Follow Slashdot blog updates by subscribing to our blog RSS feed

 



Forgot your password?
typodupeerror
×
AI Businesses Software Technology

Left To Their Own Devices, Pricing Algorithms Resort To Collusion (popularmechanics.com) 128

Reader schwit1 shares a report: When you're browsing online, who sets the prices? An algorithm, most likely. A study from 2015 showed that a third of all items on Amazon [PDF] had prices set by an algorithm, and chances are that percentage has only risen. A new study shows how easy it would be for price-setting algorithms to learn to collude with each other and keep prices at a disadvantage for customers.

This sort of collusion would stem from a certain type of algorithm, the researchers say. Reinforcement algorithms learn through trial and error. In the simplest terms, a walking robot would take a step, fall, and try again. These algorithms have often been used to teach algorithms to win games like Go.

"From the antitrust standpoint," say professors Emilio Calvano, Giacomo Calzolari, and others from the University of Bologna in Italy, "the concern is that these autonomous pricing algorithms may independently discover that if they are to make the highest possible profit, they should avoid price wars. That is, they may learn to collude even if they have not been specifically instructed to do so, and even if they do not communicate with one another."

This discussion has been archived. No new comments can be posted.

Left To Their Own Devices, Pricing Algorithms Resort To Collusion

Comments Filter:
  • Collusion? (Score:4, Interesting)

    by XanC ( 644172 ) on Wednesday February 13, 2019 @03:53PM (#58117530)

    "and even if they do not communicate with one another"

    Well that can be the case with people, too, can it not? Except that when people do not communicate with each other, I don't believe that it can be said to be colluding, by definition. So why is it collusion when algorithms do it?

    • the 1970s called... (Score:5, Informative)

      by goombah99 ( 560566 ) on Wednesday February 13, 2019 @04:03PM (#58117580)

      There's case studies many economics classes use of how the airlines in the 1970s would evade the collusion in restraint of trade laws. Basically, they developed a process for proposing fare changes. They would announce a scheduled fare increase a month in advance. If the other company followed suit they would enact it if they didn't they would retract it or post a fare decrease shortly after that.
      It worked quite well till the regulators figured it out. Braniff (deceased high end airline) was the ringleader.

      These pricing algorithms just accomplish this one time scales of hours rather than months.

      but it's just old wine in new bottles. No one invented anything or changed anything. But the amplification and universality of it made the problem worse.

      But what actually makes this interesting is that it may also be an emergent behavior as opposed to either intentional programming or an untended artifact of some algorithm. THat is, if an AI is simply asked to maximize profit in a multi-agent system it's entirely possible it will learn this tit for tat strategy. Humans do in cooperation-games in behavior theory.

      • by Anonymous Coward

        They would announce a scheduled fare increase a month in advance.

        That is literally the exact opposite of not communicating.

        These pricing algorithms just accomplish this one time scales of hours rather than months.

        The algorithms are posting publicly (well, publicly to other algorithms) their intent to raise the price of a product within the next hour? And if no other algorithm also posts intent to do so then the first algorithm does not change their price?

      • It makes me think of those items they list for a penny with a $15 shipping and handling charge but it is usually $12.99. You know they just want to get it to the top of the list and will just make up all the money in shipping and handling.

      • IIRC, they had an FU fare code for when a competitor posted a trally low gare and they matched it. Braniff, the home of the great pickle.
      • Actually, I suspect it isn't even worse. The airline situation in the 70s was discovered and proved because it was so obvious.

        For instance there are a number of studies now showing the cross ownership (one mutual fund owning many companies in a market) substantially decreases competition even when there is never any easily observed collusion or signalling like you describe.

        This suggests humans have a wide range of subtle communications channels they can use and don't even need to be aware that they are coll

        • So I strongly suspect that such price collusion is already happening and algorithms make it no worse.

          The first half of your statement is reasonable.

          Care to explain how you support the second half wherein you posit that automating a process so that it happens much faster, more ubiquitously, with far more data than a human can process causes you to conclude that it must be "no worse"?

        • "So I strongly suspect that such price collusion is already happening and algorithms make it no worse."

          Super intelligent machines are no worse than a few bumbling airline employees from the 70s?

      • But what actually makes this interesting is that it may also be an emergent behavior as opposed to either intentional programming or an untended artifact of some algorithm

        This is quite common in machine learning task optimization. Algorithms often find out that the best ways to "win" is to game the system - for example discovering flaws in models of the world and exploiting those instead of solving the intended problem.

    • Re:Collusion? (Score:4, Interesting)

      by Layzej ( 1976930 ) on Wednesday February 13, 2019 @04:08PM (#58117598)

      "and even if they do not communicate with one another"

      Well that can be the case with people, too, can it not? Except that when people do not communicate with each other, I don't believe that it can be said to be colluding, by definition. So why is it collusion when algorithms do it?

      In most countries (including Europe and the US) such ‘tacit’ collusion, not relying on explicit intent and communication, is not currently treated as illegal, on the grounds that it is unlikely to occur among human agents and that, even if it did occur, it would be next to impossible to detect. The conventional wisdom, then, is that aggressive antitrust enforcement would be likely to produce many false positives (i.e. condemning innocent conduct), while tolerant policy would result in relatively few false negatives (i.e. excusing anticompetitive conduct). With the advent of AI pricing, however, the concern is that the balance between the two types of error might be altered.

    • by lgw ( 121541 )

      There's a classic example in economics: two gas stations across the street from each other. Adjusting your prices based on the publicly displayed price you see across the street is simply not collusion. Every gas station in town raising their prices immediately when the price of oil goes up (even though their costs won't be affected for weeks) is also not collusion. Each seller still has made no promises not to undercut the other guy, they're just using the same data to reach the same conclusions.

    • by ezdiy ( 2717051 )
      Perfect markets produce "cartel" like this when all participants have near equal production capacity - they'll all gauge the price to maximize profit and minimize output as that is the most optimal nash game. And they can do it with no direct collusion, just market ticker.

      However whether amazon is a perfect market with participants of comparable capacity begs question. Real world markets have a stream of newcomers who pull down the majority to their price level and would even refuse to participate in a c
      • Perfect markets produce "cartel" like this when all participants have near equal production capacity - they'll all gauge the price to maximize profit and minimize output as that is the most optimal nash game. And they can do it with no direct collusion, just market ticker.

        Perfect, working markets will use public prices as signals to adjust prices and capacity. This is the way supply and demand is supposed to work, and it can be done without direct collusion. However, there are situations where this dynamic adjustment of prices doesn't work, e.g., if there is a monopoly or if there are no disruptors (i.e., all vendors are willing to keep their market share without competing).

        The problem with the robo-pricers is not the computer aspect. It's that there are no disruptors in

    • Except that when people do not communicate with each other, I don't believe that it can be said to be colluding, by definition.

      P

      The federal government, specifically the FTC [ftc.gov], disagrees with you. Collusion just requires that two companies agree to set prices, no communication is required. This agreement can be inferred from behavior. There are lots of examples from other posters (such as the airlines in the 1970's).

      It makes sense that algorithms would collude... it's well known that algorithms can establ

      • by cwatts ( 622605 )

        In this case, it would only be collusion if the algorithms shared data privately between themselves, rather than scraping the competitors' pricing from their 'gas station signs'. If the AIs are getting the same data as the general public, it's a free market, right? Or am i missing something?

        cw

        • You're missing something. Getting information from the signs would be "inferred from conduct". The line there is tricky, but doable. Imagine an algorithm that raises prices by a penny/gallon. If all its competitors follow suit, it waits 5 minutes and then raises prices again. If not, it lowers its price by a penny (back to where it was) and sleeps for a day or until a competitor raises prices by a penny. That is colluding.

    • I had the same thought. It seems that many businesses are no longer doing their best to put their competition out of business in order to increase profits. Instead, they concentrate on ways to increase profits through internal changes without changes in sales. Many also no longer push to innovate. R&D cost money. I guess it is much easier to effect the stock on 3-month timeframes when you just silently agree to a truce with your competition and work on reducing internal costs and using accounting tricks

    • by cwatts ( 622605 )

      My sentiments exactly... Agreement sans collusion is clearly possible.

      (see what i did there?)
      csw

      google sez; secret or illegal cooperation or conspiracy, especially in order to cheat or deceive others.

  • Collusion? (Score:5, Insightful)

    by mpercy ( 1085347 ) on Wednesday February 13, 2019 @03:53PM (#58117532)

    "That is, they may learn to collude even if they have not been specifically instructed to do so, and even if they do not communicate with one another."

    I think collusion pretty much requires communication and intent.

    Independent algorithms (or people) arriving at the same conclusion from the same data and openly performing actions in their own self-interest without communication is not collusion.

    Their actions may look like collusion from the outside, but they fail on most commonly accepted definitions of the word.

    E.g. wikipedia

    "Collusion is an agreement between two or more parties, sometimes illegal–but always secretive–to limit open competition by deceiving, misleading, or defrauding others of their legal rights, or to obtain an objective forbidden by law typically by defrauding or gaining an unfair market advantage. It is an agreement among firms or individuals to divide a market, set prices, limit production or limit opportunities."

    • Re:Collusion? (Score:5, Insightful)

      by Shaitan ( 22585 ) on Wednesday February 13, 2019 @04:03PM (#58117578)

      "Independent algorithms (or people) arriving at the same conclusion from the same data and openly performing actions in their own self-interest without communication is not collusion."

      If it comes to the same anti-consumer result and the same result as monopoly behavior in a free market it is collusion. Or insert a new word for what is bad and can't be tolerated if we are to have a sound economy.

      It doesn't matter if it is one monopoly fixing terms and pricing in an anti-competitive way, 8 companies with common interest doing the same, or 10,000 doing the same. The active communication isn't the issue. And it isn't just pricing, the major chip manufacturers engage in this sort of anti-competative behavior and coordinate on release rates by releasing roadmaps of how far they are planning to go with their capacities and features in advance via unspoken (or possibly spoken behind closed doors) agreement with each other.

      "Collusion is an agreement between two or more parties, sometimes illegal–but always secretive–to limit open competition by deceiving, misleading, or defrauding others of their legal rights, or to obtain an objective forbidden by law typically by defrauding or gaining an unfair market advantage. It is an agreement among firms or individuals to divide a market, set prices, limit production or limit opportunities."

      An unspoken agreement fits that definition just fine.

    • by Anonymous Coward

      But they are communicating via the prices they set.

      • But they are communicating via the prices they set.

        Exactly, Anonymous Coward.

        If you have a hotdog stand selling smokies for $5 and I set up a hotdog stand two blocks away, note your price, and also sell *my* smokies for $5, I'm not "colluding" - I'm just price-matching. Seems to me that's what these price-bots are doing.

        • it's true, but that also means we're down to a problem of the laws, and maybe free market in general can't fit with the times anymore. Old logic was if widgets cost me $2 to make, and I sell them for $40 a piece, someone else will start selling widgets for $20 forcing me to either drop my price or go out of business, then someone would likely also see the same thing and start selling them for $10. Unfortunately the whole concept is a bit of a sham... in that the window of time between when you drop your pr
          • Also new contenders know they are price competing against bots, if you enter the market undercutting a bot, the other bots are going to price match you in milliseconds, no customers are even going to see the window of time where you were cheaper, everyone's going to be the same price always, and by dropping the price you just threw away everyone's profit for no benefit.

            That opens the door to a new game. Consumers deliberately crashing the market. A consumer can open up a new store and advertise items at a steep discount, never intending to actually ship anything. Such a store could even take orders and refund them "later". By refunding HIS customers when he can't ship, hey, no harm, no foul there. In the meantime - if other stores match that price, then the consumer could buy product from THOSE stores at the discounted price and get some quality stuff, for cheap. In

            • That opens the door to a new game. Consumers deliberately crashing the market. A consumer can open up a new store and advertise items at a steep discount, never intending to actually ship anything. Such a store could even take orders and refund them "later".

              I assume that's why Amazon has tiers of marketplace sellers. Anyone can open a seller account and start listing products subject to a cursory review of your listings, but you are just an "also sold by" that you have to click through. It's not until you've provably sent items using real shipping companies and collected real money that eventually you get promoted to a real seller. It's only then that you get to have your prices matter to Amazon's algorithm.

              And by the time you've achieved that, well now you ar

        • But if you set your initial price to $6, and the competitor raises their price to match, that is closer to what is happening. The price that settles is one of one-upping until something gives. And since the AI bots are set to maximize profit, they will not undercut each other as a primary strategy unless sales numbers start dropping. If the competitor used a similar AI, you get the same behavior on both sides.

      • by guruevi ( 827432 )

        Not necessarily, they may simply be reacting to consumer demand, much like a store owner does when he sees there are is a lot of demand, they may raise prices, if demand suddenly drops, they may lower prices. If your price fluctuates a lot, so do consumers so in the end you come to a mutual 'agreement' across all parties that works best for everyone (customers and providers) while still guaranteeing a profit.

        What these people want is an algorithm that always goes in the customers' apparent/current best inte

    • by ljw1004 ( 764174 )

      If two algorithms are setting prices and learning from how each other is setting prices, then they are communicating. It's a fairly low bitrate communication but it's certainly communication.

      As for intent? I don't know to use that word in relation to algorithms.

      • by mpercy ( 1085347 )

        Prices do indeed constitute communication, but they are in the open. Something done openly is not collusion.

        • Re:Collusion? (Score:4, Insightful)

          by swillden ( 191260 ) <shawn-ds@willden.org> on Wednesday February 13, 2019 @05:14PM (#58118080) Journal

          Prices do indeed constitute communication, but they are in the open. Something done openly is not collusion.

          That's not true. Collusion is defined as "secret or illegal cooperation". It generally is secret because doing illegal things openly is likely to result in getting caught, but that doesn't mean that collusion can never be done in public.

        • by ljw1004 ( 764174 )

          Prices do indeed constitute communication, but they are in the open. Something done openly is not collusion.

          https://definitions.uslegal.co... [uslegal.com] "Such agreements are usually secretive". The fact that the definition includes the word "usually" implies that collusion is not always secret.

          But really we should be looking at the actual anti-trust law, rather than the definition of collusion:

          http://www.duhaime.org/LegalDi... [duhaime.org]
          Every contract, combination in the form of trust or otherwise, or conspiracy, in restraint of trade or commerce among the several States, or with foreign nations, is declared to be illegal

          When two business have algorithms, and they key off of each other's behavior to result in both picking a higher price -- (Q1) do the two business or their two algorithms constitute a combination? (Q2) does it restrain trade/commerce?

      • As for intent? I don't know to use that word in relation to algorithms.

        Why not? We all know algorithms can be racist - AmiMoJo says so - so why can't they be corrupt & greedy too?

    • Posting your prices where your competitor (or anyone) can see it IS communication.
      • by mpercy ( 1085347 )

        Sure, but it's open communication not secret and not anti-competitive, indeed it may stem from competition.

        Consider two gas stations on opposite corners. Imagine that they are owned by two people who vigorously hate each other on a personal level, to the extent that they would piss on the other if he were on fire, let alone do anything that might help the other's business succeed or even get ahead. So as soon as one station changes prices on its signs and pumps, the other owner immediately runs out and chan

        • Ok you required communication and intent. We have proven communication and now you have proven intent: "If one raises prices in hopes of getting a penny or two more per gallon of profit, the other raises his too, lest the other make more money." They intend to raise prices for profit. Collusion
  • Headline: "Left To Their Own Devices, Pricing Algorithms Resort To Collusion"
    Summary: "That is, they may learn to collude even if they have not been specifically instructed to do so"

    Classic 2019 journalism right there.
    • In game theory, you can find your way out of things like the prisoner's dilemma by assuming the other person is rational and will work with you on what are obvious next steps for both sides, to both sides.

      But this isn't collusion as nobody is making plans with anybody else.

      • by g01d4 ( 888748 )
        I think this is a type of prisoner's dilemma with each algorithm starting or arriving at the conclusion that the other algorithm is 'rational' in terms of maintaining a higher price. The issue is that the algorithm does not take into account the consumer's benefit which society, in the form of commerce regulation laws, require of it (theoretically). So while it's relatively rational to agree on maintaining a higher price it's still (theoretically) illegal and thus the agreement effectively becomes de facto
        • So while it's relatively rational to agree on maintaining a higher price....

          Maintaining a higher price than what?

          This is a key point and everyone seems to be missing it.

          All other things being equal, there is an optimal price and that doesnt mean 'higher' as although it could be higher, it could also be lower..

          Charging a non-optimal price is sub-optimal by definition, and on those grounds it is bad, but of course sometimes a company can get away with benefiting from sub-optimal pricing when the flaw in their pricing does not get exploited by other players.

          Here we are talkin

  • by Shaitan ( 22585 ) on Wednesday February 13, 2019 @03:57PM (#58117550)

    "That is, they may learn to collude even if they have not been specifically instructed to do so, and even if they do not communicate with one another."

    Sounds like the situation we have with major companies now.

  • Left To Their Own Devices, Pricing Algorithms Resort To Collusion
    Left To Their Own Devices, Corporations Resort To Collusion
    Left To Their Own Devices, Politicians Resort To Collusion
    Left To Their Own Devices,___________ Resort To Collusion ... fill in the blank with pretty much anything, and don't be surprised.
  • How do you distinguish collusion from algorithms simply figuring out the market price given buyer/seller volume, and realizing there's no benefit in deviating from that?
    • by AHuxley ( 892839 )
      A make profit number would be the point to stop most of the discount for existing and returning users.
      Then add in a loss leader algorithm for people who use another brand? Who have never used a brand? Who need to see a much lower number to consider a change.
      Once they are stuck with the brand the prices go up and the discounts are ever as good.
      Lots of ways to do the math depending on the person detected who is looking for something new.
      • by aybiss ( 876862 )

        In capitalism you charge as much as the market can bear. A make profit number would be communism.

  • It's pretty obvious Amazon adjusts at least some of their pricing based on customers past purchasing history and / or habits.

    It explains why an .mp3 download will cost me $1.25 when the exact same download will cost .99 for someone else who doesn't normally download music from Amazon.

    • Re:Collusion or not (Score:4, Informative)

      by slacktide ( 796664 ) on Wednesday February 13, 2019 @04:53PM (#58117938)
      No, it's pretty obvious that Amazon DOES NOT adjust adjust pricing on a per-consumer basis, and they do not use your past purchasing history to set pricing specific to you. If they did, then sites like camelcamelcamel would not work. Amazon does adjust pricing over time, but at any given time, all buyers are offered the same price.
      • Amazon got slapped down for doing customer specific pricing a long time ago. I don't remember it well, but I think they went the other way. They were advertising an item at $10, but if you were someone they thought *might* be interested with an incentive they'd show you $9.50 to try and make the sale whereas if you were more likely to buy it normally you still saw $10.

        Source [wikipedia.org]

  • Wrong word (Score:5, Interesting)

    by jbmartin6 ( 1232050 ) on Wednesday February 13, 2019 @04:15PM (#58117672)
    That is called equilibrium, not collusion. There once was a similar state in the US with breakfast cereals for a while, until one of the participants decided to break the equilibrium and lower prices to try to gain market share.
    • Equilibrium is when disparate forces in a market reach a balance point, not when 'competitors' refuse to compete. The cereal problem you mentioned was a classic case of refusing to compete. FTC watched those guys for YEARS trying to catch them communicating.

      An example of an equilibrium is when prices go low enough that it becomes uneconomical for some participants to stay in the market if prices go any lower, but there is also not enough profit in the market to entice any new entrant (who faces barriers t

      • That's the definition of economic equilibrium, as used by economists, sure. But the states TFA and I described are also states of equilibrium.
  • Sounds like I need to go into business with a algorithm that simply undercuts all of them by a penny.

  • I was wondering about this as I can look up the price of an item, and a few days latter after considering the purchase, the price has increased.
  • Their price for a certain brand of bottled water varies daily. Sometimes it goes up by 50% or so from the week before. They scrap amazon prices and price it the same.

  • No mention of AI, I am shocked - and pleased. I have benefited from Amazon algorithm pricing wars at least once, the 3rd party's algorithm kept dropping the price to 3 cents under Amazon, and Amazon countered to simply match that new price each time. Within a day the price fell to my price point.
  • If you set prices similar to a competitor's, that's not colluding.

    Keep in mind these people will throw you in jail if you price too low.

    Regulator: It's about fair competition.

    Senator: i got my legal donation out of it. Ease off, we are concerned with something else now.

  • Many people seem to assume that potential customers are powerless against price collusion - but they are not. "Not buying" is almost always an option, so just show sellers of things you consider overpriced the finger and do not buy.

    After some time, either those who fixed prices will notice their revenue goes downhill, or somebody will realize there is a market gap to fill when offering at a lower price.

    There are plenty of things that I never bought all my life, for the simple reason that I do not conside
    • So I shouldn't buy a washer or dryer? Dishwasher? Stove? No car, no smartphone..
      • by ffkom ( 3519199 )

        So I shouldn't buy a washer or dryer? Dishwasher? Stove? No car, no smartphone..

        If you have reason to believe that they are overpriced due to price fixing or "collusion", then yes, do not buy them.

        I own a dishwasher, a washing machine, a stove and a smartphone, and I cannot say I saw any indication of them being overpriced - in fact all these household appliances where available at prices ranging from ~ 100 to 3000 Euros, so not much "fixing" but great flexibility in terms of different offerings from many different manufacturers.

        I do not own a dryer or a car, but because I have no ne

        • Of course it is collusion. Every manufacturer makes a few price points and the features across each price point are the same. There is no manufacturer with a more full-features washing machine at the price of a basic one from other brands.
  • So exactly what happens now when competing companies price things; except on an accelerated scale.
  • by kenh ( 9056 )

    A new study shows how easy it would be for price-setting algorithms to learn to collude with each other and keep prices at a disadvantage for customers.

    Not that it is happening, not that they ARE actually colluding, only that it would be easy to do...

    Why can't the headlines match the story?

  • On numerous occasions, I have seen a product on Amazon with the same price on Walmart and/or Target. Then I'll see the price change on Amazon, and lo and behold the price also changed to that same price on Walmart/Target.
  • So now AI is already able to fully replace CFOs and CEOs. Beware, engineers, you might be next!

  • Comment removed based on user account deletion
  • Why do we keep presenting this sort of AI non-sense as a new concept? We've had price-fixing, collusion, anti-competitive behaviour, insider trading, price gouging, and myriad other names for exactly the same thing.

    It's never been difficult to rip people off.

    It's been illegal.

    That's it.

    It still is illegal.

    Start arresting people.

Top Ten Things Overheard At The ANSI C Draft Committee Meetings: (10) Sorry, but that's too useful.

Working...