University of Utah

Good Money and Bad Money: Do Funding Sources Affect Electoral Outcomes?
Author(s): Brad Alexander
Source: Political Research Quarterly, Vol. 58, No. 2 (Jun., 2005), pp. 353-358
Published by: Sage Publications, Inc. on behalf of the University of Utah
Stable URL: http://www.jstor.org/stable/3595635
Accessed: 10-08-2017 15:51 UTC

REFERENCES
Linked references are available on JSTOR for this article:
http://www.jstor.org/stable/3595635?seq=1&cid=pdf-reference#references_tab_contents
You may need to log in to JSTOR to access the linked references.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
http://about.jstor.org/terms

Sage Publications, Inc., University of Utah are collaborating with JSTOR to digitize,
preserve and extend access to Political Research Quarterly

This content downloaded from 128.110.184.42 on Thu, 10 Aug 2017 15:51:05 UTC
All use subject to http://about.jstor.org/terms
 Good Money and Bad Money:
Do Funding Sources Affect Electoral Outcomes?
BRAD ALEXANDER, EMORY UNIVERSITY

There is lively public debate about the normative impact of different kinds of money in elections. However,
there is surprisingly little examination of the practical impact that funding sources have on election outcomes.
Even if we assume that voters do not care directly about campaign finance, there may still be incentives built
into the system to discourage fundraising from some sources and encourage it from others. Therefore, I exam-
ine the actual impact of out-of-state donations, PAC donations, and self-financing on election outcomes in
open seat House elections in the 1996, 1998, 2000 and 2002 cycles. I find that some kinds of fundraising are
correlated with success, while others are correlated with failure, although t a lesser level than district parti-
sanship or total campaign spending. These results offer promise for addressing some measurement issues in
the congressional election field, as well as possible directions for future research.

t least among political elites-primarily interest vast majority of funding from numerous individuals living
groups, elected officials, and media outlets-there is within the state where they are running. PAC donations
a widespread sense politics is full of "dirty money" would be almost absent from the mix, and self-financing
that causes politicians to regularly neglect the public inter- would be limited to a relatively small share of total spend-
est and focus instead on the interests of major donors.1 ing. The House would-in this campaign finance utopia-
Whether through accounts of widespread PAC donations or be populated not by a large number of millionaires acting
stories about candidates who spend tens of millions in per- primarily on behalf of interest groups, corporations, and
sonal funds trying to win elections, there is lively public wealthy pals, but by a more representative sample of the
debate about the normative impact of different kinds of American public with financial support flowing primarily
money in elections. However, there is surprisingly little from the citizens they represent.
examination of the practical impact that funding sources Although public concern over campaign finance is not
have on election outcomes. particularly high, there may be ways in which the presence
If voters are truly outraged by particular kinds of dona- or absence of funding from sources mentioned above alters
tions to congressional candidates, one would expect to see electoral outcomes beyond direct influence on public opin-
their disappointment reflected in election outcomes. If "bad ion. If candidates who raise money in particular ways are
money" corrodes electoral linkages, we should expect can- more or less likely to win, it may be because particular types
didates who raise "good money" to benefit electorally from of fundraising are proxies for other factors (e.g., candidate
doing so. Even if we assume that voters do not care directly quality) or because fundraising has direct political impacts
about campaign finance, there may still be incentives built (e.g., grassroots organizational benefits). Candidates who
into the system to discourage fundraising from some raise money within their home state may enlist supporters
sources and encourage it from others. Therefore, I examine who can actually vote for them in the process. Self-financ-
the actual impact of funds from three suspect sources on ing candidates are not forced to fundraise, and may decide
election outcomes in open seat House elections in the 1996, not to do so, foregoing the electoral benefits fundraising
1998, 2000 and 2002 cycles. Specifically, I focus on out-of- might bring. Elected officials who rely on PAC funding
state donations, PAC donations, and self-financing. I find might be isolated from their constituencies or benefit from
that some kinds of fundraising are correlated with success, PAC support representing organized voters in their district.
while others are correlated with failure, although at a lesser Research in this area has the potential to contribute to
level than district partisanship or total campaign spending. our understanding of both legislative behavior and the
dynamics of political campaigns. On the legislative front,
THEORY AND BACKGROUND merely knowing what kinds of donations are correlated with
political success can help us understand how members
The most theoretically pleasing model candidate for might allocate their fundraising time, or which potential
those who worry about the impact of money on Congress is donors might be expected to receive the most informal leg-
relatively simple to describe. This candidate would raise the islative attention. In terms of campaigns, there are numer-
ous areas where this research might make useful additions.
For example, contribution categories are convenient proxies
The phrase "special interests" alone appears in the Lexis media database
for measuring the role of organized voter groups in deter-
125 times in the past six months.
mining outcomes, and knowing which categories correlate
Political Research Quarterly, Vol. 58, No. 2 (June 2005): pp. 353-358 with success might tell us something about the importance

353

This content downloaded from 128.110.184.42 on Thu, 10 Aug 2017 15:51:05 UTC
All use subject to http://about.jstor.org/terms
354
354POLITICAL RESEARCH POLITICAL RESEARCH QUARTERLY
QUARTERLY

of special The explanatory variables in this analysis are

interests inthe per- elec
tributions centage of donations
are received by each candidate, through-
consistentl
researchers might
out the cycle, from PACs, in-state donors, and theirbeown ab
effective bank account. These data were all obtained
measure of online through
under
ficult to either the Center for Responsive Politics,
observe, at www.opense-
such as
crets.org or the Political Moneyline, at www.tray.com. The
RESEARCH DESIGN
figures provided on both sites are tabulated directly from
campaign disclosure reports filed by candidate committees
I generate some preliminary answers to these questions with the Federal Election Commission. The PAC and self-
by looking at correlations between funding source and elec- financing figures are simply the percentage of the candi-
tion outcomes. The unit of analysis for this study is all open date's total fundraising accounted for by each source.
seat House races for 1996, 1998, 2000, and 2002. Open The in-state donation percentage is somewhat less satis-
seat races account for a disproportionate share of the newly fying, since it represents only the percentage of itemized
elected members of each Congress and are therefore the individual donations raised from within the candidate's
place one would expect to see substantial campaign finance state. It does not account for whether or not a particular
effects on both sides. From 1982 to 1992, "well over 70 per- PAC donation represents a predominantly in-state or out-of-
cent of the newcomers to Congress made it there by win- state interest, nor does it address the source of individual
ning open seats." (Gaddie and Bullock 2000: 1). donations of less than $200, which are not itemized or pub-
The dependent variable is the Democratic share of the licly disclosed. This essentially means there is no way of dis-
major party vote in each election.2 These four cycles are par- tinguishing between a candidate who raises large amounts
ticularly appropriate for this analysis, since they were not of money from small donations in their congressional dis-
highly nationalized elections with wide partisan swings in the trict, and one who raises millions through national direct
collective national outcome of House races. To capture even mail appeals. Hence, the results obtained from the in-state
impacts at the marginal level, I exclude from this analysis only variable are likely to be somewhat less conclusive than the
those open seat candidates who were unchallenged in the other two explanatory variables.
general election, or faced challengers who did not meet the In a model of factors influencing the outcome of open
very low threshold for filing FEC disclosure reports.3 This seat elections from 1984 to 1992-with the GOP share of
gives the study a total N of 157 elections (314 candidates). the congressional vote as the dependent variable-prior
Using open seat races is substantively important to this political experience, relative candidate spending, presiden-
particular analysis for at least two major reasons. First, the tial coattails, black population, Hispanic population, and
presence of an incumbent member is typically the strongest the year of the cycle all exerted substantial and significant
force in deterring high-quality challengers (Jacobson and impacts on the outcome (Gaddie and Bullock 2000: 43). My
Kernell 1983; Mayhew 1974). Hence, examining open seatapproach is largely similar, in that I include control meas-
races automatically selects a sample of the most competitive ures designed to operationalize the underlying partisanship
races in the cycle, where the enormous influence of incum- of the district, the relative spending advantage of each can-
bency is minimized and second order factors, such as fund- didate, and candidate experience.
ing sources, are most likely to matter. Secondly, the com- The relative partisanship of the district is expressed as
petitiveness of the elections tends to encourage donors to the percentage of the total vote won by the presidential
liberally support one or both candidates, ensuring a diverse nominee of the candidate's party in the most recent election.
array of different fundraising formulas that help provide So, the 2002 and 2000 races include the appropriate data
leverage on the factors being examined. In this sample, on Bush and Gore, while the 1998 data used the 1996 Clin-
PACs accounted for anywhere from 0 percent to 77 percent ton and Dole vote totals. The relative spending variable is
of total candidate fundraising, self-financing percentages simply the percentage of total spending by the Democratic
ran from 0 percent to 92 percent, candidates raised as little and Republican candidates in the race accounted for by
as 12 percent and as much as 100 percent of their itemized each candidate's spending. Inclusion of this variable allows
individual funding from within their state. So, not only does for a test of the effect of fundraising by source while con-
the sample include competitive races, but also substantial trolling for total spending. Obviously, candidates are
variation in fundraising tactics. expected to do better when they are running on friendly
ground and outspend their opponents. By holding total
2 The choice of party has no impact on the result, since all variables are
spending and district partisanship constant across races, we
calculated relatively using the same measures for candidates of both can learn whether raising money from different sources in
parties. any way disrupts the expected relationship between district
3This resulted in very few candidates being excluded. In cases where can- partisanship and total spending.
didates were excluded, the lack of competition was clearly due to an
I also include a measure of "candidate quality" in this
overwhelming partisan advantage sufficient to make the general election
a moot point. However, this does underscore the importance of model-
analysis, although such measures are problematic. First, the
ing candidate entry and primary elections, a point which I address highly subjective and personalized nature of the factors that
below. make someone likely to win elected office in a particular

This content downloaded from 128.110.184.42 on Thu, 10 Aug 2017 15:51:05 UTC
All use subject to http://about.jstor.org/terms
GOOD
Gooo MONEYMONEY
AND BAD AND
MONEY 355
355 B

TABLE 1
OLS MODEL RESU

Variable Standardized Coefficient Coefficient (Standard Error)

Pac Donations 0.11* .0006418 (.0002375)
Self Financing -0.11' -.000415 (.0001462)
In-State Donations 0.08* .0004792 (.0002106)
Presidential Vote 0.43* .0024513 (.0003232)
Relative Spending 0.46* .207834 (.0237354)
Experience Dummy 0.03 .0040944 (.0054049)
Constant 0.386
N = 157, R-Squared =

place are test

difficult
whet
variance in "challe
donations,
Jacobson 1990; A
1988). RESULT
Secondly, th
variables essential
Absent an incumbe
The signi
for variables
candidate qualit
tricts. If While
the distr
vis
trolled forousby the
presen
candidate standard
quality we
influence the outco
variance t
not makeTo
an summ
attem
simply including
zero p a
val
not a candidate
a high has
lev
The data are
this reason
analy
parties, with
ing, thre
follo
generally ance and
outperfor
points. Secondly,
level of R
st
district terrain, ste
expected,
mean presidential
ning in fr
source percentage
over their
Republicans consist
There we
ing, holding
the a mean
donat
Consistent with
exerts a th
s
share as the
ance depen
at th
variables the
are opera
positi
candidate. With
dential th
vo
(which is already
candidate
party) the explan
financing
Democratic total m
specificall
allowing analysis
in o
self-fin
didate-by-candidat
drop off n
normalizingThe the
PAC h
donation variables
statisticall
ical transformation
the major
I specifically test
signed th
(ru
campaign Inreform
other a w
aligned. Specificall
candidate
PAC donations,
win. do
In-st
paigns, and
andraise th
subst
within their distric
correlatio
ketplace. So, holdin
raise relat

This content downloaded from 128.110.184.42 on Thu, 10 Aug 2017 15:51:05 UTC
All use subject to http://about.jstor.org/terms
356
356 POLITICAL RESEARCH POLITICAL RESEARCH QUARTERLY
QUARTERLY

are less previous campaigns,

likely to community win
activities, or service in
th
to home. elected office.
Tests for The first explanation is called into question somewhat by
autocorrelatio
substantialthe generally
problems low level of public concern regarding cam- w
correlation paign finance in general. Additionally, candidates
between anyoften
the matrix attempt to turn self-financing into a PR far
falling asset by proudly b
regressions touting
ofthe inability of special interests to influence their
possibly
to reveal campaigns. For example, Senator and record self-financer co
problematic
reason to Jon Corzine recently touted himself
suspect as "Unbought and
autoco
it anyway, Unbossed"
finding in his advertising (Steen 2000:no
5). There is, inco
across the fact,
threesome truth to this argument,cycles.
at least in the sense that
predicted PACs are less likely to donate to
values self-financing candidates.
does no
with non-constant error variance. In my research, self funding and PAC donations have a rel-
atively strong negative Pearson correlation coefficient of
CONCLUSIONS AND DISCUSSION -0.51. Repeatedly framing the issue in this way may coun-
teract media criticism, making the self-financing issue a PR
The bottom line, based on these results, is that where a wash. After all, one would be hard pressed to find an expe-
candidate raises money is likely to have a second order rienced political consultant who would not gladly accept a
effect on their electoral performance. The high significance few sermonizing newspaper editorials in return for 10,000
of a relative advantage in spending on the election outcome extra gross rating points on TV
indicates that even in the most competitive group of races in The second explanation revolves around the popular view
each cycle, being able to outspend an opponent is a key among many campaign operatives that self-financed candi-
determinant of success. The political terrain of the district dates suffer from not being forced to establish connections to
also exerts the expected substantial influence over candidate thousands of voters by soliciting campaign contributions.
success. However, the standardized coefficients for the pres- However, this explanation does not completely square with
idential vote variable and the relative spending variable (.43 the reality of candidates who fund well over half of their total
and .46, respectively) indicate that normal party support spending, and still raise substantial sums from individual
may actually matter a bit less than does a spending advan- donors. For example, according to FEC records, Corzine not
tage. Given the high explanatory power of a model based on only ended up spending over $63 million of his own fortune
these two factors alone (R-square of .84 versus R-square of winning a New Jersey Senate seat, but also convinced indi-
.87 for the larger model), we can conclude they are either viduals to donate a reasonably respectable $2.8 million.
largely determinative of success or representative of other Finally, self-financed candidates may often lack the
unseen factors that influence success. It should therefore be record of public service and practical political skills neces-
emphasized that funding source impacts are of a much sary to win a competitive open seat election. The interaction
lower level of magnitude, although still relevant in the of inexperience and substantial personal spending may be
heated competition of an open seat House race when even a particularly crippling if the threat or use of personal money
slight edge matters. enables a candidate to prevent the emergence of experi-
That being said, it appears that ceteris paribus, practical enced opponents or defeat them in a primary. The party in
electoral consequences are associated with the amount a question is then left with a nominee who is long on money
candidate raises from particular sources. PAC donations are and short on skills and experience, facing a general election
a predictor of candidate success, while self-financing is cor- opponent armed with both practical experience and the
related with failure. Further, some candidates may gain an financial resources of a national political party and numer-
additional boost from having a relative advantage in the per- ous wealthy donors. At least one researcher who specializes
centage of funding they raise within their states. I discuss in studying self-financing candidates fo und that, based on
each of these three results separately in the following para- data from House primaries in four cycles, "[a]ll else being
graphs, although further research would be needed to deter- equal, non-incumbents who self-financed $1 million aver-
mine which of them may explain the findings of this study. aged about two fewer opponents with experience in a lower
There would seem to be at least three likely explanations office than those who kept their personal funds in the bank"
for the negative impact of self-financing on electoral out- (Steen 2000: B01). Regardless of which of these-or
comes in these three cycles. First, it may be the case that other-explanations one accepts, the fact remains that the
self-financing results in criticism that costs wealthy candi- normative criticisms raised about the impact of self-financ-
dates votes. Secondly, self-financing may deter candidates ing can fairly be accompanied by questions about its practi-
from gaining the political value of linkages to interest cal efficacy.
groups, community leaders, and even individual voters that At first glance the positive correlation with PAC dona-
are an important side effect of political fundraising. Finally, tions and electoral success fits with what public opinion
candidates who self-finance may also be less likely to have data tells us about public concern over campaign financing.
the political experience and skills others might acquire from As the data I examine show, virtually any candidate in a

This content downloaded from 128.110.184.42 on Thu, 10 Aug 2017 15:51:05 UTC
All use subject to http://about.jstor.org/terms
GOOD
GOODMONEY AND BAD MONEY
MONEY AND BAD MONEY 357 357

the ever-increasingrace
competitive role that incumbents play inwill
raising PAC
The funds for competitive
public is challengers under their party banner
undoub
didates accept PAC
may eliminate the distinction by funneling access-motivated
raised frequently
money to non-incumbents. in
ads, the Given the somewhat weaker strengthallega
flying of the state dona-
Of tion percentage
course, variable and the lower amount of informa-
indifferen
not imply there
tion available on it, explaining what role it may be sho
playing is
between somewhat more difficult. Some
PAC research has found that
contrib
for total incumbent
spending.
members with high ideological profiles or legisla- T
this tive influence are likely
positive to receive more out of state contri-
associat
accepting PAC
butions (Grenzke 1988). However, since thesedon
factors may
reach out havetomixed electoral groups
impacts (ideological strength may both w
to attract and repel potential reaping
represent, supporters), untangling the posi-
actual donations
tive from the negative is difficult. Further,may
since most candi-
benefits dates raise the majority of their individual dona
ranging tions from
fro
tionally, within their states, thispractice
the variable does not have an enormous
group of amount
individual
of explanatory power. Finally, determining the flow
another of causation is difficult,
way in as candidates may
whicreceive more
felt beyond actual
money from instate ch
donors because they have a higher level
and Segal 1992).
of support (rather than gaining support from the process of
A more likely expla
raising money within their state). With those qualifications,
causal process flow
however, the results from this analysis do indicate that the
donating electoral success and failure of some candidates correlates
strategicall
words, PAC
with the amount of donation
money they raise within the state where
rain or they are running. Candidates who rely heavily on
candidate funding
qua
ables in this
from study
outside their states are statistically less likely to win.
include recommen
In conclusion, these results indicate that the relationship
polling data, between fundraising andorelectoral outcomes
know
may be more
districts or candidates. subtle and complex than observations based on total spend-
There is some support in the elections literature for the ing alone would indicate. Future research looking at the
assertion that interest groups and wealthy donors are precise causal mechanisms behind these correlations might
"strategic givers." For example, at least one study demon- well unearth useful information on the electoral and legisla-
strated a link between the number of professionals con- tive processes.
ducting tasks for a campaign and its fundraising success,
implying that strategic donors supported highly profession- REFERENCES

alized campaigns in expectation of a win (Herrnson 1992).

Elsewhere, PACs were found to be strategic-but limited in Abramowitz, Alan I. 1988. "Explaining Senate Election Out-
some cases by pre-existing partisan commitments-in
donation patterns (Eismeier and Pollock 1986). There is
also evidence that PACs follow "horse-race" newspaper cov-
erage and donate accordingly (Mutz 1995).
The positive correlation of PAC donatio ns with electoral
success also raises at least some potential for improving can-
didate quality measures. If PACs intend to give strategically
to candidates likely to win (as the above literature indicates)
and are often successful (as my results indicate), then it may
comes." American Political Science Review 82: 385-403.
Abramowitz, Alan I., and Jeffrey A. Segal. 1992. Senate Elections.
Ann Arbor: University of Michigan.
Eismeier, Theodore J., and Phillip H. Pollock, III. 1986. "Strategy
and Choice in Congressional Elections: The Role of Political
Action Committees." American Journal of Political Science 30:
197-213.
Gaddie, Ronald Keith, and Charles S. Bullock, III. 2000. Elections
to Open Seats in the U.S. House: Where the Action Is. Lanham,
MD: Rowman and Littlefield.

be possible for political scientists to benefit from the private Green, Donald P, and Jonathan Krasno. 1988. "Salvation for the
information PAC directors spend a great deal of effort acquir- Spendthrift Incumbent." AmericanJournal of Political Science 32:
844-907.
ing from journalists, pollsters, incumbents, and party insid-
Grenzke, Janet. 1988. "Comparing Contributions to U. S. House
ers. This possibility comports with recent findings that inside
Members from outside Their Districts." Legislative Studies Quar-
informants in congressional districts offer accurate predictive
terly 13: 83-103.
information that is not captured by measurements tradition-
Herrnson, Paul S. 1992. "Campaign Professionalism and Fundrais-
ally used in the field (Stone, et al 2004). One might argue ing in Congressional Elections." Journal of Politics 54: 859-70.
that work in this direction is limited by continued findings Herrnson, Paul S., and Clyde Wilcox. 1994. Risky Business? PAC
that PAC giving is primarily targeted to achieving legislative Decisionmaking in Congressional Elections. New York: Sharpe.
access rather than shaping the partisan character of congress Jacobson, Gary C. 1990. The Electoral Origins of Divided Govern-
by backing winners (Herrnson and Wilcox, 1994). However, ment. Boulder, CO: Westview.

This content downloaded from 128.110.184.42 on Thu, 10 Aug 2017 15:51:05 UTC
All use subject to http://about.jstor.org/terms
358 POLITICAL RESEARCH QUARTERLY

Jacobson, Gary C., and Samuel Kernell. 1983. Strategy and Choice
in Congressional Elections. New Haven, CT: Yale University
Press.
Mayhew, David R. 1974. "Congressional Elections: The Case of the
Vanishing Marginals." Polity 6: 295-317.
Mutz, Diana C. 1995. "Effects of Horse Race Coverage on Cam-
paign Coffers: Contributing in Presidential Primaries."Journal
of Politics 57: 1015-42.
Steen, Jennifer Anne. 2000a. "Money Isn't Everything: Self-Financ-
ing Candidates in U.S. House Elections, 1992-1998." Ph.D.
diss. University of California at Berkeley.
Steen, Jennifer Anne. 2000b. "Maybe You Can Buy an Election, But
Not with Your Own Money." The Washington Post June 25: B01.
Stone, Walter, Sarah A.. Fulton, Sandy Maisel, and Cherie Maestas.
2004. "Money, Electoral Prospects and Congressional Election
Outcomes." Paper presented at the annual meeting of the West-
ern Political Science Association, Portland.
Squire, Peverill. 1992. "Challenger Quality and Voting Behavior in
U.S. Senate Elections." Legislative Studies Quarterly 17: 247-63.
Received: April 12, 2004
Accepted for Publication: May 20, 2004
blalexa@emory.edu

This content downloaded from 128.110.184.42 on Thu, 10 Aug 2017 15:51:05 UTC
All use subject to http://about.jstor.org/terms