Ballot Cues,Business Candidates,and Voter Choices in Local Elections

Abstract

American voters commonly express abstract support for candidates with a business background, yet there is minimal systematic evidence about whether it advantages candidates in actual electoral contests. We examine this question using observational data, drawing on a California law allowing candidates to designate their occupational background on the ballot, and experimental data. Candidates with a business background are prevalent in California. However, neither of our studies indicate that business candidates enjoy atypical overall electoral success (although Republican leaning constituencies have a notably more favorable view of such candidates). A political background predicts electoral success far more effectively. Further, “small business owners” have more success than other business candidates, suggesting that voters consider the specifics of a candidate’s business experience. These results advance our knowledge of decision making in low-information elections, how voters weigh private-sector versus political experience, and how they filter occupational information through a partisan lens.

Keywords

local elections voting behavior voting cues and heuristics

Consider two claims about American voters. First, they tend to make heavy use of simple cues and heuristics, including information that is present on the ballot itself such as party identification, gender, and ethnicity (Boudreau et al., 2015; McDermott, 2005). Second, Americans generally express strong support in the abstract for candidates with a business background. A 2014 Gallup Poll found that 81% of Americans thought that the country would be governed better if more people with business and management experience were in political office (McCarthy, 2014). Similarly, the 1998 Gallup Democratic Processes Survey found that nearly a third of respondents thought the political system would be better if decisions were left to successful business people (Hibbing & Theiss-Morse, 2002, pp. 137–139). As one recent study (Coffé & Theiss-Morse, 2016, p. 56) stated, “Many Americans seem to view successful businesspeople as capable, competent, and efficient, and unlikely to waste taxpayers’ money.”

While the first claim about heavy use of cues and heuristics is widely accepted by scholars, it is much less clear that identification of candidates as businesspeople provides a significant behavioral cue—it is one thing for voters to express support for “running the government like a business” and quite another to be inclined to support a business candidate when other choices are available. We know relatively little about how occupational cues affect voting choices in the United States or abroad (Coffé & Theiss-Morse, 2016), though the small body of literature in this area suggests they may be significant (see especially Crowder-Meyer et al., 2019; Mechtel, 2014). In the American context observational research typically is hampered because (outside of California) occupational background is not listed on the ballot, making it difficult to determine the impact of simple cues in this area. The limited available empirical evidence therefore often comes solely from experimental studies, not combined with observational data.

To understand the potential impact of occupational cues generally and the businessperson designation specifically, it makes sense to focus on voting choices in California. This is because California allows candidates to list their occupation on the ballot, something that rarely is permitted in other states. Candidates generally take advantage of this opportunity, allowing researchers to determine how occupational designations immediately available to citizens by simply inspecting their ballots influence voting decisions. In addition, California law prohibits candidates for local office from identifying party affiliation on the ballot, thereby removing a major competing type of explicit cue. The sheer size and diversity of the state also allows for analyzing results in a wide variety of circumstances; California has 482 cities ranging from tiny rural towns with a few hundred people to major urban municipalities with populations larger than many states.

In this article we assess how individuals with a business background perform as candidates for local political offices in California. We focus on candidates who designate a business background both because this is an especially frequent ballot designation and because public opinion surveys suggest they may be attractive to voters. We draw on two data sources to answer our research questions: observational data from over 2,500 city council and mayoral races in California from 2008 to 2015, as well as an experiment in which occupational cues given to voters varied.

American voters may express strong abstract support for business candidates, but our research challenges the notion that this broadly translates into advantages in electoral contests. We find no support for the hypothesis that designating a business background tends to be especially successful. Instead, they are on a par with educators and retired candidates and do significantly worse than candidates with a political background. Candidates who identify as “small business owners” enjoy more success than general business candidates, indicating that voters make a distinction between corporate executives and self-employed entrepreneurs, preferring the latter. We also find that Republican-leaning constituencies are more likely to favor small business candidates than Democrat-leaning ones. Overall, our evidence undermines the claim that voters have a broad preference for candidates with a business background over a political one, despite a long-documented distaste for “politicians.” Rather, the preference is for an identifiable subset of business candidates, and primarily in places where Republican voters predominate.

Occupational Cues in Local Elections

A large body of literature demonstrates that voters typically lack detailed information about candidate policy views and qualifications and rely on simple cues such as endorsements, name recognition, incumbency status, party affiliation, and visible demographic characteristics to make their choices (e.g., Lupia, 2015). This tendency is especially evident at the local level where elections tend to be low information affairs, leading some voters to distinguish among candidates using gender or ethnic identification (Crowder-Meyer et al., 2019; Matson & Fine, 2006). In elections where party labels do not appear on the ballot (such as California local elections), voters appear less likely to use party as a cue to determine vote choices even if it is possible to find candidate party preferences (Lim & Snyder, 2015; Schaffner et al., 2001; but see Bonneau & Cann, 2015). Absent any substantive cue, voters may rely on an arbitrary one, such as ballot position (e.g., Miller & Krosnick, 1998).

The research on occupation as a cue is relatively thin, although the few studies that have been conducted have found that occupation is relevant for many voters. For example, Byrne and Pueschel (1974), examining various cues that could affect voting decisions for Democratic and Republican Party county central committees between 1948 and 1970, found that candidates who listed their occupation as “professor,” “engineer,” or “lawyer” were likely to win, while those who listed their occupation as “housewife,” “salesman,” or “real estate broker” were likely to lose. McDermott (2005), using a survey experiment that presented respondents with information about actual candidates for statewide office below the governor’s level (e.g., state treasurer, secretary of state) with or without occupational cues, concluded that when occupational labels were present voters were more likely to choose the candidate whose listed experience better aligned with the elective position.¹ In addition, a recent study based on survey experiments (Crowder-Meyer et al., 2019) indicates that adding occupational information can significantly reduce the voting impact of demographic cues such as gender and race. The importance of occupational cues has also been documented in studies in other countries, such as Germany (Mechtel, 2014), Great Britain (Campbell & Cowley, 2014), and New Zealand (Coffé & Theiss-Morse, 2016).

A few studies directly assess whether voters prefer candidates with a business background. Coffé and Theiss-Morse (2016) used student experimental evidence from New Zealand and the United States to examine how occupational designations influence perceptions of competence and electoral support. Controlling for other variables, they found that the business label contributed to perceptions of competence in some areas but was associated with perceptions of less competence in others. Furthermore, the business designation did not increase electoral support among the American students and actually eroded support among the New Zealand students. Carnes and Lupo (2016) analyzed the possibility of voter bias against “working class” candidates by using a series of candidate choice experiments in Argentina, Great Britain, and the United States. Contrary to the prevailing assumption that “business owner” candidates would be advantaged over “factory workers,” they found that the average respondent in Argentina and Britain was essentially indifferent to whether the candidate had the “factory worker” or “business owner” designation, while the average respondent in the United States slightly favored the “factory worker.”

In short, the limited empirical evidence casts doubt on the idea that voters will generally prefer business candidates when facing head-to-head choices. However, this tentative conclusion results mainly from analysis of experimental data rather than actual election results, and coexists with broader survey findings about favorability towards people with a business background. Accordingly, further testing is necessary.

A recent study suggests one reason why business candidates may not be more successful: the business label may appeal to some partisans while repelling others. Kirkland and Coppock (2018) conducted a series of conjoint experiments on a hypothetical election, varying conditions including occupational and partisan labels for candidates. In the nonpartisan context, Republican survey respondents were more likely to approve business candidates. By contrast, depending on the experimental conditions the business label either made no difference for Democratic respondents or made them less likely to support a candidate. The authors found some evidence to suggest that respondents were likely to see occupational background as linked to whether candidates would support liberal or conservative policies, which would help to explain partisan differences in candidate support. Republicans but not Democrats also tended to consider business candidates as more competent. This is consistent with broader public opinion survey findings that Republicans are generally more trusting of business leaders than are Democrats (Rainey et al., 2019).

In addition, the Kirkland and Coppock study provides hints (though not clear evidence) that the “small business owner” label is more attractive to voters than the “business executive” label. This is consistent with studies showing that the American public generally favors “small business” over “major companies,” viewing the former as having more positive community effects (Newman & Kane, 2014). Such findings may suggest a demand side advantage from the small business label. In addition, the label may indicate an underlying supply side advantage, since a small business background may be especially conducive to building the type of skills needed to run a local political campaign. We will return to this idea later in the article.

Based on this literature we test three hypotheses with more and different types of data than previously used:

H1: Voters in the aggregate will not prefer candidates with a business background.

H2: Voters will prefer small business owners to other business candidates.

H3: Republican-leaning electorates will have a stronger preference for business candidates than Democratic-leaning electorates.

Study 1: Observational Data

Data and Methods

Our first analysis used an aggregate dataset of mayoral and city council elections to assess how well businesspersons do at the ballot box. The dataset includes all city council and mayoral elections in California between 2008 and 2015, over 11,000 candidates running in over 400 cities. Our main source of data was the California Elections Data Archive (CEDA), maintained by the Institute of Social Research at California State University, Sacramento. This is a unique statewide database that collects, tabulates and reports candidate and ballot measure results for all local elections. Reports from CEDA include all candidates, their ballot designations, incumbency status, and vote totals. To this information we added variables for candidate sex, election type, competitiveness, and partisanship (percent of registered voters identifying as Republican, as reported by the California Secretary of State).

California cities vary in their electoral systems. Smaller cities tend to use multimember elections for their city council while larger cities are more likely to use single-member districts with nonpartisan primaries (if no candidate received a majority vote in the primary a runoff between the top two candidates ensues). During the time period this study there were four California cities that used ranked-choice voting (RCV) for local offices. We therefore added dummy variables for both multimember elections and RCV.

The CEDA data include the text of each candidate’s ballot designation, which appears directly on the ballot under the candidate’s name. The state’s Election Code (Section 13107) allows candidates to describe their occupation, restricted by broad guidelines. Candidates can use no more than three words to describe their current “professions, vocations, or occupations” or can indicate any current elective office they hold. For example, an incumbent city council member who is also a business person could be listed as “incumbent/business owner” but not “incumbent/small business owner” because the latter uses too many words. Candidates are prohibited from using adjectives that suggest an evaluation such as “outstanding” or “eminent” and may not list a profession they no longer practice or office they no longer hold. They also cannot list a hobby, honorary titles (such as “honorary professor”) or use generic status words such as “taxpayer” or “concerned citizen.” Despite these limitations, there is a lot of room for choice of wording. For example, someone who runs a small land use consulting business and also teaches a land use course at the local college might reasonably describe herself as a “small business owner,” “land use consultant,” “businesswoman and educator,” “college instructor,” or use any number of similar terms.

California has distinct advantages in studying the impact of occupational cues as it is the only state where almost all candidates list an occupation on the ballot. In other states candidates may include such information in campaign materials but not in a systematic way, and researchers lack the ability to identify simple occupational cues they can be assured are available to all voters. As a result, the small number of prior studies in this area examining actual electoral contests have tended to focus on California races (e.g., Byrne & Pueschel, 1974; McDermott, 2005). The CEDA database is therefore especially appropriate for the type of study we are undertaking because it provides one of the most comprehensive sources of information about local elections available for any state in the United States.

We coded ballot designations that appear as text in the CEDA database with dummy variables for nine occupational groups. Regarding our primary focus, candidates who ran as “general businesspersons,” we opted for a conservative definition, limiting that coding category to candidates whose designation explicitly indicated they are a business owner or executive, excluding other business-related designations such as “consultant” and “manager.” Thus “businessperson” candidates (coded as 1 in the dummy variable) include those who identified themselves as “business owner,” “businessman,” “businesswoman,” or “CEO.” We also coded candidates as “businesspeople” if they included a business designation along with an unrelated designation, e.g., “Businesswoman/planning commissioner,” regardless of the order in which the occupations were listed. “Small business” is a separate category coded 1 for candidates who identified as a small business owner/person, local business owner/person, or who qualified the business label with the name of their city (e.g., “Oceanside business owner”).

We followed similar rules for the seven other occupational dummy variables. We created an “elected officials” category for candidates who held an elective office other than the one they were seeking (e.g., a city council member running for mayor or a school board member running for city council), which was coded separately from the “incumbent” category. The remaining five categories were attorneys, educators (both K-12 and higher education), realtors, retirees, and an “other” category consisting of candidates who did not fall into one of the previous categories. Each dummy variable was coded 1 for candidates with specified ballot designation, and 0 otherwise.

The Success of Business Candidates

Even with a relatively restrictive coding system for identifying someone as a businessperson, 23% of all candidates fit that classification (including both general and small businesspersons), far exceeding the percentage identifying as an attorney (3.9%), educator (6.2%), or retiree (9.7%). Despite their prevalence, general businesspersons were not especially successful at the polls, as indicated in Table 1. For this analysis we focused on non-incumbents, as we expected the power of incumbency to overshadow any effect of ballot designations, and only included contested elections (races where the number of candidates exceeded the number of seats up for election). Elected officials (those who held an office other than the one they were running for) were the most successful, with 44% either winning or advancing to a runoff. About 30% of general businesspersons, attorneys, retirees, and educators either won or advanced to a runoff. Business candidates did better than candidates with occupations other than the ones listed, of whom only 23% either won or advanced to a runoff. But general business candidates did worse than elected officials and about the same as attorneys, educators, and retirees. Small businesspersons, however, were more successful, either winning or advancing to a runoff 38% of the time.

Table 1.

Success of Non-incumbents by Ballot Designation.

	Percent won	Percent advanced to runoff	Percent lost	# of candidates
Elected officials (not incumbent)	37.8	6.3	55.9	463
Attorney	31.6	1.6	66.8	364
Retired	29.1	0.5	70.4	966
Business
General business	25.4	2.2	72.4	1,581
Small business	35.4	2.7	61.9	548
Educator	26.9	3.9	69.3	592
Realtor	10.0	1.4	88.7	141
Other	21.8	1.3	76.8	3,621
All candidates	25.7	1.8	72.5	7,947

Notes. Contested elections only. Because candidates can list more than one occupation and thus can appear in more than one occupational category, the sum of the occupational categories is greater than the total number of candidates.

Source. California Elections Data Archive (CEDA).

Next, we used ordinary least squares regression to test how ballot designations influence candidate vote share, controlling for candidate sex and contextual factors such as percent of Republican voters and election type. Our dependent variable was candidate vote share, measured in percent. Ballot designations of small business, non-incumbent elected official, incumbent, attorney, educator, retired, realtor and other were included in the model as dummy variables (1 indicating the candidate listed the profession and 0 otherwise) with the reference category being general business candidates. Control variables included candidate sex (1 = female, 0 = male), percentage of registered Republicans in the city, number of seats to be filled in office, number of candidates running for office, election type (mayoral and multimember city council, with single-member district council elections serving as the reference category), and whether the city employs ranked-choice voting. Our model only included contested races, and the standard errors were clustered by race.² Results are presented in Table 2.

Table 2.

Candidate Vote Share and Ballot Designations.

Candidate characteristics	Coefficient estimates (Standard error)
Small business	2.235 (0.46)***
Elected official (non-incumbent)	7.400 (0.68)***
Incumbent	9.166 (0.34)***
Attorney	1.970 (0.52)***
Educator	0.037 (0.43)
Retired	−0.018 (0.38)
Realtor	−4.536 (0.82)***
Other ballot designation	−2.021 (0.32)***
Female	1.110 (0.22)***
Contextual factors
Percent Republican	0.010 (0.01)
Number of seats to be filled	−0.731 (0.40)*
Number of candidates running for office	−2.341 (0.13)***
Election type: Mayoral	−1.835 (0.70)**
Election type: Multimember	−8.205 (0.74)***
Ranked-choice voting	−2.857 (1.37)**
Constant	40.367 (0.89)***
N	10,909
R²	0.5213

Note. Ordinary least squares regression estimates with robust standard errors clustered by race. Analysis limited to contested races only. Dependent variable is percent vote share.

Data: California Elections Data Archive (CEDA) and California Secretary of State.

Significance levels: *p < .10. **p < .05. ***p < .001.

Relative to general business candidates, candidates with small business, elected official (non-incumbent), incumbent, and attorney ballot designations had statistically significant higher vote shares, while candidates with realtor and “other” ballot designations had significantly lower vote shares. Compared to business candidates, elected officials’ vote shares were approximately 7.4% higher (p < .001), incumbents’ vote shares were over 9% higher (p < .001), and small business candidates and attorneys had vote shares approximately 2% higher (p < .001 respectively). Business candidates were less successful than small business candidates, attorneys, and elected officials, even controlling for other contextual factors, supporting H1. Although small business candidates had higher vote shares compared to business candidates in general (evidence for H2), they had almost 7% lower vote shares than incumbents (p < .001) and about 5% lower vote shares than non-incumbent elected officials (p < .001).

To test H3, we added interactions between percent of Republican voters and (1) business designation, (2) small business designation, (3) non-incumbent elected official designation, and (4) incumbent, the latter two as comparisons to determine if there are partisan differences regarding governmental experience. Figure 1 presents the interaction effect between percent Republican and each ballot designation (full model results available in the Supplemental Appendix). As the percent Republican increased, small business candidates had significantly higher vote shares, while incumbents had significantly lower vote shares. The interaction between percent Republican and business candidates and between percent Republican and non-incumbent elected officials is not statistically significant. In the aggregate, it is clear that cities with higher percentages of Republicans favored small business candidates. The evidence for H3 is mixed, however, because there is no significant vote share advantage for general business candidates in Republican leaning jurisdictions. Incumbents had significantly lower vote shares in Republican dominated cities, yet Republicans did not appear to similarly punish non-incumbent elected officials. Perhaps Republicans believed the same or current person should no longer be in office, but they were not necessarily against all or any government experience.

Our base model included multiple contextual factors, but only ballot designations and candidate sex for candidate characteristics, neglecting to control for candidate quality and campaign finance. Fortunately, we were able to include such information for a subset of our data. We include measures of candidate quality and campaign financing in our model for the four most populous cities in California: Los Angeles, San Diego, San Jose, and San Francisco. For each city we had information on total candidate expenditures, independent expenditures made in favor of candidates, independent expenditures against candidates, and independent expenditures shared across multiple candidates collected from the California Fair Political Practices Commission (FPPC) forms 460, 465, and 496. We included two campaign finance variables: (1) Expenditures per vote, that is, total expenditures by the candidate per vote cast in the race, and (2) Net Independent Expenditures (IEs) per vote, that is, net IEs in favor of the candidate (IEs for the candidate + shared IEs – IEs against the candidate) per vote cast in the race. Our measure of candidate quality was a dichotomous variable coded 1 for candidates who previously held elected office and 0 for candidates who never held previous office.³

Results for the subset of cities with spending and candidate quality information were similar to those from the full dataset (Table 3).⁴ General business candidates had significantly lower vote shares relative to small business candidates (p < .10), incumbents (p < .001), and quality candidates (p < .001), again showing that business candidates are not especially successful relative to other occupations, and are less successful than incumbents and candidates with previous officeholding experience. This finding also partially addresses a potential concern about endogeneity in the results from our full sample. Perhaps weaker candidates (notably, those without prior elective office experience) gravitate toward the business label in making choices about their ballot designation. However, analysis of the subset data mitigates against such a concern: business candidates in general do not perform notably well even controlling for quality, at least using prior elective experience variable as a measure of quality which is common in this type of research. Furthermore, small business candidates performed statistically just as well as quality candidates (p = .23), indicating that being a small business owner is as useful a cue as previous elected experience.

Table 3.

Candidate Vote Share, Ballot Designations, Candidate Quality, and Campaign Spending in California’s Four Largest Cities.

Candidate characteristics	Coefficient estimate (Standard error)
Small business	3.574 (2.13)*
Incumbent	23.578 (2.96)***
Attorney	2.594 (1.78)
Educator	2.038 (2.04)
Retired	−4.049 (2.19)*
Other ballot designation	−2.422 (1.56)
Female	0.007 (1.26)
Quality candidate	6.620 (1.69)***
Expenditures per vote	0.812 (0.11)***
Net independent expenditures per vote	0.133 (0.15)
Contextual factors
Percent Republican	0.254(0.09)**
Number of seats to be filled	−0.656 (0.95)
Number of candidates running for office	−1.631 (0.39)***
Election type: Mayoral	−2.039 (2.24)
Constant	21.889 (3.41)***
N	547
R²	0.6877

Note. Ordinary least squares regression estimates with robust standard errors clustered by race. Analysis limited to four largest cities in California: Los Angeles, San Diego, San Jose, and San Francisco. Dependent variable is percent vote share. Data: California Elections Data Archive (CEDA), California Secretary of State, California Fair Political Practices Commission (FPPC) forms 460, 465, and 496. Significance levels: *p < .10. **p < .05. ***p < .001.

In sum, the observational data indicate that having a business background is not that advantageous for local candidates; they are better off if they have experience as an elected official. Voters, however, respond more favorably to small business owners, suggesting that they make distinctions between different types of business candidates. In addition, races involving a higher share of Republican voters are generally more favorable to small business candidates than races involving a higher share of Democrats. While the usual caveats apply in terms of drawing inferences about individuals based on aggregate data, our findings at least suggest that individual Republican voters tend to view small business candidates more favorably even for officially nonpartisan offices.

Before leaving our observational study, it is worth briefly considering the topic of external validity. While our findings come exclusively from California, there is reason to believe they are broadly applicable to contests in other states. This is because our unit of analysis is each candidate in city races, and California cities are diverse in ways that make them more politically reflective of the nation as a whole than would be assumed by focusing on the state as a whole. It is well known that the Golden State is different from others in politically significant ways including high population density, dominance by Democrats, and the relatively large size of the nonwhite population. Yet even though our sample includes races in cities such as Los Angeles, over two thirds of the results come from contests in cities with less than 100,000 people and many from contests with less than 10,000 residents. Although Democratic voter registration at the state level far exceeds Republican registration, we have results from thousands of candidates running in races in which more Republicans were registered than Democrats. And even though the state is known for its ethnic diversity the sample included results from cities in which over 80% of residents were white.

Study 2: Experimental Analysis of Support for Local Candidates

Data and Methods

To further explore the dynamics of occupational label impact on voter choice, we used an experimental design to manipulate the ballot designations of hypothetical candidates running in a non-partisan city council race. Participants were presented with information on three candidates: a realtor, an elected school board member,⁵ and a business candidate. The specific designation of the business candidates varied across three experimental groups, identifying a “business person” (group 1), an “executive at a large corporation” (group 2), or a “small business owner” (group 3). The candidates were hypothetical and only identified by their ballot designation and a color rather than name (e.g., “Orange candidate”).⁶ Participants ranked the candidates by preference, evaluated their preferred candidate on personal characteristics, and answered background and demographic information questions. The Supplemental Appendix includes experimental prompt text.

Our sample consisted of 309 adults recruited on Amazon’s Mechanical Turk (MTurk) between October 4, 2018 and October 10, 2018. We requested all MTurk participants be California residents for comparability to our analysis of election results, but our sample included four participants from other states. The Supplemental Appendix includes a table detailing the characteristics for each group. The key point is that the randomization produced approximate balance across the three groups. For all groups the average or most common characteristics of respondents included (among other characteristics) being in the mid-thirties, being female, reporting obtaining a baccalaureate degree, identifying with the Democratic Party, identifying as being slightly liberal, and expressing somewhat of an interest in politics and public affairs.⁷

Our experimental design allowed us to evaluate whether individuals presented only with career information for candidates preferred business candidates or candidates with political experience, in this case a school board member.⁸ We could also distinguish between types of business candidates to explore possible differences in support among them, including possible greater support for a “small businessperson” (testing H2). In addition, we explored why participants preferred candidates with particular ballot designations and how individual characteristics, notably partisanship (testing H3), affected candidate evaluations.

Findings

Consistent with H1 and the analysis of observational data, 57.8% of respondents across all experimental groups preferred the elected school board member but only 32.6% preferred one of the business candidates (business person, executive, or small business owner). The difference is statistically (p < .001) as well as substantively significant. The realtor consistently was the least preferred candidate. Consistent with observational data results (Table 3), business candidates did not perform as well as those who held prior political office and realtors performed significantly worse than all other professional designations.

Results were more complex when dividing the respondents by experimental group. As shown in Figure 2, when the comparison was between either a generic “business person” or an “executive at a large corporation” and an elected school board member, as was the case for groups 1 and 2 respectively, respondents decisively preferred the latter and the difference in preferences is statistically significant (p < .001).⁹ Yet when a “small business owner” was compared to an elected school board member as was the case for group 3, respondents preferred the former by a margin of 51.9% to 42.5%. While this difference only is marginally statistically significant (p = .09 for a one-tailed test), it suggests that voters are more favorably inclined to small business owners versus business persons in general, consistent with H2.¹⁰ More generally, this finding indicates that variations within an occupational designation type (business) affect candidate success.

We further explored whether those who preferred a business candidate did so because they viewed a business background as a cue for management experience, a desirable “outsider” status, or both (See Supplemental Appendix D). The majority of respondents identified “budget management skills or expertise,” “management skills,” and/or “understands the economy” as a reason for their preference, regardless of experimental group. This suggests that a general preference for business candidates was largely driven by seeing them as capable and knowledgeable managers. Some respondents also based support for business candidates on their being political outsiders or bringing new ideas to government, and most respondents either assumed that business candidates did not have political experience or did not view it as important.

The experimental evidence supports our expectation that people do not generally prefer business candidates over those with political experience but candidates running with the “small business” label fare better than their corporate executive counterparts. We turn now to the expectation that Republicans and Democrats evaluate business candidates differently. Figure 3 presents the percent of participants who selected each candidate, divided by party affiliation (Democrat or Republican).¹¹ A statistically significantly (p < .05) higher percentage of Republicans, relative to Democrats, selected the business person, executive, and small business owner, consistent with H3. While only 35.5% of Republicans selected the school board member, 65.9% of Democrats selected the school board member, and this partisan difference is statistically significant (p < .00).

These partisan differences in candidate preferences are reflected in reasons participants reported for selecting their preferred candidate (See Supplemental Appendix E). We found a higher percentage of Democrats reported political experience as a reason for choosing a candidate (p < .05), while a statistically significantly (p ≤ .05) higher percentage of Republicans selected a candidate for management skills, understanding the economy, budget management skills, and being a political outsider – all characteristics a business person is more likely to possess. Given the Kirkland and Coppock (2018) findings, we expected that a higher percentage of Democrats would have chosen a candidate with political experience.

We used logistic regression to provide a more rigorous test of the impact of partisanship and type of business candidate on respondent preferences, as well as to explore the impact of other independent variables expected to influence choices. The dependent variable was coded 1 for participants who ranked the candidate first and 0 otherwise. We estimated models separately for selecting the school board member, the small business owner, and a combined measure of selecting the business person or executive. The combined business person and executive dependent variable was coded 1 if a participant in either experimental groups 1 or 2 selected the business candidate (business person or executive) and 0 otherwise.¹² Independent variables included political factors (party identification, interest in politics and public affairs, and trust in state government), demographics (age, education, race, and sex), and whether the participant or their family member had the same occupation as the candidates.¹³

As shown in Table 4, we see confirmation that in general Republicans were more likely to prefer business candidates although the partisan difference depended on the type of businessperson. A Republican was 6.3 times more likely than a Democrat to select the business person or executive (p < .001). Yet Republicans statistically were no more or less likely than Democrats to prefer the small business owner. This suggests that support for small businesspeople—but not business candidates in general—crosses party lines and provides an explanation for why small business owners performed better in our observational data.

Table 4.

Predicting Candidate Preference by Individual Characteristics.

	DV: Candidate ranked #1
	Business person & executive (Aggregate)	Small business owner	School board member
Party identification: Independent	0.369 (0.62)	−1.684 (0.85)**	0.124 (0.42)
Party identification: Republican	1.836 (0.49)***	−0.018 (0.64)	−0.586 (0.35)*
Interest in politics	0.018 (0.27)	−0.829 (0.37)**	0.121 (0.18)
Trust in state government	−0.317 (0.22)	−0.804 (0.29)**	0.349 (0.15)**
Age	0.033 (0.02)*	0.032 (0.02)	−0.021 (0.01)*
Education	−0.094 (0.13)	0.171 (0.17)	0.114 (0.09)
White	0.483 (0.44)	0.429 (0.54)	−0.216 (0.29)
Female	−0.929 (0.41)**	−0.599 (0.53)	0.768 (0.28)**
Self or family member a business owner	−0.168 (0.53)	1.882 (0.71)**	−0.491 (0.34)
Self or family member a business executive	1.138 (0.62)*	1.174 (0.89)	−0.601 (0.43)
Self or family member a realtor	0.583 (0.65)	−1.968 (0.80)**	−0.069 (0.42)
Experimental Group 2 (executive)	−1.292 (0.43)**	—	0.685 (0.35)**
Experimental Group 3 (small business owner)	—	—	−0.669 (0.33)**
Constant	−0.966 (1.29)	2.589 (1.56)*	−0.780 (0.85)
N	180	96	277
Pseudo R²	0.2372	0.2645	0.1585

Note. Cell entries are logit estimates with standard errors in parentheses. Dependent variable is whether participant ranked the candidate #1 (1) or not (0). Data: Amazon Mechanical Turk October 2018 Survey Experiment. Significance levels: *p < .10. **p < .05. ***p < .001.

Discussion

Businesspeople routinely seek political office at all levels of American government, and this tendency may be growing for at least some types of positions. A recent study finds a sharp rise in the proportion of business candidates for federal office and this is matched by an increase in the proportion of federal office holders who held a corporate executive position prior to election, with over one in five fitting this description by 2014 (Babenko et al., 2018). Business people also often run for state office, and despite common views about the prevalence of attorneys, the proportion of state legislators with a business background far exceeds the proportion who are lawyers (Fifield, 2015; Makse, 2019). Similarly, a recent study finds almost one in three U.S. mayoral candidates has a business background—a strikingly high proportion (Kirkland, 2017).

Despite the prevalence of candidates with a business background running for office, prior research offers relatively little systematic evidence about whether the business label advantages candidates. Our research takes the study of business candidates much further. We first draw on a data set of actual election results much larger than has previously been used, and that is especially appropriate for the present purpose because it includes comprehensive information on occupational labels. With information on thousands of California mayoral and city council candidates from over 400 cities, we confirm the prevalence of self-described business candidates. The proportion running for office under this label far exceeded the proportion listing other familiar labels such as attorney and educator. Yet business candidates were not notably successful attracting voters, faring worse in terms of securing electoral victories than others with prior elective experience and about the same as lawyers, educators, and retirees. The lack of distinctive business candidate success persisted even when we added a variety of control variables to a subsample of our data set, including candidate quality.

Our survey experiment provides further support for the conclusion that the business label is not generally advantageous for candidates. For example, respondents overall preferred candidates with a school board background to those with a business background. The experimental results were consistent with our findings from observational data as well as with previous experimental studies.

Beyond the analysis of overall business candidate success, we make two other significant contributions to the literature examining voter preferences for business candidates. First, our survey experiments confirm the importance of distinguishing between types of business candidates, with voters preferring candidates mentioning a small business background over those with a business background generally. This is consistent with the tendency of Americans to look favorably on small business and more suspiciously on large corporations (Newman & Kane, 2014). It would be interesting to learn more about the reasons for the distinction and the extent to which voters favor the person listing a small business background because they see that candidate as “one of us.” Alternatively, it may be that small business candidates are more effective campaigners because they tend to have useful skills such as ability to manage budgets, recruit staff, and market products. This suggests the need to compare the campaigns run by small business candidates to those run by others.

We also found that partisanship strongly influences the impact of occupational labels on voter behavior. Even though all of the elections in our observational dataset were nonpartisan, Republican leaning constituencies and Democratic leaning constituencies appear to react differently to the appearance of the business occupation label by a candidate’s name. Our experimental study reinforces the conclusion that partisanship matters, but with a twist. Those results suggest Democrats are less favorably inclined toward business candidates in general than Republicans but support for small business candidates crosses party lines. We also find that Republicans and Democrats value qualities differently in evaluating candidates based on occupational background, which is consistent with Kirkland and Coppock (2018).

The findings above suggest some fruitful areas for future research. The first is to explore in greater depth how voters use occupational cues in their voting decisions. Specifically, more research needs to be done exploring how important a business label is relative to other cues. In our experiment participants had no other information about the candidates–if they knew party affiliation or a candidate’s demographic profile would the effects of occupation diminish? Kirkland and Coppock (2018) and Crowder-Meyer et al. (2019) provide a good start to this line of inquiry, but more work needs to be done exploring the interaction and relative importance of various cues. It is possible that in more salient campaigns where voters have greater information about the candidates the effects of occupation will disappear. Given the strong preference for small business owners, it is plausible that some moderate voters may be willing to cross party lines to support small business owners running for the local city council; even if they know a candidate’s party affiliation they may rely on occupational cues to judge a candidate’s qualifications.

More work needs to be done exploring the underlying beliefs that lead voters to prefer or reject business candidates. One possibility is that voters assume business candidates are more conservative than candidates with political backgrounds. That may be especially true given the Trump Presidency; having a conservative businessperson with no political experience elected president may reinforce voters’ assumptions about the connection between a business background and a conservative ideology. There may be other assumptions that voters make about business candidates as well. For example, they may assume that business candidates have a more top-down management style or prioritize economic policy over social issues. Our experiment explored some potential beliefs about business candidates, but there may be others that researchers have not yet uncovered.

It is also desirable to approach research from a candidate perspective. In California candidates have options as to how they define their occupation on the ballot, and more generally candidates can present their occupational background in many ways on their campaign websites, advertising, and ballot pamphlet statements. How do candidates use their occupational backgrounds to appeal to voters? Do they assume that voters prefer business backgrounds and choose to emphasize that over other professional experience? Do small business candidates use their work experience differently than other business candidates? Do attorneys try to minimize their legal background under the assumption that voters do not look favorably upon lawyers? More generally, do candidates believe that their occupational backgrounds have an influence on their electoral success, and if so does it affect who decides to run in the first place? It is possible that the (largely incorrect) belief that voters prefer candidates with a business background encourages businesspeople to run for office and discourages others. In short, beliefs about the “businessperson” label may influence potential candidate decisions as well as voter choices, even if such candidate decisions are grounded in mythology.

Supplemental Material

Ballot_desig_online_appendix – Supplemental material for Ballot Cues, Business Candidates, and Voter Choices in Local Elections

Supplemental material, Ballot_desig_online_appendix for Ballot Cues, Business Candidates, and Voter Choices in Local Elections by Brian E. Adams, Edward L. Lascher and Danielle Joesten Martin in American Politics Research

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Brian Adams

Supplemental Material

Supplemental material for this article is available online.

Notes

Author Biographies

Brian E. Adams is a professor of Political Science at San Diego State University whose research focuses on democratic practices in local government. He is author of two books, Citizen Lobbyists and Campaign Finance in Local Elections.

Edward L. Lascher Jr. is a professor of Public Policy and Administration at California State University, Sacramento. He is co-author of Initiatives Without Engagement: A Realistic Appraisal of Direct Democracy’s Secondary Effects (University of Michigan Press, 2019).

Danielle Joesten Martin is an associate professor of Political Science at California State University, Sacramento. Her research has been published in journals such as Journal of Politics, Political Behavior, and Journal of Women, Politics & Policy.

References

Austin

P. C.

Hux

J. E.

(2002). A brief note on overlapping confidence intervals. Journal of Vascular Surgery, 36(1), 194–195.

Babenko

Fedaseyeu

Zhang

(2018). Executives in politics. Carroll School of Management, Boston College.

Berinsky

A. J.

Huber

G. A.

Lenz

G. S.

(2012). Evaluating online labor markets for experimental research: Amazon.com’s Mechanical Turk. Political Analysis, 20(3), 351–368.

Bonneau

C. W.

Cann

D. M.

(2015). Party identification and vote choice in partisan and nonpartisan elections. Political Behavior, 37(1), 43–66.

Boudreau

Elmendorf

C. F.

MacKenzie

S. A.

(2015). Informing electorates via election law: An experimental study of partisan endorsements and nonpartisan voter guides in local elections. Election Law Journal, 14(1), 2–23.

Byrne

G. C.

Pueschel

J. K.

(1974). But who should I vote for county coroner? Journal of Politics, 36(3), 778–784.

Campbell

Cowley

(2014). What voters want: Reactions to candidate characteristics in a survey experiment. Political Studies, 62(4), 745–765.

Carnes

Lupu

(2016). Do voters dislike working-class candidates? Voter biases and the descriptive underrepresentation of the working class. The American Political Science Review, 110(4), 832–844.

Coffé

Theiss-Morse

(2016). The effect of candidates’ occupational background on voters’ perceptions of and support for candidates. Political Science, 68(1), 55–77.

10.

Crowder-Meyer

Gadarian

S. K.

Trounstine

(2019). Voting can be hard, information helps. Urban Affairs Review, 56(1), 124–153.

11.

Fifield

(2015). State legislatures have fewer farmers, lawyers; but higher education levels. PEW, December 10.

12.

Hibbing

J. R.

Theiss-Morse

(2002). Stealth democracy: Americans’ beliefs about how government should work. Cambridge University Press.

13.

Huff

Tingley

(2015). Who are these people? Evaluating the demographic characteristics and political preferences of MTurk survey respondents. Research and Politics, 2(3), 205316801560464.

14.

Jacobson

G. C.

(1989). Strategic politicians and the dynamics of U.S. House elections, 1946-86. The American Political Science Review, 83(3), 773–793.

15.

Kirkland

P. A.

(2017). America’s mayors: Who serves and how mayors shape policy. Dissertation Columbia University.

16.

Kirkland

P. A.

Coppock

(2018). Candidate choice without party labels: New insights from conjoint survey experiments. Political Behavior, 40(3), 571–591.

17.

Knezevic

(2008). StatNews #73: Overlapping confidence intervals and statistical significance. Cornell Statistical Consulting Unit, Cornell University.

18.

Lim

C. S. H.

Snyder

J. M.

(2015). Is more information always better? Party cues and candidate quality in US judicial elections. Journal of Public Economics, 128, 107–123.

19.

Lupia

(2015). Uninformed: Why people seem to know so little about politics and what we can do about it. Oxford University Press.

20.

Makse

(2019). Professional background in state legislatures, 1993-2012. State Politics & Policy Quarterly, 19(3), 312–333.

21.

Matson

Fine

T. S.

(2006). Gender, ethnicity, and ballot information: Ballot cues in low-information elections. State Politics & Policy Quarterly, 6(1), 49–72.

22.

McCarthy

(2014). Americans say business background is best for governing. Gallup, July 21.

23.

McDermott

M. L.

(2005). Candidate occupations and voter information shortcuts. Journal of Politics, 67(1), 201–219.

24.

Mechtel

(2014). It’s the occupation, stupid! Explaining candidates’ success in low-information elections. European Journal of Political Economy, 33, 53–70.

25.

Miller

J. M.

Krosnick

J. A.

(1998). The impact of candidate name order on election outcomes. Public Opinion Quarterly, 62(3), 291–330.

26.

Newman

B. J.

Kane

J. V.

(2014). Backlash against the “Big Box” local small business and public opinion toward business corporations. Public Opinion Quarterly, 78(4), 984–1002.

27.

Rainey

Keeter

Perrin

(2019). Trust and distrust in America. PEW, July 22.

28.

Schaffner

B. F.

Streb

M. J.

Wright

G. C.

(2001). Teams without uniforms; the nonpatisan ballot in state and local elections. Political Research Quarterly, 54(1), 7–30.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.11 MB