Those who know the late Harold Demsetz’s important role in my life might be surprised that I haven’t posted on him since he died on January 4. The reason is simple: I spent my time on him writing a piece for the Wall Street Journal. It will appear on Monday.
I had limited space for the piece and so there was so much I wanted to say but didn’t have space to say. Here’s a good quote from his autobiographical essay in his 1988 book, Ownership, Control, and the Firm:
Game-theoretic constructs have, or will have, things to say about the interactions of decision-makers when their numbers are small. It is the interest in such situations that gave birth to game theory, and this remains the context in which it does best. My interest, however, is in system-wide forces that render this context less useful. The decentralized, competitive system functions to overwhelm tactical moves that may have been stimulated by interactions between small numbers of decision-makers. The analytical usefulness of the concept of decentralization derives precisely from the fact that it allows the analyst to ignore the behavior of a single individual or a small group of individuals. It implicitly asserts that the tactical measures taken by incumbent firms to bar entrants from an industry cannot long hold at bay the continuous onslaught of more efficient organizations and techniques of production. The analysis of tactical decisions is not likely to be very useful in studying the general and central tendencies of decentralized economic systems, especially when thes result from repeated “plays” in contexts that are not zero-sum in nature. Future developments may convince me of my error in holding this belief; I state it not to persuade readers of its truth, but to give them a better understanding of my interests and perspectives. (italics in original)
Here’s what I got from that. Demsetz is interested in the long-run tendencies of industries and game theory is almost useless for that. What it’s useful for is helping firms figure out how to deal with small numbers of competitors in the short run. Think of Demsetz as a modern Schumpeterian.
READER COMMENTS
David Henderson
Jan 12 2019 at 8:46pm
John,
Do you care to expand on that, or, despite my limited understanding, am I supposed to understand what you’re getting at?
Biopolitical
Jan 13 2019 at 6:44am
In the case of the prisoner’s dilemma, game theory predicts defection in one-shot interactions, and cooperation when each player (a) knows how potential partners’ have previously behaved, and (b) is free to choose with whom to play and whom to avoid. (a) and (b) define a market.
For (a) to hold you need time for experience to accumulate. For (b) to hold you need at least three potential players. Game theory does address the long run and many players.
Jon Murphy
Jan 13 2019 at 7:26am
I need to push naba a little. The Prisoners’ Delimma does not, by construction, allow people who to interact with; the prisoners have no choice over the cop in the room or the situation. The PD is a 3-person game.
Jon Murphy
Jan 13 2019 at 7:26am
*push back.
Sorry for the typos. I’m on my phone
Bob Murphy
Jan 15 2019 at 4:13pm
Just some nitpicks on your guys’ comments:
==> Biopolitical, are you summarizing an actual model (or models) you’ve read? I’ve seen a bunch of different versions of the basic PD in repeated game setups, but I’ve never heard of a variety where you can choose not to play with someone. That seems to violate the idea that you’re in a PD.
==> Jon Murphy, the canonical PD itself is a 2-person game. The “cop” is built into the payoff structure.
Jon Murphy
Jan 16 2019 at 9:16am
You are right that is the canonical description, but I think the canonical description is incorrect. I think the cop needs to be kept in as a player since the cop influences the choices of the actors, at least insofar as we want to use the PD as a predictive game. (I think Elinor Ostrom makes this point, too, but I could be wrong. I’m on vacation right now and can’t check my book, but I think she discusses it in Chapter 1 of Governing the Commons).
Biopolitical
Jan 18 2019 at 5:40am
Bob, when I wrote “In the case of the prisoner’s dilemma” I meant “When each interaction of two individuals has the payoff structure of a prisoner’s dilemma”.
Gordon Tullock describes a setup with such payoff structure, voluntary interactions and partner choice here. Tullock’s paper has been cited many times. You can pull the thread from there.
Biopolitical
Jan 18 2019 at 5:45am
Bob, by “in the case of the prisoner’s dilemma” I meant “when each interaction has the payoff structure of a prisoner’s dilemma”.
Gordon Tullock describes a setup with such structure, voluntary interactions and partner choice here. Tullock’s paper has been cited many times. You can pull the thread from there.
Biopolitical
Jan 18 2019 at 5:49am
By “in the case of the prisoner’s dilemma” I meant “when each interaction has the payoff structure of a prisoner’s dilemma”.
Gordon Tullock describes a setup with such structure, voluntary interactions and partner choice in “Adam Smith and the prisoners’ dilemma”. Tullock’s paper has been cited many times. You can pull the thread from there.
Comments are closed.