the per serule but drivers do not. The sort of cases that would support that—Appalachian Coals, BMI—also support drivers’ coordination. In fact, Appalachian Coalsmight support drivers’ coordination but not Uber’s! Manne: This is easy rule of reason territory. Again, look at Apple e-books. In a case like this, per se entails a full rule of reason analysis to decide if per se applies. Under that analysis, the court will ask whether there is any harm and any pro-competitive justifications. How would anyone show harm? There isn’t any that I can see. And even if there were, the procompetitive justifications are huge. Steinbaum: Ok, so if we’re back on this, and the plaintiff must define markets and show market power, then I’ll point out now that Uber does in fact have price- and wage-setting power, per all the evidence on firm-level supply (and demand) elasticities. Therefore, the correct market definition is that Uber is a monopsonist in the market for its own drivers. Years ago Geoff wrote a very snarky blog post about how this idea is obviously wrong and laughable—it’s funny that in the years since, eminent economics and law professors have embraced it. Manne: I actually have no idea what my post has to do with the (also obviously wrong and laughable) idea that Uber is a monopsonist. It’s one of my favorite posts ever, though, so I encourage readers to click through and tell me what Marshall is talking about. Beggs: Marshall I guess technically it’s a…duopsony? Steinbaum: No, not a duopsony. I mean, I haven’t conducted an analysis of labor supply elasticities in Uber’s case in particular, but the point of the hypothetical monopsonist testis that the right market definition is probably just one firm wide. Beggs: Maybe this doesn’t extend to all geographies, but I don’t see from here how one can consider Uber but not Lyft. Steinbaum: Depends how much wage-setting power Uber has. If it lowers wages, do drivers switch to Lyft in sufficient numbers to make that wage reduction unprofitable? That’s what should set market definition. At the end of the day, it is an empirical matter. Dube, Naidu et al find ridiculously low supply elasticities in a context where they “should be” high. Paul: Geoff, do you think per seapplies to drivers’ coordination? Manne: Think BMI, NCAA, standard setting organizations, franchises, joint ventures, etc. There are so many arrangements that couldbe classified as horizontal collusion but that the courts accept. Paul: Ahhh BMI! Here is my point. BMI justifies drivers’ coordination. There is no way to use it to immunize Uber only. Beggs: Oh now we’re in my wheelhouse do go on… Manne: I think per se is much more likely to apply to drivers’ coordination, but theoretically they could make a case to get out of it. I think it far less likely though. Paul: You are either using strong Section 1 precedent for both Uber and drivers, or you are using cases like BMI for both Uber and drivers. You can’t use different legal standards for the two. Manne: It’s not binary. BMI could justify driver coordination in theory… Paul: Geoff agrees with me again! Beggs: I do think that that answer is yes. Manne: … but it needn’t even if it justified Uber “coordination” (even though that’s not necessary because Uber is a single firm). Paul: Can I please remind everyone that the Seattle preemption lawsuit requires immediate per se treatment and if BMI applies the suit should go away. If we are in BMI territory, considering justifications for driver coordination, we are out of per se land. Manne: As I said earlier, I have some sympathies with the argument that we should permit more contracts than we currently do! Firm or not. Paul: Federal antitrust preemption requires showing that the per se rule applies to the conduct authorized by the state or local law at issue and that the statute therefore conflicts with federal antitrust law on its face. See, e.g., Rice v. Norman Williams Co.(1982). By definition if we are addressing benefits and harms under BMI, the statute isn’t preempted. Manne: I disagree. Rice v. Norman Williams also finds preemption when a statute “places irresistible pressure on a private party to violate the antitrust laws.” The fact that there might be a theoretical exception to what is typically a per se violation can’t be enough to immunize a statute, or else no anticompetitive statute would ever be preempted by the Sherman Act. Here, the theoretical, BMI–like benefit is only theoretical, because the municipal statute (unlike the state statute) clearly contemplates collective bargaining for the purpose of lowering prices, and imposes a structure that indeed makes a violation irresistible because it requires all drivers to be bound by the collusive agreement once only a majority assents to unionization. Paul: The BMI-like benefit is not theoretical. In fact you yourself are committed to the position that it is not theoretical, to get Uber out of per se territory. Once a majority of those voting vote to give themselves coordination rights to (partially) balance Uber’s, indeed the almost-certain improvement in pay, working conditions, and quality of life will flow to all drivers. Manne: I don’t rely on BMI to get Uber out of per se. I rely on there not being any agreement. And also, even if BMI applies in one place it just doesn’t mean, as you suggest, that it must apply in another. BMI requires a valid procompetitive benefit to remove conduct from per se land. Either its existence means nothing is ever facially per se, and thus makes the preemption case law irrelevant, or (as I see it) it means coordination to benefit consumers by uniform pricing, etc. matters, but coordination to raise input prices does not. Paul: I don’t agree there is no agreement in the Uber case but not in the driver case, and neither did Judge Rakoff (and I think the ancillary restraint argument would also eviscerate the per se rule). Also, BMI applies in both places because both theories essentially involve coordination among drivers as to ride services. I’ll leave it there for now for Hal’s sake. Singer: Geoff thinks a showing of harm is essential to any price-fixing case against Uber. So let’s go there and get the econs involved. Continuing with this hypothetical fare-fixing case against Uber, how would one go about showing that Uber passengers paid higher fares as a result of the alleged price fixing? How would fares be determined in the absence of Uber’s alleged coordination of prices to be charged by its drivers? Via an auction by Uber drivers? Is this even feasible? Steinbaum: Taxi companies don’t set fares. They’re set by regulators. That’s the counterfactual. Or Uber could employ its drivers. Singer: That’s super-interesting. So the but for world is one in which the regulator sets fares, and those are below Uber’s fares? Manne: The key is that with non-standard contracts, some hypothetical “market,” pre-conduct price isn’t the proper benchmark. Especially if your benchmark “market” price is one set by regulators! Marshall will like this: there are tons of product market failures (to quote Williamson). Referring back to the price (if there were one) in a crummy, heavily regulated “market” isn’t sensible. Steinbaum: The but-for regulated price is not necessarily below Uber’s, but it is non-discriminatory. Uber’s business model on the customer side is price-discrimination. Singer: Ah, so the harm is not higher prices Marshall, but instead discriminatory prices? Steinbaum:Yes. Plus geographic discrimination, probably. Of course taxis are pretty bad at this too, but one thing we know now is that Uber’s propaganda about how they “fix” that problem is false. Manne: You mean the benefitis not higher prices; it’s price discrimination! Paul: Why are we all assuming we should show consumer harm only? Also, there is absolutely no basis for a firm to treat all service providers as independent contractors and then set the price of that commodity. Manne: No basis in what? Paul: Principle? Logic? Manne: In law there is; in economics there is. Maybe not in what you think the law should be. But that wasn’t the question (although that is what we were discussing before when Hal so rudely moved us on). Paul: If a firm disclaims the employment relationship, the entire justification that has been proffered for the differential treatment of firms and cartels goes away. Manne: Not necessarily. Why should the distinction between worker/employee for employment law purposes map onto anticompetitive/competitive for antitrust law purposes? Paul: The basis for the firm exemption is, ultimately, precisely the employment relationship. The idea, going back ultimately to Coase, is that what happens inside the firm is organized by command rather than contract and that organization minimizes costs. This is the tradition Bork picked up on, in arguing for loosening up on vertical restraints and mergers while cartels remained per se illegal. If you convert the firm into a collection of contracts, as the independent contractor-based firm is, you lose the ultimate justification for the firm exemption. Beggs: Marshall, why would Uber (in the but-for world) have to classify drivers as employees when say, the tutoring company my friends run that sets prices, classifies its tutors as independent contractors? Steinbaum: Your friend’s tutoring business is in violation of the Richfield Oilstandard, alas. Manne: A standard that, at best, applies in the Southern District of California, in a substantially different context, and (I believe) hasn’t been extended since it was adopted by that court in the 1950s. Singer: In the last few minutes, I’d like to hear what policy levers you would pull if you were in charge to rectify this great injustice. Assuming antitrust law is helplessly constrained by the Chicago School teachings and the consumer-welfare standard, what are some legislative fixes that would restore coordination rights to workers or to smaller firms? Steinbaum: My policy proposal is to revive Richfield Oil: either you’re an employer, or you can’t tell people what to do. Beggs: Hot take policy idea: The party that doesn’t have direct pricing authority should get to collectively bargain. Paul:I like that! Steinbaum: I would rephrase “direct pricing authority” as “market power.” Manne: Everyone should get to collectively bargain! Unions are legal and—another shocker from me—often efficient in the collective bargaining sense. Go back to Marshall’s market power screen (i.e., what the law used to be). Remove the part that says market power means no coordination. But without market power, why would it ever be presumptively inefficient to coordinate? (The reason it might be sensible as a matter of lawis administrative efficiency. Which I would say is the reason why the law is like it is not some grand Marxian conspiracy. Paul: I’m increasingly convinced this is key to the advantages the firm exemption gives to those who already have power (power in the sense of wealth, access to capital). However, I would like to hear a definition of “market power.” If it doesn’t include factors like property rights or access to capital, it is not sufficiently broad, and we need a broader concept of economic power to replace it for purposes of allocating coordination rights. Manne: Wow—agreement all around. If no market power, then collective bargaining is ok. Right? We all agree? Beggs: 100%. Walras would be so mad at us, but yeah. Singer: Sanjukta, any final policy preference here? Paul: Broaden access to coordination rights, including looser horizontal coordination beyond firm boundaries that does not involve managerial hierarchies or concentrated ownership, with appropriate public oversight. [This includes reversing the direction of precedents like Sealy and Topco.]Also, more stringent application of antitrust to maintain decentralized markets would help to maintain a more balanced allocation of coordination rights. This is true in itself of course, but it would also blunt some of the investor-facing dynamics within firms like Uber that we discussed earlier. Those dynamics not only influence pricing behavior but also exacerbate imbalances with alternative forms of organization like producer cooperatives. Singer:I like that one too! That’s a wrap, gang. Thanks so much for participating!!~washBytes
