How many sellers are needed for markets to become competitive?

by on August 29, 2017 at 1:34 am in Data Source, Economics, Medicine | Permalink

…competitive conduct changes quickly as the number of incumbents increases.  In markets with five or fewer incumbents, almost all variation in competitive conduct occurs with the entry of the second or third firm…once the market has between three and five firms, the next entrant has little effect on competitive conduct.

That is from Bresnahan and Reiss, “Entry and Competition in Concentrated Markets.

Part of their method is to compare doctor and dentist pricing practices across towns of different size, and thus across different numbers of providers.  Then they see where bigger numbers makes a difference in terms of pricing.  Plumbers and tire dealers are considered too.  One lesson seems to be that market concentration has to rise to very high levels to make a big difference in outcomes.

If you are wondering, the “sweet spot” for a town to have a single dentist or doctor is population between 700 and 900, at least circa the early 1990s.

1 Right Wing House Music August 29, 2017 at 1:58 am

Now let’s see the behavioral economists derive this result from first principles.


2 Jonathan August 29, 2017 at 8:59 am

Well, to the extent behavioralists and experimentalists are the same, they did so back in 1995. (This is one of my favorite papers, by the way, which has never gotten the acclaim it deserves, IMO. Underrrated!)


3 Bill August 29, 2017 at 10:09 am



4 prior_test3 August 29, 2017 at 1:59 am

‘Part of their method is to compare doctor and dentist pricing practices across towns of different size’

In or out of network, cash or credit? How did the prices from HMOs like Kaiser Permanente stack up? (For those unfamiliar Kaiser Permanente – ‘As of December 31, 2015, Kaiser Permanente had 10.2 million health plan members, 186,497 employees, 18,652 physicians, 51,010 nurses, 38 medical centers, and 622 medical offices.’ Their doctors are salaried, making it impossible to compare their pricing practices. ) Oddly, this web site just skips over things like for profit HMOs when talking about American health care, such as Kaiser Permanente’s rigorous pursuit of preventative health care to reduce costs over the longer term (you know – HMO, health maintenance organization). Though if the past is any guide, as state employees, it is likely one or both are Kaiser Permanente members. As long as it is accepted that one’s health care provider is Kaiser Permanente, and not what you go out on the free market to purchase.


5 Matthew Young August 29, 2017 at 2:06 am

The post verifies material we already know.
Most of the gain comes with the second sample as the second sample reveals a band limit and the sigma of the process, it is what gives the deja vu effect. The third adds portions to the skew of the distribution.


6 jdap August 29, 2017 at 8:51 am

…why then did Tyler now choose to post this quarter century old study here?

why is this “study” suddenly significant in Tyler’s view of the universe?


7 Bill August 29, 2017 at 9:09 am

He went to the dentist yesterday and didn’t like the bill?

It’s ok with me. The paper is timeless and illustrates some truths.


8 JFA August 29, 2017 at 9:22 am

Tyler has been talking up this paper for years (just take a look at his IO syllabi). He wants to give context to the recent claims about market concentration.


9 Ray Lopez August 29, 2017 at 9:54 am

Right. And not only that, for years the US DOJ Antitrust department had an informal rule that they would not use certain antitrust laws unless the market concentration dropped below 33%, that is to say, you had three or less competitors in a field. Of course that was an informal rule from what I understand, not something on the law books, but it did seem to hold up when the DOJ did a “rule of reason” analysis (they would define the market so a RoR analysis applied to the market having three or less competitors; they would try to define the relevant market as having three or less competitors, not more).


10 John Hamilton August 29, 2017 at 2:50 am

My wife is in dental school, and her sister just graduated this past spring. The professors tell them the money is in the small towns, since you’re the only dentist around. The salary offers of other newly-minted dentists I know also confirm this. Not much has changed in this respect since the early ’90s.


11 Antitrust in God August 29, 2017 at 3:36 am

Do we no longer read the Antitrust Paradox because Robert Bork didn’t make it onto the Supreme Court (even though its cited in DOZENS of supreme court cases)?

Bork argued that the original intent of antitrust laws as well as economic efficiency make consumer welfare and the protection of competition, rather than competitors, the only goals of antitrust law.[3] Thus, while it was appropriate to prohibit cartels that fix prices and divide markets and mergers that create monopolies, practices that are allegedly exclusionary, such as vertical agreements and price discrimination, did not harm consumers and so should not be prohibited. The paradox of antitrust enforcement was that legal intervention artificially raised prices by protecting inefficient competitors from competition.


12 Steve Sailer August 29, 2017 at 4:53 am

Seems like Bork largely won.

I took a class on economics of anti-trust at Rice U. in the late 1970s that was very Borkian. The prof considered himself very dissident and cutting edge.

Twenty years later, we had ExxonMobil.


13 Bill August 29, 2017 at 8:22 am

I would argue that Bork channeled Aaron Director and George Stiglitz, that Oliver Williamson and others had changed vertical enforcement and theory well-before Bork’s book (Antitrust Division Attorney General’s Don Turner and Tom Kauper wouldn’t consider vertical cases unless they had a horizontal effect (even Bork would acknowledge this when he represented Netscape against Microsoft). What Bork did was popularize and put in English where vertical theory had gone and been accepted by most IO economists.

Bork is so old fashioned. He was before the use of game theory and strategic analysis and before the fuller development of network economics. But, it was a good critique of the 1950’s and vertical theory.


14 Bill August 29, 2017 at 9:05 am

In support of my point above–that Bork came after many of the changes had already occurred in economics and enforcement–I offer the following:

“Part of the explanation for Bork’s mischaracterization of the state of anti- trust law may lie in the delay between the book’s conception and its realiza- tion. Bork drafted his book in the late 1960s, but did not actually publish it for almost a decade. At the time that Bork began writing his book, the per se rule was ubiquitous and merger law was extremely pro-government.9 But much had changed in the ensuing decade.10 Because these changes undermined the premise of his book, Bork downplayed them. For example, Bork invoked Schwinn11 to prove that antitrust has run amok. But the Supreme Court in Sylvania12 had explicitly reversed Schwinn before Bork published The Anti- trust Paradox.13 Bork’s persistent attacks on Schwinn smack of scalping tick- ets to a concert that happened years ago. Sylvania was not an isolated case. By the time that Bork actually published The Antitrust Paradox, antitrust law and the relationship between the antitrust agencies and business interests had already begun to change.14 As Bill Kovacic has noted:
In important respects, the antitrust environment Bork described in The Anti- trust Paradox fit 1969 more closely than it did 1978. The stark portrayals of a deconcentration-minded Congress, an economically backward Supreme Court, unconstrained government enforcement agencies, and a lethargic, overmatched business community would have been more convincing had the book appeared five years earlier. Measured against Bork’s ideal vision of antitrust policy, the world was less dismal, and its outlook less discouraging, than The Antitrust Paradox indicated.15”

Here is the link:


15 Steve Sailer August 29, 2017 at 5:11 am

An interesting article I heard about from MR on the work of economist John Connor who makes a bundle as an expert witness on international cartels:

It was widely believed by economists that international cartels were unfeasible, but the Archer-Daniels-Midland lysine cartel, as in Soderbergh’s “The Informant!”, opened Connors eyes.

He keeps his lucrative database of international cartels secret, however, to preserve his monopoly power in his expert witness business.


16 Tom T. August 29, 2017 at 8:46 am

Plus he’s working to stop the creation of Skynet.


17 Adam August 29, 2017 at 2:15 pm

@Tom T.

Thanks for the lolz.


18 Bill August 29, 2017 at 10:08 am

So, you want John to dedicate his work to the public. Booz Allen, McKenzie, and Lexicon should do so as well.

Would you like to share your hard drive with me?


19 August 29, 2017 at 5:24 am

Previously I asserted that the locally sustainable number of medical doctors most probably preferred by the local medical doctor association in the developing and developed world is 1.76/K, i.e. 1 md per 568 population. Social or govermental pressures might pushed that to higher values by bringing in foreign born doctors.

More data from the TLA WFB which includes least developed countries and gives the global average NMEDK to 1.42/K, i.e. 1 per 704 population. According to the WFB Cuba is known to produce many medical doctors and is rank at number 2 at 7.2/K. You will not believe that the top rank is CongoDR at 9.0/K. Sudan came in at number 39 at 3.06/K. US came in at no. 58 with 2.55/K, i.e. 1 md per 392 pop. Liberia came last at no. 182 with 0.01/K, i.e. 1 md per 100,000 population, omitting rounding off error.


20 Matthew M August 29, 2017 at 8:03 am

That’s partly what made Liberia so vunerable to Ebola. After their most recent civil war most medical staff in the country were from international organisations like Doctors Without Borders. I believe they started pulling out a couple years before the epidemic, as the country appeared to be stabilising, against the wishes of a lot of the staff on the ground there. When the virus hit what little medical staff were left in the country bit the dust and it spread uncontrolled until MSF came back and essentially became the medical infrastructure for the country again.


21 Larry Siegel August 30, 2017 at 3:15 am

You’re right, I don’t believe that nearly 1 Congolese in 100 is a doctor. This screams “data error.”


22 Alistair August 29, 2017 at 5:41 am

Wouldn’t we expect this strength of result from Shapely analysis? Or several other basic microeconomic methods?

Also see Elon Musk vs The ULA. File under “a little competition goes a long way”.


23 rayward August 29, 2017 at 6:36 am

Doctor and dentist pricing: Doctors don’t charge the price patients are willing to pay, doctors charge the price third party payers are willing to reimburse. And that applies to an increasing percentage of dentists as more patients have dental insurance. Here’s an anecdote to emphasize how reimbursement affects pricing. Medicare does not cover routine eye care and not that many have health insurance that covers routine eye care. To attract insureds, many Medicare advantage plans cover routine eye (or dental) care. So why don’t all seniors enroll in Medicare advantage plans? Because they are managed care plans, meaning that seniors with above average health issues may be denied treatment that Medicare will cover, the consequence of which is that Medicare advantage plans typically attract a healthier pool of insureds. The healthier insureds in Medicare advantage plans accept the risk in return for coverage that Medicare doesn’t provide (such as routine eye care). But here’s where pricing comes in: the prices of routine eye care charged by optometrists vary widely, depending on whether the optometrist’s pool of patients has a Medicare advantage plan or health insurance that covers routine eye care or not. When I say widely, I mean the difference between $100 and $1,000. Pricing in health care has little to do with competition, unless one includes (for example) the techniques used by Medicare advantage plans to attract a healthier pool of insureds.


24 Evans_KY August 29, 2017 at 6:39 am

Many things have changed since 1990. Most notably, the rise of the HMO and direct billing. Not to disparage the authors, but is there anything more recent on this subject?


25 TMC August 29, 2017 at 8:24 am

The paper is about competition in general. Maybe older medical studies are the ones to use as competition in that area has gone down.


26 john August 29, 2017 at 7:37 am

It’s not really the numbers that matter I suspect. What matters will be the cost of the pricing information — so if it were easy to get full price information on the 100 dentists in fairfax couny (probably underestimating grossly) then that implies increased compitition as it’s more likely that at least one will be trying to get more customers based on a price advantage. If that information is clostly then that one of the 100 will likely be missed by many so tht dentists pricing will have little impact on the “market” price.

Clearly a case of one is a different industry setting — monopoly. In this case it’s the cost of going to some other market and the incofmaiton costs that will probably infuence the affect the number of porducers has on prices.

Exactly how well focusing on number does in sheading light on the question I suppose remains to be seen. Ultimately I suspect it masks as much as it illuminates but then that will depend on just what question one is attempting to answer.


27 rayward August 29, 2017 at 8:00 am

Speaking of supply, Texas will need a significant increase in the supply of labor to rebuild. “There aren’t a lot of qualified, idle construction workers.” Where will the increase in supply come from? I suppose Trump will have to build his wall in his second term. Necessity is the mother of invention.


28 Tom T. August 29, 2017 at 8:50 am



29 msgkings August 29, 2017 at 1:55 pm

Yes, Trump is pretty obsessed with that wall.


30 Larry Siegel August 30, 2017 at 3:17 am

The supply will come from Mexico and Poland, as usual. If we don’t let in immigrants, maybe in 30 years we will have grown a native supply of construction workers, but by then Houstonians won’t need them any more.


31 Bill August 29, 2017 at 8:32 am

Tim Breshnahan is an excellent IO economist

The paper covers more than dentist: it covers doctors, dentists, plumbers, drug stores and tire dealers.

If you want to understand why some counties in the US have few health insurance carriers (who rely on provider discounts for their offering) you will understand why some counties have no carriers willing to include those counties in their exchange offering.

It would make an interesting research project to use the Breshnahan doctor data and correlate that with counties which have one or fewer carriers offering a product on the exchange that covers that county.


32 mm August 29, 2017 at 9:58 am

in expensive medical devices I have been told by suppliers the price drops when the 3rd firm enters the field. With 2 suppliers they each try the unique advantage gambit- but with a 3rd firm price becomes paramount.


33 clayton August 29, 2017 at 9:58 am

I worked in a large industry where the insolvency (due to side bets) of the fourth major player let to a quadrupling of the gross margin of the primary product. A counterexample to this study.


34 Bill August 29, 2017 at 10:06 am

Explain. If a firm left, that would explain the margin increase but that wouldn’t be a counterexample. Also, are you analyzing competition in a market where one of the players, about to go bankrupt, is pricing either very low (below marginal cost) or very high (and not getting sales); either way, the before/after comparison doesn’t mean much because you could not say it was in equilibrium. All in all I don’t understand, so if you would elaborate that would be helpful.


35 Harun August 29, 2017 at 3:52 pm

Quadrupling sounds big. But it could be a 1% margin becoming a 4% margin.

Also, sometimes a market has crazy competitors who truly are not very rational. We had a competitor who would have tiny margins if not losses, and then he’d go bankrupt, have his family sell off a million dollar piece of land, and then go right back into loss-making business.

China is rife with companies who quote at cost and hope they can figure out how to make money later. Its frightening.


36 Bill August 29, 2017 at 5:29 pm

Harun, Are you speculating for clayton.. Only he can answer.


37 clayton August 30, 2017 at 10:52 am

“…once the market has between three and five firms, the next entrant has little effect on competitive conduct ..”

My example had three firms but showed significant anti-competitive behavior upon losing the fourth. The industry is the travel stop industry, when in 2008 Flying J filed for bankruptcy due to losses in its refining and pipeline unit. Immediately upon Flying J entering financial difficulty in the summer of 2008, gross margins on a gallon of diesel went from roughly 7-8 cents to 30 cents. Pilot purchased Flying J out of bankruptcy, leaving Pilot-Flying J, Love’s, and Travel Centers of America as the three large players in the game. Those high margins remained at least until I left the industry in 2015.

Pilot had significant legal trouble during this time, took on significant debt to buy Flying J, and the owner purchased the Cleveland Browns — it had no interest in competing on price. TA (publicly traded) was teetering on bankruptcy due to the onerous rent payments it paid to the REIT that owned its locations (the firm was split into a REIT and an operating company when it went public).


38 Hazel Meade August 29, 2017 at 10:54 am

I would apply this to increased competition among ISPs.
If we can get one or two major cable operators to enter markets in other parts of the country, then we can increase the number of competitors from 1-2 to minimum 3.
Right now I can choose from Comcast or Verizon (not counting a few smaller, low bandwidth options).
But if Cox and Time Warner expanded into the area, there would be four major ISPs to choose from.

The policy questions should be why the competitors in the regional markets remain segregated instead of expanding into each other’s territory and perhaps what policy changes would lead to that happening.


39 Adam August 29, 2017 at 11:58 am

Which is consistent with modern merger enforcement.


40 James Babcock August 29, 2017 at 1:27 pm

Seems like this might differ significantly between commoditized markets versus markets with more highly differentiated products, with the high-differentiation markets acting like multiple smaller markets with fewer firms each.


41 James Anderson September 1, 2017 at 1:49 pm

Frankly speaking, I care very little about this, as I believe the most important aspect is how we can make profits. As that’s the most important aspect for me and nothing else matters. Luckily, I am doing well which is a lot to do with broker like OctaFX since they are awesome having lowest possible spreads from 0.1 pips, high leverage up to 1.500, zero balance protection and then there is even 50% deposit bonus that’s usable, it’s just fantastic!


Leave a Comment

Previous post:

Next post: