1. Skip to navigation
  2. Skip to content
  3. Skip to sidebar
Source link: http://archive.mises.org/20628/research-shows-under-certain-market-conditions-cartels-arise-naturally-without-collusion/

Research shows that under certain market conditions, price cartels arise naturally without collusion

January 21, 2012 by

Recently reported in the MIT Technology Review was the results of research modeling billions of interactions between buyers and sellers. The data would appear to undermine the argument used to defend invasive regulation, price-fixing, and anti-trust laws, which operate under the assumption that all inconvenient price increases must be the result of shady conspiracies by industry leaders.

The results make interesting reading.  It turns out that a crucial factor is the speed at which buyers and sellers react to the market. When buyers react quickest, sellers are forced to match the best possible value for money and prices tend to drop. By contrast, when sellers react quickest, they are quick to copy others offering poor value for money. This reduces the number of sellers offering good value for money in a vicious cycle that drives prices as high as possible. This is the emergence of a cartel and it happens in these guys’ model without any collusion between sellers. Instead, it is an emergent property of the market place that happens when the sellers outperform buyers in the way they react to market conditions. “This cartel organization is not due to an explicit collusion among agents; instead it arises spontaneously from the maximization of the individual payoffs,” say Peixoto and Bornholdt.

I particularly loved the echoes of “Spontaneous Order” in this quote. Of course, the writers of this particular article are still scratching their heads, unable to conceive of a world without regulatory and punitive market controls. Simply leaving market agents alone would be too radical, but at least they acknowledge that even new strategies would still be subject to the law of unintended consequences:

But this work muddies the waters somewhat. If cartel-like behaviour is an emergent property of an ordinary market, how should it be controlled, regulated and punished? The good news is that various strategies could easily be tested using this kind of agent-based model. The bad is that new strategies may themselves lead to emergent properties that are hard to spot in advance.

The abstract is available at the Cornell University Library. Here’s a direct link to the PDF of the paper published by Tiago P. Peixoto and Stefan Bornholdt of the Institute for Theoretical Physics in Bremen, Germany: No need for conspiracy: Self-organized cartel formation in a modified trust game


Rory Carmichael January 21, 2012 at 7:25 pm

You seem to come to some weird conclusions. Just because a bad result can occur in conditions previously thought to be safe doesn’t mean that the things we know to cause the bad result are magically okay. That’s like saying that because people can get cancer without smoking cigarettes we should no longer avoid smoking cigarettes. You see evidence that the free market can have clear negative outcomes as support for a belief that the free market should never be interfered with?

Israel Curtis January 21, 2012 at 9:29 pm

First of all, your characterization of prices rising as a “bad result” or even “clear negative outcome” is entirely subjective. Obviously buyers would prefer prices to be lower, but such sentiments when combined with state power lead to price controls, which entirely ignore the other half of every economic calculation: the seller. What is “good” for one party or “bad” for another is for them alone to decide. This is precisely what prices are great at: communicating the preferences of both parties.

Second (and most pertinent to this study) is the issue of blaming price increases on nefarious collusion amongst sellers. It is widely believed among supporters of anti-trust regulation that without the “wise hand” of government forcibly preventing collusion (or even the appearance of it), such conspiracy would always happen – and price increases in certain industries (like fuel in this example) are often blamed on rich sellers colluding to “artificially” fix prices.

The assumption is that such price increases simply wouldn’t happen if regulators could stamp out all collusion, thus justifying the ever-growing reach of regulatory intrusion in the marketplace (like past attempts to prohibit alcohol consumption).

What this research demonstrates is that the fluctuations often solely attributed to cartel-pricing-schemes can actually have completely organic sources – in other words, none of the agents in the marketplace have to act nefariously in order for such price increases to occur. They are a natural result of the cumulative actions of millions of agents all reacting to each other.

I already believe that the free exchanges of individuals should never be interfered with. I never believed that price increases (inconvenient for some, beneficial for others) were ever cause for intervention. Yet those who favor such action have always claimed that without intervention we would all be drowning in criminal price collusion.

This research demonstrates that even in a model where all collusion has been eliminated, severe price fluctuations can still occur. Thus ascribing such market behavior to “immoral” or “evil” intentions (as a justification for laws against such market behavior) is nothing more than speculation.

Inquisitor January 22, 2012 at 4:15 am

Like Israel said, “bad” result is 100% your opinion.

“Negative” outcomes for whom? I don’t even agree with the researchers’ characterisation of this as a “cartel” when there is no specific intent to collude. It is simply a market outcome. I like how they think regulation is in order. No, it isn’t.

Jonathan M.F. Catalán January 21, 2012 at 9:03 pm

I haven’t read the paper, but from the excerpt it seems as if they are confusing rising prices with the formation of monopoly (or cartel) prices — it may be the photograph of gas prices which leads me to this conclusion. Sellers may use each other as guides to decide in what direction they should move their prices, but this doesn’t imply that what is occurring is cartel behavior. What it suggests is that they are using price movements as signals to suggest in what direction they should go themselves — they’re using it as knowledge to base entrepreneurial decisions on. So, if the supply of oil decreases and some sellers see themselves increasing prices as a result, other gasoline vendors will use this as signal to mark up their prices themselves. But, ultimately, the rise in the price of gasoline wasn’t caused by some psychological trend in the behavior of vendors, rather in the falling supply of gasoline.

Joe Caten January 22, 2012 at 1:02 am

This reminds me of a quote that goes something like this:

If you set your price higher than anyone else, you’re price gouging;
If you set your price lower than everyone else, you’re predatory pricing;
If you set your price the same as everyone else, you’re colluding.

Julien January 22, 2012 at 3:05 am

Here’s the link to the TR article (the one above only points to an image): http://www.technologyreview.com/blog/arxiv/27512/

Israel Curtis January 22, 2012 at 3:47 am


Rudd-O January 22, 2012 at 7:34 am

I love how they repurpose the word “cartel” to mean something that is *outside* the definition of cartel (no collusion, no cartel).

Propaganda at its finest!

Bruce Koerber January 22, 2012 at 11:18 am

One aspect of bad economics is seeing things only in the short run. A ‘study’ that examines the ever-present disequilibrium will always find evidence of disequilibrium but only fools think that equilibrium can and should be attained by intervention.

Martin OB January 22, 2012 at 7:41 pm

Apparently, the study basically says that if sellers are looking at each other much faster than buyers look at them, then they will be able to agree on a price, and thereby behave in practice as one big seller, without having to secretly talk to each other. It works like this: If one seller sets a lower price to win buyers, the others quickly imitate it, so it gets no benefits from that. On the other hand, if one seller sets a higher price to earn more profit from each sale, the others also raise their prices before the buyers realize. In other words, they don’t have to send secret signals when they can signal in the open for free. Did we need a computer simulation for this? I’m tempted to call it junk science.

Yes, in this scenario they can raise the price for ever, or up to a point when buyers are no longer willing to pay more, or in any case as high as a single seller could. Monopoly! Of course, in practice every company has a monopoly on its products, but there are always substitutes. Apple has a monopoly on iPads, but people can buy cheaper Android tablets. Nothing new.

Martin OB January 22, 2012 at 7:47 pm

By the way, I predict that the “market failure” guys will propose legislation prohibiting big business owners from being too diligent in looking at each other’s pricing. Or making it compulsory for every company to publish their pricing on some official website, where potential customers can look it up from their smartphones.

[edit: typo]

Walt D. January 23, 2012 at 2:07 am

I don’t like the picture of the gas prices – its from July 2013 and Obama got re-elected.

Horst Muhlmann January 24, 2012 at 12:04 pm

Not to worry. Here’s the Ron Paul version:


Mushindo January 23, 2012 at 5:01 am

Duh. In a competitive context with many suppliers of similar products, , and after correcting for product differentiation ( quality, brand et al…..), market prices will naturally converge without any conscious attempts at collusion…. why this would be seen as a bad thing by anyone is beyond me. the convergence will invariably settle where price and volume sold is optimal for th edemand profile of the market. there is no way to improve on that in either direction without making both producers and consumers worse off. It its cheaper, some consumers wont be able to buy it due to short supply, and if its more expensive, fewer potential buyers will buy and producers will sell fewer than they can supply.

Of course, the price convergence thus emerging will only remain until until some entrepreneur comes up with a better or more cost-effective mousetrap. then all the existing suppliers must either adapt to th enew technology, or get out of th ebusiness and make something else.

Mike January 23, 2012 at 9:42 am

The prices of everything generally settles within a certain percentage in a given area, unless one seller’s goods or service are generally perceived as better by most consumers.

From supermarket to supermarket, unless a sale is going on, wonderbread will be priced pretty much at the same level. If you buy it at a convenience store, you pay a bit more for the ‘convenience’.

Does this imply that there is a whitebread cartel? Perhaps the most evil, yet bland, cartel in the world?

Walt D. January 23, 2012 at 1:24 pm

Time to make another prediction. We will see gas prices in the $4.50 – $5 a gallon range this summer. No doubt, the usual idiots in Congress will be crying price gouging and looking to try and impose windfall profits taxes on the oil companies.
Question: If a cartel is already in place why don’t they increase prices now? Why wait until prices hit $5. Why are they doing us a favor? Why don’t they add another $1 or $2 markup? These days it is very easy to check on the internet or my iPhone where the cheapest gas station is. However, I usually end up filling up across the street at the Shell Station and paying 4 cents more because I do not want to drive 5 miles then wait 10 minutes.

Greg January 23, 2012 at 4:59 pm

I’d take this study with a huge grain of salt. They are using what is known as “agent-based modeling” (ABM). I took a graduate course on this last semester, and have formed a very negative view of the technique. You can create an ABM to show anything you want to show. I could create an ABM that shows raising taxes to 100% leads to utopia, or to disaster, whichever I want. The idea behind ABM is to try and explain how complex behavior forms by attempting to generate that complex behavior in a system. This type of logic is terribly flawed though, as there are many ways to generate a certain behavior (infinite, in fact), and finding one ABM that seems to generate the same set of data does not mean you have learned anything about why or how that set of data was generated in real world.

Of course, ABMs give politicians cover to do what they already want to do, so I think you’ll see ABMs used more and more as “evidence” to pass this bill or that. And when an ABM comes along claiming to show something that makes sense to you, be careful! Don’t fall into the trap of using that as evidence to support your world view. It is using fabricated evidence and no good scientist should ever rely upon that.

Daniel January 24, 2012 at 7:47 am

I believe that so long as freedom holds out long enough somewhere in the world, the rest of the truths like this one will be revealed and brought to the mainstream through careful thought and research. The question then becomes if we can hold on long enough.

Comments on this entry are closed.

Previous post:

Next post: