Algorithmic collusion and price-fixing
There’s a fascinating article on the FT.com (hat tip Jordan Weissmann) today about how algorithms can achieve anti-competitive collusion. Entitled Policing the digital cartels and written by David J Lynch, it profiles a classic cinema poster seller that admitted to setting up algorithms for pricing with other poster sellers to keep prices high.
That sounds obviously illegal, and moreover it took work to accomplish. But not all such algorithmic collusion is necessarily so intentional. Here’s the critical paragraph which explains this issue:
As an example, he cites a German software application that tracks petrol-pump prices. Preliminary results suggest that the app discourages price-cutting by retailers, keeping prices higher than they otherwise would have been. As the algorithm instantly detects a petrol station price cut, allowing competitors to match the new price before consumers can shift to the discounter, there is no incentive for any vendor to cut in the first place.
We also don’t seem to have the legal tools to address this:
“Particularly in the case of artificial intelligence, there is no legal basis to attribute liability to a computer engineer for having programmed a machine that eventually ‘self-learned’ to co-ordinate prices with other machines.
mathbabe 
The manipulation just keeps on growing. Here’s a clip from a former Stanford student who just left Google as an engineer. Google bought the start up company Tristan Harris created and then he went to work for Google. This is a clip from a new VPro documentary on how to make people click. Tristan points out there,s no ethics taught. There you have it, where manipulation techniques are being taught at the Univeristy level.
http://www.tubechop.com/watch/8762854
The entire documentary is good with the exception that they forgot to put in some English subtitles with some of the folks they spoke with not doing English. There’s a link on the Tube Chop page to the entire video. Tube Chop is great by the way for a quick and easy way to share clips of entire videos.
LikeLike
“As the algorithm instantly detects a petrol station price cut, allowing competitors to match the new price before consumers can shift to the discounter, there is no incentive for any vendor to cut in the first place.”
I’m not an economist, but I find this puzzling. The idea is that the sole incentive to cut prices is the time gap between when consumers find out that you’ve cut prices and when your competitors due. How large is that time gap in the absence of this software?
LikeLike
Economic models assume that you get more customers when you cut prices, and it may make up for the smaller fees you accumulate. But if there’s an instantaneous response to your price change, then you don’t get any more customers, but you do of course smaller fees and an overall smaller profit.
In terms of the time gap, I’m assuming it’s days or at least hours in most businesses historically.
LikeLike
I’m guessing what you’re getting at is that you’ll still presumably get more customers when you lower prices, you simply won’t get other people’s customers as often, but maybe in the past that was more or less true as well.
Am I guessing correctly?
In which case, this argument makes more sense in a small market, like classic cinema poster selling, where the customer base is active but there aren’t a ton of people waiting around for the prices to go down a bit.
LikeLike
This is essentially the same thing as “price matching”, when Best Buy, for example, advertises that they will match any price anywhere, it discourages other companies from undercutting their prices. The algorithm makes the mechanism faster and more widely applicable (so a larger effect), but the scheme is the same. I’m assuming this is not illegal (based solely on the fact that I still see brands doing it openly) because it isn’t technically colluding. It’s an interesting point, and I’d be interested to see empirical evidence that this has led to price increases, in gas prices or any other types of markets.
LikeLike
I don’t think this is an algorithm problem per se; it’s really just a modern variation on the price matching issue. If a market participant has a credible price matching policy, then there’s never an incentive for anyone to reduce prices, since their competition will match. Everyone is left with the same volumes and lower profits; what algorithms do is increase the credibility of the price matching threat by automating it.
There have been attempts to regulate this in the past, but it’s been very difficult. Some attention has been focused on mandating prices to be held for a certain amount of time, for example, but it’s hard to craft a policy that is unilaterally consumer positive.
LikeLike
Isn’t this a classic exercise of market power that corporations do all the time. Of course, with real time prices being more easily available, it is easier to exercise market power.
LikeLike
I recently read a very good book on AI and thoughts about its future and fears. It is called “What to Think About Machines That Think”, John Brockman. This seems to be a worry that is being thought about and not really being accepted as already present, as this Digital Price Fixing appears to be in full bloom. I am not educated in Algorithms but from you Cathy, I have learned a little. It appears impossible that no human intent was there at the beginning when the algorithm was written based on the type the engineer chose to create from the beginning. In looking at the website http://machinelearningmastery.com/a-tour-of-machine-learning-algorithms/ there are so many types of machine learning algorithms we could chose from based on our goal, it just doesn’t look like it’s impossible to find the human hand print in all this. Maybe just the law needs adjusting or to be created, one that applies responsibility somewhere in the chain of this Digital Price Fixing. Maybe you could even write an algorithm that keeps track of the algorithm that is identifying pump prices and then sending that information out to competitors before consumers can switch pumps. Even on Whatsapp everything is trackable, there is no way we can blame machine intelligence without finding the human intelligence behind the use of the app. It is sort of like HIV transmission, we can not blame the originator since we don’t know who or what it was, but now the crime of Criminal Transmission of HIV is in place to protect individuals who were infected by someone who carried HIV and knowingly exposed that person. The law just needs to catch up in AI and business and social justice…before the users of the app are too rich to care.
LikeLike
Not on-topic, but I thought you’d find this interesting. The Guardian this morning has a piece about the consequences of an EU law barring gender discrimination in pricing for car insurance premiums. The gender gap between men and women in such pricing is now 4x what it was before the law: https://www.theguardian.com/money/blog/2017/jan/14/eu-gender-ruling-car-insurance-inequality-worse
wg
LikeLike