Do Multimember Districts Lead to Free‐Riding?

• DOI: http://doi.org/10.3162/036298007782398503
• Date: 01 November 2007
• Author: MICHIKO UEDA,JAMES M. SNYDER
• Published date: 01 November 2007

649Multimember Districts

LEGISLATIVE STUDIES QUARTERLY, XXXII, 4, November 2007 649

JAMES M. SNYDER, JR.

Massachusetts Institute of Technology

MICHIKO UEDA

California Institute of Technology

Do Multimember Districts

Lead to Free-Riding?

We studied the effects of districting on intergovernmental aid by state govern-

ments to local governments in the United States. We found that metropolitan areas

receive relatively more aid when represented in the state legislature by an at-large

delegation than when divided into single-member districts. This suggests that the

free-riding that may occur with at-large representation is more than counterbalanced

by other factors. The estimated effects are robust to the effects of other confounding

factors as well as the choice of estimators.

What effects, if any, do electoral institutions have on political

representation and policy outcomes? This is a key question in political

economy and political science, and it has been explored both theoreti-

cally and empirically by many scholars. For example, one significant

body of research has explored the differences between plurality-rule

and proportional-rule systems. Another has studied the effects of

malapportionment. A third has focused on the special case of the U.S.

presidential electoral college. A fourth has analyzed the differences

between single-member and multimember districts.1

One of the main hypotheses in the literature on single-member

districts (SMDs) versus multimember districts (MMDs) is that SMDs

lead to closer ties between voters and their representatives, better

electoral accountability, and, consequently, better performance by

representatives. Consider a large city with, say, 15 state legislators. If

these legislators are all elected from one at-large, 15-member district,

then they will face a collective-action problem. It is difficult for voters

to know whom to reward for bringing home the bacon, or whom to

blame when no project appears. The result is low-powered incentives

and the tendency for each of the city’s representatives to shirk. The

city may suffer and not obtain its fair share of state spending.

650 James M. Snyder, Jr. and Michiko Ueda

There is some empirical work supporting the free-riding

hypothesis, but it is not overwhelming. Jewell (1969) and Cooper and

Richardson (2006) have found that representatives elected from SMDs

are more likely to describe their roles as “delegates,” while those elected

from MMDs are more likely to describe their roles as “trustees.” Scholl

(1986) has found a similar pattern in his survey of British (SMD) and

French (MMD) members of the European Parliament: British members

are more likely to emphasize constituency service and regional issues.

Scholl has also found evidence of an effect on the percentage of regional

projects passed, but since he does not report the sizes of these projects,

it is difficult to draw firm conclusions. Freeman and Richardson (1996)

report conflicting results about legislative roles: in their survey, legis-

lators from MMDs claimed to spend more time performing constitu-

ency service than legislators from SMDs. Finally, a few papers on the

incumbency advantage have provided indirect evidence about shirking

and incentives—a smaller incumbency advantage, suggesting a weaker

personal link with voters—but even these results are mixed.2

Moreover, several forces work in the opposite direction. We

believe these forces have been underappreciated in the theoretical and

empirical literature and deserve more attention.

First, if a city is cut up into a number of small districts, then

many citywide projects will entail significant externalities from the

point of view of each individual representative. Legislators who

represent only a small fraction of the city’s area and population have

little incentive to work hard to obtain state aid for such projects—

highways that connect different parts of the city, a highway that circles

the city, or a citywide mass transit system. Public utilities—electricity,

gas, water, and sewage—will also be unchampioned, since utilities

tend to be citywide systems. Even school districts are typically citywide

or countywide systems. State aid for education does not go to indi-

vidual schools but to the school district as a whole; the allocation of

the funds within the school district depends not only on legislative

decisions but also on decisions by city or county officials.

Second, districting may introduce collective choice problems.

What types of projects should a city seek from the state? Alternatively,

why not seek tax cuts instead of projects? When all representatives

from a city have the same geographic constituency, they will tend to

agree on such questions, especially if they are all members of the same

party, a common outcome with the nearly universal use of plurality

rule. If the city is carved up into many small SMDs, however, then its

representatives will have different and often conflicting interests. Some

delegates will be from poorer parts of the city that seek basic services,