According to The New York Times (23 August), The Justice Department “filed an antitrust lawsuit on Friday against the real estate software company RealPage, alleging its software enabled landlords to collude to raise rents across the United States.” I am not an expert in law or anti-trust, but there is another (systemic risk management) angle to this story that the NYT missed, and worth spelling out.
Here’s how the Times summarizes the case:
RealPage’s software, YieldStar, gathers confidential real estate information and is at the heart of the government’s concerns. Landlords, who pay to use the software, share information about rents and occupancy rates that is otherwise confidential. Based on that data, an algorithm generates suggestions for what landlords should charge renters, and those figures are often higher than they would be in a competitive market, according to allegations in the legal complaint. By Danielle Kaye, Lauren Hirsch and David McCabe
There’s more detail in a piece (here) the Times did earlier in the year (July 19, 2024) written by Danielle Kaye. And for a lot more background see this piece in Propublica (here, October 15, 2022). One wonders why the Times waited so long until there was government prosecution to report on this topic.
Before I get to my interest in this story, the coverage highlights two important political angles: (i) this may be a contributing cause to rent inflation because the software allows landlords, which are fairly large companies that use the software, to collude; (ii) there is an interesting issue how this case fits under existing anti-trust law because the collision is kind of indirect mediated, in part, tacitly through a third party. The company itself brags they have “purposely built” their platform “to be legally compliant.” I leave the first issue to economists and the second to lawyers.
There is another important angle that I haven’t seen mentioned yet. It’s not so long ago that we had a Great Recession caused by a financial crisis after a Wall Street meltdown. One of the root causes of that crisis were unexpected correlations in assets that risk models assumed to be uncorrelated. In particular, back then a key assumption was, for example, that local real estate markets are uncorrelated with the stock market and that local housing markets are relative uncorrelated. This is still a quite common conventional wisdom; a fairly random search, immediately found this explanation of uncorrelated asset classes at a website directed at the public at Kubera (a financial services platform): “That coupled with the long-term nature of leases and mortgages make the value of real estate assets less reactive to economic news than traditional assets.” (I have no relationship at all to Kubera and don’t use their platform!)
It’s pretty clear that YieldStar’s algorithm will induce (stronger) correlations where there had been weaker ones or almost none before. If the government is right about the effects of this software then this will be most evidently so within local markets. Ultimately, this very bad news for smaller and regional banks because it will make their portfolios even more sources of concentrated risk. As the Bank for International Settlements (BIS) puts it (December 2023), when it reflects on macroprudential risk of mortgages to banks “correlated defaults can result in large losses.” (p. 2) I like to call BIS the ‘central bankers’ bank.’
But the real systemic risk here is that the algorithm will also (accidentally) induce correlations among different housing (and so mortgage) markets. Undoubtedly, that’s hard to prove and may well involve the details of the algorithm. But the crucial issue is that the algorithm itself has become a common cause to all outputs it generates in individual markets. It becomes a systematic, structural background factor in all the markets where its users have a dominant market share. So, for example, while it will spit out different rent-levels in different housing markets, it may reduce the variance of rent-levels in each housing market, and it will do shaped by a common cause (the software).
Financial regulators have not been very worried about Yieldstar (and products like it) presumably because it actually reduces likelihood of default of big landlords and so makes mortgages more secure than expected, and even safer going forward in the medium term. By contrast, the central bankers’ bankers are, in fact, much more worried about price competition among mortgage lenders “Excessive price competition among lenders for market share can result in underpriced risks, especially as mortgages are relatively homogeneous products.” (2) This reminds us that the profitability of the banking sector is always in the background of central bankers.
But this is why I am a little bit worried. Industry leading algorithms in all kinds of businesses will, as a foreseeable side-effect, induce correlations within and among asset classes that the historical data — and the macroprudential risk models (and not to mention the regulations for capital holdings of banks) — had suggested were uncorrelated. In much of the economy this need not generate any new systemic or macro-prudential risks. But as we learn time and again, local housing markets are a source of fragility because of the maturity mismatch on bank balance sheets that long-term mortgages generate against short term liabilities (deposits and current accounts).
This algorithmically induced macroprudential risk is already undoubtedly a fact of life. I hope this case isn’t settled out of court. It would be good to have a public record on how Yieldstar’s algorithm really works. For, armed with that and all the price data of housing markets (which are relatively well understood by applied economists) it would allow us (well smarter people than me) to begin to estimate how much of a common cause it really is. At the very least it may alert central bank regulators and bankers that the economy is changing in potentially dangerous ways.*
- An earlier, slightly modified version of this post was published (here) at Digressionsnimpressions.
*Research on this post was sponsored by the Dutch Research Council’s Grant 406.18.FT.014
{ 15 comments }
Trader Joe 08.26.24 at 11:19 am
A few observations:
I agree fundamentally about the potential for algorithmically caused systemic risk, what I would question in this instance is whether these systems are pointed at the right risk. Housing failure is usually the result of excess debt or inadequate occupancy. These systems do nothing about the former and little about the latter.
I’d also assert that landlords have never lacked for knowledge about what rents were being charged in their area. In the old days we did this with telephones and newspaper listings. While RealPage along with dozens of others from CoStar, Apartments.com and any number of local rental apps have speeded up the process, the ability to collude based on this information is not as easy as you expect since most rental markets are very fragmented.
I’ll be curious how justice makes their case. I don’t doubt such software can be collusive, I question whether any actual collusion can be found in markets where price discovery is inherently easy.
When three gas stations on a corner all charge the same price for unleaded its not so much collusion as that they all have 3 foot tall signs listing their price. The same is true for local appartment listings, the signs just can’t be seen at 60mph.
Eric Schliesser 08.26.24 at 11:35 am
Hi Trader Joe,
Thank you for your well-informed skepticism.
First, I doubt that it is information that is creating collusion, but rather that the software functions as a kind of tacit contract not to compete on price.
Second, my claim is not that will lead to housing/mortgage failure. Rather, my claim is that it is creating correlated risks, where the market and regulators assume uncorrelated assets. So, that when failure arrives (for the usual cyclical reasons) it will be more costly.
Eric
OneEyedMan 08.26.24 at 12:05 pm
The earliest example I know of algorithmically induced macroprudential risk is “portfolio insurance”. Portfolio insurance was a cause of the 1987 stock market crash.
J, not that one 08.26.24 at 5:56 pm
After the housing crash of 2008, it became public that banks were using ZIP codes to set interest rates and assessments of ability to pay. The use of statistics going back two hundred years or so could be called algorithmic but there seems to be a claim that this is something new, apparently based on (1) increased granularity and use of unintuitive correlations acquired through data mining, and (2) the ubiquity of software that supposedly is intended to do something else, or that comes to be considered mandatory in an industry.
en passant 08.26.24 at 7:25 pm
Any reasoned consideration on sytemic price setting would be out of my league, but I can maybe contribute something which will interest the more competent commenters.
One could relate the case (market wide price calculating software) to another innovation of antitrust in Norway, where the three big grocery chains who are basically an oligopole where just slammed with a really big fine. I will try to summarise the case as well as I understand it, so beware, but it goes more or less like that.:
Basically they were found guilty of creating and using automated exchange of their price listings, which allowed them to react fast to any move of the concurrents, be it at the local or national level.
Which was allowing them to drive up the price, according to the findings.
In short, the old system of observing the retail price by the concurrent which would be a precondition of the free market efficiency was automated cooperatively and resulted according to the findings in…. being against an efficient market, or at least penalising the consumers.
Trader Joe 08.26.24 at 7:42 pm
@2 Eric
As to the second point on asset correlation, I guess my take would be that the same sorts of algorithms are being created by market participants to try to detect these sort of relationships – certainly to a much greater degree than was true in 2007-09.
We had a whiff of this starting in 2022 when signs emerged that office occupancy was unlikely to rebound to pre-Covid levels and there was widespread selling of CMBS and similar securities (both equities and bonds). That selloff if anything (so far) has been an over-reaction and at least initially failed to distinquish between office properties located in hollowed out central business districts relative to suburban campuses.
That said, whether the market would be similarly astute in multi-family remains to be seen. No doubt, at some point, your assertion will get the proverbial real-time market test.
On the first point – that’s certainly Justice’s contention that the software provides some sort of ‘tacit contract’ What I would be interested to know in support of that point, is whether RealPage (or any comparable) actually has a high enough market share of participants for such a contract to have any heft. I suppose in some isolated places – say a small college town – they could have such penetration. In a good size metro area though I doubt they regularly have as much as 5-10% of the market (if that).
Alex SL 08.26.24 at 10:05 pm
Maybe I misunderstand and miss some subtle point here, but why would most assets, and in particular property markets, not at any rate be strongly correlated in any crisis worth worrying about? Could we have a serious recession where, say, single family homes see their prices crash, but office space appreciates strongly? I guess a scenario could be constructed, but it would read far-fetched to me. Gold comes to mind as a supposedly crisis-safe asset, but even there people will to some degree have to liquidate their gold when they run out of money in a recession; that is, after all, the point of having it, so that you can sell it when you really need money, and then if lots of people have to do that, its price will go down too.
What I really take from stories like these is how self-defeating libertarianism is, in the sense of those people who believe that a perfect free market will self-organise if the government is removed. Without a strong government regulating it, any market will immediately self-destruct, as, exactly like the landlords in this story, whichever side of the market that has the superior leverage will collude to undermine free competition. Many people believe that competition is good at a collective level, as applied to everybody else, but nobody wants its downsides to apply to them specifically, if they can somehow weasel out of it (and the same for meritocracy, equality before the law, and a variety of other ideals).
KT2 08.27.24 at 12:26 am
Eric @2 “So, that when failure arrives (for the usual cyclical reasons) it will be more costly.”
Zephyr Teachout seems to agree;
“Price-gouging laws also protect against volatility and instability. During the immediate aftermath of COVID, unchecked price increases made an already-bad inflation problem even worse, contributing to a dangerous spiral that harmed the macro economy as well as individual consumers.”
From “Sometimes You Just Have to Ignore the Economists” “Kamala Harris’s proposed price-gouging ban might irritate academics, but it makes sense to everyone else.”
By Zephyr Teachout
AUGUST 22, 2024
https://www.theatlantic.com/ideas/archive/2024/08/economists-kamala-harris-price-gouging/679547/
Housing won’t succumb to, Uber Air Bnb or traffic type surge pricing – my better angel whispers. Yet the food sellers are edging toward a distopian “live pricing” and Minority Report. Or the Tyrell Corporation of housing.
“Digital price tags can change the cost of groceries 6 times per minute
“Electronic shelf labels bring Uber-style dynamic pricing to stores like Walmart.”
BY MACK DEGEURIN
POSTED ON JUN 20, 2024
https://www.popsci.com/technology/electronic-price-tag-groceries/
MisterMr 08.27.24 at 9:31 am
@en passant 5
The problem is that there is a belief that an “efficient market” will drive profits (rents in this case) down to a minimum, but this isn’t really true.
Trader Joe 08.27.24 at 10:37 am
@7 Alex
You are correct that real-estate within a market is correlated to some extent because they all enjoy the conditions of the local economy. What its not generally correlated to is other asset prices ranging from gold and digital currency to equities. In a really harsh global financial crisis type market, you are correct there is reduced co-variance, but its still not correlated.
Perhaps the piece missing is asset duration. In apartment (for example) there’s about a 3 year lead time from when interest rates are low (2020-2022) to when projects triggered by those rates hit the market. This was elongated due to worker and material shortage in the Covid environment.
The upshot being, incumbent property owners had huge pricing power because A) general price and wage rises provided an umbrella, B) housing was in shortage relative to demand (a largely national phenomenon) and C) there was an air-pocked in new supply. Accordingly rental prices (in most markets) rose at rates above general inflation. Whether this was collusive or just responding to normal market cues will be the crux of the case at hand.
Appreciate that Justice has focused on this period because prices rose. There are any number of examples where the same “collusion” if it exists, would work in the other direction (see SF office rents presently). Indeed I rather suspect as significant supply comes into many markets over 2024-2026 and interest rates fall (hopefully) that apartment rent will be a deflationary factor or at least not a contributor to CPI.
Alex SL 08.27.24 at 9:35 pm
Trader Joe,
Thanks, good point regarding duration, that would work against all asset classes being too closely correlated. Still, the underlying question here is if collusion like that described in the OP increased correlation specifically in the property market, and it remains difficult for me to see how property prices wouldn’t be closely correlated, as you wrote yourself @1.
mw 08.28.24 at 12:41 am
I’d argue that the ‘algorithm’ aspect here is a bit of a distraction. The real problem is the pooling of non-public information from all the participating landlords allowing RealPage to spit out recommendations based on that information. The algorithm which it uses to process this information isn’t really all that relevant.
I remember there was a case some years ago where Ivy League universities had been colluding in ‘harmonizing’ financial aid to offer students who were applying to more than one school, and the courts, quite reasonably, forced and end to this practice. Actually, now I see that this case is only wrapping up just now.
In any case, pooling non-public data via RealPage should be no more permissible than it would be if the landlords had a secret direct data sharing conspiracy. That said, RealPage could probably modify their approach to use a given client’s private data in conjunction with only public data from other landlords and still provide similarly tailored recommendations, and I really wouldn’t see a legal problem in that case. I also doubt that RealPage played any really significant role in the rent increases in recent years (in comparison to inflation, high-interest rates, and the reduced availability of home for sale due to the mortgage lock-in effect). Yes, pooling of private data via 3rd parties (if that is what is happening) should be banned, but I doubt it will make much of a difference in rents.
KT2 08.28.24 at 2:35 am
“(ii) there is an interesting issue how this case fits under existing anti-trust law because the collision is kind of indirect mediated, in part, tacitly through a third party”
Steve Winn, founder of RealPage submitted to SEC on listed share transfer to private Thoma Bravo LP, his shares turn into an “equity interest” not beneficial or controller. The Law is 1934 & 1940!
“STATEMENT OF CHANGES IN BENEFICIAL OWNERSHIP
“Filed pursuant to Section 16(a) of the Securities Exchange Act of 1934
or Section 30(h) of the Investment Company Act of 1940
“Name and Address of Reporting Person* WINN STEPHEN T
“Issuer Name and Ticker or Trading Symbol RealPage, Inc.”
…
2. “The reporting person is the sole manager and president of Seren Capital Management, L.L.C., which is the general partner of the partnership that directly owns the reported securities. The reporting person disclaims beneficial ownership of the securities reported except to the extent of his pecuniary interest, and the inclusion of these securities in this report shall not be deemed an admission of beneficial ownership of all the reported securities for purposes of Section 16 or for any other purpose.”
via “By Seren Capital II, Ltd” listed as the now “7. Nature of Indirect Beneficial Ownership (Instr. 4)”, which the SEC has accepted.
sec dot gov /Archives /edgar/data /1286225/000120919121028378/xslF345X03 /doc4 dot xml
Oh, and Steve Winn is by my reading, but NOT the SEC’s reading, THE BENEFICIAL OWNER of Seren Capital II, Ltd, and its “Other Companies for Seren Capital Management, L.L.C.” “Seren Capital Management, L.L.C. is listed as an officer in twenty-one other companies.”
“Key People Who own Seren Capital Management, L.L.C.
Name Stephen T. Winn”
corporationwiki dot com /Texas /Carrollton/ seren-capital-management-llc/30272482 dot aspx
Which leads to the fall guy.
KT2 08.28.24 at 2:40 am
RealPage may be fined for hording the Maltese Falcon jewels – the algorithm, yet Gutman says it is only a 50 month sneeze.
RealPage has a potential, if convicted, of “Agreeing to fix prices is punishable with up to 10 years in prison and a $100 million fine”.
See Atlantic below
The US has about 44m rental units. Defenders RealPage and property managers seem exposed in the lawsuits of multi family building units (?) of about 2m MFB units.
$100m fine is just a pimple.
If 44m units rent raised $1 per month.
$100m / (44m x $1 a mth)
= approx 2.25 mths to recoup $100m fine.
2.25mths x 22 (44m/2m units of defenders RealPage and property managers)
RealPage can recoup fine at $1 rent increase per month in 50 months.
And I assume book $100m fine as cost of doing business loss against earnings. Ah, Delaware. Looks after us everytime.
###
Then we have the Maltese Falcon conundrum. Who is the fall guy? 10 years jail potentially. Gutman eventually opts for his “like a son” henchman Wilmer. Greed overides love. Who knew.
Gutman at RealPage is not the founder Steve Winn anymore. Thoma Bravo LP is now holding privately ‘the bag’ which is “AUM US$138 billion (2023) [4]” so say Wikipedia. So who will get Wilmer’s fall guy role for RealPage?
I’d suggest one the minions, or better Jeffery “… there’s way too much empathy going on here,” Roper. “The software is designed to counteract the downward pressure on prices caused by leasing agents having “too much empathy,” a YieldStar developer named Jeffrey Roper told ProPublica””.
Roper has been showing his empathy bypass for over 40 years… ‘including Alaska Airlines, all of which agreed to change how they used the technology. “At one point, federal agents removed a computer and documents from Roper’s office at the airline. He said he and other creators of the software weren’t aware of the antitrust implications. [Ignorance is no defence] “We all got called up before the Department of Justice in the early 1980s because we were colluding,” he said. “We had no idea.”
“The Origins of YieldStar
“RealPage hired Roper as its principal scientist in 2004 to improve software it had bought from Camden Property Trust, ”
“Such agents sometimes hesitated to push rents higher. Roper said they were often peers of the people they were renting to. “We said there’s way too much empathy going on here,” he said. “This is one of the reasons we wanted to get pricing off-site.” “This was just a ripe business,” with lots of money and lots of opportunities for technological improvement, Roper said.”
From “Rent Going Up? One Company’s Algorithm Could Be Why.” by Heather Vogell, ProPublica, with data analysis by Haru Coryne, ProPublica, and Ryan Little Oct. 15, 2022
Roper may yet play Wilmer.
###
‘UNITED STATES DISTRICT COURT
SOUTHERN DISTRICT OF CALIFORNIA
[5 humans] Individually and on
Behalf of All Others Similarly Situated,
Plaintiffs,
vs.
REALPAGE, INC.; GREYSTAR REAL
ESTATE PARTNERS, LLC; [… other entities]
Defendants.
Case No. “22CV1611 WQHMDD”
CLASS ACTION COMPLAINT”
Key phrases in above Class Action Complaint –
– “collusive price over volume”
– “adopted a philosophy of economic occupancy”
– “notwithstanding market conditions and tolerating any reduced physical occupancy that might engender.”
pg10, line 17: … “The Lessor Defendants’ Outsource Price and Supply Decisions to a Common Decision Maker—RealPage—Which Eliminated Competition”… Items 44 & 45.
In The Atlantic, “We’re Entering an AI Price-Fixing Dystopia AUGUST 10, 2024 “… the late Justice Antonin Scalia once called price-fixing the “supreme evil” of antitrust law. Agreeing to fix prices is punishable with up to 10 years in prison and a $100 million fine.”
Yet as “Luis Quintero says this does not fit the classic definition. Luis is an economist at Johns Hopkins University who studies real estate competition” says (NPR below);
“RealPage are essentially behaving as a cartel. But Luis Quintero says this does not fit the classic definition.” NPR
So searching the SEC for “RealPage” over the last 5 years returns 10,000 documents mentioning “RealPage”. We would need a real page ai to sort “RealPage” SEC filings!
NPR “The lawsuit that could shake up the rental market” JANUARY 11, 2024 says; … “Like, the plaintiffs allege that the landlords and RealPage are essentially behaving as a cartel. But Luis Quintero says this does not fit the classic definition. Luis is an economist at Johns Hopkins University who studies real estate competition.
“LUIS QUINTERO: So this is a challenging instance because there is no smoky backroom. But the algorithm in itself, it’s the smoky backroom. And it might be true that the company owners are not aware that they were colluding, or they’re not actively colluding, and therefore not satisfying the traditional definition of a cartel. But the impact on the market is the same. And so we have a situation where technology is changing faster than both our legislation and regulation and even our understanding of economic concepts.”
###
So the potential Wilmer – Jeffery “… there’s way too much empathy going on here,” Roper – will NOT be going to jail.
And the fine will be written off as cost of doing business.
At $1 a month.
Paid for by multi families.
Exceptional.
Trader Joe 08.30.24 at 11:33 am
@KT2
Good research. I think your conclusion is likely pretty close to the pin.
2M of 44M is miniscule market share – though certain sub-markets could be much higher. That would be Justice’s best hope is that they can get them on a sub-market having enough concentration to move the needle.
Comments on this entry are closed.