Streetwise Professor

January 12, 2019

A Great Passes: Harold Demsetz

Filed under: Economics,Regulation — cpirrong @ 1:40 pm

Last week, the great economist Harold Demsetz passed away at age 88. Harold (whom I knew slightly) was truly a giant, who made seminal contributions to industrial organization, property rights economics, transactions cost economics (especially his early recognition of the bid-ask spread as a cost of transacting, well before it became a focus of research in finance), information economics, and the theory of the firm.

He also coined the memorable phrase “nirvana economics,” which skewered the then-prevalent (and alas, too often currently prevalent) tendency to compare imperfect market outcomes with perfect ones, soon followed by a prescription for regulation to correct the “market failure.” He noted–and this can not be emphasized enough–that the true “relevant choice is between alternative real institutional arrangements.” That is, there are government failures too, and it is necessary to evaluate those in order to make policy choices. Nirvana is not a choice.

Like many great economists of his era (e.g., Coase), Demsetz’s work was literary rather than formal, but that definitely does not mean it lacked rigor. Demsetz wrote well, and could present tightly reasoned and impeccably logical theoretical arguments without resorting to a single equation. His article on entry barriers is a great example of this. There was a great deal more economic logic and insight in a typical Demsetz paper than in the typical modern densely mathematical work.

Demsetz’s biggest contribution to my economic education was his work that confronted, and largely demolished, the prevailing structure-conduct-performance paradigm in industrial organization, and the related empirical work on the relationship between industrial concentration and profits, which concluded that a positive correlation was the result of market power. End of story.

Demsetz demonstrated (as Sam Peltzman formalized shortly afterwards) that cost-concentration correlations could give rise to profit-concentration correlations even in the absence of market power. A simple story that illustrates the point is that a firm that experiences a favorable cost shock when its competitors do not will expand at their expense, and earn a profit commensurate with its greater efficiency. This tends to increase industry profitability and concentration.

Demsetz also showed in a famous paper (“Why Regulate Utilities?” that structural monopoly (e.g., a “natural monopoly” due to extensive scale economies) does not necessarily convey market power. Further, in
“Industry Structure, Market Rivalry, and Public Policy” he argued that competition for the market could be quite intense, and even thought it might result in a firm obtaining a large market share, (a) the firm’s ability to exercise market power might be limited, and (b) competition for the market could be an engine for progress, including notably product and process innovation.

In this work, and that of his contemporaries primarily at Chicago and UCLA, Demsetz undermined the prevailing paradigm in industrial organization, with its simplistic equation of market structure and market power. This resulted in a revolution in economic science, but also public policy, and in particular antitrust policy.

Alas, there is a counter-revolution afoot, and quite depressingly Chicago is one of the leaders in this. In particular, Luigi Zingales and the Stigler Center (!), and the Center’s Promarket blog, are among the leaders in resuscitating the notion that concentration is per se objectionable, and creates market power. In my perusal of this literature, and the voluminous writings about public policy it has spawned, I find no real intellectual advance, and indeed, perceive severe retrogression. In particular I find little effort to confront the Demsetz critiques (and the critiques of others that followed).

It is very sad that Harold Demsetz passed, although after a long and very productive life. It is sadder still that many of his most perceptive insights predeceased him.

Print Friendly, PDF & Email


  1. Indeed a great loss to the world. I contend that the cyclic nature of [economic] theories will see a return of many of his ideals and economic philosophies, despite their having predeceased him. Sometimes in the demise of one creates an inertia in those left behind to explore and expand on little known works of the departed.

    Comment by John — January 12, 2019 @ 3:23 pm

  2. David Henderson wrote this great tribute to him in today’s WSJ:

    Comment by Tom Kirkendall — January 13, 2019 @ 6:23 pm

  3. Armen Alchian passed in 2013, and now Harold Demsetz. Neither received a Nobel Prize, but both made economics immeasurably richer. RIP

    Comment by DrD — January 14, 2019 @ 10:39 am

  4. …thoroughly enjoy your blog and classic Chicago analytical approach. Keep up the good work

    Comment by Bill Coorsh — January 14, 2019 @ 3:36 pm

  5. “Neither received a Nobel Prize”: nor did any other economist.

    Comment by dearieme — January 15, 2019 @ 1:40 pm

  6. @Bill–Thanks. I’m doing my best to keep the tradition alive.

    Comment by cpirrong — January 15, 2019 @ 3:39 pm

  7. A brief but comprehensive tribute. Thanks for this.

    Comment by Global Super-Regulator on Lunch Break — January 16, 2019 @ 2:35 am

  8. Quite a bit of the game-theoretic IO of the 1980s and 1990s represented efforts to show circumstances where the Chicago/UCLA school’s generalizations did not apply, and so antitrust or regulation could make things better. The difficulty for this literature was that these models required knowledge inputs not available to courts or regulators (or frequently to the players themselves), often depending delicately on various assumptions about the order of moves, what variable (e.g. price, quantity, quality, etc.) was to be treated as “committed” in a Nash equilibrium sense, details of information transmission, the relative shapes of different prior distributions of types, etc.

    One response to this conundrum of the delicacy of the new theory is represented by John Sutton, who realized that accepting lower-precision predictions could allow for more robust conclusions and considerable descriptive insight. His work, however, did not create a rationale for government intervention and, IMO, is one reason why his attempted reformation of old-school cross-industry industrial organization didn’t excite the field.

    Another response to this problem was faith that smart people could figure it out on a case-by-case basis. I was at one conference where a very smart I.O. economist presented some results showing that various interventions in a particular market could either make things better or worse, depending on a number of inherently unobservable parameters. He concluded by emphasizing that he was not advocating “policy nihilism,” whereupon I spoke up and stated that I wanted to put in a word in favor of such nihilism, which did not go over particularly well. But now I see that members of this school often possess considerable common sense, as with Carl Shapiro here (in Bay Area opposition to the Zingales et al zealotry–how things can turn around!):

    Comment by srp — January 16, 2019 @ 6:08 pm

  9. @srp–Thanks. Your point about the 80’s/90’s models is spot on, and one I’ve made in other contexts. These tended to be contrived models engineered to produce a particular result. They were completely lacking in robustness–for precisely the reasons that you mention, namely, their sensitivity to the fine details of the game structure. The fact that these fine details are unobservable in practice just adds insult to injury. These models are completely non-operational.

    My take on these models was that if that’s what you have to do to overturn the basic Chicago results, it’s a testament to the robustness of those results.

    Sam Peltzman’s review in the JPE of the Handbook on Industrial Organization eviscerates this era’s theoretical work. Sam ridicules the City of Theory, and rightly shows (a) how fragile it is, given the dependence on the fine structure of the models, and (b) how little it has actually informed us about the way the world really works. I can’t recommend his piece enough.

    Thanks for the pointer to Shapiro. I’ll read with interest.

    Comment by cpirrong — January 16, 2019 @ 7:41 pm

RSS feed for comments on this post. TrackBack URI

Leave a comment

Powered by WordPress