Fat Cats and Thin Kittens - Cato Institute · Fat Cats and Thin Kittens Are People Who Make Large Campaign Contributions Different? ... Fat cats are prime suspects in the conspiracy

Fat Cats and Thin KittensAre People Who Make Large Campaign

Contributions Different?by John McAdams and John C. Green

John McAdams is an associate professor of political science at Marquette University. John C. Green is a professor of politi-cal science and director of the Ray Bliss Institute of Politics at the University of Akron.

No. 76

Critics of campaign finance in the United Statesoften direct their fire toward contributors who makelarge donations. Critics charge that large contribu-tions are unfair, unrepresentative, and undemocrat-ic. Accordingly, they push for “reforms” that wouldfavor small contributions over large, and publicmoney over private donations.

Survey data on contributors contradict thatstereotype of contributors of large amounts andtheir effects on American politics. Overall, “fatcats” differ less from contributors of smalleramounts than critics have alleged. The differ-

ences that do exist are mostly unsurprising andgenerally small in magnitude. Survey resultsshow that both policy liberalism andDemocratic partisanship are well representedamong contributors of large sums.

The supporters of McCain-Feingold arguethat new restrictions on large contributions willprofoundly alter American politics for the bet-ter. Their claims have no basis in fact. New lawsaimed at restricting large donations in favor ofsmaller ones will have little effect on practicalpolitics.

September 25, 2002

Introduction

One of the most enduring stereotypes inAmerican politics is the “fat cat,” the wealthycampaign contributor. From the robberbarons of the gilded age to the corporate richof the mid-20th century to the purveyors ofsoft money in the 1990s, those who donatelarge sums have a firm place in the mytholo-gy, if not the reality, of American elections.Fat cats are prime suspects in the conspiracytheories of the left and right, potent propsfor cynics of every persuasion, and the greatbane of all reformers. Political scientists fre-quently focus on fat cats as well, assigninglarge contributions a prominent place inexplanations of political power and partycoalitions. Such assertions are rarely basedon more than anecdote and conjecture.1

This paper juxtaposes the fat-cat stereo-type with an empirical picture based on sur-vey data. We have collected a national sampleof donors to competitive congressional racesand public records of their subsequent feder-al donations. We use those data to describethe political and social correlates of donorsacross the gamut of contributions.

We conclude that new laws aimed at restrict-ing large donations in favor of smaller ones willhave only modest effect on practical politics.

Criticisms of Contributorsof Large Sums

Critics of fat cats in American politicsargue that contributions influence cam-paigns and the policies proposed and passedby elected officials. Money in politics, theysay, has the same effect it has in the market-place: the more money people spend, themore influence they can “buy.” If that is so,contributors of large amounts are especiallyinfluential in politics by virtue of their finan-cial largesse.

Critics decry this influence because theybelieve democracy and fairness require all citi-zens to have an equal opportunity to influence

policymaking.2 In this view, large donationsare unfair and undemocratic because personalwealth is distributed unequally among citi-zens, which gives the wealthy more opportu-nity to affect the political process.3 Critics alsobelieve that money has relatively more influ-ence on politics than other resources, whichare more evenly distributed, such as votes,time, energy, or wisdom.4 When push comesto shove, the critics say, the special potency oflarge donations can overcome other efforts byless affluent citizens.

Critics also believe that fat cats give inpursuit of private, selfish interests. “Bigmoney” pursues material concerns, allowingdonors to seek to maximize personal returnsfrom their political “investments” ratherthan pursue a more public-spirited agenda.5

Moreover, large donations make the politicalsystem less representative of the public as awhole, even if they are made with a broaderagenda in mind. The wealthy are differentfrom you and me, the critics claim, and notjust because they have more money. Theyhave different class interests, are more con-servative, and are likely to support theRepublican Party.6

Because large contributions are unfair,undemocratic, corrupt, and unrepresenta-tive, critics have sought and obtained limitson the size and level of individual contribu-tions. They believe that smaller donations aremore fair, democratic, public spirited, andrepresentative. Government financing ofcampaigns supported by voluntary incometax check-offs represents the logical exten-sion of such thinking: individual contribu-tions in such systems are nearly equal andtoo small to threaten other forms of partici-pation, garner special favors, or overrepresentspecial interests. Those who give small dona-tions, “thin kittens,” are thus seen as muchbetter for the functioning of the political sys-tem than fat cats.

What does the scholarly literature have tosay about these assumptions, conclusions, andcorollaries regarding fat cats? First, scholarsoffer a series of qualifications to the critics’broad assertions about the power of money in

2

Fat cats are primesuspects in the

conspiracy theo-ries of the left and

right, potentprops for cynicsof every persua-

sion, and thegreat bane of all

reformers.

politics. For example, studies reveal that moneyaffects electoral outcomes ambiguously andindirectly; campaign spending often mattersless for well-financed candidates than for thosewho are less well funded.7 Similarly, studiesshow that money affects legislative behavior incomplex ways; one of many influences, moneyis rarely the most important and is often incon-sequential.8

Second, research on campaign contribu-tors reveals a more complex picture. On theone hand, donors as a group differ substan-tially from the general public; they are farmore affluent and politically active.9 Indeed,giving even a small amount puts an individ-ual in the activist corps, and other kinds ofactivists also differ from the general public ina similar fashion.10 On the other hand,donors are quite diverse politically, represent-ing a wide variety of attitudes, motivations,and forms of activism.11

This analysis looks at the criticisms of largecontributors by testing three hypotheses:

• Hypothesis 1: Givers of large donationsshould be much wealthier than contrib-utors of smaller amounts (and thustheir participation is potentially unfairand undemocratic).

• Hypothesis 2: Givers of large donationsshould be motivated more by materialand instrumental concerns than con-tributors of smaller amounts (and thustheir participation is potentially venal).

• Hypothesis 3: Givers of large donationsshould have different attitudes thancontributors of smaller amounts.Specifically, we would expect givers oflarge donations to espouse conservativeeconomic views and be Republican par-tisans (and thus their participation ispotentially unrepresentative of the pub-lic as a whole).

Findings

We sampled campaign contributors frompast elections and surveyed them to test

those conjectures. The Appendix describesour sampling approach and the subsequentsurvey of contributors that underpins thisstudy. We turn now to assessing each hypoth-esis in light of our data.

Hypothesis 1: Givers of Large DonationsAre Wealthier Than Contributors ofSmaller Amounts

Table 1 presents demographic differencesbetween contributors of small, moderate, andlarge amounts. Small contributions were lessthan $500, moderate contributions were from$501 to $5,000, and large contributions were$5,001 or more. The contributors of largeamounts are more likely to be male, less likelyto belong to a labor union (although very fewcontributors in any category do), and morelikely to be over 60 years of age. The occupa-tional data are interesting. Contributors oflarge amounts are a bit more likely to beemployed in private-sector businesses and abit less likely to be employed in government.They are also a little more likely to be lawyers.

One variable, however, overwhelms all theothers: income. Ninety-four percent of con-tributors of large amounts make more than$100,000 annually. Campaign contributionsare, in other words, what economists call a“normal good”: you consume more of themas your income increases. As the termimplies, this finding is about the most unsur-prising one imaginable.12 As might be expect-ed, contributors of large sums are wealthierthan donors of smaller amounts and thepublic at large.

Hypothesis 2: Givers of Large DonationsShould Be Motivated More by Materialand Instrumental Concerns ThanContributors of Smaller Amounts

Table 2 explores the motivations of contrib-utors by looking at what respondents saidprompted their contributions to political cam-paigns. We asked whether various factors were“very important,” “important,” “somewhatimportant,” or “not important” to their deci-sion to contribute. Not all the alternativesoffered are very interesting or theoretically

3

Studies show thatmoney affects leg-islative behaviorin complex ways;one of manyinfluences,money is rarelythe most impor-tant and is ofteninconsequential.

important. We wanted to give respondentsplenty of opportunity to offer widely acceptable“goody two-shoes” answers, in the hope thatthey would then be more forthcoming aboutself-interested motivations.

In fact, the givers of small contributionswere a bit more likely to admit self-interestedmotivations. They were more likely to say that

they would contribute to a candidate whowould “do a good job of watching out for theinterests of the district” (the “state” in Senateelections), more likely to admit that they sup-ported incumbents, and more concerned that acandidate would, if elected, “be accessible toconstituents.” Givers of large donations were abit higher on “the candidate is reasonably

4

Givers of smallcontributions

were a bit morelikely to admitself-interested

motivations.

Table 1Demographic Characteristics

Small Moderate LargeCharacteristic Contributors Contributors Contributors Tau-b Significance*

Male 69.3% 78.0% 84.8% .12 .001Black 1.6% 1.3% 0.9% .01 noHispanic 0.6% 1.0% 0.6% .01 noAsian 1.3% 1.7% 1.9% .01 noUnion member 5.1% 0.7% 0.6% .11 .001Occupation:

professionalor managerial 76.4% 81.6% 80.1% no

Sector of employmentSelf-employed 44.7% 49.6% 46.6% —Government

employee 14.0% 4.7% 2.3% —Private-sector

businessemployee 34.9% 42.4% 48.0% —

Private,nonprofitemployee 6.2% 3.3% 3.0% .001

At least somegraduate training 56.3% 57.5% 59.9% .05

Family income over$100,000 per year 50.1% 80.5% 94.1% 0.33 0.00†

ReligionProtestant 45.2% 47.5% 46.3% .01 noCatholic 23.6% 19.7% 17.4% .05 .01Jewish 11.5% 18.0% 18.3% .08 .001Consider self “not

really religious” 25.5% 23.5% 25.9% .00 noConsider self

“fundamentalist” 5.0% 3.5% 1.9% .05 .01Age: over 60 34.7% 39.3% 42.7% .06 .001A lawyer 15.9% 26.2% 25.2% .11 .001

*Significance of Tau-b coefficient, which indicates the probability that this relationship with ordinal data ismerely the result of chance.†Significance of chi-square, which indicates the probability that this relationship with nominal data is merely theresult of chance.

5

Table 2 Motivations for Contribution (percentage who said particular factor was “very important”)

Small Moderate LargeItem Contributors Contributors Contributors Tau-b Significance*

Candidate “would be effective in getting things done in Congress” 53.4 52.0 47.5 .03 no

Candidate “generally agrees withme on policy questions” 32.6 30.7 28.3 .02 no

Candidate “has the very highestethical standards” 67.6 69.9 69.7 0.02 no

Candidate “will do a good job of watchingout for the interests of the state [district]” 33.9 31.6 26.5 0.05 0.01

Candidate “is reasonably friendly toward the interests of people in my industry or line of work” 17.4 18.6 20.3 0.02 no

Candidate “will exercise independentjudgment on policy questions” 49.8 53.1 41.9 0 no

Candidate “is running against an opponent that I find unacceptable” 31.5 39.0 23.4 0.02 no

Candidate “is an incumbent” 3.3 4.0 1.3 0.07 0.05Candidate “will (if elected) be accessible

to constituents” 33.5 29.4 19.3 0.08 0.01Candidate “agrees with me on one (or a few)

very important issues.” 33.0 32.5 35.1 0 no

*Significance of Tau-b coefficient.

Table 3Circumstances of Contribution


Got request in mail 37.3% 32.1% 36.0% .03 noSomeone working for the campaign

(not known to me personally) calledto ask for a contribution 6.0% 6.0% 4.7% .01 no

I attended a fundraising event (suchas a dinner, coffee, or reception) 24.6% 28.5% 25.5% .02 no

An economic interest group (union,corporation) or trade association supported the candidate 38.9% 39.9% 30.7% .03 no

An interest group promoting a particular cause supported the candidate 6.4% 5.5% 5.9% .01 no

People where I work were encouragedto contribute to [name of candidate] 11.5% 9.1% 8.1% .04 .05

I made the contribution on my own initiative 4.9% 3.5% 2.5% .04 .05


friendly toward the interests of people in myindustry or line of work,” but the difference isnot statistically significant. On net, we have lit-tle evidence that givers of large amounts andcontributors of smaller sums have fundamen-tally different motives for giving.

However, thus far the analysis has beenbased on respondents’ own, perhaps self-serv-ing, accounts of their motivations. A differentapproach is to ask respondents about the con-crete proximate circumstances that accompa-nied a particular campaign contribution. InTables 3 and 4, we see respondents’ accounts ofthe circumstances surrounding a single contri-bution they made in 1988 to a single candidate.

Little in these data supports the notionthat fat cats are any different from people whomake smaller contributions. The only two dif-ferences that reach statistical significance arethe fact that the latter were more likely to saythat “people where I work were encouraged tocontribute” and that they were also more like-ly to say that they made the contribution ontheir own initiative. In regard to the first dif-ference, one imagines apolitical business exec-utives dutifully making a contribution tosome candidate the company’s chief lobbyisturges them to support. In regard to the sec-ond, one assumes that the people who makesmaller contributions are on fewer lists of like-ly contributors and thus not so relentlesslysolicited as the fat cats.

Hypothesis 3: Givers of Large DonationsShould Have Different Attitudes ThanContributors of Smaller Amounts

We would expect givers of large donationsto be more conservative and Republican thancontributors of smaller amounts. The datapresented in Table 5 yield a rather clear pic-ture. Fat cats are a bit more economicallyconservative, perhaps a bit more conservativein foreign policy, and not significantly differ-ent on social and racial issues. Two differentforms of the questionnaire asked about poli-cy preferences in two different ways. Form Aasked people whether they were liberal orconservative on four different issue areas: (1)economics, (2) social issues, (3) foreign poli-cy, and (4) racial issues. As can be seen, the fatcats were about ten percentage points lesslikely to call themselves “liberal” on econom-ic issues13 and about nine percentage pointsless likely to call themselves “liberal” on for-eign policy issues.14 Neither social issues norracial issues show any statistically significantpattern of differences. Fat cats were alsomodestly more Republican than the donorsof smaller amounts—by about eight percent-age points. Looked at another way, there arealmost as many liberals and Democratsamong donors of smaller amounts as amongfat cats, despite the greater conservatism andRepublicanism of the latter.

We should keep in mind, however, that

6

Table 4How People Who Attended a Fundraising Event Came to Do So


I got an invitation in the mail 53.2% 53.1% 50.5% .01 noA friend or acquaintance invited me 42.2% 43.2% 41.4% 00 noAn economic interest group or tradeassociation encouraged me to attend 7.5% 6.2% 13.1% .02 noAn issue-oriented interest group thatI belong to sponsored the event 8.0% 6.2% 10.1% .00 noPeople where I work were encouraged to attend 3.8% 4.6% 2.0% .00 no


contemporary liberalism is more than govern-ment activism in the economy and dovishnessin foreign policy. About one-third of fat cats inour sample identified themselves as liberal onsocial and racial issues. In general, they werenot significantly less liberal on social andracial issues than givers of smaller amounts.

On the whole, we might conclude that con-tributors of large sums are slightly less liberalthan donors of smaller amounts. That cau-tious conclusion, however, flies in the face ofthe rhetoric and expectations of people whosupport limits on “big money.”

If we turn from people’s general labels to

Table 5Partisanship and Policy Preferences


Democratic Party identification 39.7% 33.9% 34.6%Republican Party identification 43.7% 50.5%5 51.7% .06 .01Claims to be “liberal” or “very liberal” on

“economic issues” 28.6% 18.3% 17.4% .12 .001Claims to be “liberal or “very liberal” on

“social issues” 40.9% 38.1% 34.5% .04 noClaims to be “liberal” or “very liberal” on

“foreign policy issues” 30.0% 18.3% 21.1% .09 .001Claims to be “liberal” or “very liberal” on

“racial issues” 40.5% 33.3% 34.7% .04 noA Ronald Reagan supporter 45.8% 51.8% 53.7% .06 .01A Jesse Jackson supporter 7.5% 2.5% 2.5% .09 .001A Pat Robertson supporter 2.8% 1.7% 1.2% .04 .05A labor union supporter 10.8% 7.5% 7.8% .04 .01Agrees that “the United States is spending

too much on defense” 59.0% 59.3% 57.0% .02 noSupports school prayer constitutional

amendment 30.7% 31.4% 27.9% .00 noWants “affirmative action programs . . .

pursued much more vigorously” 32.3% 29.5% 26.6% .03 noSays the United States “should never

have supported the Contra rebels” 37.8% 29.0% 27.5% .07 .01Says government should “take strong action”

to “limit or abolish the practice of abortion” 20.6% 16.4% 17.0% .00 noAgrees that “public policy should discourage

the ownership and use of handguns.” 56.4% 59.8% 60.9% .03 noSays that “a comprehensive national health

insurance program would be a serious mistake” 35.7% 44.6% 50.3% .09 .001Says that “busing to integrate the public schools

has done more harm than good” 43.5% 49.0% 48.8% .03 noFavors “a substantial increase in the federal

minimum wage” 34.7% 25.5% 25.2% .09 .01Agrees that “what this country needs is more

governmental planning of the economy” 20.1% 3.5% 0.0% .10 .01


7

their positions on various specific issues, wefind the same pattern. Fat cats are a bit moreconservative on government economic plan-ning, the minimum wage, and national healthinsurance. They are a bit less likely to labelthemselves “union supporters.” Responses tospecific foreign policy questions paint a mixedpicture. Fat cats are a bit less likely to say thatthe United States “should never have supportedthe Contra rebels in Nicaragua” but no less like-ly to say that the United States “is spending toomuch on defense.” All groups seem to share inthe belief that the military establishment builtup to fight the Cold War could be downsizedfollowing the collapse of communism.

Finally, we turn to the intensity of opinionamong different strata of contributors. Arethe contributors of large amounts more likelyto have strong political views, to resist com-promise, and to insist that their party musttake strong positions that satisfy them, or arethey (conversely) more inclined to compro-

mise than other classes of contributors?Table 6 shows some evidence that contribu-

tors of smaller amounts are more adamant intheir political views. For example, they are a bitmore likely to say that they would rather “losean election” than “compromise [their] basicphilosophy.” By similarly modest margins, theyare less likely to be able to “overlook policy dif-ferences” in order to “support [their] party’scandidate.” Conversely, they are more likely toinsist that they must “agree with the candidateon major policy issues.” On the other hand,they are less likely to claim “strong politicalviews.” They are also more willing to say that“the major political parties are run for the ben-efit of powerful special interests.”

Conclusion

Are the fat cats different from you andme? Yes, they have more money. And to the

8

Table 6Attitudes toward the Political Process


Would “rather lose an election” than “compromise my basic philosophy” 71.5% 68.9% 63.5% .05 .01

Says that “the best thing” about Americanpolitical parties is that they “have room for every point of view” 45.5% 45.6% 44.6% .01 no

Says that “parties ought to take strong, clearstands on important public issues” 89.9% 91.1% 87.1% .03 no

Says that “the American political system makesreal change much too difficult” 53.7% 47.1% 51.0% .05 .01

Says that “the major political parties are mainlyrun for the benefit of powerful special interests” 42.1% 30.3% 20.2% .16 .001

Leans toward “almost always [can] support . . . party’s candidate and are willing to“overlook policy differences” 10.2% 10.7% 18.0%

Center position 44.4% 44.8% 45.3%Leans toward “must agree with the candidate on

major policy issues” 45.4% 45.5% 36.9% .05 .01Claims to have “strong political views” 58.4% 61.1% 73.0% .08 .001


extent that their greater income allows themto be more active in politics, the critics mayhave a point. It is possible, though far fromproven, that higher income may bring someunfairness and undemocratic participationinto the political process.

But the systemic consequences of suchparticipation by fat cats are far from clear.

It is true that fat cats are a bit more eco-nomically conservative than contributors ofsmaller amounts. But an analysis of some ofour findings suggests that the differencesmay not be very important. For example,contributors of large sums are more opposedto government planning of the economy, butvery few people in any group of contributorslike such government planning. Fat cats aremore opposed to increasing the minimumwage, but an increase in the minimum wagesailed through Congress in 1996 because ofits public popularity—fat-cat opinionnotwithstanding. Fat cats are less likely tofavor government health insurance, butClinton’s health care proposal went down todefeat when public opinion turned againstit—something that was probably inevitablewhen the complexities were aired.

It is difficult to see, however, how peoplewho make frequent and rather large contri-butions to political candidates, parties, andpolitical action committees differ very muchfrom those whose gifts are fewer and smaller.Reducing the power of the fat cats in favor ofcontributors of smaller amounts would notchange the balance of power between liberalsand conservatives much.

Our findings do, however, undermine theconventional stereotype of the maker of largecampaign contributions. On the whole, suchcontributors may be slightly less liberal than mak-ers of smaller contributions, but fat cats hardly fitthe mold of conservative Republicans dead setagainst social and racial progress; giving more tocampaigns does not preclude giving to advanceDemocratic candidates and liberal causes.

Compared with people who make small con-tributions, people who make large donationsseem a bit more pragmatist than purist in theirorientation—although the differences are mod-

est at best. That is not because fat cats are venal-ly motivated, while contributors of smalleramounts are motivated by purposive concerns.As we have seen, the distribution of motivationsseems to vary only trivially across groups.

Our findings—like those of almost anyinquiry into this issue—have some implica-tions for recently passed campaign financereform legislation. The rhetoric from manyreformers stressed the need to limit the powerof “big money.” It’s an interesting questionwhether the Bipartisan Campaign Reform Actwill actually reduce the power of fat cats,because while it targets “soft money” it actual-ly eases restrictions on the more useful andversatile “hard money.” But even if we assumefor the sake of argument that it will actuallyredistribute power from those who contributelarger amounts to those who contribute less—or vice versa—the actual policy implicationswould likely be minor. Restricting the contrib-utors of large sums in favor of donors ofsmaller amounts will not the change the polit-ical landscape very much, if at all.

Appendix: SurveyingCampaign Contributors

Our study starts with a sample of cam-paign contributors that was obtained fromthe files of the Federal Election Commission.We surveyed contributors by mail, using themailing addresses in FEC records. The pri-mary sampling unit was a 1988 race (for theHouse or Senate), and we sampled contribu-tors to both the Democratic and theRepublican candidate for each race.

For the Senate races, we chose 10 states bya random process. This yielded 20 campaignsthat were part of the sample. For the House,we chose 20 races by a random process. Thisyielded only 35 campaigns (since in 5 of thechosen races an incumbent was runningwithout major party opposition). In addi-tion, an oversample of 10 competitive Houseraces was drawn. This was done to guaranteereasonable representation of an exceptionallyinteresting class of contributors.

9

Little supportsthe notion thatfat cats are any different frompeople who makesmaller contributions.

For each candidacy in the sample, we drewnames from FEC records of contributorswho gave money during 1988. We set an over-all target of 100 names for each Senate can-didacy. For each House candidacy, the targetwas 30 names. We examined all preelectioncontributor reports and drew a roughly equalnumber of names from each report of 1988contributors. However, since the sample wasdrawn during December 1988, no postelec-tion reports were included.

If four reports filed during 1988 con-tained the names of contributors, the targetfor a particular campaign was divided byfour, and this became the “report target”: thenumber of names to be drawn from eachreport. In a typical Senate case, 25 nameswere drawn from each of four reports. Ifinsufficient names were found in earlyreports to reach the target, additional nameswere drawn from later reports. A few cam-paigns simply had fewer total contributorsthan the target.

For sampling purposes, a cluster consist-ed of the contributors listed on a single pageof an FEC report. If a report contained sevennames per page, and the report target was 25names, four pages would be chosen by a ran-dom process. As this example suggests, therewas a bias during the sampling process tooverfulfill the quota somewhat. This provedto be advantageous, since a substantial num-ber of the names sampled proved to be dupli-cates (as we will discuss below).

A Purposive, Rather than Representative,Sample

As this description makes clear, this wasnot a representative sample of contributors.For example, to get a representative sample,the target number of contributors to eachcampaign would have to be in proportion tothe total number of contributors, or else theprobability of inclusion of a campaign in thesample would have to be proportional to thetotal number of contributors. Rather, thenumber of contributors selected was roughlyequal for each campaign. Further, a represen-tative sample would have to draw names from

each report in proportion to the total numberof contributors listed. Finally, the inclusion ofan oversample of competitive House races is adeviation from representativeness.

The sampling design used was chosen tobalance practical and theoretical considera-tions. Data available from the FEC could beused to determine the number of contribu-tors giving $500 or more. But we wanted toinclude contributors down to the $200 level.On the other hand, it would have been quitepractical to set the report target in a way thatgave early contributors (typically fairly few innumber) a probability of inclusion equal tothat of late contributors (typically numer-ous). It was decided that early contributorsconstituted an important analytical categoryand that it was important to get a reasonablylarge absolute number.

Much the same logic was applied to sever-al categories of contributors. The sample wasdesigned to get a reasonably large number ofcontributors in highly competitive races, aswell as a reasonable number of contributorsto hopeless, “not a snowball’s chance” cam-paigns. It was, in other words, a purposivesample, drawn with an eye to the interestingquestions it was intended to illuminate. Thismeans, however, that we must be very carefulabout saying that “X percent of campaigncontributors” did such and such.

As the data in Table 7 suggest, our sam-pling procedure served us well for the pur-pose of studying fat cats. A representativesample of contributors would have beenrather problematic, since among the ranksof all contributors fat cats are few in num-ber. Our sample, in contrast, was prettymuch a sample of contributions. Relative to asample of all contributors, it heavily over-sampled the population of people whomake multiple, frequent contributions.

Given our purposive sample, readers maywant to be a bit wary of taking the percentageshere as absolute gospel. But the relationshipscan be accepted with a good bit of confidence.There is no reason to believe we have oversam-pled liberals who give a lot of money or conser-vatives who give little (or vice versa).

10

Contributors ofsmaller amounts

are moreadamant in their

political views.

The SurveyThe sampling procedure described above

yielded a total of 4,789 names. These wereentered into a database and screened forduplicate entries and apparent spouses ofother contributors in the database. The elim-ination of 528 duplicate entries and 116 con-tributors who were spouses reduced thenumber of potential respondents to 4,145. Afurther 60 names had to be eliminatedbecause of bad addresses; 20 were eliminatedbecause the respondent was dead or disabled;and 55 were eliminated when mail wasreturned stamped “forwarding orderexpired.” Two people insisted that they hadnot made a contribution and were includedon the list by mistake. Thus, our sample ofpotential respondents totaled 4,008.

Mail questionnaires were sent to potentialrespondents during the period May 1989–April1990. An initial mailing and as many as threefollow-ups (if necessary) were sent. A total of2,340 usable questionnaires were returned, for aresponse rate of 58.4 percent. Nine respondentsremoved the identifying information (a serialnumber) from the questionnaire and could notbe used in this study.

The questionnaire asked contributors for avariety of information, including their party

identification, their political opinions, the cir-cumstances surrounding the contribution,and their relationship with the candidate.Further, respondents were asked to rate sever-al “factors someone might take into accountwhen deciding to contribute money to aparticular candidate for the U.S. [chamber].”

Another set of questions asked about theproximate conditions surrounding the con-tribution. Where the contribution was givenat a fundraising event (such as a dinner, cof-fee, or reception) respondents where askedhow they got invited to the event. Theseseries of questions were designed to get at theall-important question of the incentives thatmotivate contributions.

The next stage of the project involvedsearching the digitized records of the FEC forthe names of persons in our sample. The FECdigitizes campaign contribution records as amatter of course. For the 1988 election cycle,persons who made contributions of as littleas $500 are included in the FEC database. Incontrast, in the 1990 and later election cycles,contributors of as little as $200 are includedin the FEC data. Because of this, and becausethe questionnaires were actually adminis-tered during the 1990 election cycle, wedecided to use FEC data from the 1990 elec-

Table 7Distribution of Contributions

Contributions Number Percentage

Count of Contributions0 1,166 50.0%1 242 10.4%2 180 7.7%3 108 4.6%

4–6 212 9.1%7–9 124 5.4%10–15 128 5.4%16 or more 171 7.3%

Aggregate total of contributions$500 or less 1,405 60.3%$501 through $5,000 604 25.9%$5,001 and above 322 13.8

11

Our findingshave some impli-cations forrecently passedcampaign financereform legisla-tion.

tion cycle and to add 1992 data to get abroader base of contributions and a longertime span to provide a more reliable assess-ment of contribution activity.

Bob Biersack of the FEC conducted a com-puterized search of the agency’s records andsupplied to the authors the records of personsin our sample. This opens up the possibility ofanalyzing those contributors who make mul-tiple contributions. Thus, if an individual fellinto our sample by virtue of having made asingle contribution, data from the FEC allowus to see what other contributions he or shemade. One complication here arises from thefact that our sample was based on hard-copyrecords and many campaigns voluntarilyreported contributions of $50 or $100, andwhere available, we used these names to drawour survey sample. This raises the possibilitythat some persons making multiple, smallcontributions will show up in our data asmaking (for the 1990 and 1992 election cycles)zero contributions and contributing zeroaggregate dollars when in fact they are makingcontributions. We doubt that any of thesepeople belong in our fat–cat category (seebelow) and doubt that more than a tiny hand-ful would make it into our moderate contri-butions category. In sum, our focus was onfat–cat contributors. Our ability to pull out ofour sample people who made multiple, andrelatively large, contributions allowed us toidentify these people effectively.

Data ProblemsMatching contributors in our sample with

FEC records has proven to be a tediouschore—as anyone who has worked with thesedata would expect. The same individual mightbe listed under two or three different forms ofhis or her name. If records list “Joe Smith” asmaking several contributions, and “Joe Smith,Jr.” as having made additional contributions,we may have to wonder whether this is two dif-ferent persons or only one. Further, contribu-tors will frequently have both a residence anda work address—and some contributions willbe reported to the FEC under one address andothers under the other. Happily, multiple

pieces of data exist to help sort out theseissues. Our hard-copy records included themailing address of the contributor (which wasincluded in our database), and FEC recordsinclude either the occupation or the place ofwork of contributors.

Notes1. See Robert Biersack, Paul S. Herrnson, andClyde Wilcox, eds. Risky Business? PAC Contributionsin Congressional Elections (Armonk, N.Y.: M. E.Sharpe, 1994).

2. Richard Briffault, “Public Funding and Demo-cratic Elections,” University of Pennsylvania LawReview 148 (December 1999): 566.

3. Robert Biersack et al., “The Allocation Strategies ofCongressional High Rollers: A Preliminary Analysis,”Paper presented at the annual meeting of the MidwestPolitical Science Association, Chicago, 1994.

4. David Nichols, Financing Elections: The Politics ofan American Ruling Class (New York: NewViewpoints, 1974).

5. Thomas Ferguson, Golden Rule (Chicago: Uni-versity of Chicago Press, 1995).

6. Dan Clawson, Alan Neustadt, and James Bearden,“The Logic of Business Unity: Corporate Contri-butions to the 1980 Congressional Election,” AmericanSociological Review 51 (1986): 797–811.

7. Gary C. Jacobson, “The Marginals NeverVanished: Incumbency and Competition in Elec-tions to the U.S. House of Representatives,1952–1982,” American Journal of Political Science 31(1987): 126–41.

8. John Wright, “PACs, Contributions, and RollCalls: An Organizational Analysis,” AmericanPolitical Science Review 79 (1985): 400–14; and LarrySabato, PAC Power (New York: Norton, 1984).

9. Ruth Jones and Warren Miller, “FinancingCampaigns: Macro Level Innovation and MicroLevel Response,” Western Politics Quarterly 38(1985): 187–210.

10. Sidney Verba, Kay Schlozman, and HenryBrady, Voice and Equality: Civic Voluntarism inAmerican Politics (Cambridge, Mass.: HarvardUniversity Press, 1995).

11. John C. Green and James L. Guth, “Big Bucksand Petty Cash: Party and Interest Group

12

Restricting thecontributors of

large sums infavor of donors

of smalleramounts will not

the change thepolitical land-

scape very much,if at all.

Activists in American Politics,” in Interest GroupPolitics, ed. Allan Cigler and Burdett Loomis, 2ded. (Washington: CQ Press, 1986), pp. 91–113.

12. The reader may be wondering whether any of thebivariate relationships in Table 2 might go away (orperhaps become more important) in a multivariateanalysis. We addressed that issue with a multivariateregression (OLS) model. The dependent variable wasthe natural log of aggregate total contributions, andthe independent variables were those in Table 2,with a change or two. Certain relationships did goaway. Union membership and employment in thenonprofit private sector ceased to show any statisti-cal significance. No religious category seemed to bedistinctive. Four variables continue to show robusteffects. Males were more likely to be large rather thansmall contributors, as were those with high incomesand older people. And lawyers were, even after con-trolling for all the other variables, more likely to belarge contributors.

13. Further defined as issues “such as regulationof business, the minimum wage, social welfareprograms.”

14. Further defined as issues “such as Contra Aid,arms control, or the defense budget.”

15. This oversample was chosen by a randomprocess from a list of races listed as “marginal” inthe Cook Political Report of August 17, 1988.

16 Further, the relevant FEC data were not avail-able when the sample was drawn.

17. Invariably, contributions from apparentspouses appeared to have been coordinated, i.e.,were to the same candidate, given on the same ornearly the same date, etc.

18. Two different forms of the questionnaireasked for this information in different formats.

13

Published by the Cato Institute, Cato Briefing Papers is a regular series evaluating government policies andoffering proposals for reform. Nothing in Cato Briefing Papers should be construed as necessarily reflectingthe views of the Cato Institute or as an attempt to aid or hinder the passage of any bill before Congress.Additional copies of Cato Briefing Papers are $2.00 each ($1.00 in bulk). To order, or for a complete listingof available studies, write the Cato Institute, 1000 Massachusetts Avenue, N.W., Washington, D.C. 20001, call(202) 842-0200 or fax (202) 842-3490. Contact the Cato Institute for reprint permission.

Documents

Fat Cats and Thin Kittens - Cato Institute · Fat Cats and Thin Kittens Are People Who Make Large Campaign Contributions Different? ... Fat cats are prime suspects in the conspiracy