Calculus of Consent — The Review

Calculus of Consent — The Review

Buchanan and Tullock (BT) wrote The Calculus of Consent: Logical Foundations of Constitutional Democracy in 1962, but their analysis is still worth considering.

Although the book is written for academics, it is not hard to read. Reasonable concentration and patience is all that’s required to understand BT’s exposition.*

Here’s their big point: Constitutions should be designed from the perspective of individuals seeking their own interest, not a society seeking the greatest good for the greatest number.** They say this not because they think that people are selfish bastards, but because governments of/by/for the people need to be designed for the choices that individuals make. It’s a question of “appropriate technology.”

Here are a few notes that I made while reading:

  • Economics “works” because people are different. That means that they can trade — one man’s trash is another man’s treasure. The same holds for politics, where people have different needs and interests, and these change over time.
  • Political choice is subject to uncertainty (NOT risk) over time, over multiple decisions. That means that a constitution should be designed to maximize “average” benefits (wins less losses) across many decisions (ex-ante unknown), allowing for trades in votes over many issues (logrolling). Decisions are made on two levels: At the ex-ante constitutional stage, general rules for making decisions are decided. On the current-event legislative stage, particular decisions are made. An individual may approve a constitution that allows him to be overruled, knowing that he will still gain net benefits from constitutional protections over time. On individual issues, the individual will vote for his own interest — or trade his votes for other issues of greater importance.
  • There are three ways to make policies: Private individual decisions, private group decisions, and public group decisions. Private groups need unanimous agreement; public groups do not. The most-efficient decision mechanism recognizes two costs: The cost to the individual of the mechanism (i.e., where the majority tells the minority what to do) and the cost of making a decision (i.e., the larger the majority required, the larger the cost). There is a sweet spot in the aggregate of these costs, where the cost to the individual and the cost of making a decision are minimized. You want to design a constitution to hit that spot.
  • Note that a rule of unanimity has the lowest cost to the individual, since no individual will allow a policy that hurts himself. Unanimity has a high decision cost, since it requires that everyone agree on a policy.
  • The literature of collective action tends to focus on the cost of not doing something; they miss the cost — to the individual — of doing something. This cost is perhaps a “choice externality.”
  • The more-disaggregated the decision authority, the lower the cost of a decision (subsidiarity).
  • A single-issue referendum is inefficient because it does not allow logrolling — vote trading that takes issue intensity into consideration.
  • BT make a major mistake here, I think, in ignoring (or missing) the problem of logrolling a series of bad policies into place.*** Because they assume that voters — not representatives — are making decisions, they assume that logrolled-policies are beneficial on net. This assumption falls apart when self-interested representatives trade their votes for policies that benefit special interests. In the resulting circle of value-subtracting, robbing Peter-to-pay-Paul policies, they make everyone worse off, in multiple ways.
  • Public projects need only benefit the proportion of votes necessary to get the project enacted. It’s clear that these projects can be less efficient than private collective projects.
  • BT’s theories match observed constitutions and legislative processes.
  • Representative voting means that a minority of 1/4 — 51% of the voters in 51% of the districts — can make decisions. Beware!
  • An individual may accept private costs (e.g., from allowing prostitution to continue) if the alternative (collective control of sex) is worse.
  • The best way to allocate a collective good is to give every individual an equal share (adding to 100% of the good) and then allow trading. That’s what I have said for all-in-auctions and human rights and water!
  • As government has expanded its range and allowed for narrowly-defined programs (remember that this was written in 1962!), the benefits to special interest groups have increased. From this, we can see that special interests will thrive as government’s size and scope increases.
  • Tullock says that game theory accepts the rules as given while economics allows the rules to change (or be ignored). This useful classification was discarded when GT was merged into economics. I make this point when discussing “conflict theory” — where rules are endogenous — but I wonder how many economists fail to consider what scenarios when rules can be broken.

Bottom Line: I give this provocative and useful book FIVE STARS. Every political scientist should read it. Anyone running an organization should read it. Citizens should read it. People upset about BP, or Iraq, or the DMV should read it. (Or maybe just read this review a few times and think about the difference between what you want government to do, what it can do, and what it does do.)

* I tried recently to read Keynes’s General Theory and got lost in his witty erudition. I left it on the bus.
** Something that Professor Wantrup, my benefactor, also understood.
*** I do not know if they updated their theory in the past 50 years (!) to account for this problem. Help?