Feb 18, 2026
On February 18, The Capitol Forum held a conference call with David O. Fisher, Senior Counsel at the American Antitrust Institute, to discuss DOJ’s proposed settlement with RealPage and assess whether its terms effectively prevent coordinated pricing through algorithmic software. The full transcript, which has been modified slightly for accuracy, can be found below.
TEDDY DOWNEY: Hello, everybody and welcome. I’m Teddy Downey, Executive Editor here at The Capitol Forum. Today, I’m joined by David O. Fisher, Senior Counsel at the American Antitrust Institute, to discuss his recent commentary, “Closing Costs, a Critical Examination of the DOJ’s Proposed RealPage Settlement.” And David, thanks for doing this today.
DAVID O. FISHER: Yeah, thanks so much for having me.
TEDDY DOWNEY: While I have the audience, if you have questions, please answer them in the questions pane or email us at editorial@thecapitolforum.com, and we’ll get to your questions. So, maybe we could start with the RealPage litigation that the DOJ brought. Why did the DOJ sue RealPage, and what did they allege? And then, after that, maybe we can get into the settlement.
DAVID O. FISHER: Yeah, yeah. So, DOJ brought a complaint alongside several state attorneys general, and that’s important to note right up front, that the settlement that we’re talking about is just with DOJ and RealPage. So, the state AGs’ case will continue. But the complaint alleges that essentially RealPage is serving as what we call an algorithmic cartel manager. And that’s sort of an antitrust term for an algorithm which functions to collect data and set a common set of pricing rules that competitors follow.
So, the allegations are as follows: Real page is an algorithmic pricing software provider for multifamily landlords around the country. And they enter into licensing agreements to provide pricing recommendations to these landlords, owners of multifamily housing units, where the landlords give up some commercially sensitive data, and RealPage uses that data both to train its algorithm and as an input that it uses to make a pricing recommendation to the landlord. So that’s the basic business model.
Landlords can accept or decline the algorithm’s price recommendations, but there’s allegations in the complaint that RealPage actually disincentivizes them from declining and encourages them to adopt an auto-accept feature, such that the recommendation is immediately incorporated into their pricing. And DOJ alleges that the widespread adoption of RealPage’s software has increased landlords’ market power in real estate markets across the country and actually raised rents above the competitive level in those markets.
So, just to take a step back, this is a set of facts that we’ve seen in a couple of cases now. Most of them are brought by private plaintiffs, but it very much follows the standard set of allegations that we’ve seen in algorithmic pricing cases.
TEDDY DOWNEY: So, we have this case. It’s basically that this software is helping landlords increase rent prices. It’s creating a monopoly problem of increasing these rent prices across the country where RealPage is providing services to these landlords, sort of helping them collude, allegedly. And so, DOJ comes out and settles. How unusual is it for DOJ to settle, but the states don’t also settle? I mean, it happens every now and again, but it seems pretty rare historically. What do you make of that? Or is that atypical?
DAVID O. FISHER: I mean, like you said, it happens. I don’t think there’s necessarily a standard playbook for the way that these things go. It’s really much a case-by-case sort of situation. And I will also say that there’s a lot at play here that we’re not privy to.
So, obviously, the parties have a much better understanding of the sort of risk assessment involved in each of their cases. DOJ has a complete other set of concerns that a regular litigant might not about allocation of resources. It’s got a lot of other business going on. It spends taxpayer money to prosecute these cases. And so, there’s a costbenefit analysis there in terms of the settlement.
From my view, looking at the outside in, it’s an interesting opportunity. Because we can sort of get the best of both worlds where we have DOJ, a really key player here, setting out a list of terms, which, in its view, is a reasonable resolution for the case. But we also get the benefit of continuing the litigation with the states at the wheel. And I think I fully expect the states to continue the litigation vigorously. And so, it’ll be interesting to see how that plays out. But it does make for an interesting set of facts here. There’s a lot to pay attention to moving forward.
TEDDY DOWNEY: One interesting thing, before we get into your analysis of the settlement, is that when you bring things up that are potential weaknesses in the DOJ language, they could take steps going forward to fix that. And then that could also come up in the litigation from the state. So, this is kind of a very timely, high-impact analysis from that perspective. Or is that not the right way to think about it?
DAVID O. FISHER: I’m happy to see it that way. I’m glad you see it that way. I mean, from our perspective at AAI, we’re kicking the tires. We’re looking at the settlement, like I said, as outsiders to the litigation, but as wellinformed parties who’ve thought and written a lot about these issues in the past. There’s utility to it, I think, in a lot of ways.
Most obviously there’s the standard public comment process that happens under the Tunney Act, pursuant to which we filed comments. We appended my commentary to comments that we filed with DOJ, and they will respond to those comments under that process. And they might answer our questions, elucidate some of these issues. They might make amendments to the terms of the settlement.
And then, moving forward, if and when DOJ gets a settlement approved, the states can use that as a baseline for the kind of things they want to see or address in the litigation moving forward. And then I think there’s like the sort of broader implications here, which is not just about RealPage, but more about how the courts and policymakers and business leaders are viewing this kind of algorithmic pricing software and this business model, and how we should be treating it moving forward.
So, I think there’s a lot to learn from this situation. And there’s a sort of a wait-and-see element to the degree to which the settlement terms actually sort of accomplish what they say they’re going to
TEDDY DOWNEY: And I know the paper’s about sort of the details. But before we get into the details, I have one more process question. Because if you look at a lot of what’s going on at DOJ’s Antitrust Division, it is kind of unseemly and unprecedented in terms of there is another Tunney Act proceeding going on with the HPE merger, where there are allegations from states that the DOJ entered a settlement for political reasons, not for competition reasons. There’s a lot of scrutiny there. We have that overhanging this Ticketmaster litigation that is about to go to trial as well. Is that coming up in this settlement at all?
There were some questions at the time of the settlement that the DOJ leadership and the White House went above the Antitrust Division’s head, and they have ties to private equity in this space, and maybe there was a little bit of deal-making there. Does that come up at all? Or is there any prospect of that coming up, since you have this state on the other side of the federal government on an antitrust settlement?
DAVID O. FISHER: Yeah, I mean, I’ll say that our comments are not at all about the process by which this settlement came about, and I’m not aware of any specific reason to question that process in this particular case. I mean, you mentioned HPE Juniper and some of the other high-profile activity that’s going on at the department right now. I think that’s sort of a different matter from our perspective, and the commentary is very much focused on the terms of the agreement itself.
And look, I think it’s important to note that the agreement appears to be aimed at exactly the right thing, right? I think its aims are the right ones. And this is a much more nuanced question about whether or not the terms will, in practice, have the effect that it’s clear that DOJ intends them to have. And so, we’re not really talking about the sort of bigger questions of governance and process that are implicated in those other cases that you mentioned.
And I just want to make clear that in filing these comments, AAI is sort of engaged in the regular public comment process of any settlement. We’re not challenging the validity of the settlement in any way like you’ve seen happen, of course, with the states and the HPE Juniper situation. So, this is a much more sort of targeted assessment of the specific terms.
TEDDY DOWNEY: It sort of seems like they’re at some level operating in good faith with the overall scope of what the settlement terms were. And what were those? I mean, obviously, this business model is, on its face, just a little questionable, just based on what we’ve seen in terms of rent increases across the country, in terms of articles discussing landlords now. Instead of focusing on having as small amount of vacancy as possible, they’re trying to maximize profit by having some vacancy and keeping prices high. That is sort of a new way of operating as a real estate owner.
So, we have some really heavy allegations here, really problematic market outcomes. And then you have the settlement, which, as you mentioned, seems to be at least addressing the right things. What were those kind highlevel goals of the settlement?
DAVID O. FISHER: Yeah, that’s a good question. So, when I’m thinking about this, sort of on the theoretical level, we use this term algorithmic cartel manager. And that’s a term that we’ve sort of taken from the existing literature. And it’s a meaningful one because it ties this kind of conduct back to sort of long-standing, kind of old school Section 1 conspiracy cases. And I think that’s a good framework for looking at the terms of the settlement.
There’s two main functions of an algorithmic cartel manager. And I think each of these functions are problematic on their own under Section 1 of the Sherman Act.
The first is the function of collecting competitor data in a way that allows competitors to benefit from each other’s nonpublic data. And then the second is developing and setting a common set of pricing rules that competitors can follow. And so, when we look at the settlement, we can see that all of the terms are aimed at one or the other of those two sort of problematic ways that algorithmic cartel managers function.
So, just to sort of put the terms in buckets, I think it’s easy to think of them as sort of three different buckets that the terms can fall into.
The first is the ways that the settlement limits RealPage’s ability to make pricing recommendations to one landlord using the nonpublic information of another landlord.
The second bucket are limits on the kinds of nonpublic data that RealPage can use to train its pricing algorithm, to train its models that it uses when it’s developing its algorithm.
And the third is restrictions on the features of RealPage’s software that are designed to give users more control over the final pricing and output decisions.
So, I think to just sort of take a step back, we can look at all the terms in those three buckets. And yeah, I mean, the commentary sort of digs into the specifics of whether those limitations are meaningful and we can dive into those, but that’s sort of the three main areas.
TEDDY DOWNEY: Yeah, because really what you dig into is how these terms could be interpreted in a way that’s in the spirit of the settlement, which is dealing with this pricing collusion problem, or sort of just actually creating a loophole for RealPage to just continue to do what it does. And obviously, that’s where the rubber meets the road. If it’s comprehensively making RealPage change their business model and do something different that is sort of not going to raise prices, but actually going to make the real estate rental market more efficient, when there’s high demand, prices go up, when there’s low demand, prices go down, more price discovery, not collusion oriented. Obviously, that would be the goal of the settlement if we’re taking this whole thing on good faith.
Let’s just go through all the key terms that you think need to be defined more specifically so that there’s no ambiguity in terms of how RealPage interprets this. Because we know if it’s left to them with not enough oversight, they’re going to take the most generous interpretation they can, I would imagine.
DAVID O. FISHER: Yeah, so I think there’s one key ambiguity that we identify, and it’s kind of a textual argument. But what it comes down to is the scope of the limitation on the data that RealPage can use in making the price recommendation.
So, this is the first bucket. And just to sort of set the groundwork here, these terms are focused on data that are used in what’s called runtime operation. That’s the term that’s used in the settlement agreement. And that’s distinct from training. And to understand that distinction, I think it’s helpful to talk in terms of ChatGPT as an example, because I think that’s a technology that we all have some degree of interaction with at this point.
So, if you think of ChatGPT, it uses an algorithm that’s trained on a massive amount of data during what’s called the training phase. And that’s before you or I, as the users, ever come in contact with it.
And then in runtime operation, that’s when I as a user will put in a question for ChatGPT to answer. And then it uses my question as an input and sort of generates a response. That’s what’s referred to as runtime. And it’s sort of distinct from the training, which goes on kind of in the background away from the users.
And so, I think the ambiguity arises in the scope of the data that RealPage is allowed to use under the proposed settlement agreement in that runtime operation. And it revolves around the term subject property data. And the settlement agreement limits RealPage to using only subject property data in the runtime operation.
And that term subject property data is defined as a property kind of simply, and maybe a little circularly a property to which RealPage makes pricing recommendations, right? And so, I think there’s an ambiguity there. Because subject property can refer to only that one property that the algorithm is making a price recommendation for in any given one runtime operation. Or it could refer to any property that RealPage gives pricing recommendations to more broadly.
And there’s a big difference there, right? Because if it means any property to which RealPage gives pricing recommendations, then the term would appear to allow RealPage to use all of its licensees’ non-public data in making a pricing recommendation to any one licensee. And that’s only a slight modification of RealPage’s business model.
It would prevent RealPage from using the non-public data of other entities that use other products, for example, its benchmarking products, which is also a source of non-public data. But it wouldn’t prevent the sort of keepitin-the-hub information exchange that a lot of people have identified as problematic.
TEDDY DOWNEY: And also, it’s already pretty ubiquitous in a lot of places. So, it wouldn’t help matters if you use any RealPage customers’ data. That can be 100 percent of a market.
DAVID O. FISHER: According to the complaint, there’s markets where 80, 90 percent of the market is using RealPage’s software. So, it’s not much of a restriction, if that’s what the restriction means.
Of course, DOJ has been very clear in its competitive impact statement, and other documents when it released this settlement, to say that RealPage is only allowed to make pricing recommendations based on the non-public data of the licensee to which it’s currently giving price recommendations. And, of course, that more restrictive meaning would be much more consistent with the goals of the settlement agreement, as we’ve just talked about.
But as we write in our commentary, we think this sort of alternate meaning is also plausible. And I think, as you’ve alluded to, we’ve seen disputes like this arise in the past. We talk about the Live Nation/Ticketmaster situation, where the government approved the merger of these two parties with a consent decree. And the consent decree contained a term that said I think it was that the combined entity couldn’t threaten venues by saying that if they went with another ticket sales provider, they would lose shows provided by the combined entity.
And there was a seemingly clear term preventing that kind of threat. But the government got bogged down in a years-long dispute over what that meant. And Ticketmaster took the position that it was able to tell venues they might lose shows if they went with another contractor, but that didn’t amount to a threat under the actual words of the agreement.
And so, that just serves as an example of the fact that, down the road, when other lawyers, not the ones who negotiated this actual settlement, are arguing before other courts about the meaning of these terms, they’re incentivized to go with the interpretation that’s most favorable to the defendant. And I think our commentary is aimed at making sure that the DOJ is not assuming too much about what these words mean on their face and making sure that the interpretation that we think it’s clear that DOJ means, which is the more limiting one, is the one that gets adopted.
TEDDY DOWNEY: And also, these monitors they appoint often aren’t like the strictest oversight bodies. My experience with them is they don’t really lay down the hammer on the defendant on these types of disputes.
DAVID O. FISHER: I think more importantly than that is that the monitor is only bound to the terms of the agreement and the terms are what the terms are, right? And so, I think David Fisher sitting here as a lawyer sort of wondering about which interpretation is the best one here, that’s the kind of opening that I think any contract lawyer would worry about needing to be interpreted later in the future.
And there’s all sorts of complicated arguments that you could make about the spirit of the agreement or what the parties intended at the time and that sort of thing, but that’s not the kind of dispute that DOJ wants to get into down the road. So, yeah, that’s one of the key recommendations in our comment is that the DOJ clarify those terms.
TEDDY DOWNEY: And what were some of the other recommendations you made where DOJ could strengthen the terms of the settlement?
DAVID O. FISHER: Yeah, and this is still on the point of runtime data, right? And beyond the sort of ambiguity that I just mentioned, I think the issue of user inputted data is potentially a problem area from my perspective. And so, this is data, as defined under the terms of the proposed settlement agreement, that licensees collect on their competitors and then are able under the proposed settlement to input manually into RealPage’s software, and which RealPage can use during runtime to make pricing recommendations to that landlord.
Which just sort of on its own strikes you as a little strange. And I think it’s even stranger in the context of the markets that we’re talking about. I mean, DOJ and the states describe in the complaint concentrated markets around the country, local real estate markets, in which only two or three landlords may own all of the multifamily housing units in that given market. And in which information sharing between those players is regularly occurring. They’re regularly getting together and talking about pricing strategy and sharing non-public information or competitively sensitive information with each other.
And just sort of, again, looking from my outsider perspective at the agreement, I don’t see anything in the proposed settlement agreement which prevents RealPage from being used to facilitate a price-fixing arrangement that the parties themselves are able to carry out. They exchange the information with each other. They feed it into RealPage’s algorithm. And it’s essentially being used as a highly sophisticated calculator to tell them, based off of that competitor information, how they should be pricing.
Of course, that kind of an information exchange can itself be illegal under Section 1. But it’s not clear how DOJ would be able to police all of these local markets and the ways that the local competitors are meeting with each other. And it does seem odd that such a settlement would allow RealPage to sort of serve that facilitating function.
I did sort of have a chance to look at other settlement terms. DOJ has also settled claims against Greystar and some of the other larger property management companies that are also named in the complaint. And those seem to have terms which would prohibit the property owners themselves from collecting that kind of data and using it for pricing purposes. But of course, DOJ hasn’t sued every landlord out there, nor reasonably could they. And it seems to me on the face of the agreement that these other landlords that use the product would continue to be able to do this.
TEDDY DOWNEY: It’s almost like you’d be better off not addressing that at all because you’re sort of allowing for something that seems like it should be illegal, right? You’re like kind of explicitly exempting something. It seems pretty unusual, to be honest, to your point.
But just going back to the ChatGPT analogy, it’d be you’re getting your own sensitive instead of RealPage is banned from putting this sensitive data in, in an aggregate way. But if you want to go out and get it, you can just input it yourself and then ask the ChatGPT. It’s like getting your own data and plugging it into ChatGPT and saying what should the price be? You’re just doing that with RealPage. Is that a fair way to kind of characterize it, going back to the ChatGPT analogy?
DAVID O. FISHER: Yeah, I think that’s right. It’s sort of putting the data collection aspect of it in the hands of the actual competitors and using that common algorithm as a facilitator more for crunching the numbers and providing the optimal output based off of that sort of already collusive data. Like an accessory, if you will, if you want to use the terms of a criminal law.
But I’ll say again, as I have with these other terms, there may be a perfectly reasonable explanation for why that’s included. There may be sort of data-based, market power-based arguments why this level of competitor information doesn’t rise to the sort of degree of concern that will give rise to price changes within a given market. But as I’ve said before, if we’re taking all of the allegations in the complaint as true, it’s hard to see why in an oligopoly market with only a couple of players that are exchanging significant information with each other how that wouldn’t be a problem. And I think that’s something that we’re hopeful that the DOJ will elucidate in its responses.
TEDDY DOWNEY: And we’ve talked a lot about how the buckets where they’re using data and how they use data. What about just serving as the sort of conduit for price increases? I know that comes up in your recommendations as well. I think that sort of stands out to me as really problematic in terms of one thing I’ve seen from people who want to do more price fixing is they try to say, hey look, the only thing that’s illegal is if we get in a smoke-filled room and have an agreement, right?
And that’s what they’re doing with non-public data. They’re saying, hey, non-public data is the only thing you should be worried about, you know? And we could still just do this as a software provider. As long as we’re just not doing anything with non-public data, that means we’re not doing anything shady, right? That’s the digital version of being in a smoke-filled room. But obviously, just serving as a centralized price recommendation allows for potential just consistent just, hey, we’re just going to recommend a higher price all the time and you’ll make more money that way.
So, what are your recommendations around things sort of outside how they input non-public data for the, when it comes to this?
DAVID O. FISHER: Yeah, and I think just to contextualize, I think what we’re getting at here is sort of the second function of an algorithmic cartel manager, which is not the data collection and use. It’s the developing and setting common pricing rules. And the Supreme Court has been really clear on this, that Section 1 is violated when agreements disrupt the independent centers of economic decision-making that competition requires.
So, in order for competition to work and promote the best products at the lowest price, competitors need to be making their own decisions on pricing and output. And agreements are illegal under Section 1 when they make those pricing decisions, joint decisions, rather than individual ones. And so, I think that’s sort of what you’re getting at when you’re saying that notwithstanding the data, it’s the adoption of common pricing rules that is itself problematic under Section 1.
So, the way that the settlement agreement deals with that is to require software features that give landlords more control over the actual pricing and output decisions. And one of these features has to do with what’s called guardrails, which are the limits on the amount that rents can go up or down over a given term. The proposed settlement requires users to have more control over those guardrails. For example, that prices can shift downward as often and to the same extent they shift upward, if that’s what the market demands.
And I think that those provisions are sort of aimed at making sure you can get some of the same potential benefits of pricing algorithms with respect to responding quickly to market conditions while limiting its tendency to drive prices upwards.
And then another key element here is that the proposed settlement prohibits RealPage from requiring landlords to accept its recommendations or pressuring landlords to accept its recommendations or making it more difficult for landlords to decline its recommendations. And that’s aimed squarely at this function of the cartel manager.
A lot of people say that there’s advocates out there who argue — that there’s not a Section 1 problem as long as the competitors maintain the final decision about how to price the unit. And that if the recommendation is just a recommendation, that’s not an illegal agreement for purposes of Section 1.
At the same time, though, the data shows that shared pricing algorithms tend to drive prices up even when users make the ultimate call on pricing. And I actually think the complaint is very good at explaining this and explaining it in some detail.
So, even if a landlord doesn’t accept the price recommendation, they’re still influenced by it in a way that affects the ultimate price. So, let’s say, for example, that the software recommends to every competitor in a given market that they increase their prices by 20 percent. Even if all the landlords don’t accept that 20 percent price increase, they’re still influenced by it in a way that may cause them to increase their prices by a different, smaller amount. And you’ll see market prices go up by, say, 10 percent. And that increase in price is something that you wouldn’t have seen under normal competitive conditions.
And that coordination of pricing decisions can still be anti-competitive even if the actual acceptance rate of the algorithm’s pricing recommendations is actually quite low. And that’s made very clear in the complaint. I think the parties did a great job of claiming that.
So, that raises the question then, if the only way that we’re addressing this sort of pricing-rules aspect of an algorithmic cartel manager is to increase users’ ability to set their own prices, how is it that those provisions are actually going to keep prices from going up in practice?
And I think that’s something that we’ve asked DOJ to sort of respond to. But I think there’s a more important point here, which is that the precise dynamics of the way that this functions are still up for some debate. And there’s not a lot of data to work with here.
And so, I think the chief thing that we recommend in our commentary is that if and when the settlement is approved under whatever terms are finalized, it’s critical for the DOJ to continue to monitor prices in these markets. Because if they’re right and these sort of data restrictions and the sort of maintaining the freedom to make your own pricing decisions are sufficient to keep the software from having an anti-competitive effect on prices, we should see prices change in the future. So long as RealPage is complying with the terms of the agreement, we should see prices start to stabilize, even go down in some market conditions.
But if we’re not seeing a meaningful change from before the settlement to after in the markets that have been alleged to be effected, I think there’s some serious questions about whether these terms are enough. And then that’s where enforcers, policymakers, start to build from that and make better terms. But I think there’s a lot of, like I said, sort of a wait to see element with that.
TEDDY DOWNEY: Yeah, they can look at pricing, but also marketing, right? If RealPage really fundamentally has to change their business model to be more about actual efficiency in the market with prices going up and down, they’re going to have to stop saying, hey, we’re going to guarantee the ways that you’re going to make more money, right? You’re going to make more money by lowering your vacancy levels, right? Sort of the traditional way that you’ve seen the real estate market function. So, we should see more indicia of the market kind of changing as well, correct? I mean, if these rules are working as intended.
DAVID O. FISHER: Yeah. And so, you’re talking about occupancy rates, right? And this goes back to something that you said at the beginning too. I think when we’re looking at tacit agreements to collude, which aren’t sort of written down, it’s sort of when, in this case, the landlords that are all using the software, never really talk to each other about how the fact they’re all using it, but there’s sort of an understanding they are, those agreements to collude are illegal under Section 1, and they always have been, but they’re very difficult to prove. And courts rely on circumstantial evidence to infer that an illegal tacit agreement exists.
One of the key pieces of circumstantial evidence that is critical in these cases is if the individual competitor is acting in a way that’s inconsistent with its own interest and only consistent with the common interest. And reducing output is a key fact that plaintiffs want to plead in these cases. Because if I’m acting on my own as a competitor and I’m trying to get more draw customers away, it makes no sense to leave apartments vacant.
But if I know that my competitors are going to keep the same price, or at least not undercut me, keeping vacant apartments can keep the market-wide price high, and we both benefit from that. And I think the vacancy issue is a critical one when we’re looking at whether or not there’s a potentially illegal tacit agreement there.
Just by comparison, there’s another set of cases that arise out of the casino hotel context. There’s sort of two sets of litigation there on the Las Vegas strip and in Atlantic City. And there the plaintiffs pled that the incentives of the business model is typically heads in beds. Because whatever money you lose at renting a hotel room for under the market price, you make back on the casino floor. And you want to fill up those rooms at whatever price you can, give them away if you have to, because you’ll still make money in the casino. When you sort of see those casinos that are using the same pricing algorithm depart from that and leave rooms empty in order to keep the price high, that’s a classic indicator of an illegal agreement here. And yeah, I think that’s a key fact in a lot of these cases.
TEDDY DOWNEY: So, while we’re monitoring all of these provisions, let’s assume DOJ does that obviously, there’s still just the risk of doing an agreement like this with the potential for all these loopholes. If you’re the states, let’s say they ultimately prevail, what kind of remedy would they ask for that wouldn’t get them stuck in this seeming brutal hellscape of monitoring and trying to get a monopolist to obey the law when it’s not in their interest?
DAVID O. FISHER: Yeah, I mean, that’s the multi-million dollar question, isn’t it? I’ll say this, it’s not an easy problem to solve, partly because it hasn’t been done before. The reason why this settlement, proposed settlement agreement, is so interesting is because none of, we’ve seen a flurry of cases on the whole government pricing issue, but none of them have developed past the motion to dismiss phase yet, right? So, we haven’t gotten a summary judgment record or an opinion on the merits, let alone a remedy. And so, it’s sort of not clear what’s needed here.
I think that the data suggests that there’s a potential competition problem when the same thirdparty pricing software is making pricing and output recommendations to multiple competitors in the same market, which combined have market power. What that means about a suitable settlement term, I’m not sure. It’s very possible that if we sort of address these issues that we’ve identified in our commentary, that the settlement gets approved and we see a miraculous change in the way that the markets are working. And we find that this approach of addressing data and addressing user ability to accept or reject pricings is an effective one. But again, we have to wait and see.
TEDDY DOWNEY: I mean, is there any kind of I mean, normally I would say, oh, divest this or structural remedy, but in this it’s getting to the heart of the business model, right? So, what is the structural remedy that you would even seek? I guess we’ll just have to wait and see what kind of more draconian measures could be implemented than what we’re seeing here. But I’m certainly interested to see that as well.
We’ve got some listener questions here. Do you have anything else you want to add on what you want to see, David, before we get to the listener questions actually from DOJ in terms of fixing the remedy or clarifying it?
DAVID O. FISHER: Yeah, I encourage your viewers to read our commentary, which is posted on the AAI website. There’s a whole other element here, which we just didn’t have time to discuss, which has to do with the limitations on the non-public data that RealPage can use to train its algorithm. Just quickly, RealPage can only use non-public data to train its pricing models if that data is 12 months old and cannot be identified with geographic specificity more than statewide. So, it has to be a year old and it’s got to be state level. It can’t be more local than that.
I think that raises a whole host of questions about, to the extent this software is supposed to make prices more reactive to real world market prices, like wouldn’t you expect those terms to make the algorithm worse at what it says it is good to do? And if that’s the case, then why has RealPage sort of agreed to those terms? And we get into the commentary more specifically about what the workarounds might possibly be in terms of using public data to supplement that old training data and other potential sort of things that RealPage may be thinking.
But we can’t make any firm conclusions without being privy to RealPage’s sort of business thinking and sort of DOJ’s deliberations. So, yeah, that’s another area where we hope that DOJ’s responses will provide more clarity.
TEDDY DOWNEY: It also just brings up I think you mentioned Zillow and there are other public data feeds where you can get a lot of transparency around rental pricing from scraping the web. And now you have AAI and Zillow. And it just seems like Whac-A-Mole them all, right? Like, oh, you take away this. And then they’re like, oh, yeah, sure. Yeah, sure, 12-month delay? Sure. Yeah, that’s okay. Because they just turn around and do a deal with Zillow and the data is just as good, right? Or what have you. And it’s like public data, because like anyone can buy it. Or they’re scraping it, what have you. So, then you’re in a 10-year fight over the definition of public versus non-public.
DAVID O. FISHER: Yeah, I mean, I’ll say that we’re living in a new era where machine learning is making data much more valuable and also much more accessible. And in that era, I think this sort of red line distinction that a lot of folks have tried to make between public and non-public data is kind of losing its significance. And what we should really be looking at here are the numbers. Like, how are market prices being affected?
And that continues to be the core focus, of course, under Section 1, and really the bottom line for consumers here, and not sort of getting bogged down in specifics about what kind of data is being exchanged.
TEDDY DOWNEY: This is such an important issue. I want to get to our listener questions now. First listener question is will courts look to the DOJ settlement when deciding other RealPage cases, particularly RealPage’s First Amendment suits against states and cities? As a media company myself, I’m deeply offended that they would entertain such an argument in the first place. But go ahead, David, what are your thoughts here? Will they look at this when they’re looking at other RealPage cases?
DAVID O. FISHER: Yeah, I mean, the short answer is absolutely. What we’re getting to here, and the reason why all eyes are on this proposed settlement, I think, is that this is the first definitive statement from really the leading enforcement experts here on this issue, the DOJ, listing out a set of terms which it deems suitable to address the harm alleged in this case.
That gets to sort of the fundamental question about what it is in this business model that is problematic and what can potentially be saved. And I think that’s a core question in all of the litigation against RealPage about algorithmic price fixing. So, the short answer is yes.
With respect to the New York state case, just for context, New York has passed a ban on there’s a lot of proposed legislation out there. I can’t recall if it’s a blanket ban or more of a targeted ban on the use of algorithmic pricing in real estate, which RealPage has challenged on First Amendment grounds. I think RealPage’s First Amendment challenge, specifically presents a whole other set of issues related to the scope of the First Amendment and how it applies to law enforcement, which is outside the scope of my expertise.
But in terms of whether and how RealPage’s conduct violates Section 1 of the Sherman Act, yeah, this agreement is going to be part of the courts’ consideration when thinking about those cases.
TEDDY DOWNEY: Yeah, again, as a media company, other companies using the First Amendment to protect themselves from a state wanting to put rules on their society seems gross, gross, gross maneuver, in my opinion.
DAVID O. FISHER: I’ll let you editorialize on that. But I do want to just add one fine point to that, what I just said, which is all the more reason why DOJ should articulate clearly the terms that it has adopted and the connection they have to the alleged harm.
Because I think there’s an important public record to be made here that will be used as a benchmark in the future, that courts will be using when they craft remedies, that litigants will be using when they come up with their own settlement terms or request remedies in cases, and that businesses are going to be using to determine which of their business practices are acceptable. And so, I think all the more reason why we need transparency here.
TEDDY DOWNEY: Yeah, yeah. I mean, there’s just so much here. Like, why aren’t they fighting DOJ on First Amendment challenges to put these rules on them? But here we go. How do the effects of algorithmic pricing vary between supply constraint and open housing markets?
DAVID O. FISHER: That’s a good question. I think it’s a market question, right? And I think I can sort of respond generally. I think the crux of these technologies is they’re really turbocharging competitors’ ability to collect and use information on market forces, supply and demand. And public housing, the availability of public housing in a market, the portion of the population which is eligible for that housing, all of those things are going to affect supply and demand in a given market.
So, I think that’s a great example of when we talk about, oh, RealPage could be supplementing non-public data with publicly available data that it can use to train its algorithm. That’s exactly the kind of information we’re talking about, information about the number of units available in a community and the number of people that are likely to apply for those units. The availability of public housing is going to affect all of that greatly. And it’s one of the things that makes these tools attractive is because there’s a lot at play here. And having the computing power to be able to make meaning out of them is really valuable. So, I think of it as one more market force that RealPage is able to sell to its clients that it can help them account for.
TEDDY DOWNEY: And how is the DOJ likely to proceed on its crusade against algorithmic pricing? Will the settlement guide new DOJ policy statements?
DAVID O. FISHER: I have the same question. Yeah, I mean, look, as somebody who’s been watching DOJ and paying attention in these kinds of cases, it’s not a stretch to say that this agreement is going to give important insights into the way the government is thinking about these cases, both in terms of the conduct that they view as problematic and they’re likely to go after in the future, and in terms of its view of the risks of litigation in these cases and the reasonable use of government resources.
Do I know beyond that what’s going on behind the curtain? No. I suspect this will continue to be an issue. And I’ll just say that DOJ is not the only player here. I think the states are critical in the litigation landscape here. Private plaintiff suits are critical in the litigation landscape here. And while DOJ’s conduct, actions, are an important sort of bellwether for how those other players are thinking about things, it’s not the be-all and end-all. And that’s why I think all eyes are sort of now on the states and on the private plaintiffs who are continuing these cases.
TEDDY DOWNEY: Well, another question here. If the court denies class certification in a RealPage multidistrict litigation, do you think this crusade it sounds like this person doesn’t like this algorithmic pricing stuff do you think this crusade against algorithmic pricing dies because the lead case has now fallen apart and we don’t have precedent on the merits of these cases?
DAVID O. FISHER: The word crusade has been used twice now. I don’t know if it’s the same person.
TEDDY DOWNEY: This might be the same person, just wanted very much wanted to get a point across here, but it’s interesting.
DAVID O. FISHER: Yeah, so without endorsing that terminology, no. I think my short answer is no. I mean, the MDL decision is often meaningful in terms of that’s what they’re asking about, right? Whether the court’s going to sort of approve consolidation of these cases against RealPage or allow them to continue in separate venues around the country.
TEDDY DOWNEY: Yes.
DAVID O. FISHER: No, I don’t think that sort of litigation lives or dies on that decision. Often, it’s the defendants in the position of arguing in favor of consolidation because it makes it easier to defend cases in one jurisdiction rather than in multiple.
So, I do think that at AAI, we’re looking more at the upcoming anticipated Third Circuit opinion in Cornish v. Caesars, which is going to be a critical opinion, I think, in terms of instructing lower courts on how to deal with allegations of algorithmic price fixing. We’ve already got an opinion out of the Ninth Circuit in Gibson v. Cendyn on this issue that we’ve said at AAI publicly we think came out the wrong way. We’re hopeful that the Third Circuit will get it right. But I think those opinions are really going to be what district courts are looking to. And whether or not these cases against RealPage are heard in one venue is sort of an ancillary issue.
TEDDY DOWNEY: We’ve got another question here following up. The question was about class certification, not consolidation.
DAVID O. FISHER: Yeah, I mean, class certification is undeniably important in terms of the economies of scale, right? And that’s why cases often, you know, why it’s such a critical issue in litigation.
But again, I think that we’re going to continue to see challenges to this business model under any form. And the law is going to continue to develop here because it has to, right? This is a business model that’s increasingly pervasive in our economy. It’s showing up in a wide array of markets and it presents some serious issues on both sides. And so, I think the courts are going to have to deal with it. Legislators are already beginning to try to deal with it. So, I don’t think it’s going to go away anytime soon.
TEDDY DOWNEY: Just to put a focus on that point, from a legislative standpoint, even if the courts do nothing and bless this technology at some level or create these sort of broad get out of jail free arguments, you’re still going I mean, it’s not like the issue of affordability of your housing is going away. In fact, most newspapers are saying this is the number one political issue. The White House says it. Congress, people running for election are saying it is this.
Do you agree with that? That this is going to continue to be a focus and is in many ways sort of epitomizes this. Are you going to do anything about high prices? The biggest cost for everyone is their housing costs. How could the focus of colluding on increasing your rent or alternatively, just the companies that do have market power, not just the software, but the underlying consolidated real estate owners. I mean, even the President has said he wants to ban institutional ownership of rental housing. I mean, it’s not like it’s hard to see how this could go away, even if you get one bad court decision.
DAVID O. FISHER: Yeah. Look, reasonable minds can disagree about the extent to which algorithmic pricing is the cause of the affordability crisis in housing. And there’s a lot of pressure on policymakers to address that issue, that bigger issue of affordability.
But just sort of taking the algorithmic pricing part of it, we’ve already seen a slew of legislation coming from cities and states and even bills floated in Congress. At the federal level, Congress is specifically targeted at this issue. And I think that’s indicative of the sort of seriousness with which the voters are taking this issue. So, yeah, I think the proof is in the pudding and we’re seeing all sorts of folks from courts to legislatures, to legislators, to policymakers, taking this issue seriously.
TEDDY DOWNEY: Well, David, you got some great questions. This was a truly interesting conversation. I can’t thank you enough for doing it. It’s great to see you. I’m so happy you’re at AAI. You had a great stint at the FTC. You’re doing great work. This was tons of fun. Thank you so much for doing it. Thank you, David.
DAVID O. FISHER: Thanks so much, Teddy. Always a pleasure.
TEDDY DOWNEY: Thank you to everyone for joining us today. And this concludes the call. Bye-bye.
