Has Industry Concentration Increased Price and Restricted Output?

That is part of the title of a new paper by Sharat Ganapati, here is the abstract:

American industries have grown more concentrated over the last few decades, driven primarily by the growth of the very largest firms. Classical economics implies that this should lead to hikes in prices, reduction in output, and decreases in consumer welfare. I investigate forty years of data from 1972-2012 using publicly available market shares and price indices for both the manufacturing and non-manufacturing sectors and find mixed evidence. Manufacturing concentration increases are indeed correlated with slightly higher prices, but not lower output. However concentration increases are correlated with increases in productivity, offsetting a large portion of the price increase. In contrast, non-manufacturing concentration increases over the last twenty years are not correlated with observable price changes, but are correlated with increases in output.

In other words, the output restrictions are not there.  The amazing thing is that, over the last few years, I have seen a few dozen journalists and also economists handle this question, without ever asking much less trying to answer this question (Noah Smith being an exception).


'In contrast, non-manufacturing concentration increases over the last twenty years are not correlated with observable price changes, but are correlated with increases in output.'

Why, Facebook's productivity has been staggering, right? Not to mention that of Google. whose ability to deliver ads has grown amazingly over the past 15 years.

I think at some point fb/google get broken up. They are incredibly efficient firms and we've seen others in that space falter (snap a little, twitter a lot), but they probably have some of the highest economic rents of any firms right now. I think a good argument can be made that most of their profits are due to pricing power (in an oligopolistic industry) versus being more efficient.

I have a friend with $1 million in sales in consumer products paying fb 20k per month in ads. He still makes money (looks like 100k), but I think a lot of small/medium sized businesses are paying out a lot of their profits to these firms. In essence for online advertising its mostly a duoply.

How much would there pricing power fall if fb was forced to divest instagram. Google was forced to divest youtube? they would still be highly profitable firms, but then you would significantly increase the ecosystem of firms competing for advertising dollars.

I don't think this happens today, but 3-5 years from now I think people seriously look at it.

I have no problems with capitalist making money on strong business strategies, but I think these firms may be making more off of industry consolidation and pricing power than being efficiently run.

All this means is that firms have been getting more efficient. As the firm gets more efficient, some of these gains are passed on in the form of non increasing prices and in the form of increased production. The measure of how much of the benefit is captured is the increasing profit margins of companies that is associated with increasing concentration of companies. However, its very possible that productivity would have been even greater and production even higher if these companies did not have market power. In fact, we have seen a slowdown in overall productivity! Might these things be related?


So, one might conclude there has been a great stagnation, actually? Sounds almost as if someone should write a book concerning the subject.

@Steven Wolf - makes a metaphysical argument (i.e., one that cannot be disproved) that I was going to make as Devil's Advocate.

Bonus trivia: "Metaphysics" literally means in Greek "after (meta)" "physics (physics)". Why is it called "After Physics"? Because Aristotle's (an ancient Greek scientist philosopher) book that deals with metaphysics was the book whose sequence number is the book just after Aristotle's book on physics. Hence the designation, which became synonymous with the subject matter itself. Simplicity.

How could one factor in all the variables that have affected output and pricing since 1972? For example, increased concentration means fewer firms which means the remaining firms produce greater output offsetting the reduction in output by the firms no longer producing. As for pricing, increased imports from low-cost firms in places such as China have limited the ability of firms here to increase prices even in highly concentrated industries. After PCs were introduced, there was an explosion in the number of producers but prices were very high whereas today there are only a handful of producers yet prices are a whole lot lower. As for non-manufacturing firms, how does one even measure pricing and output of firms such as Facebook and Google? And, of course, Amazon is sui generis, selling goods at prices that resulted in losses (or no profits) because the aim was volume not profits, any losses incurred made irrelevant by an ever rising price of Amazon's stock, the success of Amazon's strategy destroying many firms in the retail sector both large and small. It's complicated, isn't it?

Facebook and Google sell advertising. That makes them compete with newspapers,TV and radio.

But isn't advertising fairly easy to measure?

The author is trying to correlate concentration with prices. Concentration measures require you define a product and geographic market to measure concentration. Advertising is a good example: what is the concentration in the advertising industry. What is the relevant geographic market you are measuring it in.

This paper screwed up its data pretty badly, and I am surprised it was not picked up on review.

The author uses NAIC (census defined) industries to compute market shares. The author does not combine NAIC codes where those manufacturers produce the same output. So, if sugar beet manufacturers produce sugar and sugar cane processors produce sugar, and each has its own NAIC code, the production of each is not added to produce a sugar market because the author did not combine codes. The author did adjust for imports

NAIC codes are often based on manufacturing processes, not outputs.

Second, non-manufacturing industries are even more difficult because service industries are often local and NAIC data does not go to very local levels, i.e., counties, cities, etc.

Given this, it is a little surprising the author did not get just noise.

@Bill- sorry, but you are screwed up, not this Big Business is Good author. As to your first point, using NAIC codes is conventional and combining the codes will actually *further increase* output, since two categories is greater than one, making the author's conclusions even more valid. Second, and equivalently, adjusting for local non-manufacturing output that does not appear in the NAIC data would likewise broaden the net and capture even more output.

Do you want to debate me on why Koch Industries is humankind's friend? Challenge Accepted (SM).

Ray, Sorry, Ray, your comment makes no sense. Standard IO practice is to combine NAIC codes when the product is the same or similar. Peaches in a can are the same as peaches in a glass bottle; steel made with electrical arc furnaces is the same as steel made with a blast furnace, which is the same as steel made from scrap. When an author says they use straight NAIC codes, and thens cites in the bibliography papers which combine NAIC codes in their research, or earlier research which included SIC codes, you have to wonder. You also have to wonder when you have non-manufacturing codes (service industries) and use them for support, when many of them are collected at state, and not county, levels, you have to wonder as well. Furthermore, you can't use NAIC codes when the market is regional or local.

Regarding your Koch and other comments: I don't work backwards from a conclusion that you apparently want to make. Whether something is larger or smaller is irrelevant: what is relevant is that you have a relevant product and geographic market. You should take the evidence...and only good evidence...and follow that. Recognizing the deficiencies, many IO folks study INDUSTRIES, which will deal with consolidation of codes and substitute products and also imports, rather than just regressing unadjusted numbers.

Sorry Bill, but you failed to address my arguments. If A plus B is bigger than A or B alone, then trying to show that somebody has a monopoly in A or B alone would fail if there's no monopoly in A plus B and they are compliments, 'peaches' to use your example (the point of the paper and my point). If you know anything about antitrust, and how it works, it's all about defining the relevant market. If no monopoly is a big market, then no monopoly is the smaller market, given A and B are substitutes.

You must work for anti-Koch forces, dark forces? Who's paying you to post here Bill?

Ray, I know quite a bit about antitrust and defining relevant markets. Your comment makes no sense. If A and B are complements, you do not add them. Ray, c'mon. If A takes business from B, A and B are in the same market. Furthermore, your claim that adding A and B necessarily reduces market share is equally silly. If in A a firm has a Aprime has a 90% market share, and in B the firms each have say 30% shares, and NAIC market A is say 10 times bigger than NAIC market B, and A and B are together in the same market, after addition Aprime still has has a very high market share of the relevant market, market AB.

Ray, go read; maybe start at the Antitrust Division merger guidelines; look up some articles by Greg warden or Dave Scheffman, and quite pretending. NAIC is not a market. It's a starting point.

This seems to be a lose-lose for free market thinkers. If our understanding of large market actors is right, the 'cost' of doing more socialist things should be going down with increased concentration. If we are wrong about the effects of large market players, that should lower our certainty in a lot of related economic conclusions.

@KevinH - actually, 'what is good for General Motors is good for America' is an old Libertarian argument. If you read B. Doherty's meaty tome on Libertarian-ism (where TC makes a cameo), you'll see that libertarians would often debate that John Rockefeller's Standard Oil should have been left alone and not broken up by the government, since it was a sort of natural monopoly. The same argument has been made about breaking up AT&T, and to a degree history shows judge Harold H. Greene's decision to break up Ma Bell into Baby Bells was flawed, given the subsequent consolidation in that industry.

Do you want to debate me on why Koch Industries is humankind’s friend? Challenge Accepted (SM).

It doesn't hurt free market thinkers as much as you think. The most important part of the free market is that a firm can fail. Planning is actually easier at scale because you have more information and can control more of what is otherwise wasteful and costly uncertainty. But privately owned large firms and publicly backed or owned large firms are still completely different scenarios. There has been a lot of research in east asian economies that even a model with publicly championed mega firms work fine as long as the government allows multiple of such firms to compete and win/lose business based on commercial success.

So it hurts the free market thinker who thinks distributed decision-making is actually the key to capitalism. But if you understand that the decision-making is just as hard (or harder) for the firm as the government and the real benefit in the capitalism we see most commonly is simply market discipline, this kind of finding is not a problem at all.

Someone please help me with this line:

"However concentration increases are correlated with increases in productivity, offsetting a large portion of the price increase."

Increase in productivity and increase in price seem incident on different groups. In what way would they offset?

@Anonymous - correlation is not causation and you are confusing how the sentence was worded. The author is saying that since it's well known that "Big Business" is more productive (as correlated for example to patents issued) than Mom-and-Pop stores, then Big Business' increased productivity means they will produce more output for less input, which is good for society, and this offsets (to society) any price increase. You're smart enough to figure it out once you re-read the sentence given the above.

Do you want to debate me on why Koch Industries is humankind’s friend? Challenge Accepted (SM).

I considered that, but it too leads to incidence on different groups.

We know from past MR discussions that superstar firms enjoy greater market concentration, higher productivity, and pay higher wages, but to a small and shrinking fraction of the total workforce.

Thus more workers face higher costs without the benefit of working for such a superstar.

As an example, more low wage workers buy iPhones than high wage workers eat panini at Apple Inc.

(Koch Industries, I don't know.)

Was "panini" best there? There has to be a better Apple Cafeteria food symbolizing the affluence and ideology of the place .. organic soy sushi?

Interesting, but no article on 3% GDP in the 3rd quarter?

Google "3% GDP krugman" to see video after video of Krugman saying "it will not happen". Where are the rebuttals? Now that the yolk of obama's tyranny if slowly being lifted it looks like the economy is doing quite well.

Getting back to Krugman, why isn't this story being played 24-7 on the socialist business channels? Wait, there really is only one. Why isn't it playing the clip?

Oh I know, they are only talking about how the tech earnings of yesterday require us to break up americas most trusted brands. That makes sense. Liberals love to use their guns!

But, wouldn't one would expect more concentration in industries where marginal costs are very low. As is obviously the case with software, and industries largely based on the application of software. A larger user base produces more revenue, making it possible to spend more on development and marketing, producing a rising barrier to market entry ...

Software is a natural monopoly by product, not by company. Solidworks has a near monopoly because of file compatibility. Microsoft has a monopoly on spreadsheets but not on browsers.

Well the answer could be Yes and No:

1) Modern firms are better at managing and taking advantage of economies of scale.

2) In terms of tech concentration, we have to remember the marginal cost of a byte is next to nothing.

3) But the high concentration economy long run does limit not just competition but people are competing hard to get the best jobs which changes the age of family formation and size.

The industries that have seen the steepest increases in prices correlate much more strongly to being heavily subsidized and regulated industries than they do to industry concentration level (e.g. health insurance, higher education and telecommunications). In contrast, some of the most highly concentrated industries in the market today are also some of the most consumer friendly areas of the economy with low, competitive pricing (e.g. search engines, generic drug distributors, cloud services, consumer staples products).

I notice he picked the series from 1972, the Nixon shock an start of the new monetary system. So, he is finding a possible natural trend in monetary cycles. Follow the trend.

Government obligations have increased, imposing an increasing fixed cost on labor. Larger corporations can move labor within, without suffering the labor market uncertainty that comes from government wage taxes.

Manufacturing has seen increased international competition. Weak players exit or merge with more efficient firms until the rate of return equals the risk-adjusted market rate of return. Explains increase in price. Output increase because of an outward shift in demand curve due to increase in population, growing international markets, greater ability to price discriminate across markets, noise in markets has more effect if you have fewer players, Or something else, like new distribution channels (related to the ability to price discriminate.)

Non-manufacturing firms may have seen an increase in non-price competition i.e quality differences. Or a greater ability to price discriminate. Medical service prices are all over the place, and quality and services can vary. Data mining allows for greater tailoring of services. More effecient delivery channels can increase reaching new markets but competitive pressures keep down prices.

Or the noise in the data gives you misleading information .

Comments for this post are closed