Campaign Contributions in an Unregulated Setting: an Analysis of the 1984 and 1986 California Assembly Elections

• Published date: 01 September 1992
• Author: Jay K. Dow,Janet M. Box-Steffensmeier
• DOI: 10.1177/106591299204500304
• Date: 01 September 1992
• Subject Matter: Articles

CAMPAIGN CONTRIBUTIONS IN AN
UNREGULATED SETTING: AN
ANALYSIS OF THE
1984 AND 1986 CALIFORNIA ASSEMBLY ELECTIONS
JANET M. BOX-STEFFENSMEIER, University of Texas at Austin,
and
JAY K. Dow, University of Austin at Texas and
University of Missouri-Columbia
he
extent to which a received majority at the ballot box legit-
t
imizes the actions of elected officials depends on the shared
belief that no single individual has disproportionate influence
in the electoral process. Campaign contributions, however, are com-
monly perceived to provide the means by which some citizens become
more politically equal than others. This is despite the fact that there is
no clear relationship between money and electoral outcomes (Crain
and Tollison 1976; Glantz et al. 1976; Jacobson 1978, 1980, 1985;
1990; Silberman and Yochum 1978; Welch 1981; Feldman and Jondrow
1984; Ragsdale and Cook 1987; Green and Krasno 1988, 1990; Grier
1989; Snyder 1990), or legislative voting behavior (Chappell 1982;
Kau and Rubin 1982; Wright 1985, 1990; Wilhite and Theilmann
1987; Grenzke 1989a; Hall and Wayman 1990).
Our understanding of the connection between interest groups, rep-
resentatives, and money is primarily informed by the investment model
of the contributor-legislator relationship, the origins of which may be
found in the early economic analysis of business demand for govern-
ment regulation. The model specifies that contributors allocate resources
among recipients such that the expected return of the investment,
calculated in relation to the political outcome in the absence of contri-
butions, is maximized. Empirical attempts to substantiate the eco-
nomic theory of campaign contribution allocation seek to predict the
distribution of resources across incumbents seeking reelection to the
United States Congress based on a candidate’s possession of specific
Received: April 4, 1991
Accepted for Publication: November 8, 1991
1
NOTE: A previous version of this paper was presented at the 1990 Meeting of the
Public Choice Society in Tucson, Arizona. We thank James Endersby, James
Granato, Kevin Grier, conference participant John Londregan, Michael Munger,
Brian Roberts, Peverill Squire, and an anonymous reviewer for helpful
comments. All errors are our responsibility.

---

610
legislative attributes. This paradigm remains the single most impor-
tant component of the campaign finance literature.
We believe the acceptance afforded the economic model of cam-
paign finance is unwarranted by existing empirical studies. This is
because the theory is predicated on the assumption of a perfect market
for public policy. In particular, the model assumes there exist no
restrictions on campaign contributions. This is clearly antithetical to
the federal electoral arena where imposed contribution limits are extra-
ordinarily low relative to the resources of major contributors.
The existence of ceilings on campaign contributions suggests inter-
est groups are unable to optimize the allocation of their resources and
the distribution of contributions across candidates will be skewed rel-
ative to that predicted by theory. Therefore, recovered characteristics
which predict the allocation of campaign contributions are likely arti-
facts of regulation. As a result, little is learned about the robustness
of the investment model or of the true (non-regulated) value of legis-
lative attributes. Further, we are given a distorted view of interest
group activity and gain little insight relevant to more general issues of
electoral reform.
Our study tests the investment model in an unregulated context -
the campaign finance arena of the 1984 and 1986 California Assembly
elections. This setting allows us to capture completely the tradeoff fac-
ing interest groups between resource use in the public and private sec-
tors. Most important, the use of an unregulated environment allows
us to recover the extent to which investors target contributions toward
candidates with attributes most relevant to their interests. Consequently,
we provide a more appropriate test of the model than that found in
previous studies. Our analysis, viewed in conjunction with previous
research, also provides a bench mark by which the merits of campaign
finance and institutional reform may be assessed. Finally, the study of
campaign finance in a context other than United States congressional
elections has intrinsic benefits as well.
The paper is separated into six sections. Section 2 places our research
in the context of previous studies on the political economy of cam-
paign finance. Section 3 motivates the empirical specification with a
brief overview of the salient aspects of the California institutional
structure. Section 4 presents the estimated model. Sections 5 and 6
contain the results and subsequent discussion. The paper concludes
with a review of the implications of our work.

---

611
PREVIOUS RESEARCH
The past decade has witnessed the development of a significant
literature on the political economy of campaign finance. This research
is largely predicated on the &dquo;Chicago&dquo; model of economic regulation
(Stigler 1971; Peltzman 1976; Becker 1983), and treats campaign con-
tributions as the medium through which the market for policy clears.
Theoretical extensions of the model (Denzau and Munger 1986; Austen-
Smith 1987) predict that interest groups seek to equate the marginal
value of the last dollar contributed, defined in terms of a legislator’s
ability to provide desired policy, across recipients.
Attempts to substantiate the investment model of campaign finance
seek to isolate the determinants of legislative productivity and predict
the distribution of contributions across individuals. Empirical efforts
to recover factors which influence the provision of campaign contribu-
tions have emerged primarily in the last decade. These studies char-
acterize legislator productivity and, consequently, the distribution of
resources, as a function of three classes of legislative attributes. First is
ideology or constituent preference as captured in, for example, party
membership or an interest group vote score. Second, a measure of
candidate need, such as margin of victory in the previous election, is
frequently included. Finally, institutional assets such as committee
assignments and seniority are incorporated because, presumably, these
provide the primary means by which the marginal cost of supplying
legislation is differentiated across individuals.
Not surprisingly, interest groups are most likely to provide money
to ideologically sympathetic representatives, or to those possessing insti-
tutional assets most relevant to their policy objectives. Refinements in
this literature seek to determine if ideological or pragmatic concerns
dominate the allocation decision, and to recover differences in the con-
tribution patterns of distinct classes of interest groups.
There exists supporting evidence on both sides of the question of
whether contributors tend to weigh ideological or pragmatic concerns
most heavily in the allocation decision. Gopoian (1984), Poole and
Romer (1985), Poole, Romer and Rosenthal (1987) find ideology to be
the significant predictor of the distribution of contributions. Conversely,
Grier and Munger (1986, 1991), Evans (1988), Hall and Wayman
(1990) and Endersby and Munger (1991) argue that committee assign-
ment is the major predictor of political action committee (PAC) con-

---

612
tributions. Candidate need is positively associated with contributions
while seniority results have been mixed (Grenzke 1989b).
The extent to which a given class of attributes dominates recovered
estimates is primarily a function of: (1) the level to which the contrib-
utor source has been aggregated (e.g., &dquo;corporate&dquo; or &dquo;labor&dquo; contrib-
utors versus individual PACs associated with specific industries or
unions); (2) the institutional setting studied; and (3) the specific inter-
est under consideration.
Ideological considerations, for example, are often predominant in
analysis conducted at a high level of aggregation (Poole and Romer
1985). Studies conducted at the level of the individual PAC recover
more tangible considerations as the significant predictors of resource
allocation (Munger 1989). This suggests individual contributors base
contributions, for example, on the specific policy property rights asso-
ciated with committee membership. Differences across otherwise sim-
ilar (e.g., corporate) PACs are lost in aggregation. The remaining
common characteristic is ideological orientation. With respect to insti-
tutional differences, Grier, Munger and Torrent (1990), find that,
unlike in the House of Representatives, committee membership in the
Senate is insignificant in the allocation decision of special interests.
Finally, there are differences among contributor classes. For example,
petroleum contributors tend to be more ideological than other indus-
tries in the disbursement of contributions (Evans 1988; Gopoian 1984).
The defining characteristic of this literature, unfortunately, is the
relatively low explanatory power of the estimated models. Regressions
of interest group contributions on considerations such as chamber voting
record, seniority, and committee assignments generally explain about
30 percent of the allocation decision, and rarely exceed 40 percent.’
I
We argue weak results are obtained because the estimated models
are designed to evaluate a theory predicated on the existence of a per-
fect market for policy, i.e., an unregulated setting. The...