*The Great Reversal: How America Gave Up on Free Markets*

That is the new book by Thomas Philippon, and perhaps the title is a bit misleading, as the book covers both regulatory barriers and natural economic forces behind higher concentration levels.  I am a big fan of Philippon’s work, but I am not so convinced by his arguments in this book.  Most of all, he is trying to argue for systematically greater monopoly power in the American economy, but he is reluctant to provide much evidence for output restriction, the sine qua non of market power.

First note that market power does not seem to be up at the level of actual market competition.  And capital’s share of income does not seem to be rising in a manner consistent with the monopoly theory, see here and here.

I agree with him about health care, and also (highly regulated) cable television and thus internet connections.  I agree with all of his suggestions for removing regulatory barriers to entry, for instance by allowing foreign airlines to serve domestic U.S. markets.  From a policy point of view, I am quite close to his perspective.

But when it comes to monopoly power too much of his evidence is circumstantial. OK, there is greater stability for market leaders in many sectors, and weak investment aggregates, but all the time antitrust suits find evidence for output restrictions — so why doesn’t this book offer more of such evidence?  Here is one passage (p.39) that caught my attention:

…we see a sharp increase in concentration in the airline industry after 2010.  That is enough to trigger our interest, but not enough to conclude that competition has weakened.  We must first check that concentration has also increased at the route level.  We find that it has.  We can further show that it came together with higher prices and higher profits.

I have only a pre-publication copy, and perhaps some of the book is missing in my edition, but I don’t see the cited evidence presented, nor is it in the airlines section starting on p.137 (which does document increasing concentration at the national level).  To consider the contrary evidence, here is an excerpt from an earlier MR post:

As for output restrictions, here is the DOT series on aggregate miles flown.  No doubt, there are problems around the time of 9/11 and also the Great Recession, with 2008-2012 being a period of slight quantity contraction.  But in 1985 there were 275,864 [million] total miles flown, in 2006 it was 588,471, and 641, 905 in 2015.  I’ll ask again: if there is so much extra monopoly, where are the output restrictions?

Or look at the price index.  Overall prices are down considerably since 2008, and from about 2000 to 2016 they run from about 250 (eyeballing) to about 270, noting 1998-2010 saw a huge run-up in oil prices.

Since I wrote that post there is clearer evidence for a steady price decline since 2012 (he is claiming higher concentration since 2010), just look at the price index, which is FRED channeling BLS.  Now maybe those are the wrong numbers for some reason, but I don’t see anything in the Philipson book to counter them.  I don’t see output restriction considered at all.  I don’t see a price series presented at all.

That is only one sector, but it reflects my deeper worries about the book.  I just don’t see the evidence for output restrictions, or, in many cases I don’t see the evidence for higher prices.

The most sustained discussion of prices comes on pp.114-122, where it is shown that PPP-adjusted prices are higher in America than in Europe, and furthermore the gap is growing.  That is far too much aggregation for my tastes (“Europe”), PPP adjustments are not exactly scientific, it is not very direct evidence for market concentration being the culprit, and furthermore if I understand him correctly, the Big Mac index also has the United States becoming relatively more expensive, even though McDonald’s clearly has faced massive competition in recent years.

To be sure, if you believe in a productivity slowdown, as I do, you also have to feel that America’s economic sectors, in some counterfactual sense, could be much more dynamic, more prone to disruption, and yes more competitive.  It is a great disappointment to me that is not the case.  But that is far from the view that monopoly power is increasing in the American economy in an economically significant manner, across a wide variety of sectors (health care caveat noted, and even that is selective, as there has been a significant cost slowdown).

So I remain skeptical about the main claims in this book.


'but he is reluctant to provide much evidence for output restriction, the sina qua non of market power'

Somehow, one wonders if John Rockefeller would agree. Or Thomas Edison, for that matter. Of course, neither of them were economists.

' But that is far from the view that monopoly power is increasing in the American economy in an economically significant manner'

The interesting thing is that a dramatically better word would be 'cartel,' not monopoly, but then, that would be based on Adam Smith's observation of how business works.

'So I remain skeptical about the main claims in this book.'

Well, it is not a love letter to B-B.

I did not see anything that would support the headline. Clearly over the last 60 years most countries have taken advantage of America's effort to have free trade. When you are the only country trading "freely" all the other countries benefit. Finally America is putting some teeth into free trade and demanding that the trading nations walk the walk and the pundits suddenly discover that there is no free trade.

'and the pundits suddenly discover that there is no free trade'

Dean Baker never gets any attention, does he?

Shouldn't the marginal product/service price matter more than output when analyzing these monopolies? In the broken markets of healthcare, cost is extravagant while output is somewhat steady.

"Output" is increasing with costs. Prices of "outputs" have not increased overall much in decades beyond overall cost of living.

Subsector outputs, eg some drugs, have seen price increases driven by rent seeking, eg diabetic drugs, but most drug cost increases are due to output increases with Purdue Pharma paving the way for many providers increasing outputs by a factor of five to ten in about a decade for a generic drug by function/substitutes and pricing.

Medicaid expansion in Kentucky increased outputs by about 50% in Kentucky with clear benefits in employment increases beyond health care employment.

Kentucky before Obamacare had a huge amount of output restriction. Tens of thousands of people were denied outputs they needed to be healthy enough to function in society as workers, students, consumers.

Contrary to the argument. Implicit in conservative price theory, entire communities of poor people with no government welfare benefits did not result in health care outputs in those communities falling in price to near zero, the maximum price most could afford.

Few workers will pay to work at zero pay. Ie, few doctors will pay for supplies in order to treat patients and charge nothing. The only doctors practicing in those communities had two costs, the cost of printing prescription pads, and a store front where he charged $20 to write a script for 60 oxy which was filled for $30 and then half the pills sold for $5 each and the rest used to feed the near zero cost output by the doctor: the price of the paper plus the time to look at the "patient" and judge him not DEA and then fill in the form, and exchange for the $20. The pills cost the drug company a penny to make and maybe 10 cents to deliver to the drug store for 35 cents who sold them for 50 cents each. The output restrictions on such "health care" have not had much impact on total health care outputs.

The treatment of opiate addiction is severely output restricted. The number one constraint is treatment can't cost zero, so the price can't be low enough for more than 1% of addicts. The only way to increase outputs to treat even half of problem addicts is by government payments for more outputs, which then runs into the skilled worker and facility supply costs.

Health care costs in the US do have high operating costs related to all the efforts to cut outputs by increasing prices as an economic policy. Eg, higher copays, deductibles, are intended to cut outputs to treat, for example, essential hypertension. Ie, by cutting doctor visits and prescription drug outputs, the result is higher ER outputs and uncompensated hospital bills. To cut those outputs, hospitals have been closed in communities, generally closing ER, but that simply forces people to vacate such communities for cities where paid for outputs can support unpaid for, but very costly outputs.

Conservatives want output restrictions, but refuse to demand creative destruction, euthanizing the people who can't pay costs and selling body parts for scrap. Instead they point to EMTALA which mandates outputs at zero price no matter the costs.

Isn't the issue with using prices to assess if an industry has become less competitive the knock-on impact of cheap Chinese capacity?

Even with output restraints, you would expect some price reduction/volume growth from any imperfect monopoly, passed on from the lower cost of employee living driving down wage rates.

Since we can't see the counterfactual of how far prices might have fallen/output risen in e.g. airlines had there been no consolidation or restraints, looking at price/output moves without that reference point leaves us in the dark.

Has anyone tried looking at returns on invested capital over time as a proxy for competitiveness? Since competition should bid these down over time, no?

I would spell it 'sine qua non'

Well, that would probably represent an output restriction, and an attempt to monopolize correct spelling.

Meaning that if we simply redefine GDP measurements (see previous posts), it is clear that spelling mistakes do not represent anything to be worried about.

Q.E.D. or QED, as the market for other Latin terms has yet to be threatened by monopoly power restricting output.

And now, so would Prof. Cowen, proving that there is no monopoly on correct spelling.

Went into McDonalds last week for the first time in a while and was shocked at how much prices had gone up. Breakfast burritos were over $2.50 each and a breakfast meal was over $6. When did the dollar shrink to just drinks? Don't know if this is because of a lack of competition or monopoly power but I doubt that the Big Mac index showing price increases in the USA are inaccurate (or is it the case that incomes have gone up that much over the past few years that on a PPP basis it's still cheap? I can't believe that as the price appears to have doubled at least).

Fast food has been shockingly expensive for some time. I've discovered that part of the difference is price-discrimination* but knowing how to order just makes the cost reasonable, not cheap.

* The cheapest items are no longer displayed on the menu, you have to know what to order. Also the good deals tend to be restricted to the app.

I think I was so accustomed to really cheap fast food that now it's a shock to pay over $5 for breakfast. To add to my comment above, nothing on the $1 or $2 menu in floor is food- it's all drinks- which defeats the purpose of it. Panera as well you can't get out of there under $8. And remember this is Florida, one of the cheaper states in the country.

Wages have been increasing.

"We can further show that it came together with higher prices and higher profits."

1/ choosing 2010 as a starting point for airline prices - coming out of the Great Recession - is what's doing the work here, the long run story of lower real airfares has largely continued and first class fares are lower, too.

2/ the drivers of increased airline profits are (1) lower fuel prices, and (2) richer co-brand credit card deals. Consolidation has improved airlines' bargaining position vis-a-vis banks more so than consumers.

The airline industry has certainly consolidated, creating stronger players that aren't serial candidates for bankruptcy. But what's the theory here on how many competitors exist in a claim about monopoly? Delta, United, and American are all about the same size (United has the most seats, American the most planes, Delta the most revenue in the latest quarter) while Southwest Airlines carries the most passengers. And hasn't Alaska's acquisition of Virgin America just made the Seattle-based airline a more robust competitor in more markets, rather than reducing competition?

Moreover it's the ultra low cost carriers - Spirit, Frontier, Allegiant - that have been the driving forces in the U.S. airline industry.

Unfortunately foreign competition on U.S. routes (or foreign ownership of U.S. airlines) is an insufficient remedy because the scarce resource is gates and in a few cases slots at major airports.

Government-owned and run airports effectively grant perpetual property rights in gates and slots to incumbent airlines, locking out competitors. Indeed a driving force behind Alaska's Virgin America acquisition was access to congested airports where they couldn't replicate themselves.

Air travel is not an industry I would choose to defend as being competitive, or even acceptable; indeed, on a scale of 1 to 10, with 10 being excellent, I'd put air travels in a low single digit. Besides America's only manufacturer of airliners producing a defectively designed aircraft and thinking they can get away with it (Boeing knew the design was defective - that's why they installed the sensor to prevent stall), the experience of flying for the customer is terrible. No, I'm not referring to death from a crash, but abuse. Cowen's misplaced obsession with output restrictions as the only indication of monopoly power actually fits the air travel industry: Cowen wouldn't know it since he mostly travels in and out of large cities with lots of flights and airlines, the bulk of Americans are stuck with smaller airports that have been mostly abandoned by the airlines. And even if those smaller airports haven't been abandoned, the prices have exploded. Try flying in and out of a smaller city, and I suspect Cowen will observe output restrictions.

Just to make the point close to home, for many years I flew weekly between two medium-sized (about a million population in one, several million in the other) sunbelt cities. In the beginning there were 14-15 direct flights per day with service offered by three or four airlines. Today, there are none. Today, I have to fly north in order to go south, trudge my way through the airport, wait for the connecting flight, board, and fly to my destination, a process that makes flying grossly inefficient and unpleasant - and takes forever even without the usual delays. It makes more sense to drive. Or it did. All those passengers who used to fly between the two cities now drive, a drive that at one time took about five hours now takes seven to eight hours because of the increase in traffic. The drive is such a horrible experience I would be tempted to fly in a 737 Max. Tempted, but I don't have a death wish.

"But in 1985 there were 275,864 [million] total miles flown, in 2006 it was 588,471, and 641, 905 in 2015."

Backing up Rayward's point here, I have relatives in Nebraska and my wife has relatives in Oklahoma. I've noticed recently that while there used to be one or two daily direct flights from Dulles or National to Tulsa and Omaha (on airlines that no longer exist), it is now necessary to book connecting flights.

Not sure how useful increased mileage is as an indicator, if some large portion of those miles are due to an increase in circuitous connecting flights (Dulles to Denver to Tulsa!) for the same trips that used to be direct.

True, the costs seem to be about the same as they used to be, but the relevant output here is not "miles", it's convenience of the trip, which has diminished considerably for consumers, though presumably works better logistically for the consolidated airlines.

They don't have trains or buses in America anymore?

Trains account for < 1% of inter-city passenger miles and I'd wager bus service is similar. There's only one notable carrier in passenger rail and only one in bus service, though I think there may be scads of modest charter services.

Greyhound is the only scheduled US nationwide bus carrier, so probably a fair proxy for bus traffic. They carry about 16 million passengers 5 billion passenger-miles a year. There are also a bunch of local charter companies (away games, casino trips, tour groups, etc.).

In 2017, domestic airlines carried about 750 million passengers for 684 billion passenger miles.

So busses accounts for 2% of the passengers and .7% of the passenger-miles of domestic travel.



It's been a very long time since I was on a train or intercity bus in the US. But then I started driving at 15 :)

There is very good bus service between the Washington DC area and New York City. My wife uses it all the time to visit her mother in NYC. We went up in May for the 95th birthday and it took just under 3 hours. Of course it is subject to the vagaries of traffic but for us it is as fast as the train.

A former top corporate lawyer (a long-time partner at Simpson Thatcher no less) has written an essay in which he concludes that under current law corporate executives “are legally obligated to act like sociopaths” and suggests that corporations be required to “adopt a binding set of ethical rules, approved by stockholders and addressing the key ethical dimensions of corporate life”. This isn't some Harvard Law School professor, but a career defender of corporate America. https://www.nytimes.com/2019/07/29/business/dealbook/corporate-governance-reform-ethics.html

Given that companies have no human parents, children, spouses, political representatives or, it is not clear in what sense they should be expected to behave as social actors. If they are, then they should have formal political representation to enfranchise them in the system that legislates for society, shouldn't they?


'Corporations are people, my friend', Mitt Romney, 2012

Oh, and there are almost 12,000 registered lobbyists in DC

Read much?

CEOs vote Republican. It all makes sense now.

In the modern political landscape there really is no one who is for free markets all the time. The left puts their finger on the scale to favor solar, the right coal.


Still, I'd say if you are going to monkey with markets, do it for progress.

Progress towards hell?

That is beautifully inverted. How many images of hell are of a smoking inferno? How many include crisp solar panels?

All of Philipon's analysis of "competition" is based on national concentration ratios for Census NAICS "industries." He won't accept that Census "industries" are NOT antitrust markets. Well known to serious IO economists, which he is not. See https://one.oecd.org/document/DAF/COMP/WD(2018)46/en/pdf.

Even for airlines, Werden & Froeb report data showing that "the airline industry experienced great change since deregulation in the late 1970s, including many mergers, but examinations of market concentration—at the route level—have not found systematic increases. Over the period 1984–90, the weighted average route-level HHI on domestic U.S. routes decreased slightly. During 1995–2009, the HHIs “for the largest 1000 short-, medium- and long-haul routes revealed a general downward trend in concentration.” And for 2007–12, “a slight reduction in concentration in the highest-traveled markets” was observed." See https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3156912. Philippon must be more serious and careful in his analysis.

Comments for this post are closed