Introduction

Marketing theories and concepts continue to be applied to political marketing (Milewicz and Milewicz, 2014; Newman, 1994). The development of marketing concepts and principles in this field is based on the practices of many professional political consultants and marketing scholars working in the political arena (Davies and Newman, 2006; Henneberg, 2004; Henneberg and O’Shaughnessy, 2007). As expected, the term “political marketing” has gained widespread usage within the political field. The acceptance of its use and conceptual validity is in large part due to the interaction between the academic fields of marketing and political science (Speed et al., 2015). This interaction is also responsible for the development of a number of emerging theoretical perspectives focused on the context of political marketing (Jain et al., 2018).

Within the political domain, a political candidate is analogous to a brand or a brand’s product. Specifically, a candidate is an important signaling device that facilitates marketplace exchanges (Parker, 2012). The literature has increasingly acknowledged the concept of the candidate and political party branding, leading to greater dissemination of the branding concept (Harsandaldeep and Seerat, 2022; Schofield and Reeves, 2015; Van Steenburg and Guzman, 2019). Such work has legitimized the use of Aaker’s (1991) model of brand equity based on customer perceptions as an important conceptual framework for understanding political consumers, that is, voters. According to Needham (2005), in an era characterized by the “permanent campaign,” multimedia-driven strategies, and passive audiences, candidate branding becomes a necessity.

The current study focuses on candidate brand equity and its relationship with voter intention. Independent of the party, candidate brand equity provides consumers with an informational shortcut that may impact vote intention (Schneider, 2004). Singer (2002) drew a parallel between a political party and a corporate brand, considering the candidate as the product. Similarly, Smith and French (2009) point to the role of the political leader in shaping associations within a political party brand. Building on this perspective, Milewicz and Milewicz (2014) presented evidence highlighting the complexity of defining the boundaries of the candidate brand construct. Moreover, other authors argue that political brands encompass a multifaceted ecosystem consisting of interconnected yet often distinct sub-brands (Pich and Armannsdottir, 2022). The significance of these studies cannot be overstated, as they provide valuable insights into how the electorate’s emotional connections with candidates influence voter intention (Banerjee and Chaudhuri, 2021; Van Steenburg and Guzman, 2019). The existence of a relationship between the higher number of votes and the higher value of brand equity does not seem likely. The objective of this survey is to identify the brand values and characteristics of the candidates by examining the dimensions of the brand. In this way, it will be possible to know which dimensions allow an improvement of the brand equity and on which ones an action should be taken (Cuesta-Valiño et al., 2021a) and, ultimately, the number of votes obtained.

Background

Brand equity

According to the American Marketing Association, the definition of a brand is as follows: “a name, term, design, symbol, or any other feature that identifies one seller’s goods or services as distinct from those of other sellers” (AMA, 2017). Additionally, Kotler and Armstrong (2010) note that a brand is more than a name and logo, stating that “a brand represents a consumer’s perceptions and feelings about a product and its performance—everything that the product or service means to consumers” (p. 236). Moving beyond the definition of brand, Aaker (1991) articulated the components and influence involved in the formation of his brand equity concept. Aaker proposed and defined consumer-based brand equity as a multidimensional construct comprised of five components: “perceived quality, brand loyalty, brand awareness, brand association and other proprietary brand assets, such as patents, trademarks, and channel relationships”. Using two of Aaker’s brand equity components, Keller (1993, 2001) introduced the distinctive notion of brand identity, which is founded on brand awareness and brand associations. Work by other researchers found that Keller’s dimensions exhibit a positive correlation with his brand concept (Barwise, 1993; Yoo and Donthu, 2002; Yoo et al., 2000), providing evidence of validity to Kellar’s work. In addition, Konecnik and William (2007) suggest a relationship between Keller’s dimensions and the cognitive, affective, and conative components of attitudes, suggesting that Keller’s model may be a way to influence consumers’ knowledge, feelings, and behaviors.

The brand equity concepts proposed by Aaker and Keller serve as foundational principles for the research of other authors, neither operationalized the components of their conceptual models. This work was left to other researchers, including Park and Srinivasan (1994), Srinivasan (1979), Kamakura and Gary (1993), Swait et al. (1993), Pappu et al. (2005) and Yoo and Donthu (2002). Such operationalizations of brand equity have been classified as either direct or indirect. Chatzipanagiotou et al. (2016) note that in the first case measure brand equity by assessing consumers’ preferences or utilities. Conversely, indirect operationalizations of brand equity concentrate on physical indicators to measure brand equity (Pappu et al., 2005; Yoo and Donthu, 2002).

In the present study, brand equity is measured utilizing Aaker’s (1991) model and brand association was defined as the perceptual attributes associated with the brand, collectively known as a brand image (Aaker, 1991; Keller, 1993). For its part, brand awareness was defined by the level of brand knowledge, i.e., recognition of the brand name. Perceived quality was measured as a subjective judgment of a product’s overall excellence (Li et al., 2012). Brand loyalty was assessed as the extent to which an individual experiences a sense of allegiance toward the branded product (Wymer, 2012). Finally, brand assets refer to proprietary elements and, in comparison to the other dimensions, this factor is considered objective, as it refers to tangible or behavioral aspects. In other words, it cannot be measured as a perception or psychological concept. For this reason, it is conceptually different from the other four (Aaker, 1996). Thus, the current study presents a psychological concept of brand emotion. Perceptions of strong brand equity indicate that consumers possess high brand-name awareness, maintain positive brand associations, perceive the brand as high quality, and exhibit overall brand loyalty (Ahmed et al., 2015; Stahl et al., 2012).

From brand equity to voter-based brand equity

Atzger et al. (2020) state that political party affiliation, policy initiatives, and the political leader constitute the key elements of the political brand. This is fundamental to the creation of a voter-based brand equity (VBBE) model that was derived from the CBBE model (Atzger et al., 2020; Milewicz and Milewicz, 2014). The VBBE concept is comprised of two separate components, party-based brand equity (PBBE) and candidate-based brand equity (CBBE). This division revolves around establishing a political party and its candidate as two distinct entities with which voters can develop different levels of knowledge and feelings, both of which can impact voting behaviors.

The separation between a political party and a candidate is supported by other researchers. The basis of creating the VBBE model from the CBBE model was the result of researchers studying the mental maps that voters construct when prompted to contemplate a political candidate (Smith and French, 2009; Hayes, 2005; Phipps et al., 2010; Van Steenburg and Guzman, 2019). Emotions are part of any brand, but in the political domain, the importance of this dimension is significant, as the politician is characterized by a strong level of commitment and emotional engagement from the consumer-voters, which presents opportunities for building brand equity (Gilal et al., 2022; Sutton et al., 1997). Voters often see a politician as an extension of themselves, generating a degree of passion and irrational relations with related parties and candidates (Richelieu and Couvelaere, 2005; Kaur and Sohal, 2019). As a result, the “candidate brand equity” construct emerged as an evaluative model for assessing the strength of a candidate’s brand among potential voters, independent and unique from the candidate’s party. With this understanding of the nature of CBBE, it was later used to assess the strength of brand equity for presidential candidates (Atzger et al., 2020; Parker, 2012).

The basis of VBBE and how it separates a political party from its candidate is based on the idea that candidate brands are like other individuals who transform into brands, and act as a source of identity (Atzger et al., 2020; Mitsis and Leckie, 2016; Parker, 2012). This idea is captured in the concept of human brands. A human brand is delimited by “persons, well-known or emerging, who are the subject of marketing, interpersonal, or inter-organizational communications” (Close et al., 2011, p. 923). Even if a human brand is associated with or is the “product” of a particular organization, those associations can be independent in the minds of consumers (Osorio et al., 2020; Speed et al., 2015). When voters go to the polls, it is the candidate (human brand) for whom they are casting their vote. This research suggests that the voters’ associations with the candidate are what defines the brand, but that these associations are independent of the candidate’s political party (Phipps et al., 2010; Singer, 2002; Smith and French, 2009; Van Steenburg and Guzman, 2019). For this reason, the present study focuses only on CBBE.

Voting intention

Consumer behavior theory has been applied to electoral behavior (Ben-Ur and Newman, 2010; Newman, 1999). An increasing body of research provides evidence that suggests a relationship between product performance, like the brand, and behavioral intention (Kaur and Sohal, 2019; Oh, 2000; Matzler et al., 2008). Different studies relate voting intention with variables based on political issues, image of social and candidate image, situational contingency, and personnel or brand (Bennett et al., 2019; Ben-Ur and Newman, 2010; Newman, 1999b; Newman and Sheth, 1985). Candidate brand is a vital variable in assessing voting intention, as other factors such as political involvement were not found to have a significant impact on intentions to vote for specific candidates (although they do predict election participation). This indicates that political involvement does not necessarily reflect individuals’ regard for specific candidates and political party affiliations (Schofield and Reeves, 2015). The connection between candidate brand dimensions and voting intentions suggests that generalized perceptions of a politician can lead voters to biased assumptions about their intentions and abilities if elected to public office. Nevertheless, it is precisely these generalized perceptions that consistently shape human beliefs and actions (Bennett et al., 2019), suggesting that emotion is an important dimension of voting intention.

The Spanish Presidential Election system

Following the d´Hondt rule, the Spanish Parliament (Congress) has 350 members, each of whom serves a maximum of 4 years. Spain is geographically divided into 52 electoral districts, which include 50 districts along with the cities of Ceuta and Melilla. A minimum seat count is allocated to each electoral district and is used to apportion the remaining seats among the provinces, using provincial total populations as weights. One of the most important functions of the Congress is electing the president, which requires a simple majority from the parliament.

The shift from a two-party system that lasted for three decades to a multi-party system occurred during the 2015 general election. Traditionally, the two main parties, “Partido Popular” and “Partido Socialista Obrero Español”, were able to secure over 80 percent of the seats. However, in the 2015 election, their combined share dropped to only 51 percent. The Spanish political system consists of two types of political groups: statewide parties and non-statewide parties. The non-statewide parties do not field candidates in all electoral districts (Orriols and Cordero, 2016). For this reason, the present study analyzes the presidential candidates of statewide parties.

Hypothesis development

To accomplish our aim of understanding the role of candidate-based brand equity on intention to vote, we examine the correlation between these two constructs (Fig. 1).

Fig. 1
figure 1

Theoretical model of relationships.

Full size image

Political parties understand the importance of convincing voters to select their candidate. One method of achieving this is to take a CBBE approach. This approach allows for the creation of engagement between voters, candidates, and parties. Some authors consider the candidate brand as the main component of voting behavior (Blackston, 1992) because it is associated with personal relationships (MacLeod, 2000). As noted earlier, candidate-based brand equity is the reference to voter-based brand equity. Specifically, a positive assessment of candidate brand equity essentially reflects a positive perception of the brand as a whole (Laroche and Brisoux, 1989; Li et al., 2012). Furthermore, the combination of the candidate brand equity dimensions (awareness, association, quality, loyalty, and emotion) can serve as indicators of voters’ overall subjective assessment of the relevant brand information (i.e., knowledge confidence) and relative date in relation to competitors (i.e., brand familiarity). Each candidate will have a unique distribution of results for each dimension. This variation in values for each dimension by the candidate may provide a unique profile that candidates can use to manage their brand equity. This suggests hypothesis 1 (H1) and its stated sub-hypotheses (H1a, H1b, Hic, H1d and H1e):

H1: The CBBE model will provide a unique personal profile for each candidate through the brand dimensions.

H1a. The Candidate Perceived Quality influence on candidate brand equity.

H1b. The Candidate Brand awareness influence on candidate brand equity.

H1c. The Candidate Brand association influence on candidate brand equity.

H1d. The Candidate Brand Emotion influence on candidate brand equity.

H1e. The Candidate Brand Loyalty influence on candidate brand equity.

Electoral polling is a main tool for measuring voting intention. Despite its common use, election polling can only estimate overall voting intentions, but it cannot explain or provide direct insight as to the “why” of voter choice. CBBE profiles, however, have explanatory power as to the “why” of voter choice. In addition, the CBBE profiles provide a roadmap for how candidates can improve their election outcomes. Unlike traditional election polling, the CBBE approach provides candidates with tools that may be used to influence voter knowledge, feelings, and behaviors. Within the political context, this means that candidates can positively impact voters’ knowledge, which can then influence how voters feel about the candidate, and in turn, influence how they vote.

Based on this reasoning, the stated hypothesis 2 (H2) is

H2. The Candidate Brand Equity has a positive effect on voting intention.

Method

Sample and data collection

A survey was conducted in person and included demographic and general political-related questions. Surveys were conducted in Spanish. Data collection was conducted across seven different Spanish cities. The survey was conducted on a voluntary basis, and participants did not receive any form of compensation for their participation. The sampling used is non-probabilistic discretionary sampling by quotas. A questionnaire was pre-tested on a representative sample of the Spanish population, consisting of 56 individuals aged between 16 and 64 years. The sample was carefully distributed by gender and age to reflect the proportions of the Spanish population. The purpose of this pre-test was to assess the adequacy of the scales used in the questionnaire. The final total sample consisted of N = 375 individuals, which is considered an appropriate size for survey research using questionnaires (Malhotra, 2009). All participants indicated that they were aware of all candidates who were listed in the questionnaire, meeting the criterion for inclusion in the study.

The candidates listed were from the five most important statewide parties for the 2015 Spanish presidential election. This included Mariano Rajoy (R), Pedro Sánchez (S), Albert Rivera (Ri), Pablo Iglesias (I) and Alberto Garzón (G). Participants were asked to respond to each survey item with regard to all five candidates, resulting in 85 item responses (17 items × 5 candidates).

Just over half of the sample were male (51.4%). The average age of the participants was 30 years, with ages ranging from 18 to 81 years. Table 1 presents a comprehensive overview of the participants’ demographic data, revealing a well-balanced and representative sample in terms of socio-economic and demographic characteristics.

Table 1 Sample information.
Full size table

Measures

Candidate brand equity and voting intention (VI) were measured with 17 items, 14 for CBBE and 3 for VI. All items were derived from scales used previously in the literature. All items were evaluated using a five-point Likert-type scale, where 1 represented “strongly disagree” and 5 represented “strongly agree”. Table 2 lists all items and their corresponding psychometric properties.

Table 2 Measurement model (reliability and validity of scales).
Full size table

In order to adapt Aaker’s original brand equity items (1991, 1996, 2011) for the context of political candidates, certain modifications were made to the wording. The original model had 22 total items, 19 for CBBE and 3 for VI. The final model had 17 total items, 14 for CBBE and 3 for VI. The CBBE items included three items for candidate awareness (“I am aware of this candidate”, “I can recognize this candidate among other competing candidates” and “It is on my mind the party of this candidate”), three items for candidate brand association (“some characteristics of this candidate come to my mind quickly”, “I can quickly recall the logos or symbols associated to this candidate” and “I haven’t difficulty imagining this candidate in my mind”), three items for perceived candidate quality (“the likelihood that this candidate would be a good president is extremely high”, “the likelihood that this candidate would be a workable president is extremely high” and “the likelihood of a party winning the elections are increased with this candidate”), three items for candidate brand loyalty (“I (would) prefer that this candidate (was) is in my favorite political party against other candidates”, “I (would) vote this candidate if he (stood) stands for election” and “it is smarter to relay on this candidate in my party better than another candidate”). Two items for the new dimension, candidate brand emotion, were included to measure emotional aspects. The items were “this candidate transmit yourself emotions when he is talking” and “it is not easy to explain why this is a candidate with a great value.” The voting intention variable consisted of three items (“my willingness to vote the party where this candidate is involved is high”, “I would trust in voting a party promoted by this candidate” and “the likelihood of voting the party where this candidate is involved is high”).

The initial phase of the data analysis plan involved examining the factor structure of candidate brand equity and voting intention (Aloisi et al., 2018). The survey comprised a total of 22 questions. Following the elimination of items with factor loadings below 0.40, the findings indicated a viable model consisting of 17 scale items to measure the constructs. After the model was developed, the candidate brand equity was comprised of 5 dimensions measured with 14 items. The voting intention was one dimension measured with 3 items. The model fit and reliabilities for the scales suggest a well-fitting model (see Table 2).

Brand equity has been proposed as a first-order formative construct in past studies. This means that individual concepts are weighted and summed to define brand equity. However, other studies suggest the adoption of higher-order models to capture the complexity of constructs (Podsakoff et al., 2006) because these models allow individual dimensions to be treated as important components of the larger model, thus improving the representation of the construct (MacKenzie et al., 2005). For this reason, candidate brand equity is defined by reflective first-order models (that defined the measurement of the individual dimensions) in combination with a second-order formative model to form the candidate brand equity score. As the dimensions did not necessarily share a common underlying theme, the model assumed that the individual dimensions were not correlated (Jarvis et al., 2003). The main objective was to calculate scores for the individual dimensions and the overall level of candidate brand equity. This is akin to constructing an index that serves as a composite latent variable (LV), where its computation necessitates the utilization of formative indicators rather than reflective ones. In the case of a formatively measured LV, the indicators cause the LV (see Fig. 1) (Arnett and Hunt, 2003; Jara and Cliquet, 2012; Krystallis and Chrysochou, 2014; Wang et al., 2011; Wang and Finn, 2012).

Data analysis and results

Partial least squares structural equation modeling (PLS-SEM) was employed for data analysis. Similar to structural equation modeling (SEM), PLS-SEM allows researchers to simultaneously examine the structural component (path model) and measurement component (factor model) in a single model (Gefen et al., 2000). While covariance-based structural equation modeling is widely used, the present study opted for PLS-SEM, a variance-based form of structural equation modeling, due to its strong predictive emphasis (Hair et al., 2011). PLS-SEM is particularly suitable for prediction-oriented analyses, as it maximizes the explained variance in outcome measures. PLS-SEM accommodates formative measures that are assumed to cover the entire construct domain (Diamantopoulos, 2011), which is the assumption made for the proposed model of candidate brand equity. Each part of the proposed model, including the measurement model, structural model, and overall model, requires validation using PLS-SEM (Esposito Vinzi et al., 2010). PLS-SEM has demonstrated robustness when applied to non-normal data commonly encountered in survey research (Cassel et al., 1999; Mooi and Sarstedt, 2011; Reinartz et al., 2009). Assuming a medium effect size as defined by Cohen (1988), a significance level of 0.05, and statistical power of 0.8, the proposed model would require a minimum sample of 91 cases. Therefore, the sample size of the current study (N = 375) exceeds the minimum required to detect a medium effect size.

Aggregation of data

Respondents provided their assessment of candidate brand equity and voting intentions for each of the five candidates. Table 3 provides the mean scores, separately for each candidate, of these constructs, including all dimensions of candidate brand equity. Rivera obtained the highest mean score for candidate brand equity (3.42), followed by Iglesias (3.18), Sánchez (2.77), Garzón (2.72) and Rajoy (2.60). In the same order, the five candidates have a candidate-related voting intention, with Rivera (2.83) leading, followed by Iglesias (2.43), Sanchez (2.22), Garzon (2.21), and Rajoy (2.20). In preparation for hypothesis testing via structural equation modeling (hypothesis H1 and its sub-hypotheses), responses relating to individual candidates were aggregated into a mean score for each participant, representing a general assessment of candidates based on these constructs.

Table 3 Coefficients and scores by dimensions and candidates.
Full size table

Measurement model analysis

A two-stage covariance-based SEM process was used to evaluate the measurement models of the individual dimensions of candidate brand equity (Anderson and Gerbing, 1988). In order to assess the reliability of the measure, it is necessary to examine the relationship between each item and its corresponding latent construct. All estimated factor loadings linking the items to their respective constructs exceeded the minimum threshold of 0.71 and demonstrated stronger associations with their intended construct compared to other constructs included in the model (Hair et al., 2011) (see Table 2). These findings provide robust evidence for the reliability of the reflective measurement models for each individual construct. To further evaluate the reliability and validity of the formative second-order measurement model, the variance inflation factor (VIF) was utilized. The VIF assesses the collinearity of the indicators. A formative model is based on a multiple regression framework. Thus, collinearity between indicators would hinder the identification of the effect that each indicator has on the construct. Specifically, for candidate equity, a linear regression analysis was performed that used measures of perceived value, perceived quality and attitudinal loyalty as explanatory variables and items for candidate brand equity as outcome variables. Here, VIF values ranged between 1.5 and 3.0. As recommended by Ringle et al. (2013), VIF values between 0.20 and 5.0 are considered acceptable.

Scale composite reliability and Cronbach’s alpha were utilized to assess internal consistency. A reliability level of 0.70 is generally considered as a benchmark for “moderate” reliability, while a stricter interpretation of reliability in basic research is set at 0.80 (Nunnally and Bernstein, 1994). The composite reliability for each set of reflective measures within each factor exceeded 0.87, and all Cronbach’s alpha scores were well above 0.70. Consequently, the latent constructs demonstrated acceptable reliability, and all items were retained for the study.

Discriminant validity was assessed through two steps. The first step involved examining the average variance extracted (AVE). The AVE represents the proportion of variable variance captured by the construct’s measurement in relation to the variance attributable to measurement error (Fornell and Larcker, 1981). AVE values should exceed 0.50. In this study, all AVE values exceeded 0.50 (see Table 2). The second step involved comparing the square root of the AVE with the correlations among the other constructs. This comparison ensures that each construct is more closely related to its own measures than to the measures of other constructs (Fornell and Larcker, 1981). In the current investigation, this criterion was met, indicating discriminant validity for all constructs. The collective results provide strong support for the discriminant validity of all constructs.

Structural model analysis

Consistent with the conceptual development, Candidate brand equity was operationalized as a second-order construct comprising five latent first-order formative constructs. This conceptualization is depicted in Fig. 2 and the indicators for candidate brand equity, candidate name awareness (0.22), candidate brand association (0.20), perceived candidate quality (0.34), candidate brand loyalty (0.34), and candidate brand emotion (0.22) suggest they comprise, to different degrees, the construct. The findings of the analysis are presented in Fig. 2, showing the standardized path coefficients for each of the direct paths in the model. To test the hypothesized relationship, we specified a direct path from candidate brand equity leading to candidate-related voting intentions (H2). The analysis indicated a strong relationship between candidate brand equity and voting intention (β = 0.62). The results also indicated that the construct candidate brand equity explained approximately 39% of the variance in voting intention.

Fig. 2
figure 2

Results.

Full size image

Additional analysis was performed to examine the first hypothesis and its corresponding sub-hypotheses. Table 3 and Fig. 3 provide the coefficients between candidate brand equity dimensions and candidate brand equity for each different candidate. Additionally, scores for each candidate and dimensions are included in the results. These can be used to develop the importance-performance map (IPMA) and provide for an analysis of whether a candidate is in a different position and if they need to work to manage their brand in a different way, providing support for hypothesis (H1) and its sub-hypotheses (H1a, H1b, Hic, H1d, and H1e).

Fig. 3
figure 3

Importance-performance map analysis by parties and candidates.

Full size image

The findings indicate that the components with high influence or importance (constructs demonstrating a strong total effect) also exhibit relatively low performance (low average latent variable scores) (Ringle and Sarstedt, 2016). Specifically, loyalty (0.344) and perceived quality (0.342) demonstrate the highest influence or importance while having relatively low performance (2.296 and 2.671 respectively). Both variables have the potential to boost candidate performance, which is relevant for political strategists. Candidate brand association (0.201) and candidate brand awareness (0.218) have low importance and relatively high performance (3.961 and 4.053). Finally, candidate brand emotion (0.222) has a low importance and a low performance (2.499).

The results for Rajoy and Sanchez exceed those of the total model in importance (coefficients), candidate brand loyalty, candidate perceived quality and candidate brand emotion. When considering performance (scores), the results for Rajoy and Sanchez are lower than the overall model scores for the same variables. The results for Garzón are like those for Rajoy and Sanchez for brand awareness and brand associations. They differ, however, in terms of importance and performance for the other variables in the model. Finally, Iglesias and Rivera are lower than the result of the total model in importance and performance for candidate brand loyalty, candidate perceived quality, and candidate brand emotion. These results for these candidates are larger than those for the total importance and performance of candidate brand awareness and candidate brand association (Fig. 3).

Discussion and conclusions

The measurement of candidate brand equity in the weeks leading up to the Spanish general election exhibited a strong correlation with future voting intentions. That is, the candidates perceived by voters as strongest (Podemos and Ciudadanos) achieved more than one-third of total votes (20.66 and 13.93 respectively). Both candidates were from new statewide parties. Candidate-perceived quality and candidate loyalty dimensions were most influential on candidates’ brand equity. The candidates best valued for their brand equity are those appearing as new parties with impressive results in the 2015 elections in Spanish elections. They became the third and fourth most-voted political force. In any case, the present study is exploratory due to the size of the sample, the methodology used as well as the need to conduct longitudinal studies on the topic. It can be deduced that the endorsement of brand equity in political campaigns is a crucial factor in determining the likelihood of success or failure for a political party or candidate in an election. However, it is important to note that this may not always be the case, as individuals hold varying perspectives on each party, and their support may not solely rely on the affiliation of their preferred candidate. Further research is required to delve deeper into this area of study (Singh and Banerjee, 2018). Similar to previous research, the findings demonstrate that voter preferences for candidates during elections are influenced by the application of political marketing research and practices, presenting opportunities for future research (Yalley, 2018).

The influence of party-based brand equity implies the presence of a partisan bias in the formation of candidate brands. Rajoy and Sanchez are the candidates for the traditional “two party-system”, as they had the highest values for brand loyalty and brand equity as compared to the other candidates. The findings indicate that loyalty might have a greater significance for the major parties, highlighting the necessity for the other parties to focus on enhancing these specific dimensions of candidate brand equity (CBBE). The results for candidate quality perception were like those of brand loyalty. The influence of brand loyalty and candidate quality perception for these two candidates, however, did not differ in relation to the other candidates as in the case of brand loyalty. Findings indicate that brand loyalty and perceived quality might have a more significant impact on the Spanish presidential election, particularly for the major parties.

During the data collection period, all five candidates had a high level of familiarity, resulting in substantial candidate awareness. This factor provided limited value in comparing candidate brand strength, particularly among members of the “two-party system” where its influence was low. However, this does not diminish the significance of candidate brand awareness during different stages of a campaign. Having a high level of name recognition has always been a crucial strategic advantage or obstacle for lesser-known candidates, which must be addressed early in the campaign cycle. Nevertheless, for most candidates, the role of candidate brand awareness in candidate brand equity and voting choice is not a significant concern. Only Garzon has room to increase his brand awareness to improve his overall brand equity.

Candidate brand association is the dimension of brand equity that strategists exert the most control over and allocate significant resources toward establishing and managing. It holds great importance as it reflects the structures of consumer brand knowledge and encompasses the comprehensive image of a brand. The influence on candidate brand equity is limited and the scores of all candidates are very high. This makes it difficult to efficiently manage activities with this dimension. Considering the actual information associated with candidate brands, particularly their favorability (i.e., whether they are liked or disliked), maybe the most suitable approach. Like brand awareness, Garzon is the only candidate who can possibly increase his brand association to benefit his overall brand equity.

Candidate brand emotion was relatively smaller/weaker than the other dimension. However, given the novelty of this new dimension (Wang et al., 2011), it may be too soon to suggest that it can’t be used to improve overall brand equity.

This study contributes to the field of political marketing by treating political candidates as brands and employing a theoretically grounded, multidimensional brand equity model to assess their potential success in future elections (Atzger et al., 2020; Parker, 2012). The proposed model offers a framework that aids in evaluating and managing political brands, as it encompasses voter-centric dimensions such as awareness, association, perception (quality evaluation), loyalty, and emotion, all of which are linked to future voting choices (Kaneva and Klemmer, 2016; Van Steenburg and Guzman, 2019). By examining candidate brand equity as a multidimensional construct, this study significantly advances the understanding of political branding and sheds light on related research in the field (Bennett et al., 2019; Jain et al., 2018; Kaur and Sohal, 2019; Osorio et al., 2020).

Practically, this study offers a valuable alternative to polling data through the application of candidate brand equity valuation. This approach proves advantageous not only due to the established connection between brand equity and voting intentions but also because it enables the management of various brand equity dimensions for the benefit of a candidate. From a business perspective, the candidate brand equity measurement scale developed in this study equips political research practitioners with a method to assess voter perceptions of candidate brand equity based on the dimensions that truly matter (Cuesta-Valiño et al., 2021a). It suggests that political marketers should employ market targeting, marketing intelligence, and demographic segmentation, while also devising positioning strategies utilizing political personalities and brand ambassadors in line with voters’ demographic preferences for candidates (Yalley, 2018).

Study limitations

In any case, the present study is exploratory due to the size of the sample, the methodology used as well as the need to conduct longitudinal studies on the topic, several limitations should be mentioned. The study design employed a sample of residents from five specific cities, and it is important to note that they may not be representative of all voters or the general population of the country. Furthermore, the sample exhibited a bias towards Ciudadanos’ voters, with a higher proportion (23%) compared to voters of other parties. Therefore, caution should be exercised when attempting to generalize the findings of this study beyond the sampled group of voters.

Additionally, it is worth noting that the study design was limited to utilizing candidate brand equity measures specifically developed for a political context. This narrow focus may restrict the comprehensive understanding of other factors that can influence voter behavior in elections. Although candidate brand equity seems to have its own identity, considering a party brand equity variable as a moderator might be beneficial. The construct of voting intention is closely associated, but there are varying theoretical perspectives among scholars regarding this relationship. In this study, candidate brand loyalty was conceptualized based on Aaker’s (1991, 1996) theory and operationalized through candidate brand equity, treating loyalty as one dimension of brand equity. However, Keller (1993) argues that brand loyalty is an outcome of brand knowledge. This discrepancy in brand equity theory is present in numerous brand equity studies (Cuesta-Valiño et al., 2021b).

Future research

Future research could consider a test of the proposed model using data collected during election cycles in other countries and in other political levels (i.e., state and local). A more robust measure of voting intentions would also be of value. Because candidate brand equity is moderated by other variables, future research could explore the identification and examination of additional variables to gain a deeper understanding of a causal multidimensional model. It is also important to investigate the role of party bias in relation to candidate brand equity. By studying these aspects, we can enhance our comprehension of the complex dynamics involved in shaping candidate brand equity. More work on brand emotion might also be beneficial. Finally, a longitudinal approach would be used for understanding the trajectory of CBBE, perhaps allowing candidates the ability to manage it before, during, and after elections.