Back to top Jump to featured resources
ILSR Press Room filed under Independent Business

Will Amazon-Whole Foods Survive Antitrust Scrutiny Under Trump?

| Written by Nick Stumo-Langer | No Comments | Updated on Jul 3, 2017 The content that follows was originally published on the Institute for Local Self-Reliance website at https://ilsr.org/will-amazon-whole-foods-survive-antitrust-scrutiny-under-trump/

Forbes’ Washington Bytes – July 3, 2017

Written by Washington Bytes 

Graphic by Daniel Kleinman, Forbes StaffMANDEL NGAN/AFP/Getty Images

Graphic by Daniel Kleinman, Forbes Staff

This installment of the #BytesChat focused on the pending merger between Amazon and Whole Foods. Special guests included Stacy Mitchell, co-director of the Institute for Local Self-Reliance and the co-author of the report Amazon’s StrangleholdGeoffrey Manne, the founder and executive director of the International Center for Law and Economics; Ryan Radia, research fellow and regulatory counsel at the Competitive Enterprise Institute; and Marshall Steinbaum, fellow and research director at the Roosevelt Institute. The chat was moderated by Hal Singer, editor of Washington Bytes. The conversation covered the potential merger-related harms that are traditionally recognized by antitrust law, as well as some merger-related harms that are not. It concludes with a discussion of counterstrategies by other online platforms and merger remedies. The transcript has been edited lightly for readability.

Hal Singer: Good afternoon!

Stacy Mitchell: So you know Amazon is eyeing Slack to buy?

Marshall Steinbaum: I did not! More data for them. How nice.

Geoffrey Manne: You mean more data for consumers. How nice!

Mitchell: Yes, works nicely with their Amazon Web Servicesbusiness division.

Singer: Let’s go. Is the proper lens with which to analyze the merger vertical integration by a distributor (Amazon) into content (Whole Foods)? Or is there some better lens? I’m looking for a framework.

Manne: And right off the bat you have an error, Hal!

Singer: How can a question be in error?

Manne: You say Whole Foods is “content” and Amazon is distribution.

Singer: So what is Whole Foods if not content, i.e., food?

Manne: Whole Foods is a channel of distribution. The whole premium-natural-and-organic-supermarkets market was always bunk—it should have been (if anything) premium natural and organic food. Whole Foods sells food. It isn’t food itself. If the Federal Trade Commission (FTC) had properly defined the market, Wal-Mart, and everyone else would have been deemed in the market and it never would have challenged the Whole Foods/Wild Oats merger. Instead they made up this “market” based on the type of store. Your framing is not a proper framework because Whole Foods isn’t content; it’s far more a horizontal merger between two channels of distribution.

Mitchell: Amazon has monopoly power in online retail, and it’s looking to acquire a company that would help it sustain and expand that monopoly. Last year, Amazon captured nearly half of online spending, and enforcers should view this deal in the context of that market.

Singer: So Stacy, do you agree with Geoff that this merger is horizontal and not vertical?

Mitchell: It has elements of both. But what I disagree with is the notion that this should be viewed as being just about the grocery market.

Singer: I would think that a merger proponent would prefer the vertical lens, as that makes it harder (compared to horizontal mergers) for antitrust agencies to intervene.

Manne: I prefer the appropriate lens, regardless of how much harder or easier it makes it for antitrust agencies to intervene.

Steinbaum: I think the main threat of the merger is that Whole Foods has a very valuable supply chain in organic (and other) groceries.

Ryan Radia: I agree that Whole Foods has a valuable supply chain, but I see that as a plus. Is there a viable online delivery service for fresh organic groceries in most U.S. cities? AmazonFresh is available only in a handful of cities, and doesn’t have the same range of offerings as Whole Foods. (Update: Whole Foods delivers through Instacart, but not in cities such as Pittsburgh, Cincinnati, and Memphis. I also notice Instacart is missing many store-made Whole Foods goods, such as marinated raw meats. And Instacart’s delivery fees can add up.)

Manne: Very valuable supply chain? Hardly. Whole Foods is trivial in any relevant product market you could reasonably define to assess this merger. Hard to see how it expands anything, least of all online. Whole Foods is struggling, precisely because it can’t lower costs and has to lower prices to compete, which eats into margins. Amazon can cut costs, which will lower prices and be a win for everyone.

Steinbaum: I agree with Stacy. It’s about acquiring a competitor under pressure.

Manne: Sometimes known as a “failing firm” argument, which is a mark in favor, not against, a merger.

Mitchell: Whether Amazon will lower prices is speculative. Their private-label food products (coffee, snacks, etc.) are premium-priced. Whole Foods pricing is aligned with what they are already doing.

Steinbaum: My view is that this merger is a no-brainer to approve under existing antitrust doctrines, which is exactly why those doctrines are flawed. We, but not the law, know that Amazon is anticompetitive.

Manne: No, you think that, but have nothing but pure conjecture to support it. Definitely not “we.” And in any case, even if true, a merger is no place to address it.

Mitchell: My view is that the laws tell us that we should say “no,” but the regulatory process is organized to say “yes” to this deal.

Steinbaum: Stacy is right. The law of course does recognize vertical threats to competition, but various precedents and enforcement decisions simply ignore the law.

Manne: The law says we should say no? Not hardly. The law is clear that a merger with about a 500 Herfindahl-Hirschman Index(HHI) in the likely relevant market (if it’s a horizontal merger), or that offers efficiency benefits of vertical integration (if vertical), should be approved.

Mitchell: A monopoly looking to acquire another company to expand its monopoly is not outlawed? By what reading?

Manne: The precedents. They are the law.

Singer: Now that we’ve settled on a lens. Kidding. Let’s see if we can identify any harms flowing from the merger that the antitrust laws will detect. Please identify specific output or input markets in which the merger presents a threat to competition. Stacy, do you want to start?

Mitchell: Whole Foods offers assets that allow Amazon to expand its online monopoly. 46% of online spending last year went to Amazon.

Steinbaum: Exactly. Whole Food’s supply chain will enable Amazon to strengthen its online retail monopoly. As Stacy has shown, an enormous share of online retail purchases start on Amazon.

Manne: Sure, but how many of those online purchases were for organic kale?

Mitchell: Most importantly, Whole Foods gives Amazon logistics infrastructure. And delivery is key to sustaining a monopoly online. That’s why Amazon has been investing so heavily in delivery. And why it’s going after UPS and FedEx.

Steinbaum: And that share will go up as a result of this merger. In other words, Amazon is a two-sided market, and strengthening its upstream domination with Whole Food’s supply chain and infrastructure will cement its downstream monopoly.

Singer: Geoff/Ryan, care to respond?

Manne: Whole Foods has approximately 0% of the online market. It represents a tiny share of the overall grocery market and a declining share of the organic market. And compared to Amazon’s supply chain and logistics capabilities, Whole Foods doesn’t add much. I don’t see how the merger changes Amazon’s incentive and ability to engage in anticompetitive behavior. The deal seems good all around.

Radia: Regarding the logistics infrastructure, Amazon’s purchasing Whole Foods may allow the company to deliver fresh foods faster and/or at a lower price, and in more cities. I guess that’s potentially bad for UPS/FedEx (and perhaps 7-Eleven, which houses Amazon lockers), but it seems good for consumers.

Mitchell: Whole Foods brings 18 million square feet of fresh-food warehousing space, plus 460 stores for last-mile delivery plus pickup. That’s a big addition to Amazon’s logistics.

Singer: Ok Stacy. What happens next? Tell us the story. When Amazon gets to deploy these new assets from Whole Foods, who gets harmed and how?

Mitchell: Amazon adds groceries to its mix, which further cements its position as a monopoly platform where most online commerce occurs. This gives Amazon more power to extract rents from suppliers and competing sellers that need to use its platform. At the same time, Amazon builds out its logistics, allowing it to control rapid-package delivery, weakening UPS, forcing more competitors to use Amazon’s third-party services, which makes them more dependent on Amazon, which gives it the ability to extract more, and so on.

Steinbaum: Lina Khan has shown that Amazon has pressured would-be acquisitions, lowering the sale price (a distressed company, as you describe it), then assimilates them like the Borg. In the case of Diapers.com, consumer prices eventually went up.

Manne: Well, it’s a false story about diapers.com. Amazon spent six years trying to make that company work and couldn’t.

Radia: Regarding the diaper example, Khan’s claim is that Amazon raised prices after it had been offering diapers below-cost. But given that 80 percent of baby products are sold off-line, I’m not sure that Amazon has ended up charging supra-competitive (or even especially profitable) prices for diapers since it bought diapers.com

Steinbaum: I thought Geoff was going to start where Ryan went—that all of this consolidation and market power ends up benefiting consumers. Not that consolidation doesn’t exist, which seems to me to be indisputable.

Manne: Oh, it’s self-evidently true that consumers benefit. But before we even get there, we have to start where Hal is trying to get us—is there even any prima facie case that would be cognizable under the antitrust laws — let alone one that could overcome procompetitive benefits? That’s how antitrust works: the initial burden is on the plaintiff to make out the prima facie case.

Singer: Well I think Stacy has articulated a theory. She said the merger would allow Amazon to “extract rents from suppliers.” So it seems like the harm manifests not in the form of higher prices, but instead in below-competitive prices for input providers.

Steinbaum: So the prima-facie case is about using upstream consolidation to (1) force other “content providers” to use its logistics network, and (2) cement a downstream monopoly. I think a secondary case is that 500 Whole Foods stores will become experimental laboratories with which to perfect price discrimination.

Singer: Marshall, you embraced my lens! Two debate points!

Manne: So it’s a monopsony theory?

Mitchell: It’s that and more. Amazon has selectively raised some consumer prices, so that may come too. Also on the consumer end is the loss of product diversity.

Singer: Well let’s focus on squeezing input providers, which might result in a loss in output. Geoff/Ryan, what is wrong with this theory?

Manne: How is it a competitive harm for a potentially powerful buyer to enter a market, whether it does so by buying Whole Foods or just starting to sell produce itself? Especially given that Whole Foods is a tiny share of the grocery market and Wal-Mart is already in it? The merger doesn’t give Amazon power to extract monopoly rents. Adding Whole Foods to Amazon adds almost nothing. If it has that power, it’s there already and the merger won’t appreciably change it.

Radia: If Amazon were buying another large online retailer (say, Overstock.com), then the monopsony argument would make at least somewhat more sense. But with WF, I am with Geoff: I don’t see a monopsony problem.

Singer: Can’t Amazon/Whole Foods combined more efficiently squeeze food suppliers than just Whole Foods can do alone?

Manne: I suppose, but “squeeze” is a loaded term. Arguably any increase in size would accomplish this. But that’s a good thing. Perhaps Whole Foods has too little power.

Mitchell: Consolidation is so extreme in the grocery sector that suppliers have seen their share of the food dollar go down, and supermarkets and processors have seen their share go up. Adding another powerful player certainly doesn’t make that a more competitive market. But, as I said at the start, this is also about Amazon’s dominance of online retail. Buying Whole Foods would amplify its platform position and give it more power vis-à-vis all suppliers.

Manne: Consolidation is extreme? The nationwide HHI is under 500! The biggest retailer has only 17% of the market, and Whole Foods, the ninth largest, has only 1.7%. There’s no concentration here.

Mitchell: Walmart has greater than 50% of grocery sales in over 40 metro areas. National market for groceries is consolidated, but it’s the local level where it’s especially extreme.

Steinbaum: Exactly, if we define the relevant market at the local level, as Stacy has done, then we do detect signs of market power.

Manne: No. You apparently detect signs of concentration. But in the monopsony context you’re going to have a tough time with that. Even produce can travel around the world and be sold fresh. So a local “monopsony” doesn’t grant market power.

Mitchell: Walmart’s power in groceries has triggered several other supermarket mergers. The effects on suppliers are pretty well-documented. And consumer prices have trended up. There are other reasons for that, but consolidation is one.

Manne: There are tons of other reasons for that (oil prices, most notably), and I’ve seen no credible evidence tying it to consolidation at all, let alone somewhat. The short-term trend in food prices is mixed, at worst.

Radia: Regarding those 40 metro areas, I suspect Wal-Mart’s share might decline if more consumers were willing to buy groceries online for home delivery. About 60 percent of Americans are reluctant to purchase perishable foods online, according to a new MorningConsult poll. Amazon plus Whole Foods might mean more competition in those metro areas if the company is able to create a more palatable experience.

Manne: Right, a big point is how Amazon plus Whole Foods could actually mount a real challenge to Wal-Mart in certain places. Seems like win for all! (Except Wal-Mart…).

Steinbaum: But that would in turn strengthen the monopsony argument. A larger share of e-commerce would take place through Amazon’s platform and logistics network, pressuring other suppliers.

Singer: Let’s switch to perhaps an easier story than monopsony. The next question is inspired by a new merger piece from Capitol Forum: Does the merger reduce actual or potential horizontal competition in grocery delivery between Amazon Fresh and Whole Foods/Instacart?

Mitchell: It’s hard to imagine that this won’t torpedo investment in competing grocery delivery startups. This deal would let Amazon own grocery delivery without even having to work for it.

Steinbaum: Right, not that this is dispositive, but Blue Apron’s value took a big hit when Amazon-Whole Foods was announced.

Manne: “Without even having to work for it”? What does that mean?

Singer: Geoff/Ryan: Wouldn’t competition in the market for grocery delivery be stronger if AmazonFresh competed against Whole Foods/Instacart (in lieu of acquiring Whole Foods)?

Manne: That’s not a relevant market! We’re back to defining markets by channels of distribution. Don’t you think “grocery delivery” competes with grocery stores?!

Radia: I agree this deal is bad for Instacart, but I don’t think that matters by itself.

Manne: So you think antitrust should require to Amazon to build its own stores and put Whole Foods out of business at greater cost, rather than buying Whole Foods?

Singer: No. I think antitrust considers the reduction in actual or potential competition in a relevant antitrust market. Stacy/Marshall: Why can’t Instacart, after losing the Whole Foods business, strike a deal with (say) Wegmans and go after Amazon/Whole Foods?

Steinbaum: Well, Instacart has nowhere near the existing infrastructure or access to capital to make that viable. That’s the whole argument Stacy has been making for years. There’s increasingly no plausible way around Amazon.

Manne: So no one but Amazon has access to capital? I’m pretty sure if monopoly profits were actually on the table, Wegmans would be more than happy to front Instacart the capital.

Radia: AmazonFresh has been slow to expand. It launched in 2007, and it still doesn’t deliver in major cities such as Portland, OR. Who knows if Amazon would really build grocery stores in more than a handful of affluent cities, if that?

Steinbaum: Wegmans is not going to front an all-out assault on Amazon in e-commerce. Walmart is only now doing it, and only just. Amazon is already dominant and already anticompetitive. There’s also, dare I say it, the threat of antitrust! Look at Apple e-books. I think that case was decided correctly, but I think it’s fair to say the agencies have been favorable to Amazon in the past and would-be competitors might assume they will be going forward.

Radia: It’s interesting that more companies haven’t mounted an all-out assault on Amazon. Is it because would-be competitors are scared of Amazon, lack the capital to compete, or fail to see the profit potential? To the extent that it’s the latter, that suggests we should be less worried.

Mitchell: And Walmart has struggled mightily despite all those resources. One other note: Part of the issue here is that online markets are different in critical ways from offline markets. So this framework about defined markets that are distinct is rather analog. It misses the fluid relationship between online and off.

Manne: Totally agree with Stacy there’s a fluid relationship between online and off — which is why there is no real separation between the two segments (online/offline) most likely. So we’re talking about products, not channels of distribution. And the relevant market includes online and offline sales of those products.

Mitchell: That’s not what I mean. More retail spending is going to be digitally driven, even offline retail, and Amazon increasingly has a lock on consumers online. I don’t think antitrust authorities have a handle on these dynamics. And the best we can hope for with this scenario of Walmart successfully challenging Amazon is a couple of dominant players. That’s not my idea of a competitive market that has the right ingredients for generating innovation and value. Seee.g., Wilfred Dolfsma & Gerben van der Velde, Industry Innovativeness, Firm Size, and Entrepreneurship: Schumpeter Mark III?, 24 J. Evolutionary Econ. 713 (2014) (showing that a mix of firm sizes is good for innovation) and Germán Gutiérrez and Thomas Philippon, Investment-less Growth: An Empirical Investigation, NBER Working Paper No. 22897 (2016) (linking declining corporate investment with increasing concentration). There’s also a lot of reporting, by me and others, that suggests that diversity in retail channels helps new products find their way. Small producers, new authors, etc. find a few places that initially sell their work and it grows from there. In the absence of that retail diversity, we lose pathways to market.

Manne: I’m not familiar with a literature that says that that market structure can’t be innovative or compettive. I am familiar with the literature that shows, for example, that vertical integration is consistent with innovation and consumer welfare benefits (See, e.g., LaFontaine & Slade, Vertical Integration and Firm Boundaries: The Evidence, 45 J. Econ. Lit. 629 (2007)); that large firms are in fact innovative (See, e.g., Knott & Vieregger, Reconciling the Firm Size and Innovation Paradox, US Census Bureau Center for Economic Studies Paper No. CES-WP-16-20 (2016)); and that we don’t actually know what market structure is most conducive to innovation (See, e.g., Katz & Shelanski, Merger Policy and Innovation: Must Enforcement Change to Account for Technological Change? in Innovation Policy and the Economy (Jaffe, Lerner and Stern, eds., 2005)). You seem to have a preference for industry to “look” a certain way. That’s not sufficient.

Singer: Let’s shift gears. What are the potential harms flowing from the merger that are not readily cognizable under antitrust law, but nevertheless could be important from society’s point of view. Seems like consumers—the focus of the antitrust laws—may not be worse off in the short run, but other economic actors could be.

Steinbaum: I think Stacy and Lina have pretty definitively shown that Amazon has harmed upstream competition and innovation.

Mitchell: Consumers are only the focus recently, of regulatory policy and case law. Not the statutes.

Steinbaum: Since we’re talking about upstream markets, I’m not sure why we assume consumer welfare is the appropriate standard. Seems ill-targeted toward the empirical question.

Manne: So what other than consumer welfare would be an operationalizable, non-arbitrary standard?

Singer: I want to focus on harms that might fall outside of the narrow consumer-welfare lens. Let’s start with the accumulation of data on personal shopping patterns that could serve as an entry barrier for distribution rivals. Stacy/Marshall, why is big data a barrier to entry? And how does the merger make things worse?

Steinbaum: Stores in affluent areas, already frequented by upscale shoppers, are great laboratories for gathering data.

Manne: And laboratories are bad?

Steinbaum: Everyone will pay their own personal, customized price.

Mitchell: Amazon uses the data it gathers about customers buying from third-party sellers to undermine those sellers as competitors.

Manne: Price discrimination of course has ambiguous welfare effects. So why are you so sure it’s bad?

Singer: Not first-degree price discrimination, Geoff. That’s unequivocally bad for consumers. Third-degree price discrimination generates ambiguous welfare effects. Sounds like Amazon is doing something closer to the former.

Manne: Actually, (assuming it becomes possible) even first-degree price discrimination has ambiguous effects. See, e.g., Shiller, First-Degree Price Discrimination Using Big Data (2014).

Steinbaum: Why so sure price discrimination is good? When it turned out Kaplan test prep was charging Asian customers more for its services, that certainly caused a stir.

Singer: I get that Whole Foods offers Amazon a treasure trove of new data. But doesn’t Wegmans possess the same data for its affluent shoppers?

Steinbaum: Wegmans has not been running price-discrimination experiments for decades now.

Radia: Amazon already knows a ton about long-time Prime customers, among other customers. In theory, it could combine that info from whatever it learns about those same customers whenever they walk into a Whole Foods. But translating that into harmful behavior seems difficult.

Manne: This is not a cognizable antitrust harm.

Singer: Perhaps. But I want to see if we can agree that the merger could impact things (like employment of low-skilled workers, the demand for brick-and-mortar office space, etc.) in ways that might escape traditional antirust scrutiny.

Manne: OK, but no one answered my question of what otherstandard you’d use. Just describing what happens and saying it’s bad isn’t sufficient.

Steinbaum: Okay, here’s another candidate standard: racial discrimination is bad.

Manne: Sure, Marshall, racial discrimination is bad. Are you saying Amazon engages in racial discrimination now!? Anyway, we don’t need antitrust to deal with that.

Steinbaum: Antitrust law did once deal with that.

Singer: I think what Marshall is saying is (1) Amazon prioritizes or discriminates in favor its own wares, and (2) antitrust is not good at policing this sort of behavior. Is that fair, Marshall?

Mitchell: More than half of online shoppers start at Amazon. Online is a sticky ecosystem and Amazon has perfected, through Prime and now Alexa, luring people in so they stop shopping around. That’s the thing that becomes very hard for a challenger to contend with. On Prime, Amazon loses money because it knows that once you sign up and pay that $99, your tendency to look around drops considerably. This is a winner-take-all set up with network effects. And the conflict between Amazon as platform and Amazon as direct seller—that’s where there anticompetitive effects are. We need to split those two pieces apart. And we need to make the platform a common carrier. That’s where you end up if you look at this from the public interest standpoint.

Manne: Why does treating Amazon as common carrier promote public interest? You want Amazon to look like the local power company? That seems unlikely to be a net positive. Are you familiar with the amount of innovation in public utilities? You’re joking, right? Also, doesn’t that mean Kroger should be a common carrier as well? And certainly Wal-Mart.

Mitchell: Amazon will sometimes give itself the buy-box even if it’s not the cheapest option. Within weeks of a new seller coming on to the platform, Amazon brings 25% of their best-selling items into its own inventory. It owns third-party sellers customers and controls how they may communicate with those customers. It shuts off seller accounts overnight for no cause and with no recourse. What Amazon has done is turn an open market into a privately controlled arena where it sets the terms by which other players may buy and sell. That’s why its platform has to be broken off from its direct retail and made to comply with common carrier rules.

Manne: Let me rephrase to make that more accurate: What Amazon has done is create a market where it sets the terms by which other players may buy and sell. That’s why its platform has been so successful, so beneficial for consumers. Common carrier rules would ensure it never got created in the first place.

Steinbaum: Seems to me Amazon is, technologically speaking, a pretty straightforward application of advances that would have existed whether or not that company did.

Radia: I agree the technology would have existed in another world, for the most part—but not many other aspects of the company. Zero-price same-day shipping of a ton of products for a modest annual membership fee in most major U.S. cities? I am skeptical that investors would have tolerated Amazon’s creation of this offering despite the company’s minuscule earnings for so many years under any sort of common carriage regime.

Steinbaum: You may well be correct about investors’ tolerance. But that’s precisely because they (correctly) perceived the threat of Amazon’s market power.

Singer: I’m not a fan of common carriage (as you probably know). But I get how web upstarts can be harmed if the playing field is slanted towards the platform owner.

Radia: There are plenty of small-time manufacturers who have beef with Amazon. But some of those manufacturers, after initially catching fire through Amazon, have moved to direct-to-consumer sales. For example, The Ripple Rug, a very popular cat toy, took off through Amazon’s platform. Now the company sells the new version of the rug only on its website.

Singer: But that’s a huge drop-off in views relative to being seen on Amazon.

Mitchell: As recently as 2015, most shoppers started at Google or another search engine.  Now 55% of them start at Amazon. That’s only going to climb with Alexa’s integration into our lives and into many products. So those alternate channels are going to be smaller and smaller. For the report we did last fall, I talked to many small and midsize manufacturers and competing retailers who had thriving web traffic and have recently seen that plunge.

Manne: I’m still waiting to hear anyone propose a standard that would distinguish between discrimination or use of data or anything else to harm competition (or decrease welfare in a measurable way) and those that improve outcomes. So far it’s just a presumption of harm. But the economics we have refutes that, and a non-standard standard is just a recipe for politicized enforcement. You sure you want to encourage “antitrust in the public interest” when, say, the Trump administration gets to decide what that is?

Steinbaum: The standard we currently have is extremely political. It looks the other way as enormous companies have achieved dominance over the economy, driving up profits and driving down wages, investment, and growth.

Manne: It isn’t so political, Marshall, because the courts that enforce the laws y’all want to dismantle ensure that there is a limit to discretion. Of course there is some politics in it, but it would only be far, far worse under your standard-less standard.

Singer: Sticking with non-traditional effects, Diane Lim wrote in Washington Bytes about how the merger could reduce some low-skilled jobs at Whole Foods store, like cashiers. Is that a problem for antitrust?

Mitchell: With the new research on the link between consolidation and wage inequality, I think it should be.

Steinbaum: I think that’s a good question. I think we’ve seen that other consummated mergers, including vertical ones, have had adverse effects on employment and wages.

Mitchell: Amazon has cost about 300,000 job loses at brick-and-mortar. And created about 150,000 jobs. In the long run, the economy might make those up elsewhere. But it’s a big impact for the one in ten Americans working in retail.

Manne: Stacy, as you know, I think your assessment of the causal relationship between job losses and Amazon is flawed. But even if it weren’t, the right way to deal with that is fixing the employment problem, rather than harming all consumers (including the 90% who aren’t also working in retail) and long-term innovation (which would result if any firm that reaches a disfavored size is broken up).

Steinbaum: The employment problem is the market power problem per my latest research.

Manne: We don’t have space here to debate that paper, but, as you know, I disagree whole heartedly. Among many places to look, see David Schleicher’s recent paper on labor mobility (Schleicher, Stuck! The Law and Economics of Residential Stability, 127 Yale L. J. __ (2017)). And Hsieh & Moretti, Housing Constraints and Spatial Misallocation (Working Paper, 2017).

Steinbaum: I take issue with Schleicher in an article in Democracy, and Hsieh and Moretti is irrelevant: It says nothing about the decline in labor mobility over time—it is merely cross-sectional.

Mitchell: One last thing on jobs. One of the biggest problems in the economy is the deterioration of wages and job quality. Both Amazon specifically and consolidation more broadly have played a big role in driving that trend. Antitrust needs to consider it.

Manne: We should definitely care about people’s livelihoods. But we should be smart about helping them without doing more harm in the process.

Mitchell: We looked at Amazon’s wages in eleven metro areas and found that they pay 15% less on average than the prevailing wage for the same type of work in the region. A possible sign of monopsony power over labor rates? They also rely heavily on temp works who have no job security.

Singer: But Stacy, does the merger make things worse for workers? Or is this a complaint about Amazon generally?

Mitchell: We can assume that Amazon will bring more automation to Whole Foods, both in distribution and retail. And we can assume that Amazon will want to apply its signature wage-lowering and subcontracting approaches at Whole Foods. Much of the labor model Amazon is bringing to logistics consists of gig-economy arrangements and very low-wage fly-by-night subcontractors. We should remember that UPS and the postal service are one of the few surviving corners of the working middle class in this country. That’s a big deal.

Manne: Stacy, the real problem is that your approach here assumes a market structure that is conducive to innovation and other welfare benefits, but the literature does not support it at all. Perhaps the literature is flawed; it’s a tough issue to get a handle on. Unfortunately, the initial burden of proof is on the proponents of breaking up firms.

Singer: Let’s conclude with counterstrategies. What is to stop another company from replicating Amazon’s platform as an e-commerce company? Perhaps incumbent distributors like UPS could do so, either by integrating itself or on a non-integrated basis? Why can’t Wegmans or Walmart, for example, strike a deal with FedEx to replicate and thereby neutralize Amazon-WholeFood’s advantage?

Radia: My best guess is that potential entrants don’t see a clear path to earning a reasonable return in the online retail market. But I would welcome Wal-Mart acting more aggressively online. And there are interesting potential combinations of other brick-and-mortar retailers that specialize in specific markets (e.g., Home Depot, Best Buy, Staples, Kohl’s, Kroger). If Amazon starts earning supra-competitive prices, that could jump-start a lot of interesting joint ventures and mergers.

Steinbaum: Mergers with Amazon I fear. I think Stacy has already articulated the right counter-strategy: Break up Amazon’s retail from distribution, place a common carrier-type regime on its distribution, and take a critical look at price discrimination to see whether it actually reduces or increases consumer surplus. I do not think that any of the company pairings proposed in this discussion as a means of disciplining Amazon would ever work, or even happen.

Singer: But Marshall, your side has the burden to argue that market forces cannot be counted on.

Steinbaum: Look at the parade of firms resisting Amazon, only to eventually capitulate. Walmart is the only one left that could conceivably challenge it meaningfully, but Walmart is also anticompetitive, and my vision of the economy is not one in which two vertical monopolists battle it out and crush the rest of us in the meantime.

Manne: That’s just it though, Marshall: “your vision of the economy.” Why should we implement your vision? And if antitrust is just about implementing someone’s “vision,” there’s nothing to ensure that it’s your vision that will be implemented. Are you comfortable if I’m the one deciding what vision to implement?

Mitchell: Here, here, Marshall. I want to live in a world where entrepreneurship is still possible. I think antitrust has a responsibility to create open and fair markets. I think Amazon may be uncatchable at this point, absent proactive intervention. The public interest would be well served by that intervention.

Manne: It’s not that simple and you know it, Stacy. Let’s not score political points rather than address the complexity. You would close Wal-Mart stores, too. But that pulls the rug out from other people. Consumers would be harmed, and suppliers, and, in the case of Amazon, the retailers who access global markets, and on and on. I am advocating not adopting massive industrial policy programs on a hunch.

Mitchell: This is political. It’s about power and how it’s distributed.

Steinbaum: Well the existing competition policy isn’t exactly doing a great job. Maybe go back to the one that we had back when the economy used to grow? I’m not saying it was perfect, but it sure seems like we made huge policy changes based on theoretical models that we never actually tested.

Mitchell: Marshall, I agree. We need to be doing more to address market structure.

Manne: You want to distribute massive political power to whomever happens to be in the White House. Because without economic standards, that’s what will happen. It’s a huge mistake. Notwithstanding the fact that it isn’t clear it would accomplish even what you want in this instance, it’s always preferable to rely on evidence and actual effects and not to fall prey to the Nirvana fallacy.

Singer: Wow. Didn’t expect us to go there. Anyhoo, I think we’ll have to make that the last word. Thanks to everyone for joining.

Read the full story here.