Public Finance, 2nd Edition

compared with having a single government was that it permitted individuals to

``vote with their feet,'' as they search for the combination of local services and

taxes that maximizes their utility. Tiebout believed that if all people were free

to search in this fashion, and packages of services and taxes were replicable,

then social welfare would be maximized. This was so for two reasons. First, the

ability to search for one's most preferred level of public goods avoids the free-

rider problem associated with nonexclusive goods in the single-government

model. People naturally reveal their preferences as they search among local-

ities. Second, people with the same tastes will congregate together,

2

thereby

providing a better match of preferences to the level of public services provided.

Also, the public services will be oVered at minimum cost. No cost diVerences

can persist across localities oVering identical services because people will

naturally gravitate from high-cost to low-cost towns. In eVect, the market

for local public services will be perfectly competitive.

Tiebout spawned a huge literature that tested his conjecture using formal

models, both positive and normative analysis. The positive analysis considers

how people sort themselves among the localities. The normative analysis

judges the outcomes of the sorting process using the standard eY ciency and

equity norms.

The literature has generally not supported Tiebout's conjecture; the prob-

lem of forming optimal jurisdictions turns out to be much more subtle than

Tiebout had imagined. The positive analysis has shown that the sorting

process may not reach an equilibriumÐsome people always want to move to

another locality. Normative judgments are moot absent an equilibrium. Fur-

thermore, even if the sorting process does reach an equilibrium, the outcome is

often not optimal. The ability of people to move in response to government

policies introduces another avenue for ineYciency even though it may produce

a better match of preferences for the local public services. Tiebout's conjecture

that federalism produces a social welfare optimum obtains only under highly

specialized conditions that are unlikely to hold in most practical settings.

The literature on mobility following Tiebout is among the largest in all of

public sector economics, so large that we cannot hope to do it justice here.

Our more modest goal is to highlight some of the principal modeling tech-

niques and results in the literature.

THE MODELING DIMENSIONS

Models of mobility under federalism vary along at least eight dimensions that

inXuence the results predicted by the model. The dimensions include the

2

As George Stigler put it, people would choose among high service±high tax, medium

service±medium tax, and low service±low tax communities. See G. Stigler, ``Tenable Range of

Functions of Local Government,'' in Federal Expenditure Policy for Economic Growth and

Stability, Joint Economic Committee, Subcommittee on Fiscal Policy, Washington, D.C., 1957.

856 THE MODELING DIMENSIONS