3 ms·

This sounds like price fixing as a service, but I wonder if the law is prepared to handle it. The FTC guidance on price fixing[1] only discusses "agreement amon

This sounds like price fixing as a service, but I wonder if the law is prepared to handle it. The FTC guidance on price fixing[1] only discusses "agreement among competitors", and never mentions each party outsourcing pricing to the same service. If the competitors never meet or agree on anything, is it still illegal? 1: https://www.ftc.gov/advice-guidance/competition-guidance/guide-antitrust-laws/dealings-competitors/price-fixing https://www.ftc.gov/advice-guidance/competition-guidance/gui...

paulryanrogers 4y ago
Collusion by proxy is still collusion. My guess is their marketing will take care never to put in writing that effect, yet still communicating it with a wink and a nod.
hammock 4y ago
It could be. There is something called tacit collusion, I am not immediately aware of case law involving a third party algorithm though. Here is a paper discussing the exact scenario: https://mdpi-res.com/d_attachment/sustainability/sustainability-13-00951/article_deploy/sustainability-13-00951-v2.pdf?version=1611207775 https://mdpi-res.com/d_attachment/sustainability/sustainabil...
dasil003 4y ago
> If the competitors never meet or agree on anything, is it still illegal? No idea how the laws were written, but the vast majority were conceived before modern big data systems were possible, and even today most regulators are unprepared for arguments like this (from the article): > Using software like YieldStar is “taking what we used to do manually on a yellow pad and calling people on the phone and putting it on a codified system where you take the errors out of the pricing,” he said. Modern big data and ML capabilities mean that not only is the human removed from the details of the process, but even understanding the mechanics and relative weight of different inputs becomes all but impossible. Large-scale human-based processes can still feel opaque and arbitrary from the outside, but at least there are entry points for human empathy and (more importantly) accountability. As we saw in the 2008 financial crisis, it's a very dangerous state of affairs when the risks and rewards of financial innovation are decoupled across different parties by means of opaque and unintelligible instruments. Computers are not the same as paper.
- treis 4y ago
  >Modern big data and ML capabilities But for the most part this stuff doesn't work. And I think in this specific case if there were really some secret sauce they'd make a lot more money as property flippers. Buy underperforming property for cheap, apply secret magic, and sell much better performing property for much more would make you way more money than selling a SAAS.
  - richbell 4y ago
    >>Modern big data and ML capabilities > But for the most part this stuff doesn't work. Yeah, not a great example when ML is frought with discriminatory biases or bizarre edge cases like Twitter banning people for posting innocuous pictures[0]. People get so blindsided by the magic they forget that these systems are often designed poorly, trained poorly, and cannot be easily explained. [0] I can't find the link, but it was something like a picture of a bear that resulted in your account being immediately suspended.
drdec 4y ago
I would submit that you are making a common mistake in applying an engineer's or programmer's mindset to the law. What I am thinking of is that company that tried to get around copyright by having their customer's buy TV antennas located at their HQs and then they would stream the feed from the antenna to the customer. Every piece was technically legal but because the entire scheme resembled a patently illegal activity (profiting from distributing copies of copyrighted material) the courts shut them down. Similar reasoning will be applied in this case. If enough of the market is using the same service to set prices, that will be struck down as collusion. (All IMHO and IANAL)

paulryanrogers 4y ago

Collusion by proxy is still collusion. My guess is their marketing will take care never to put in writing that effect, yet still communicating it with a wink and a nod.

hammock 4y ago

It could be. There is something called tacit collusion, I am not immediately aware of case law involving a third party algorithm though. Here is a paper discussing the exact scenario: https://mdpi-res.com/d_attachment/sustainability/sustainability-13-00951/article_deploy/sustainability-13-00951-v2.pdf?version=1611207775 https://mdpi-res.com/d_attachment/sustainability/sustainabil...

dasil003 4y ago

> If the competitors never meet or agree on anything, is it still illegal? No idea how the laws were written, but the vast majority were conceived before modern big data systems were possible, and even today most regulators are unprepared for arguments like this (from the article): > Using software like YieldStar is “taking what we used to do manually on a yellow pad and calling people on the phone and putting it on a codified system where you take the errors out of the pricing,” he said. Modern big data and ML capabilities mean that not only is the human removed from the details of the process, but even understanding the mechanics and relative weight of different inputs becomes all but impossible. Large-scale human-based processes can still feel opaque and arbitrary from the outside, but at least there are entry points for human empathy and (more importantly) accountability. As we saw in the 2008 financial crisis, it's a very dangerous state of affairs when the risks and rewards of financial innovation are decoupled across different parties by means of opaque and unintelligible instruments. Computers are not the same as paper.

treis 4y ago
>Modern big data and ML capabilities But for the most part this stuff doesn't work. And I think in this specific case if there were really some secret sauce they'd make a lot more money as property flippers. Buy underperforming property for cheap, apply secret magic, and sell much better performing property for much more would make you way more money than selling a SAAS.
- richbell 4y ago
  >>Modern big data and ML capabilities > But for the most part this stuff doesn't work. Yeah, not a great example when ML is frought with discriminatory biases or bizarre edge cases like Twitter banning people for posting innocuous pictures[0]. People get so blindsided by the magic they forget that these systems are often designed poorly, trained poorly, and cannot be easily explained. [0] I can't find the link, but it was something like a picture of a bear that resulted in your account being immediately suspended.

drdec 4y ago

I would submit that you are making a common mistake in applying an engineer's or programmer's mindset to the law. What I am thinking of is that company that tried to get around copyright by having their customer's buy TV antennas located at their HQs and then they would stream the feed from the antenna to the customer. Every piece was technically legal but because the entire scheme resembled a patently illegal activity (profiting from distributing copies of copyrighted material) the courts shut them down. Similar reasoning will be applied in this case. If enough of the market is using the same service to set prices, that will be struck down as collusion. (All IMHO and IANAL)