Lawsuit filed against rent-setting software RealPage
propublica.org
propublica.org
The result is all the companies in an industry are basically pooling their salary decisions and then carefully sticking close to what the others are paying. If the companies did this directly, it would 100% be wage-fixing. But somehow laundering it through an intermediate makes it legal.
It probably is illegal. This case with rents will be very interesting to watch.
In the insurance industry, a special law (McCarran-Ferguson Act) had to be passed to allow companies to pool loss data to set rates. Otherwise, it was seen as an anti-trust violation. The law was passed because the public interest in making sure insurance companies are solvent and can pay claims outweighed the anti-trust issue.
I think these tools are on very shaky legal ground.
This certainly seems like one of those cases worth watching.
I'm trying to think through this and wondering how companies should set salaries. I mean, I get your point about wage fixing, but I get my information about how much to ask for from shared storage that try to determine "market reference points". Is the idea that it should somehow be done in a vacuum? Seems like a chicken and the egg problem.
"Wage fixing occurs when companies in the same industry conspire or agree to suppress employee earnings below the the market rate."
According to "https://www.nka.com/practice-areas/employee-rights/wage-fixi...". So, as long as they are at market rate and not below market rate, it's not wage fixing? On the flip side it could be that the shared nature of the information makes it easier for companies to compete with eachother on salaries and raises rates.
It's just in this case salary info is (mostly) not public in the US.
This level of information asymmetry (in the US at least) is bs.
If everyone brings their salary info to a secret meeting and writes it in a chalkboard and then uses that chalkboard to set their salaries it’s obviously collusion, and cases like that have plenty of legal precedent.
If instead of walking the numbers over to the secret lair the communicate via an API and the chalkboard is a digital database and GUI has anything substantive changed?
And your premise is ridiculous. Of course how they came to know each other’s salary info isn’t irrelevant, it’s the collusion part.
There’s a long long history of this kind of stuff with prices that are completely public. There’s a Matt Damon movie about it. It’s not some novel concept it’s a time honored way to break the law.
> If everyone brings their salary info to a secret meeting and writes it in a chalkboard and then uses that chalkboard to set their salaries it’s obviously collusion, and cases like that have plenty of legal precedent.
The how of "set their salaries" is the difference between wage fixing or not. ie a company sharing / receiving of salary info (aka getting market reference points) and deciding that it wants to be at the Nth percentile is not wage-fixing. Wage-fixing would be if that company got together with another company (or companies) and agreed to stay at the Nth percentile.
No it’s not.
Privately communicating with direct competitors about pricing is invariably the first step of an illegal price fixing conspiracy.
Sure, and having a drink is invariably the first step to drunk driving. That doesn't make having a drink = drunk driving.
To be clear, I'm not defending the practice or saying it's never illegal. I'm saying sharing is not sufficient in and of itself. This article may be more convincing than an internet rando: https://news.bloomberglaw.com/daily-labor-report/wage-fixing...
> While wage fixing is easy to spot, a more subtle and related antitrust problem springs from wage disclosure: the mere agreement among employers to exchange wage information, even if they don’t agree on the actual wage itself, CAN lead to liability IF the information exchange negatively effects wages.
> Antitrust can be complex, but even a simple agreement on wages is a violation. Proof can come by the agreement itself without proof of competitive effects.
> An agreement between or among employers to exchange wage information, even if employers do not agree on the amount, can be illegal if that exchange resulted in lower wages.
Well yeah, but it sure as hell ain't orthogonal.
If you see a bunch of people sitting around drinking heavily at a bar with their cars parked outside and someone says hey what does any of this have to do with drunk driving you might raise an eyebrow.
Kind of like when you see a bunch of people who are supposed to be in direct ruthless competition with each other sharing lots of detailed price information.
The link you posted just reinforces my point that this is also how the law tends to look at these things.
They were fined a small fraction of what they saved for the wage fixing (and essentially nothing to the total impact which included people far beyond these companies due to market distortion-- when the scheme collapsed and my pay went up something like 20%, even though I didn't work for any of the involved companies).
So why shouldn't anyone else do the same?
To greater and lesser extents, we know what rents are and we know what salaries are. Companies and landlords might have slightly better or more complete information, but it's not like we're totally clueless. A key point of the RealPage algorithm isn't "this is what rent is in your area." The key point is "this is how much you can charge before you start losing renters to other landlords."
The problem is when RealPage becomes a market mover - a company that has so much market influence that it doesn't just participate in the market, but exerts control over the market. If every landlord used RealPage and every landlord obeyed RealPage's suggested pricing, it wouldn't be about telling landlords the market price or even predicting how much they could charge before losing renters to other landlords. It would be about predicting how much they could charge before people exited the housing market (eg. by moving away form the city). Of course, if you're doing that in all the cities, then there isn't even other cities that people can move to.
Let's say that RealPage is a simple ML program or even a genetic algorithm. It makes some random moves and starts learning "if I do X with rent, Y happens." "Y" might be the landlord losing 1% of their tenants, but with rents going up 10% they're making more money than they're losing. That's "fine" for some definition of fine that accepts money-maximizing capitalist competition as fine (I'm not really interested in debating the housing market in this comment).
Now, the issue is that RealPage's algorithm will quickly learn that it can raise rents much higher if it controls the market in a city or neighborhood. If they tell one landlord raises rent by 30%, that landlord is likely going to lose most of their tenants to other landlords. If they tell every landlord to raise rent by 30%, then tenants are no longer in a competitive market where they can rent from someone else. Essentially, the landlords have formed a cartel.
Housing activists don't object to open data about what rents are. In fact, housing activists generally support requiring landlords to register rents with the city. Labor activists don't object to open salary data. In fact, labor activists want open salary data because we know that leads to more negotiating power for workers. The issue with RealPage isn't just that it's supplying and collecting data. It's that it is setting prices that, in some markets, a majority of landlords are following. That means that it's no longer a competitive market and (intentionally or not) they've formed a cartel where the algorithm is basically learning that it has the pricing power of a cartel.
It's one thing to say "a software engineer at level X makes $Y-Z." It's another thing if 75% of the companies hiring software engineers used the same SalaryAlgorithm which told them "you can lower your engineer salaries by 15% (because we're telling everyone to lower salaries by 15% and we control 75% of the market)." There's actually been a lot of research more recently on monopsony and oligopsony (https://en.wikipedia.org/wiki/Monopsony). Basically, it's the opposite side of monopoly and oligopoly. With a monopoly, you have one company selling something. With monopsony, you have one company buying something. Often this can be companies hiring people (buying labor) and having an impact on wages. If there's only one company hiring software engineers, they're going to control the wages for software engineers. If there's thousands of companies hiring software engineers, they don't have much control - unless they all agree to use the same algorithm for setting salaries and that algorithm quickly learns that it has monopsony power.
I think a key part of an algorithm like RealPage's effectiveness can be that it can achieve monopoly/oligopoly power and then learn how much it can charge based on that power - rather than learning how much it can charge based on market rates in a competitive market where landlords aren't colluding (even if the collusion is unintentional). I'm not even suggesting that RealPage's intent was to form a cartel, but if their algorithm is even the tiniest bit smart, it's going to learn pretty quickly that it can rase rents higher when it has monopoly/oligopoly power.
Which is I suspect why RealPage encourages calls with software users and "strongly discourages" pricing things differently than their software says. The official line is because it's an "educated estimate", and the unofficial line, I suspect, is "because the number we have given you is one part in a much larger equation that others are relying on, too".
Your whole post is insightful and correct but it's actually even more basic than that.
RealPage doesn't have to be a market mover, it doesn't even have to be an organization, it can just be a concept, it could literally be a math equation. It could be the words "don't lower rents by more than X% ever" spray painted on a piece of plywood. It could be a trade association that is a non profit.
It doesn't matter what the mechanism is. The collusion comes from everyone using it. When everyone is following the same script then it doesn't matter what color paper the script is written on, what matters is everyone following it is engaged in a conspiracy.
When a real estate firm signs up for the software they've made a decision to join an illegal price fixing conspiracy. Focusing on the mechanism of exactly what's going on inside the black box of the software itself is a red herring. As long as it involves fixing prices, then it's price fixing. A
Clearly, that's exactly what it is.
If this is just a tool, as the authors claim, to help estimate best prices, then it's just an estimate. But "strongly discouraging" any deviation from the model sounds very much like "because this number is what everyone else is relying on, too".
Monopolies and cartels are ubiquitous in the American business landscape. They aren’t highly visible, as they operate in the operational layers between stages of industry: rental unit pricing (as highlighted in this story), niche goods distribution, other services. They aren’t the classic top-down monopolies observers expect to see, yet they have the same impacts of exploiting their peripheries, eliminating competition, crippling innovation, and inducing inflation.
I believe eliminating cartels and monopolies is the single greatest economic opportunity presented to the US.
Rent going up? One company’s algorithm could be why - https://news.ycombinator.com/item?id=33224502 - Oct 2022 (279 comments)
I think there were other threads also?
I know a lot of HN posters live in hope of developing something that will be acquired by a FAANG, but corporate acquisitions are usually bad for everyone but the person running the company being acquired.
There is no strict need for this software to be anti competitive, now that prices are online they can do plenty of experimentation day to day on their own prices.
I also think the claim that they have some secret source of data from negotiated rents is not realistic, large companies don't really negotiate on rent.
However its bad news for the company that the lead developers of their two big software platforms appear to have gone on record in a pretty negative way.
Probably they got to greedy and have exposed themselves to anti trust by commingling data.
Sharing your own negotiated prices may or may not be legal, but it's often bad manners at the very least and damages your business relationships.
My submission today "Culture at Google: Part One, the Movies" deliberately does not say how much I paid for PPRs, because I was asked not to tell other distributors that.
Now take what RealPage is doing with apartments and imagine if every major airline used the same revenue management solution (most use their own which are proprietary) to set airline fares. I hope RealPage (and these investor-owned apartment owners) pay dearly for this. Imagine what the size of a class action suit would look like.
That will be interesting. If they are thinking that "legally compliant manner" is rooted in nothing more than having permission to the data (for instance, because users of their software sign off on it in the licensing agreement), they may be in for a surprise.
I was looking at NYC apartment prices in 2021 and I still get price estimate emails on properties I looked at almost two years ago.
Im thinking if you’re an independent landlord, this feedback loop is driving your notions of “value”.
A passive version of the OP?