Stevens: The economics of collective choice

Stevens. 1993. The economics of collective choice. Boulder: Westview Press, Inc.

In Brief

While WikiSummary rarely summarizes textbooks, this book provides excellent reviews of the economic theory of the firm and its applications to political science. It connects economic theories to political ones in ways that are often assumed but rarely made explicit.

Ch 1-4

The first four chapters are a textbook-style review of basic economic concepts. In keeping with WikiSummary policy of not summarizing textbooks, we present only a few brief notes here.

Ch 1-2

A review of supply and demand, utility curves, Edgeworth boxes, and introductory microeconomics.

Ch 3: Efficiency

Market failure type I: Efficiency failures.

• Public goods, externalities, conditions for the market to work properly

Ch 4: Equity

Market failure type II: Equity failures.

• Four philosophies of what is just. Libertarianism, additive utilitarianism, maximin (Rawlsian) utilitarianism, egalitarian utilitarianism

Ch 5: Voluntary Solutions

Voluntary solutions to market failure. It is possible for non-coercive solutions to market failures to be produced.

• Joint supply of a public good: see page 104. But this situation is a bit contrived. Transactions costs in particular get in the way.
• Coase theorem: in a world free of transactions costs, any distribution of property rights would yield the same degree of externality production. Example: If a firm has to buy pollution rights, the amount it would pay would depend on the cost of the next least-costly alternative method of disposing of waste. But if the firm already has pollution rights, its decision of whether to give them up would depend on the cost of the next least-costly alternative method of disposing of waste. So you'll get the same amount of externality (pollution) whether the firm owns pollution rights or somebody else does.
• Private provision of public goods. People don't free ride. Why not?
• Selective incentives (Olson)
• Altruism
• the real answer is somewhere between these two extremes
• Experiments with public goods:
• Economists free ride more often than others do in experiments
• Money-back guarantees and fair-share systems both facilitate cooperation (during prisoner's dilemmas) considerably. A fair-share system is a coercion system: I pay union dues whether I want to or not. A money-back guarantee says, "We need to raise $1million to hire a lobbyist. Please contribute. If we don't get enough, we'll return it all." (That way your donation isn't lost for nothing if the target isn't reached.)
• Willingness to pay (for a good) and willingness to accept (i.e. sell out): should be the same, but prospect theory says they are not. E.g. one study found that Wisconsin hunters were willing to pay $20 for a goose-hunting permit, but those who already had one wouldn't sell it for less than $100.
• Club provision of goods (since many goods aren't purely public goods, but are excludable): efficient club size is defined by intersection of marginal congestion costs (rises with size) and marginal supply costs (falls with size)
• Local governments as clubs: Tiebout (1956)

Ch 8: Regulation and Taxation

Looks at 5 main models of regulation/taxation.

1,2,3: Begins with nice summaries of Stigler, Peltzman, and Becker, which are individually summarized in separate entries (follow the links).

4: Hettich and Winer give a model of tax systems (how many tax bases, what rate structure, and what special provisions) based on marginal tax administration costs and marginal tax discrimination benefits.

• Different tax bases: If all income is taxed the same (meaning capital gains and labor wages are taxed using the same rate structure), that's one base.

5: Shepsle and Weingast (1984) give a model based on pork. When legislators have geographic districts, they vote for local benefits, even at the expense of national welfare. Since benefits of federally-supported local projects are concentrated, but costs are dispersed, this rarely leads to an efficient outcome.

• Sub-district constituencies matter: the people within your district who actually vote for you.
• Supra-district constituencies matter: people who make campaign contributions, e.g.

On the median voter theorem:

• Politicians tend to work with part of their constituency, not to rush toward the median voter
• Political preferences CAN be represented along a single dimension (ideology)
• Voters tend to be distributed normally when it comes to ideology, but politicians are distributed bimodally. This is to be expected, since (as Downs and Olson might predict) people farther from the center have more to gain from getting into politics, so the benefits are more likely to exceed the costs.

Legislative markets: five characteristics

1. Outcomes are anticipated, not actual; (2) Expected outcomes are favorable for both demanders and suppliers; (3) Outcomes are uncertain (other legislators might oppose the one you buy off); (4) Outcomes may be short-lived (e.g. overturned next year); (5) Legislative activities vary greatly in time and in importance (having a senator hear your argument is nice, but getting her to support you on the floor is better).
2. PAC money is the main currency given to legislators in exchange for legislative activities. Ideology predicts PAC contributions fairly well: highly polarized interest groups give legislators money with which to compete for moderate votes. (The interest groups are polarized b/c only those with more extreme views have any incentive to contribute.)
3. Legislative activities (i.e. what legislators supply to lobbyers): access (hear you out), consideration (think about it, hold hearings), voting.
4. Legislative committes (i.e. how legislators trade votes with one another): Weingast and Marshall 1988. It is in the legislative committee that legislators implement deals made in legislative markets.
5. And legislative committees make it easier to provide pork, making more districts "safe." Combine that with the high entry costs of running for office, and you see why the House in particular has so little turnover.

Ch 9: Administrative Government, part 1

Begins by considering a few common views of congressional-administrative relations:

• "Iron triangles," subgovernments: the government is really a bunch of subgovernments interested only in one issue area; each subgovernment is an iron triangle. E.g. the defense subgovernment consists of the DoD, lobbyists [the defense industry], and Congressional defense committees [and legislators with large bases in their districts]. Main problem: forgets that legislators have diverse constituents, and that every issue area has lobbyists on both sides.
• Agencies are dominant (Niskanen 1971). Agencies are budget maximizers who are able to extract higher-than-efficient budgets from Congress because the agencies have private information about how much "output" they can really produce at a given price. Example: An agency knows it can maintain 100 national parks for $100 million. It knows that Congress is willing to pay $120 million for 100 parks, and it knows that Congress doesn't really know how much 100 parks costs. So it says it cannot produce 100 parks for less than $120 million. It uses the extra $20 million to buy gold-plated toilet seats for the agency head. 3 main Xs: bilateral monopoly, asymmetric info, budget maximization.
• Congress is dominant (e.g. McCubbins and Schwartz 1984). By using fire alarm oversight, congress shifts the costs of oversight to those with the greatest interest in the issue at hand. Kiewiet and McCubbins (1991) give four solutions to p-a problems: screening and selection of agents, contract design, monitoring and reporting, and institutional checks. Shepsle (1979) and Sheplse and Weingast (1981) point out that Congress overcomes cycling problems with structure-induced equilibria (i.e. committees create agenda control rules tha end cycling).

Observations on whether the median voter is as effective a principal as a legislator (he isn't):

• Reversion budgets can allow agencies to extract higher-than-optimal budgets from voters (pg 295).

Ch 10: Administrative Government, part 2

Ch. 9 considered whether agencies or principals are in charge (an open debate). Ch. 10 considers how we can analyze the effects of an agency's policy. There are several DECISION FRAMEWORKS for doing so. Most of the chapter examines the most common framework: benefits-cost analysis. Some of these frameworks are empirical; others are normative.

• Benefit-cost analysis: seeks to identify all the benefits and costs of a policy/regulation. It's not as straightforward as it sounds, though. [pg 304 summarizes the 7 general things such an analysis should do.] Difficulties: It's hard (but not impossible) to measure costs/benefits of non-market goods (e.g. clean air, national parks). [Two methods: travel cost method (pg 313) and hedonic methoc (314).] It's hard to know how heavily to discount future benefits (e.g. costs of a dam are short-term; benefits are long-term). It's hard to know which secondary effects to include (e.g. an irrigation project will boost agriculture; do you also count the secondary benefits that accrue to food processors?). And it's hard to place a value on human life (e.g. a safety program reduces highway deaths; how many costs are worth it?)--although there is an interesting method for doing so (a hedonic method) on page 321.
• Technology-based standards: use the best available technology, even if it isn't most efficient (e.g. treat the sewage as well as possible, even if the costs of doing so are higher than the utility that society gets from additional cleanliness).
• Cost-effectiveness analysis: minimize the costs of achieving a given goal. [weakness: if the goal is to reduce cancer deaths, this analysis won't say how many lives should be saved by lowering the cancer rate as opposed to saving lives in some other way.]
• Risk-benefit analysis: An analysis that is intentionally more vague than benefit-cost or cost-effectiveness analysis; the intention is to produce a whole lot of information that can then be debated in a democratic context.
• Environmental impact statements: like a risk-benefit analysis, it produces lots of information. Downside: "lack of commensurability among outcomes" (hard to know the value of the environmental impact).
• Multiple-objective programming: Assign weights to certain outcomes (e.g. how important is reducing crime against people relative to crime against property). Then run benefit-cost analysis to get the most benefits, subject to these weights.
• Program evaluation; Go back and look at an implemented project and see what it did, who was benefited/hurt, etc.

Ch 11: Federated Government

Reviews themes from the whole book, and introduces problems that arise from having federated government.

Governments perform three functions: allocation (how to use resources, what to produce), distribution, and stabilization (fiscal/monetary policy to drive general growth).

Citizen mobility across federal units implies that stabilization should be done by the central government. Stevens also claims that mobility implies that distribution done centrally, but his argument lacks any logic whatsoever (see notes on pg 331). Allocation should be shared by all levels of government.

Some public goods are local (e.g. a fire department)--the farther you get from it, the less you can consume it.

Tiebout (1956): a summary and criticism. Main problem: even if Tiebout's system is efficient, it will doubtful be equitable.

Intergovernment grants: Stevens reviews the effects of various types of grants to local governments: block grants, categorical grants, formula grants, matching grants, and so on (pg 335).