Introduction
A growing literature is uncovering substantively meaningful effects of the place-based connections that elected officials have with the geographic areas they represent. The consequence most scholars have concerned themselves with is the electoral advantage that candidates enjoy when they possess deep local ties to their home districts or states. These ties encompass various forms of personal biographical roots, such as being born and raised, attending school, or having worked or raised in a family within the geographic boundaries of the jurisdiction a candidate is running to represent. These effects have been demonstrated at the experimental (Munis Reference Munis2021; Schulte-Cloos and Bauer Reference Schulte-Cloos and Bauer2021) and observational levels (Evans et al. Reference Evans, Arzheimer, Campbell and Cowley2017; Hunt Reference Hunt2021a, Reference Hunt2021b; Reference Hunt2022; Stevens et al. Reference Stevens, Islam, de Geus, Goldberg, McAndrews, Mierke-Zatwarnicki, Loewen and Rubenson2018) at multiple levels of government, and in both American and international settings. However, to the best of our knowledge, no scholarship has spoken to the effects of local candidate roots at the state legislative level.
Alongside traditional theories as to why local roots should positively impact legislators’ electoral fortunes generally, we propose here a novel framework for understanding why local ties should be just as, if not more, impactful in state legislative races. We pair this framework with an original dataset of nearly 5,000 sitting state legislators from all 50 states, for all of whom we have collected unique measures of their local roots, such as where they were born, and where they went to college and postgraduate school. We use these measures to create a Local Roots Index (LRI) modeled after Hunt’s (Reference Hunt2022) measure, used originally for the US House of Representatives, with the expectation that incumbents with higher scores on this index will be more electorally successful than their more “carpetbagging” counterparts.
Our results are in line with previous findings for other offices and levels of government. Given the notoriously high proportionFootnote 1 of uncontested state legislative seats (Squire Reference Squire2000), we first model the likelihood of incumbents running unopposed in their general elections, conditional on the depth of their local roots and other key variables like partisan balance of the district, chamber seniority, prior political experience, and race and gender. We find that although state legislative incumbents with few to no local roots in their districts run unopposed about a quarter of the time, their counterparts with deep local roots do so more than 40% of the time. This is an effect size that rivals that of district partisanship, and speaks to deep local ties as a major factor in discouraging potential challengers to state legislative incumbents. We also find that even among those incumbents who do attract a challenger from the opposing party, local roots offer statistically significant advantages of as many as three percentage points of vote share, a finding consistent with federal offices like the US House. These combined effects of local roots at both the candidate emergence and general election stages demonstrate not only that local roots are highly impactful in state legislative races; but that they continue to influence them even amidst increasing nationalization and polarization of state legislatures that has characterized the modern era of American politics.
Local roots and electoral advantages
It is well established by the literature that local ties are desirable attributes for candidates that voters appreciate and reward on Election Day. Early work demonstrated what V.O. Key dubbed the “friends-and-neighbors” effect, in which candidates for office pull higher-than-expected voter support in their area of residence compared with other areas in their jurisdictions (Aspin and Hall Reference Aspin and Hall1987; Key Reference Key1949; Parker Reference Parker1982; Tatalovich Reference Tatalovich1975). Other work has captured these effects via home state advantages in presidential elections (Garand Reference Garand1988; Lewis-Beck and Rice Reference Lewis-Beck and Rice1983). More recent work has confirmed these findings in a variety of ways (Campbell et al. Reference Campbell, Cowley, Vivyan and Wagner2019; Panagopoulos, Leighley, and Hamel Reference Panagopoulos, Leighley and Hamel2017; Put, von Schoultz, and Isotalo Reference Put, Von Schoultz and Isotalo2020), indicating that the electoral power of geographic closeness remains a factor in modern American elections.
Other literature has demonstrated jurisdiction-wide advantages associated with local ties and, conversely, the pitfalls of candidacy in an area with which one has no background, commonly referred to as “carpetbagging” (Galdieri Reference Galdieri2019). Previous work has shown that voters’ deep geographic ties to one’s jurisdiction is a desirable candidate trait (Munis Reference Munis2021), resulting in consistent electoral advantages for local candidates in both congressional primaries (Hunt Reference Hunt2021b) and general elections (Evans et al. Reference Evans, Arzheimer, Campbell and Cowley2017; Hunt Reference Hunt2021a; Stevens et al. Reference Stevens, Islam, de Geus, Goldberg, McAndrews, Mierke-Zatwarnicki, Loewen and Rubenson2018).
In Home Field Advantage (2022), Hunt offers a framework for understanding how these electoral benefits emerge from deep local ties. One set of mechanisms is practical in nature, focusing on heightened local name recognition; more extensive social, economic, and political networks in the jurisdiction; and a homegrown knowledge of the community, including the issues its voters prioritize the most. A second set of mechanisms is more symbolic, drawing on classic findings on phenomena like home styles, representational trust, and the “personal vote” (Cain, Ferejohn, and Fiorina Reference Cain, Ferejohn and Fiorina1987; Fenno Reference Fenno1978; Fiorina and Rohde Reference Fiorina and Rohde1991). This work posits that voters are largely in search of representatives who are “like them,” and thus can be trusted to have their best interests at heart once in office. Local roots in a particular home area, when shared between candidate and voter, can create such a connection and imbue the relationship with trust. More recent work has drawn out these symbolic mechanisms via what has come to be called “place identity” – the representational connection individuals feel with particular places or types of places – as the bedrock of why voters consistently choose homegrown candidates at higher rates (Jacobs and Munis Reference Jacobs and Munis2018; Munis Reference Munis2021; Schulte-Cloos and Bauer Reference Schulte-Cloos and Bauer2021).
Local roots in state legislatures
Somewhat surprisingly, the literature on local roots or place identity, to date, has focused largely on federal or statewide offices. As noted above, this work has found that local roots are an important component of candidate assessment in these contests. In this paper, we argue that state legislative elections can also be fruitful venues for observing the influence of local roots.
First, state legislative elections encompass smaller constituencies (districts), which are more parochial and provide a fertile environment to sow the seeds of stronger place-based attachments. Second, the platform of state legislative candidates is much more focused on state and local issues that often uniquely or disproportionately affect that particular constituency. The ability to address these issues may be more contingent on the place-based attachment a candidate has to that community. Third, it is much easier for state-level candidates to personalize their campaigns and emphasize their local credentials, particularly when they can directly reach many or most constituents through door-to-door campaigning or townhall meetings. Furthermore, voters are more likely to personally know a state-level candidate and be familiar with them outside of politics; most candidates elected to state legislatures continue to hold nonpolitical jobs, since serving as a state legislator is mostly a part-time commitment. Finally, while politics at all levels has become increasingly nationalized (Hopkins Reference Hopkins2018), this effect is still less prevalent at the state level. Candidates for state legislatures can deliver a more localized message, rather than exclusively having to tow the party line.
The effects of local roots do not operate in an electoral vacuum; rather, they also work to translate electoral success into representational benefits (Hunt Reference Hunt2022). This may be even more conspicuous at the state level because legislative candidates can benefit from more familiar personal connections to garner a coalition of loyal voters. Constituents are more receptive when they are actual participants in a mutual relationship with their representative who they feel has specific qualities that will make them more responsive to their needs (Germany Reference Germany2008). Therefore, deep local roots act as a form of descriptive representation wherein a candidate makes their place identity as “one of them,” a group connection that could have a similar influence to that of shared gender, race, and ethnicity (Bratton and Haynie Reference Bratton and Haynie1999; Rouse Reference Rouse2013). State legislators are also more responsive in this relationship and work to communicate this responsiveness (Jewell Reference Jewell1982) because they see it as valuable, not only for present but future electoral and representative benefits; state legislative service provides ample opportunities for progressive ambition (Maestas Reference Maestas2003).
The factors above indicate that the attachment to place and homegrown candidates should have a significant effect in the electoral success of state legislative candidates. Therefore, based on previous work on congressional elections that has uncovered the importance of local roots and the work on state-level elections that may predict a similar if not stronger effect on state legislative contests, we consider the following hypotheses:
H1: Locally rooted state legislators will be more likely to run uncontested in their general elections.
H2: If they do face opposition, more locally rooted state legislators will receive higher two-party vote share than their less-rooted counterparts.
Data and methods
In this study, we utilize a new dataset compiled by the State Legislators Data Service from KnowWho, a commercial data analytics firm that collects and sells background information on state lawmakers.Footnote 2 Our dataset uses KnowWho’s available data on serving state legislators as of 2018, when the data were obtained by the authors. Although KnowWho provided substantial baseline of data for most sitting state legislators, additional coding and data collection efforts on the part of the authors were required for several independent variables, most notably the measurements for legislators’ local roots. The result was a cross-section of nearly 5,000 state legislator observations, which represents just under 70% of all sitting legislators at the time.Footnote 3
To capture the depth of legislators’ local roots, we employ a modified version of Hunt’s (Reference Hunt2022) LRI, which is a summed index of several legislator-specific local roots indicators. We use four indicators in this analysis: whether the legislator was born in their home state; whether more specifically they were born within their district boundaries; went to college in their home state; or obtained postgraduate education in their home state.Footnote 4 Although most prior work has used single indicators like birthplace or current residency, the use of an index more comprehensively captures local roots at many points across a legislator’s life prior to their service.Footnote 5
KnowWho’s data contained the requisite information for some of the local roots indicators (about 50% of sitting legislators), but extensive candidate-level research was necessary to gather more complete information on these indicators, and to capture the broadest possible cross-section of sitting legislators. We were able to increase this sample to just under 70% of sitting legislators using their campaign websites, social media pages, news articles, and official biographies on their official state legislature websites. The combination of KnowWho’s data and our own coding efforts yielded city/state locations for birthplace, undergraduate, and postgraduate education for each of these legislators. Footnote 6 To determine whether a legislator was actually born in their district, we used GIS tools to intersect this city/state location with state legislative district shapefiles; if the city intersected with the district the legislator represented, they were coded as having been born in their district (see Hunt Reference Hunt2022 for further details).
Our legislator-level data also include each legislator’s most recent election results,Footnote 7 which allowed us to parse two separate dependent variables corresponding with Hypotheses 1 and 2 respectively: first, whether or not the incumbent legislator ran uncontested in their last general election (H1); and second, the legislator’s eventual share of the vote in that election.Footnote 8 We argue that the higher an incumbent’s score on the LRI, the more likely they will be to run unopposed; and the higher the general election vote share they will receive if they do face a challenger. We also include a covariate for district magnitude (operationalized using the logged total district population) on the suspicion that local roots may be more or less meaningful depending on the size of the district in question.
We also must control for factors that independently condition the likelihood of challengers emerging to face incumbents, and the eventual outcomes of the general election race. Chief among these is the partisan balance of the state legislative district, captured here using DailyKos’s measures of the vote share for the incumbent party’s most recent presidential nominee. Higher figures for this variable, therefore, indicate a friendlier partisan environment for the incumbent. We also include control variables for state legislator seniority in their chamber, scaled across states from 1 to 100; whether they are a Democrat, female, or nonwhite; whether they had elected experience prior to their state legislative service; whether they previously held nonelective roles in the party organization or other campaigns; whether they represent a multimember district; and the level of professionalization in their state legislature (Squire Reference Squire2007). We also include Hinchliffe and Lee’s (Reference Hinchliffe and Lee2016) statewide measure for whether the state has a traditional party organization system. For our modeling techniques, we utilize standard logistic regression (for H1) and ordinary least squares regression (for H2), with standard errors clustered by state in order to account for any nonrandom uncaptured likeness between legislators from the same state.Footnote 9
Results
We first investigate whether state legislators with deep local roots are more likely to run uncontested in their general elections. The logistic regression results in Table 1 strongly suggest that this is the case. At high levels of statistical significance, legislators with higher LRI’s are far more likely to run unopposed in their districts.
Likelihood of running uncontested in general election
Results found using standard logistic regression; SEs clustered by state.
* p < 0.1
** p < .05
*** p < 0.01
Figure 1, which generates predicted probabilities based on the model in Table 1, tells us that these effects are substantively as well as statistically significant. State legislators with the deepest local roots are predicted to run uncontested in the general election a little over 40% of the time. They are nearly twice as likely to do so than their unrooted counterparts, who run uncontested less than 26% of the time. Figure 1 also offers important context for the size of this effect. Although local roots and carpetbagging are by no means as impactful as district partisanship in predicting electoral fortunes or the emergence of potential challengers, the effects are in the same ballpark: the safest state legislative incumbents, based on presidential performance in the district, run unopposed about 56% of the time, compared to 11% for those running in districts that heavily favor the opposing party.
Predicted probability of state legislator running unopposed in their general election based on five-point Local Roots Index (left) and partisan safety of the district (right).
Although incumbents can put themselves in strong positions to deter potential challengers from running, the latter’s decisions to do so are ultimately out of the incumbent’s hands. And so, are incumbents’ local roots still impactful even when they do attract a general election challenger? The results in Table 2 again suggest that they are. The finding is more substantively modest, but still statistically significant: deeply rooted incumbents accrue on average about three additional percentage points of vote share in their general elections compared to their “carpetbagging” counterparts.
Effects on general election vote share
Results found using standard linear OLS regression; SEs clustered by state.
* p < 0.1
** p < .05
*** p < 0.01
This finding is consistent with the effect sizes found by Hunt (Reference Hunt2022) in his investigations of the US House. A vote share effect of three percentage points represents a six-point spread in terms of vote margin. For many incumbents, this will not likely change the actual outcome of the election. However, nearly 600 incumbents in our sample of state legislators fell within this six-point margin of victory or defeat (that is, they garnered between 47% and 53% of the vote). In addition, a several-point improvement in electoral fortunes has positive effects for incumbents on the amount of campaigning, fundraising, and spending required of them to feel electorally comfortable. Even if the electoral boost gained by local ties is not decisive in the race, it can grow the margin of victory such that the outcome scares away future potential challengers. In this sense, this result can partially help explain the results on uncontested races found in Table 1 and Figure 2. Beyond candidacy and campaign effects, parties and outside groups are also invested in these margins. With more homegrown candidates on their side, they can afford to reroute crucial campaign dollars and other infrastructure to more competitive state legislative races.
Average percentage of state legislators who ran unopposed in the general election (left) and percentage of the general election vote received if opposed (right) based on state legislators’ Local Roots Index.
Finally, attention should be paid to the district magnitude covariate in both models. Specifically, should local roots really matter equally in both small districts of only a few thousand constituents, versus larger ones that approach millions? Although fuller theorizing about these interactions is beyond the scope of this short article, we note here that including logged total district population as a covariate has no tangible impact on the power of local roots – that is, results are virtually identical whether the covariate is included or not. In addition, although we do not include it in the core models, using this variable as an interaction with the LRI produced null interaction terms, indicating that the impacts of local roots on both the probability of running unopposed, as well as total vote share, are both unrelated to the magnitude of the district.
Conclusion
Local candidate roots have been shown as impactful in modern congressional elections, and in the context of measuring the effects of “place identity” as a meaningful representational connection between voters and elected leaders. We have demonstrated that these effects are substantial in state legislative races as well. In addition to proposing novel theory as to why state legislative elections are fertile ground for local candidate effects, we have shown that local roots produce significant disincentives for potential challengers to incumbents; and that even when they are challenged, locally rooted incumbents are in better electoral positions than those without local ties.
These results have important implications for state legislative elections. The findings for the impact of roots on running uncontested are particularly instructive in the area of candidate emergence and recruitment. Parties and outside groups looking for new candidates for a seat that is being vacated by a retiring member of their party would clearly do well to find a candidate with deep local roots, who can discourage potential challengers in future elections. On the flipside, parties looking to make headway against potentially vulnerable state legislative incumbents in the opposing party might look to deeply rooted candidates of their own to potentially attenuate the influence of partisanship and challenge these incumbents.
These findings also raise questions about whether the impact of local candidate roots is limited to elections, or plays out in more complex ways within the legislative process. It is possible that candidates with local roots (because they can achieve greater cross-party appeal) could have more moderate voting records, leading to a less ideologically polarized state legislative chamber. Future work could examine this and other potential intersections between legislator roots and their lawmaking behavior in the chamber.
Supplementary materials
To view supplementary material for this article, please visit http://doi.org/10.1017/spq.2023.5.
Data availability statement
Replication materials are available on SPPQ Dataverse at https://doi.org/10.15139/S3/FMYG8Z (Rouse Reference Rouse and Hunt2023).
Funding statement
The authors received no financial support for the research, authorship, and/or publication of this article.
Conflict of interest
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Open access
