I recently joined Liberty Fund’s No Due Date economics book club, where over the next year, participants will read one book per month selected by GMU economics professor Peter Boettke. Pete will also lead group discussions and provide other resources. January’s selection is the first volume of James Buchanan’s collected works, The Logical Foundations of Constitutional Liberty, which collects many of his better-known papers from throughout his career. Buchanan was one of the cofounders of public choice theory, and won the 1986 economics Nobel.
This post, the second of three, collects my notes from those readings. I’m posting them here mostly for my benefit, so I can easily find them during the discussions, and can refer back to them later if I cite them in the future. Readers new to Buchanan or curious about the major themes of his work might benefit from skimming these notes, though I highly recommend reading the primary source. I may or may not do this for future months’ readings, depending on how useful it is.
Note that I copied and pasted these notes unedited from a Word document I kept open while reading. These notes do not always distinguish between as-is descriptions of Buchanan’s arguments, and my opinions and original thoughts about them. Reader beware.
WEEK 2: PUBLIC FINANCE IN THE DEMOCRATIC PROCESS
“Individual Choice in Voting and the Market” (Journal of Political Economy, 1954), pp. 75-88.
-Buchanan builds a model of individuals making decisions. They just happen to be voting decisions. For simplification, it’s a direct democracy model without representatives.
-In markets, consumers get what they want. This is not guaranteed in voting markets. There is uncertainty. This affects voter behavior.
-People might vote to signal values, knowing it might not cost them personally. Hence the people who vote for Prohibition, then visit their bootlegger.
-Mises: people bear less personal responsibility for their voting choices than their market choices. So their political choices are more corruptible than their market choices.
-p. 81: “Choice implies that alternatives are mutually conflicting; otherwise, all would be chosen, which is equivalent to saying that none would be chosen.”
-Market choices are unbundled, and less mutually exclusive than political choices, which are take-it-or-leave-it bundles. P. 82: “As a result of this difference, individual choice in the market can be more articulate than in the voting booth.”
-Market decisions are among actual alternatives; political decisions are among potential choices. If the voter loses, they don’t get their preferred choice. Even if they do, there is no guarantee the political process will operationalize it. (he doesn’t seem to make this last point, but would likely agree with it.)
-Market “voters” can be overruled in the sense that, say, their favorite store ot product will go out of business if they don’t get enough “co-voters.”
-All of these differences would remain true under complete economic equality. Objections that “one dollar, one vote” in the marketplace is unfair is not an objection to the points Buchanan is actually making.
-Market choices are not more rational than political choices. The individuals making them are the same. Their differences are in incentives and institutions, not in rationality.
-Section VII, on when to use ballot box instead of the market, is a bit muddled. Institution-level changes gain their legitimacy through the ballot box. Political choices should be made when private choices harm the goals of a majority, or when they are obviously inferior—and it is worth the tradeoffs in choice and liberty to use political means. I would add in something about transaction costs.
-A language problem: current language does not differentiate between market freedom and market power. This semantic point leads to a lot of avoidable confusion.
“Social Choice, Democracy, and Free Markets” (Southern Economics Journal, 1954), pp. 89-102.
-Reaction to Kenneth Arrow’s Possibility Theorem’s philosophical implications. His major argument is that cyclical majorities, which are a product of the intransitive democratic preferences that Arrow’s theorem predicts, provide a bulwark against the tyranny of the majority, and allow for ongoing policy experimentation, rather than setting the initial, “rational” result in stone.
-This highlight’s Buchanan’s subjectivism. He isn’t terribly concerned with this or that policy. He is concerned with the larger system-level processes. Normatively, he seeks to avoid tyranny and stasis, and that’s about it.
-This rests in turn on the core Buchanan theme of methodological individualism. Societies don’t reason or have preferences, individuals do.
-Arrow misuses the word “process,” which has caused confusion in both Arrow and his debaters.
-Buchanan argues that Arrow’s theorem applies to how a welfare function is derived—but not the decision-making process that reacts to that function.
-That doesn’t matter for voting behavior, but it does for market behavior, according to Buchanan. Arrow’s theorem is useful for analyzing voting, but not for markets.
-Methodological individualism: the concept of “social rationality” is incoherent. Societies do not reason, individuals do.
-Interesting side point from Buchanan: utilitarians are individualists, and are therefore philosophically inconsistent whenever they leave the individual and speak of social utility. I add that interpersonal utility comparisons are also impossible.
-Because individuals are rational and the concept of rationality does not apply to societies, we can observe intransitive “preferences” and cyclical majorities in democracies. Also, these inconsistencies can be useful as a check on power and on tyrannies of the majority.
-Cyclical majorities also allow for ongoing experimentation with new policies. The status quo is never set in stone.
-Buchanan invokes near-unanimity as a benchmark of true collective choice, prefiguring The Calculus of Consent, which would appear eight years later.
-Market decisions to tend to obey the transitive property, since they are made solely at the individual level. They are not public choices. Buchanan does allow that this is true only to the extent that an individual’s preferences are, in fact, transitive. Anyone who has spent time with a small child knows that real-life human preferences are not always transitive.
“The Pure Theory of Government Finance: A Suggested Approach” (Journal of Political Economy, 1949), pp. 119-132.
-Buchanan contrasts individualist and collectivist (organismic) approaches to costs and benefits of taxes and spending.
-It was standard practice at the time to count only the costs of government, and not the benefits. Buchanan argues that both matter, and benefits should be counted as well. Later in his career, he would have taken this in a more explicitly public choice direction—the implications for concentrated benefits and diffused costs are obvious. Here, he hints at it, but doesn’t go very far in that direction of analysis.
-One problem with the collectivist/organismic approach is that it thinks in aggregates, rather than in terms of separate individuals. Since interpersonal utility comparisons are impossible, so are accurate societal cost-benefit calculations.
-A price theory point Buchanan does not make: the technical difficulties of separating individual costs is “expensive” in terms of effort and complexity for economists. This is why they choose the “cheaper” option of thinking in aggregates. While rational from a price theory standpoint, this leads to unrealistic analysis.
-Buchanan argues that the aggregate cost of the state should equal its aggregate benefits, in which seems a fairly straightforward Marshallian calculation at the margin. He is agnostic about how those costs and benefits are distributed. That’s for the political process to decide.
-This article clearly reflects his recent study of Italian political economists. He quotes several.
-One of them raises a good point: if political benefits were to be equally spread out, a capita tax would be fair. Since that is not what most people want the state to do, that is why government costs are not equally distributed, nor its benefits.
-The “fiscal residuum” is the difference between a government’s costs and benefits. These will vary from person to person. The goal is for it to equal zero for society as a whole (Buchanan ignores transaction costs and political waste here, but for this simple model’s purposes, that is fine). A progressive tax and benefit system would have a negative residuum for rich individuals, and a positive residuum for poor individuals.
-(Not Buchanan’s point) In practice, democracies often have positive residuums for the middle class, which has the largest number of voters, and negative residuums for the rich and poor. This is for public choice reasons—politicians know want to maximize votes more than they want to maximize any distributional fairness norms they may have.
“Positive Economics, Welfare Economics, and Political Economy” (Journal of Law and Economics, 1959), pp. 191-209.
-Economic theory was developed by utilitarians, and the discipline has been taken over by positivists. Even Milton Friedman is a positivist. This is where Paretian welfare economics comes from. Most economists are not content to describe what is; part of their job is advising policymakers on what should be. Buchanan doesn’t like this.
-Clever insight about Pareto optimality: it avoids the cardinal no-no of interpersonal utility comparisons. Individuals make their own decisions about what makes them better off and worse off, so no interpersonal comparisons are needed. Kaldor and Hicks took Pareto’s approach and developed the new welfare economics.
-A newer development in welfare economics, headlined by Paul Samuelson and others, rejects Kaldor and Hicks. Samuelson, et al rely on a “social welfare function,” and thus commit the non-no of interpersonal utility comparisons.
-Welfare economists, especially of the Samuelsonian variety, assume omniscience of the observer or policymaker. Buchanan says this is unrealistic, and should not guide policymaking. It gives too much power to policymakers to make decisions on others’ behalf.
-Revealed preferences as fatal to the omniscience assumption: We don’t know other people’s preferences until they reveal them through their actions.
-An economist should not decide upon changes, because he has no way to know society’s preferences; the very concept is incoherent. Instead, an economist should present a menu of changes, upon which individuals can decide on, either individually or through the political process.
-This is another example of Buchanan’s subjectivity. His ideological priors are liberal in the sense that he cares about individual consent. But he’s neutral about which policies individuals consent to.
-A rough analogy to Buchanan’s job description for economists is as medical diagnosticians. The patient has a problem, the economist uses their tools to diagnose it and prescribe possible remedies. But ultimately the patient chooses what action to take—though in this case through political consensus, not individual choice.
-Compensation for externalities, such as pollution: Buchanan sees payment for externalities not as an ethical concern for policymakers, but as necessary for an an honest prices system, so individuals can make their own accurate decisions about Pareto-optimal changes. His subjectivity shows up again.
-A political economist’s job is to suggest possible gains from trade, not to impose them against people’s wills—the economist doesn’t know people’s preference functions, and could make non-Pareto-optimal mistakes.
-Good question on p. 203: “Unless the relevant choices are to be made by some entity other than individuals themselves, why is there any need to construct a “social” value scale?”
-Buchanan exposes the vulnerabilities of his own argument—this is the mark of a good scholar. His argument depends on people being reasonable; this is not always true. His argument depends on a contract theory of the state; many people object to this. And no large group of people will be unaminous in decisionmaking, which is the ideal. Some “relative unanimity” benchmark short of that will have to do in real-world political systems, such as a majority vote, a 2/3 majority, or whatever rule people decide on. Buchanan is agnostic on which relative unanimity rule is best.
-If a policy doesn’t gain unanimous consent, is it Pareto optimal? Tough question. Real-world societies will nearly always have to settle for something short of that ideal.
-Which also makes a society progressively more vulnerable to tyrannies of the majority, the closer the adoption rule moves to a 50-percent-plus-one majority.
-A bit of game theory: economists must think at least one move ahead. Don’t recommend what people want right now, recommend what people will want after a proposal goes through the political process.