Decentralizing Pork: Congressional Roll‐Call Voting, Decentralized Administration, and Distributive Politics

• Published date: 01 February 2018
• Author: Anthony M. Bertelli,Pamela J. Clouser McCann
• Date: 01 February 2018
• DOI: http://doi.org/10.1111/lsq.12183

ANTHONY M. BERTELLI

New York University

PAMELA J. CLOUSER McCANN

University of Southern California

Decentralizing Pork:

Congressional Roll-Call Voting,

Decentralized Administration, and

Distributive Politics

Congress packages pork-barrel spending in complicated proposals that belie theo-

ries of distributive politics. We theorize that roll-call voting on such bills depends on

grant programs’ administrative centralization, party ties with presidents or home-state

governors, and differences in geographic representation between chambers. Analyzing

votes between 1973 and 2010 using a within-legislator strategy reveals that House mem-

bers are less likely to support decentralized spending when they are copartisans with

presidents, while senators support decentralization regardless of such party ties. When

House members or senators share affiliation with only governors or with neither chief

executive, the likelihood of support rises with decentralization.

Understanding how government programs are designed and imple-

mented as well as how they distribute resources is vital to the study of

politics and public policy. Congressional scholars argue that legislators

have a compelling incentive to vote in favor of bringing federal dollars to

their districts and states (e.g., Balla et al. 2002; Baron 1991; Cox and

McCubbins 2005; Evans 1994; Shepsle and Weingast 1981). Although

many Americans decry the principle of earmarks, most voters prefer

legislators who bring government projects and dollars home to their dis-

tricts, putting local above national interests (see generally Drake 2013;

Miller 2016; Pew Research Center 2010). Indeed, voters may use federal

spending in their district or state as a proxy when assessing a legislator’s

competence (Grimmer, Messing, and Westwood 2012; Lazarus 2010).

Previous work on distributive politics focuses on the funds districts

receive rather than the legislation on which members vote. While tar-

geted spending is attractive, distributive bills are often complicated, and

to our knowledge, we provide the f‌irst large-scale study of votes on

LEGISLATIVE STUDIES QUARTERLY, 43, 1, February 2018 69

DOI: 10.1111/lsq.12183

V

C2017 Washington University in St. Louis

them. Targeted funds and earmarks are straightforward mechanisms for

bringing dollars to constituents, yet bills are rarely crafted using them

alone. Lee draws our attention to “the many ways that pork can be pack-

aged” (2004, 205), such as formula, block, categorical, and

demonstration project grants, not to mention contracts, loans, and other

means (Miller 2016; Salamon 2002). A voluminous literature has devel-

oped around the politics of distributive spending and the factors

associated with the ability to bring home federal dollars, but studies of

federal outlays focus on how administrators distribute programmatic dol-

lars, rather than how legislators voted on the bills that provided the

underlying authorizations.

1

In parallel studies of legislative coalition for-

mation, scholars argue that members of the majority party, senators from

smaller and less populous states, or House members from more populous

states are rewarded with more federal resources (e.g., Ansolabehere,

Snyder, and Ting 2003; Knight 2008; Lee 2000). These studies implicitly

assume that all legislators who benef‌it from funds would vote in favor of

bills authorizing them, and our study challenges this assumption.

Members frequently vote against spending that may benef‌it their

constituents and must justify such decisions, as in the case of Rep. Heath

Shuler (D-NC) and the 2009 economic stimulus package: “I didn’t vote

for it, but that doesn’t mean that I don’t support western North Carolina

getting its fair share” (Thrush 2009). Transportation bills, for example,

often incorporate district- and state-level distributive programs (Lee

2004), yet not every member votes in favor of f‌inal passage, even when

his or her district would benef‌it. What explains these voting choices?

Current theories fail to capture the conditions under which members

oppose laws offering federal spending in their districts (but see Sellers

1997). Given that districts would see these funds months, if not years, in

the future,

2

we make the important observations that (1) provisions

decentralizing federal programs shift inf‌luence from federal to state off‌i-

cials and (2) members are concerned about the distributional

consequences of such delegation choices and (3) vote accordingly.

We offer a theory that roll-call voting on bills with grant programs

depends on the level of administrative centralization of those programs,

party ties between a member and the president or home state governor,

and the differences in geographic representation between chambers.

When members vote on legislation, they also vote on the degree to

which administration is centralized with federal agencies or decentral-

ized to the states (McCann 2016).

3

Decentralized provisions create an

additional administrative layer between members and their constituents

that impacts the costs of ex post oversight. Centralized provisions, by

contrast, leave administration in the hands of a federal agency. We argue

70 Anthony M. Bertelli and Pamela J. Clouser McCann

that time-constrained legislators understand that they must spend their

scarce resources to ensure that their districts receive federal dollars and

that sharing partisan aff‌iliation with the president or home state governor

can moderate the ex post costs associated with overseeing the distribu-

tion of funds to constituents. Moreover, while senators can claim credit

for any decentralized funds because their constituencies are statewide,

House members must still ensure that funds administered at the state

level reach their districts.

Using correlated random effects statistical models that account for

unobservable attributes of legislators, we analyze a unique data set of

30,005 coded provisions and associated roll-call votes in the House and

Senate from 199 major laws enacted during the 93d to 111th Congresses

(1973–2010). Within-member designs such as ours are common in the

contemporary literature on pork-barrel politics (Alexander, Berry, and

Howell 2014; Berry and Fowler 2016). Beyond ideology, committee

membership, seniority, prior electoral margins, and other characteristics

that have been found to be signif‌icant in studies of pork-barrel politics

(see Grose 2011, 202–03), legislators represent different groups of con-

stituents, which exert different pressures on their votes, suggesting the

importance of legislator-level designs to account for this heterogeneity.

Our design further accounts for unobserved heterogeneity by state, con-

gress, and policy domain, insulating our hypothesis tests from a variety

of inf‌luences over pork-barrel spending on both the supply and demand

side (e.g., Bertelli and Grose 2009; Stein and Bickers 1995). We f‌ind

House members less likely to support proposals with decentralized

administration when they share partisanship with the president alone or

also with homestate governors. When House members and senators are

copartisans with only their governors or with neither chief executive, the

likelihood of support rises with decentralization. Our theory and results

suggest a new way of viewing distributive politics in policy formation

with a host of important implications for our understanding of

partisanship and administrative design on congressional voting.

A Theory of Members’ Voting Choices on Federal

Distributive Programs

We begin our argument with two reasonable assumptions. First,

we assume legislators care about re-election. Second, we assume that

individual members of Congress are time constrained in their legislative

activities (Hall 1987, 1996), which restricts the extent to which they can

oversee when and how funds are deployed to their states and districts.

For existing programs and policies, members of Congress spend

71Decentralizing Pork