May 21, 2021
On May 18, The Capitol Forum hosted a conference call with AAI President Diana Moss to discuss AAI’s letter to DOJ antitrust division head Richard Powers expressing concern that UnitedHealth Group’s proposed $7.8 billion acquisition of Change Healthcare would likely harm competition and consumers, as well as her views on the merger’s potential competitive impact. The full transcript, which has been modified slightly for accuracy, can be found below.
MR. NATE SODERSTROM: Thank you. Good morning, everyone. Thank you for joining The Capitol Forum’s Conference Call to discuss UnitedHealth Group’s Proposed acquisition of Change Healthcare. I’m Nate Soderstrom, Senior Editor at The Capitol Forum and I’m joined today by Diana Moss, who is the president of the American Antitrust Institute.
A quick note before we get underway. I have a number of questions I’ll be asking Diana, but we are also happy to entertain questions from the audience. If you do have questions for us, please email them to email@example.com. Again, that’s
firstname.lastname@example.org. And Capitol is spelled with an “o.” With that, let’s dive right in. First off, Diana, thank you so much for joining us today.
DR. DIANA MOSS: Thank you, Nate. Thanks to The Capitol Forum. Appreciate being on the program today.
MR. NATE SODERSTROM: Absolutely. So just to kick things off, we’ve seen quite a few deals announced over the past number of months. AAI, of course, doesn’t write letters to the antitrust agencies on all of them. So when it comes to UnitedHealth/Change Healthcare, why does AAI see this one as uniquely meaningful to the point you decided to come out publicly to discuss your potential concerns?
DR. DIANA MOSS: So I think it’s safe to say I could hire ten more people, economists, lawyers and others, to cover all the issues that we’d like to cover at AAI. But we do have a small staff, small but mighty, as I like to say. I think this issue we took up for some pretty good reasons. Most of all, it really highlights the intersection of competition issues in the digital tech sector with another really important sector, which is health care. So a lot of Congressional focus, agency focus, with the monopolization cases in digital tech. The media seems very enamored with digital tech for the last two years. A lot of resources are going into digital tech. AAI is a real proponent of ensuring that we give our attention to competition issues across all sectors, not just
digital tech—although we do a fair bit of work there—but also all areas of enforcement, those three major areas of enforcement.
So what we have here is really a marrying up of issues, emerging and digital tech, with competition issues in a really important sector, health care. And it’s not just health care where we’re seeing digital technology issues pop up. We’re seeing it in digital farming, for example, over in the agricultural space. We’re seeing it in real estate. AAI just did a big program on competition in real estate. So looking at those intersection points of competition issues raised by digital tech, but with applications in other sectors, I think is really important.
The other major reason we took this one up is because the whole data analytics transition, the digitization of many industries, really isn’t a success story in health care that it is in other places. And we can talk more about that later in the conversation. But it does go to what will be the likely efficiencies claim for the parties to push this transaction through, which is, wow—cloud infrastructure, data analytics, offers huge opportunities for cost savings and for innovation. But our response is not so fast. Because there’s a pretty strong record that data analytics really—the uptake for data analytics and other cloud-based technologies—is not happening in health care as it is in other sectors.
MR. NATE SODERSTROM: Right. Like you said, it’s certainly a mixed track record there. One thing you allude to in your letter, a couple of things that you allude to in your letter, that are kind of interesting about UnitedHealth Group, who is, of course, the acquirer here.
First, they’re really vertically integrated and they occupy really significant positions along basically every level of the health care delivery supply chain, whether it’s health plans, whether it’s health care providers, PBMs, et cetera.
Second, they’ve been really acquisitive in recent years, whether we call them a serial acquirer or not. But in any event, they’ve done a lot of deals. When it comes to UnitedHealth Group, is that a background where (a) they’re fully vertically integrated, (b) they’re something of a serial acquirer, that from your perspective raises particular concerns when thinking about this transaction?
DR. DIANA MOSS: I think it does. There’s a lot of talk and movement on the antitrust reform front these days obviously. AAI really supports vigorous, strong enforcement that incorporates information on past patterns of acquisition. For example, the largest digital tech companies have acquired their way to dominance in markets. So growth through acquisition, taking out smaller firms, potential rivals, creating verticals within a larger supply chain, I think is all relevant for analyzing future
transactions on the M&A front. And certainly, we’re a large proponent, AAI is a proponent of looking at past violations of antitrust laws in evaluating future antitrust enforcement actions.
So that kind of answers the question up front. I think, to dig down a little bit, we are seeing a tremendous amount of verticalization in health care. We have insurers buying PBMs, hospitals buying physician practices, insurance companies now buying digital assets which focus on claims administration, revenue payment cycles. That’s really what we have here in the case of Change Healthcare and UnitedHealth Group.
So we are seeing verticalization not only in health care, but in media and distribution, agricultural, biotech. And what that does, I think, is that it creates some troubling trends. One is you’ve got a lot of bargaining power going on within a supply chain, bulking up to be a better bargainer vis-a-vis providers or insurers. We now have very high concentration at all levels of the health care supply chain in PBMs, in hospital systems, in health insurance.
So we’re kind of at a dangerous point where the ability for smaller firms to enter markets with disruptive, innovative business models is getting harder because of higher barriers to entry created by these really reinforced vertical supply chains in health care. So the paradigm of competition is
shifting from relatively free entry at any level in the health care supply chain to just having a few vertically integrated systems. And there, you’re talking about, well, how many systems is enough to ensure robust competition so that there are benefits to consumers and that providers and other medical professionals are not harmed through the oppression of monopsony power?
So I think we’re in a very precarious position here. And it all sort of goes back to your original question, which is, how have these firms gotten so large, so tightly vertically integrated? And what does that mean for future transactions?
MR. NATE SODERSTROM: Yeah, absolutely. And then I think maybe one other question around perhaps the broader environment here before we dig in on the specific deal. But there are a couple of things you mentioned in your letter as counseling towards the view that this deal should face particular scrutiny. We’ll call them environmental factors, certainly when it comes to health care.
The first is, of course, as you mentioned, the Covid-19 pandemic, which, knock on wood, seems to be entering its final stages, and certainly has had a big impact on health care markets. You also mentioned accelerating consolidation and concentration in health care markets more broadly. Both actually, you mentioned in your letter.
What role should those dynamics play when it comes to DOJ’s review of this deal? Do they counsel towards more caution, more skepticism, these broader environmental factors you mentioned in your letter?
DR. DIANA MOSS: Yes. So there’s been a lot of talk about acquisitions during the pandemic. And I’ve certainly been on a lot of programs domestically and internationally talking about this. And I would sort of sort it into two buckets. The issue of greatest concern was acquisitions of failing firms during the pandemic. And we did see some of that. I think the agencies were on code red looking to flag those types of acquisitions under what are fairly rigorous standards in the horizontal merger guidelines. That’s kind of bucket number one.
I put this type of acquisition, in terms of its connections to the pandemic, in bucket number two, which is especially with the health insurers now coming out of the pandemic, everyone’s going back to the doctor to get procedures done, getting back on regular health visits, schedules. Costs are for sure going up. All you have to do is look at their financials to see that. So what better way to reduce your costs than to gain more market power? And, of course, what they’re seeking to do, I think, is obviously to reap what are some pretty clear economies of scale and scope in cloud computing, cloud infrastructure to do this. Obviously,
eliminating markups on purchase services is another motivation for vertical integration.
So this second bucket is a little more complex than just acquiring failing firms that didn’t make it through the pandemic. This is seeking to get larger and to create a more solid supply chain, internalizing all of these costs and externalities to reduce costs, number one, but presumably to be a little bit more innovative over the longer term, which I think should be viewed very skeptically.
So I think antitrust enforcers should be looking to this rationale for kind of post-Covid, post-pandemic, consolidation when they look at issues like efficiency claims and that sort of thing.
MR. NATE SODERSTROM: Yes, certainly, as you allude to, a dynamic environment on both those fronts. Your letter I also thought it raised a really interesting point on data, health care data. Both companies have a lot of data. And you write that post merger, UnitedHealth Group could have less incentive to protect sensitive patient data, and therefore privacy, either because you’re losing some level of head-to-head competition or simply because you have a larger health group that just maybe has more leeway. There’s less of a check on them on protecting this data around privacy.
Obviously, we haven’t seen the U.S. antitrust agencies really go to the mat on privacy concerns in the merger review context. But I did want to ask if you could talk a bit more about potential privacy and data concerns you see arising from this deal.
DR. DIANA MOSS: Sure. And I think the data concerns and the privacy concerns around user data are part and parcel of a major competitive concern here, which is potential foreclosure of rival health insurers that outsource their data analytics, for example, and rely on Optum, which will be larger and more powerful after the acquisition for their clearinghouse functionality, claims administration payment and revenue cycle services.
So the data analytics, the privacy issues, sort of all feed into theory of harm, one of theories of harm, which could be vertical foreclosure. Obviously, the other theory of harm is that you’re just horizontally combining assets in a health care data analytics market. And the DOJ obviously will be looking into where the overlaps are, whether it’s particular services or whether it’s these cluster markets within the cloud computing domain.
So I think I understand the privacy issues. You have to get to the root of what the value proposition in the digital ecosystem is. And again, here’s another example of it. You don’t have to necessarily just be in the digital tech sector and
be talking about Facebook or Google. You can be anywhere in a supply chain where there is strong cloud infrastructure and digital capability. That could be health care, farming, wherever, the value proposition is to collect data. And it’s to harness the value of that data through the application of data analytics, which often involves artificial intelligence and machine learning. Optum has that capability. Change Healthcare has that capability. They will be bigger and stronger as a result.
The very strong incentive to realize the value proposition is to keep users in the ecosystem. You don’t want people leaving the Optimum ecosystem to go get services in another ecosystem. And you can do that, as a digital ecosystem, by shaping preferences through the well-known techniques of targeted advertising and recommendations. And in the health care system, that could range from anything in terms of insurance plans to other types of health care services, preventative health services, for example. So the whole value proposition is to keep the user engaged, keep them in the ecosystem, shape their preferences through algorithmic recommendations.
We also know that users in any digital ecosystem, there is an enormous disconnect between what they say they value in terms of their privacy and how they actually act in terms of protecting their privacy. There’s a disconnect there, which is really a sort of an
So we don’t have much confidence that there are really any incentives to compete on privacy because of the value of user data and the centrality of enriched user data to the digital ecosystem value proposition. And that should give enforcers significant pause when looking at an even larger digital ecosystem that would result from this merger.
And remember, it’s not just collecting data and having more data. Ultimately, what they want to do is to simplify financial transactions on the claims front. They want to increase speed and accuracy. Those are all sort of cost reducing types of benefits. But the ultimate goal here is to use all of these data analytics and outcomes from that process to sort of improve clinical decision making, and that is a much longer-term type of benefit that relies heavily on the collection of data and the exploitation of user data.
MR. NATE SODERSTROM: Yes, that all makes sense. And then I think as you kind of alluded to at the outset of your answer there, there’s a lot going on with this deal. But I think what’s probably the crux of the matter, or very well might be, is concern that UnitedHealth Group could disadvantage UnitedHealthcare’s health insurance, health plan rivals, through its control of Chane’s EDI clearinghouse and some of the data running through that, and associated data analytics.
Your letter raises what I thought was a really interesting point, which is, and I’m quoting here, “Such an anticompetitive strategy would be very hard for rivals to detect and make it more difficult for them to compete.” I thought that was so interesting because when prices go up post merger, people see that. It’s visible. You can figure out that you’ve been harmed by a merger. Again, if that’s what happens. If you’re a health plan here and UnitedHealthcare has post-merger access to your competitively sensitive information that’s running through a clearinghouse and perhaps does something to advantage UnitedHealthcare or to disadvantage you, you’re not necessarily going to know. In fact, you very well might not know at all.
So I wonder if you could just talk about this idea that rivals who might be harmed by this deal, especially health plan rivals to UnitedHealthcare, don’t necessarily have a strong mechanism to actually detect that the deal’s harmed them. And I guess secondarily, how should that issue play into the agency’s thought process here?
DR. DIANA MOSS: Right. And so I think returning to the first principles to get at your question is really important. This merger raises a standard foreclosure concern, which is, in addition to just sort of horizontal concerns resulting from combining two very similar assets, but the vertical concerns, I think, are very, very important. So now you’re going to have a larger digital health care technology
platform by combining Optum with Change Healthcare. That is a very valuable asset and it’s going to be larger and more powerful. UnitedHealth Group is also vertically integrated into health insurance. So there’s the connection point to creating concern over vertical foreclosure.
So if you’re a rival health insurer searching around for another digital health care technology service that you can use other than UnitedHealth Group’s Optum, that might be kind of tough. There is high concentration in that market and it’ll depend on how it’s actually defined by the Department of Justice. But this poses a potential concern for health insurers. So if health insurers are purchasing services from this now larger and more powerful Optum, given that it is a digital technology platform, given the value of data, given the cloud infrastructure that is available there to process data, to keep users in the system, to steer them around into the different services that UnitedHealth Group offers through other Optum subsidiaries and certainly to keep users within or subscribers within the UnitedHealthcare insurance system. That’s a pretty ominous scenario for a rival health insurer.
And the mechanism for doing that may not be very obvious because we’re talking about digital technology. It could be, for example, that the foreclosure strategy disadvantaging rivals while favoring UnitedHealthcare might occur through how the algorithms are created and deployed
through the Optum system. That’s not easily detectable—it’s not your standard foreclosure theory in a manufacturing industry where you would cut somebody off from access to say, a gas pipeline or to content, for example, in the AT&T/Time Warner deal. This would be highly technical, in the cloud, very much based on sort of the algorithmic processes that go on and would be very, I think, very hard to detect. And that also raises questions for enforcers in terms of remedies.
So, for example, if this was teed up, this theory of harm of vertical foreclosure, how does an outside antitrust enforcer craft a remedy that is designed to get at how data is collected, how data is processed, how algorithms are constructed and that sort of thing. I mean, it raises very, very sticky issues.
But I will say on the vertical foreclosure theory, this would not be the first time—if DOJ goes down this road—where they have floated a theory like this. So in the Google/Admeld case, for example, there was concern over the extension of market power from search to display advertising. The European Union on the Google/Fitbit merger was very concerned about more market power in health and fitness data and extending that market power to the ad tech market. So for enforcers, this is not new stuff. They have taken on these issues in other digital tech mergers. So I think the same idea is in play here.
MR. NATE SODERSTROM: Yeah, it’s a really interesting question and we’ll certainly get back to remedies in a second. You’re an economist. When you think about, say, the more traditional vertical concerns about raising rivals’ costs or input foreclosure or thinking about DOJ’s suit seeking to block AT&T/Time Warner, and there you had an attempt to quantify the harm. And I forget, $750 million in consumer harm in 2021, whatever the case may have been, it’s quantifiable—again, it’s not your straightforward horizontal analysis, but it’s something you can seek to quantify. The questions we’re talking about here where it’s algorithmic discrimination or maybe access to rivals’ competitively sensitive information, they both sound like plausible theories of harm, but at the same time very difficult to quantify or predict with any real level of certainty or even likelihood.
So I guess how do you think about the vertical theories here, where it’s not really raising rivals’ costs where you can quantify it, and it seems to be these kind of other concerns, which seem like they’re very real, but in that same respect, if you’re an agency and potentially have to go to court on these concerns, how do you make that case when the economics are even less quantifiable than they would be in your typical vertical situation?
DR. DIANA MOSS: Yeah, well, that’s the million dollar question. And I think DOJ has got a big job on their hands in terms of how they frame up that type of theory of vertical foreclosure if they choose to go that route. I think a simpler version of that comes from the horizontal integration of these two platforms, Change Healthcare and Optum. There you do find some similarities with what the states, for example, put forth in the Facebook case, which is where the metric of exchange in an information market is not price. It’s user attention and information.
If there are stronger incentives post-acquisition for UnitedHealth Group through Optum to violate user privacy, that reduces quality. So privacy is a quality, is a dimension of quality. And quality is absolutely reachable under the consumer welfare standard. So that’s not new. That was put forth pretty directly in the Facebook case. The same issue could arise here in terms of less powerful incentives, weaker incentives, to protect patient privacy coming from a horizontal theory of harm.
In terms of actually proving up a foreclosure theory, as you asked, I think that’s a tough one. I’d have to actually give that some more thought. How in a courtroom, for example, the government would make their case that vertical foreclosure is likely and would result in harm to competition through the exclusion of private health insurers?
I don’t think it’s impossible. I think the government has proven itself to be creative and willing to take on these challenges. But that’s what we’re all thinking about. Like what does that case look like in practice?
MR. NATE SODERSTROM: Right. And certainly with this one, the horizontal piece around claims editing software and payment integrity. So we could see some horizontal elements to a case as well as some vertical elements adding further level of complexity to things. Speaking of which, speaking things that may or may not be quantifiable, as you’ve alluded to, this is a largely vertical deal. You would expect a key piece of what the merging parties are going to tell the agency here is that this deal would create various efficiencies. You argue in your letter that efficiencies claims here should be viewed with extreme skepticism. I’m wondering if you could walk us through your thinking there.
DR. DIANA MOSS: Yes, they should be viewed with extreme skepticism. Certainly, the announcement that AT&T is spinning off WarnerMedia as of yesterday, a mere three years after having acquired Time Warner, should be a huge red flag for the issue of vertical efficiencies. So clearly spinning off an asset three years after you acquired it is pretty proof-positive that the efficiencies that were claimed in AT&T/Time Warner did not pan out. And that merger gave rise to the vertical merger guidelines, which are really
heavy on a pro-efficiencies kind of orientation. So I think that’s a separate discussion and has enormous implications.
In this case, they will absolutely, and have made, an efficiencies argument that this will not only reduce costs, but it will allow for the delivery of better clinical decision support at the point of care. So amassing more data, processing more data, will enable better decisions, presumably to reduce costs and to improve patient mortality.
That all sounds great. It sounds like sort of your classic coordination of efficiencies between digital health care technology and health insurance. But not so fast. There’s a big literature out there on how data analytics really have not been embraced with the kind of enthusiasm that we’ve seen in other sectors, and that’s for a few reasons.
One is there’s huge patient confidentiality and privacy concerns in health care. Patients value their confidentiality. They are reluctant to give out information because they fear their insurance company will find out and their rates will go up. There is a fundamental mismatch between the objective functions of insurers, which is to minimize costs, and providers, which is to maximize quality of care. That’s a fundamental misalignment of objectives across levels of markets.
And the research has also shown that the health care profession is resistant to adopting data analytic approaches. Something like 56 percent of hospitals have no plan to adopt data analytics in any significant way in terms of the future.
So you have this sort of really interesting fact pattern in health care which should raise skepticism about the role of data analytics and arguments for why that would produce efficiencies in this particular merger. And as we know, efficiencies arguments have to be pretty powerful to overwhelm concerns over anti-competitive effects. Call it a structural vertical presumption, if you will, similar to what we have on the horizontal side. So that should absolutely factor into to an investigation and how the DOJ chooses to pursue the case or even take a remedy in that case.
MR. NATE SODERSTROM: Yeah, I do want to get to remedies in a moment. But we also do have a question from the audience. The question is—do you think concern about privacy in health care can be addressed by HIPAA so that these concerns may be viewed differently than privacy issues in digital tech for example?
DR. DIANA MOSS: Right. That’s a really good question. AAI is doing a pretty significant package of work right now, which we will be releasing in the next month or so, on digital tech, looking at issues like privacy, looking at the ability of antitrust to reach to issues raised by digital tech.
I think the general conclusion is that sort of notice and consent policies around privacy more generally are not very powerful tools to correct for the market failures we see in digital ecosystems. This goes to the issue I raised earlier of what users’ attitudes are towards their privacy versus what their actions are surrounding their privacy. There’s a real disconnect there. Users say they value privacy, but then they go and do things in the digital ecosystems that just contradict what value they placed on privacy. The opt in/opt out, consent notice, all of that stuff is just barely scratching the surface of correcting for those significant market failures around privacy.
I think the same issue applies here with HIPAA. I don’t think HIPAA is going to do it in terms of correcting those misalignments, but more important, providing any significant constraint on a larger and more powerful digital health care platform to act on incentives to exploit user data by violating or treating their privacy, treating their privacy with less respect.
So that’s a really fundamental issue in terms of remedy. And I don’t think HIPAA is a very powerful tool for addressing the much stronger incentives to exercise market power that this merger would create.
MR. NATE SODERSTROM: Yeah, which does get us to—and we’ve alluded to it a couple of times, but let’s address it head on—is this question of remedies. I think this is a deal where, at least in theory, you would see a resolution where DOJ puts in a firewall remedy that would prevent information exchange between the UnitedHealthcare and Change Healthcare’s EDI clearinghouse and associated data, and that could maybe resolve some or even most of the vertical concerns here, if we’re being charitable. I know AAI’s been skeptical of conduct remedies. And your letter notes here perhaps when it comes to the digital technology markets, there’s even more reason for caution. But I did want to ask you again if you could talk a bit about your skepticism of behavioral remedies, that kind of outcome, when it comes to this deal in particular.
DR. DIANA MOSS: Right, right. So good question. I think to start that, just to tee that up, I think we have to be clear in acknowledging that there is really no structural remedy here. Any structural remedy that DOJ could take would gut the deal, the value of the deal, to them, to the parties. I can’t think of anything that would work or be acceptable to the parties that looks like a structural remedy.
So you really are in the conduct. You’re either in the position of seeking a full stop injunction or going down the conduct route. And as we know, and you’re right, Nate, AAI has taken very strong positions on the ineffectiveness of
conduct remedies. They do not do anything to alter or change the incentive to exercise market power. They are simply rules and requirements and prescriptions for how the company can conduct its business post-merger.
And we now have a long history of conduct remedies failing, breaches of firewalls, breaches of nondiscrimination provisions. Look what happened in Live Nation/Ticketmaster. Gosh, the DOJ had to amend the consent decree there because of a violation of the remedies. We had concerns in Comcast/NBCU on conduct remedies. The list just goes on and on and on. Conduct remedies, especially when dealing with very large, powerful, highly vertically integrated companies are not very effective.
Here I would put those concerns on steroids because of sort of the fundamentals of the digital technology market. You’ve got network effects. You’ve got externalities in data. You know, a little data goes a long way. You’ve got this fundamental market failure around user data and privacy. You’ve got the enormous value add associated with enriching data through data analytics and being able to sort of keep users in an ecosystem through algorithmic processes. If you take all of that collectively and then say, okay. Well, we’re going to fix the anti-competitive incentives to discriminate against rival health insurers post merger by putting firewalls into place, by requiring data sharing, for example, by creating sort of artificial
boundaries and constraints on how the firm can behave, I think that’s a recipe for immediate failure because of the very, very powerful incentives that there are in a larger Optum to fully exploit these market failures, economies of scale and cloud, to fulfill this value proposition, I think they would run roughshod over any conduct remedies. And by the way, it would be really, really hard to detect a violation of those conduct remedies.
So unlike Live Nation/Ticketmaster, it was easy to detect violation of conduct remedies because concert venues, rival concert venues, absolutely were threatened. You’re not going to see that in a digital ecosystem. Because a lot of this will occur digitally in ways that are very, very difficult to detect. So I think it really creates a difficult fact pattern for conduct remedies and with an almost guaranteed recipe for failure.
MR. NATE SODERSTROM: Yeah, I think those are all very valid points when it comes to this one. Diana, I think we have managed to work our way through all of my questions. We have nothing outstanding from the audience, which I think it means we are wrapped up. So thank you so much for joining us today.
DR. DIANA MOSS: Thanks very much. Always happy to be on. And we’ll be monitoring what goes on, on this issue.
Vol. 9 No. 187 – May 21, 2021