1 Introduction

The COVID-19 pandemic impacted the daily routines of people around the world, resulting in a tumultuous economic climate and numerous human casualties (Giones et al., 2020; Khlystova et al., 2021; Uddin et al., 2021). In the USA, federal leaders proposed numerous recommendations for combatting the spread of the novel coronavirus during its infancy. However, the decision to implement such recommendations was ultimately delegated to state leadership (Wang et al., 2021). Indeed, state leaders were tasked with managing extreme environmental uncertainty and unforeseen demands from both citizens and organizations alike. Furthermore, leaders were given very little time to prepare before acting. As such, decisions concerning the health of the state’s citizens and the condition of the state’s economy were likely influenced by leaders’ personal and party values.

Although the crisis continues to affect many people at the time of this writing, there remains considerable interest regarding how individual differences among state leaders, and in particular governors, impacted various outcomes during the pandemic’s early stages. For example, research suggests that governor political partisanship (Neelon et al., 2021) and gender (Sergent & Stajkovic, 2020) are significantly related to the number of deaths due to COVID-19 complications. While scholars continue to investigate the important relationship between state leadership and the well-being of citizens, much less attention has been given to how these leaders impacted state-level economic outcomes — specifically entrepreneurial outcomes — during these times of extreme uncertainty. Indeed, launching a new venture during normal economic conditions is no easy feat, let alone amidst an international pandemic. Yet, during the crisis, new ventures were created by entrepreneurs to exploit opportunities, despite economic and social restrictions imposed by state governments. While anecdotal evidence suggests that states experienced vastly different economic outcomes, the impact that state leaders had in creating these disparities is generally missing from the literature.

Consequently, we fill important gaps by empirically investigating phenomena at the interface of COVID-19 and entrepreneurship. In doing so, we deliver on prior requests to quantitatively address the short- and medium-term economic effects of COVID-19 (see Belitski et al., 2022). With few exceptions (e.g., Belghitar et al., 2022; Block et al., 2022; Kalenkoski & Wulff Pabilonia, 2022), much of the literature in this area has placed its focus on theory building and commentary (e.g., Kuckertz & Brändle, 2021; Tang et al., 2021). While theory development was a particularly necessary element of research during the early stages of the pandemic — so as to help initiate the process of building knowledge on the topic — our understanding of how COVID-19 has impacted important entrepreneurial outcomes can be further captured by empirically testing these relationships.

Our study builds on prior empirical work in the area (Fairlie & Fossen, 2022) by providing a theory-driven model linking state leader characteristics to economic recovery. We support our theoretical account by empirically examining the relationship between both governor party and governor discretion (i.e., when there is political congruence between a state’s governor and its legislature) and the number of new food and restaurant ventures created in the months after the initial onset of the COVID-19 pandemic. The food service industry is an important context for entrepreneurship research, given that 9 out of 10 restaurants have fewer than 50 employees, and 7 out of 10 restaurants are single-unit operations (NRA, 2022). Moreover, the food and restaurant industry accounts for nearly 4% of annual US GDP — accounting for nearly $1 trillion — and is among the largest industries for small business owners, trailing only healthcare in number of workers (NRA, 2022; Small Business Administration, 2021). During the COVID-19 pandemic, food service was responsible for at least three times the amount of job loss experienced by other industries, as more than 40% of US restaurants ceased to operate during this time (Kaufaman et al., 2020). Indeed, the food and restaurant industry was among the most affected industries after stay-at-home mandates were imposed, causing entrepreneurs to significantly alter their traditional business models. Therefore, we posit that the restaurant and food service sector is an appropriate benchmark for entrepreneurial outcomes and that new growth in this sector during the COVID-19 pandemic would represent a considerable economic achievement.

Second, we contribute to upper echelons theory (UET) and the political science literature by enhancing our collective understanding of the traits and characteristics of political leaders. UET is traditionally invoked when examining corporate executives — contending that chief executive officers’ (CEOs’) unique values, experiences, and personalities influence their decisions and actions, thereby having an impact on firm outcomes (Hambrick, 2007; Hambrick & Mason, 1984). We extend this logic and argue that the theory’s major tenets can also be applied to governors, the highest-ranking position in state leadership (Abney & Lauth, 1983). Specifically, we suggest that not only do governor characteristics affect outcomes at the state level (similar to how CEOs’ characteristics affect outcomes at the firm level), but that environmental context also matters. Said simply, we contribute to UET by examining how governor characteristics can impact both the political and economic environments.

Within the political environment, we invoke UET’s unity of command perspective (see Finkelstein & D’aveni, 1994) to assert that when governors have discretion to act unilaterally, they are able to advance their proposed policies unimpeded. Discretion has indeed been shown to be a crucial determinant in effectively managing uncertainty (Krause et al., 2014), including for governors during the COVID-19 pandemic. For example, Wang et al. (2021) found that governor discretion — which the authors conceive as the alignment between governor political affiliation and state legislature affiliation — enabled Democratic governors to issue stay-at-home orders more quickly. Likewise, we argue that alignment between political branches enhances governor discretion and fosters expedience and consistency in policy enacted at the state level. For prospective food and restaurant entrepreneurs, we posit that unity of command conveys consistency in terms of legislation and health policies that may provide stability and enhance business optimism, as entrepreneurs may better understand what is required of them as they seek to launch their new ventures.

We also observe interaction effects between the economic environment and governor characteristics. Specifically, we show how the demand for jobs (i.e., state small business unemployment rate) serves as a salient moderator for governor discretion and the number of new food and restaurant ventures. We posit that high unemployment rates present an uncertain and dynamic environment in which to launch a new food venture. Thus, having a consistent message from one’s elected state leaders could help assuage such concerns, especially if traditional employment ceases to become a viable option for prospective entrepreneurs. In our efforts to show how governor characteristics and government composition serve as important antecedents to new venture creation in the context of the food and restaurant industry, we contribute to the existing literature that intersects UET and entrepreneurship (e.g., Boone et al., 2019; Gulati & Higgins, 2003).

2 Theory and hypothesis development

2.1 Upper echelons theory within the political context

UET (Hambrick & Mason, 1984) is an important theoretical perspective that is often invoked within the management literature to explain how corporate executives’ — typically CEOs’ — values, experiences, and personalities influence organizational decision-making. Prior work has demonstrated how CEO characteristics play a critical role in constructing both strategic choices of the firm and the firm’s subsequent performance (see Neely et al., 2020; White & Borgholthaus, 2022). Because executives are boundedly rational (Hambrick, 2007), personal values and beliefs influence the way they navigate complex and uncertain environments when making decisions. Ultimately, the primary tenets of UET contend that “(1) executives act on the basis of their personalized interpretations of the strategic situation they face, and (2) the personalized construals are a function of the executives’ experiences, values, and personalities” (Hambrick, 2007: 334). Thus, a firm’s strategies and performance strongly reflect its CEO’s inherent characteristics.

While UET is traditionally applied within the context of corporate executives, research also suggests that it is an appropriate theoretical lens through which political leaders may be studied. For many years, researchers have argued that state governors serve in capacities similar to corporate CEOs (Abney & Lauth, 1983; Ferguson, 2003), as they are both the highest-ranking officials in their organizations and possess similar levels of power and discretion in these roles (Wang et al., 2021). Political scientists have specifically brought attention to this comparison, referring to state governors as ‘chief executives’ (Ferguson, 2003), because of the similar responsibilities that governors possess in addressing various policies, programs, and stakeholders (Keena & Knight-Finley, 2018).

Similar to how CEOs’ experiences, values, and personalities have an impact on decision-making, we expect that these same characteristics will influence how governors make policy and pursue various courses of action. As such, and consistent with recent management scholarship (e.g., Wang et al., 2021), we extend the central tenets of UET to a political context. In doing so, we integrate both the management and political science literatures. Specifically, we examine how two particularly salient governor characteristics — political party and discretion — influence approaches to policy, especially with regard to the “life or livelihood” debate that became very salient for decision-makers and the public during the early stages of the COVID-19 pandemic.

2.2 Governor party and new food and restaurant ventures

UET would suggest that political affiliation is one particularly salient characteristic that impacts a governor’s decision-making (Hambrick, 2007). The US political system consists of two major parties: the Republican Party and the Democratic Party. Generally speaking, the Republican Party is considered more conservative and individualistic, while the Democratic Party has a collectivist point of view and takes a more liberal approach to policy. Building on the idea that differences between political parties are grounded in deep psychological feelings of right and wrong (Kivikangas et al., 2021), moral foundations theory (MFT) posits that Republicans and Democrats differ in the extent to which they endorse five fundamental values (Graham et al., 2009): harm/care, fairness/reciprocity, ingroup/loyalty, authority/respect, and purity/sanctity. By taking a conservative approach that emphasizes virtues and institutions binding people into roles, duties, and obligations, Republicans tend to score higher on ingroup/loyalty, authority/respect, and purity/sanctity. Democrats, on the other hand, take a more liberal approach in which greater emphasis is placed on protecting individuals from harm or unfair treatment by other individuals or social systems, resulting in higher scores on harm/care and fairness/reciprocity.

Because of fundamental differences in their foundational values, as suggested by MFT (Graham et al., 2009), Republicans and Democrats have differing views about almost all major aspects of society, including business and government. For example, Republicans tend to (1) be business friendly, arguing that tax cuts on businesses will enable them to hire more workers, thereby increasing demand and growth; (2) advocate for supply-side economics which primarily benefit businesses and their investors; and (3) support individual pursuit of prosperity without significant government interference. Rebuilding the American economy has been the top priority of the Republican Party for many years (Barrasso et al., 2016) and continued to be the primary aim of the party throughout the COVID-19 pandemic (Wheeler, 2020). Oppositely, Democrats tend to (1) support economic policies that benefit low-income families, arguing that a reduction in income inequality will spur economic growth, and (2) support Keynesian economic theory, suggesting that governments should spend their way out of a recession. Democrats place a higher priority on worker rights and safety, as opposed to business ownership and top-line economic growth (The Democratic National Committee, 2020).

Based on these tenets, Republicans and Democrats have strong preferences and opinions with regard to the role that government should play in society, specifically with regard to the government’s involvement in enacting various policies and programs (Francia et al., 2005). Wang et al. (2021) argued that states led by governors from the Democratic Party were more likely to issue stay-at-home orders at the initial onset of the COVID-19 pandemic because the Democratic Party favors increased government oversight, stricter regulations and laws, and shared responsibility among individuals (Miller & Schofield, 2008). They further proposed that the Republican Party tends to prefer individual responsibility and fewer regulations, thereby making them less likely to issue stay-at-home orders, or to do so more slowly. Based on traditional party tendencies, prior work (e.g., Neelon et al., 2021; Wang et al., 2021) has demonstrated that states led by Democratic governors had lower COVID-19 death rates than those led by Republican governors.

After COVID-19 was declared a national emergency, governors began issuing stay-at-home orders to combat the spread of the virus.Footnote 1 Existing research has shown that Democratic governors issued stay-at-home orders faster than their Republican counterparts and maintained COVID-19 restrictions for longer periods of time (Neelon et al., 2021; Wang et al., 2021). On the other hand, Republican governors were reluctant to issue restrictive COVID-19 orders and instead prioritized economic well-being and the economic livelihood of individuals. Stay-at-home policies had especially large repercussions for businesses in the food and restaurant industry because many of these businesses depend on patrons visiting their establishments in person to generate revenue (Kaufaman et al., 2020). Republican governors, concerned about the impact that stay-at-home policies might have on their respective state’s economy, and especially the impact on small businesses, delayed enacting such policies, if they issued one at all (Wang et al., 2021). Moreover, states led by Republican governors implemented relaxed policies (relative to states led by Democratic governors) relating to physical distancing and mask mandates, thereby creating an environment which was more conducive to new ventures in the food and restaurant industry. As such, we hypothesize:

  • H1 Governor party is associated with the number of new food and restaurant ventures, such that states led by Republican governors will be associated with a higher number of new ventures.

2.3 Governor discretion and new food and restaurant ventures

While prior work in the UET literature has placed substantial focus on executive — or in the context of our study, governor — characteristics, scholars have also shown considerable interest in identifying and exploring situational circumstances under which the effects of individual characteristics are more or less salient (Cragun et al., 2020). One particularly important construct is that of managerial discretion (Finkelstein & Hambrick, 1990; Wangrow et al., 2015), which relates to executives’ latitude to act as desired. If a manager has high discretion, they are said to be able to enact their agendas with less opposition, whereas low discretion could (and in many cases, would) thwart an individual’s plans. High discretion enables unity of command within an organization, as subordinate leaders follow the will of the primary leader and thus decisions within the unified command structure practically fall under the purview of a single leader. Said simply, managerial discretion likely amplifies the impact of decisions made by a chief executive (Finkelstein et al., 2009).

Discretion has emerged as a controversial topic within the literature because it often serves as a double-edged sword (Dalton et al., 2007; Finkelstein & D’aveni, 1994). Specifically, scholars have argued about the inherent trade-offs associated with unity of command. Having a homogeneous and unified structure can facilitate swift decision-making, but such a structure also makes it difficult to implement checks and balances within the system (Krause et al., 2014). Within the context of US governors, Wang et al., (2021: 4) discuss discretion in the following manner: “[W]hen a state’s governorship and legislature share the same political party, the governor may receive less opposition and pressure from the state legislature and thus have more discretion.” Alternatively, when the governorship and legislature do not share the same political party, the governor is more likely subjected to scrutiny and opposition.

Perhaps even more significant from an entrepreneurship perspective is the signal that unity of command sends to the citizenry. Having a high unity of command signals a consistent approach to policy (e.g., consistency in political affiliation between the governor and state legislature), which may lead to continuity in economic expectations and lower uncertainty about the general environment (Nelson & Gibson, 2017). Institutional uncertainty and the uncertainty related to economic policies serve as important factors for potential entrepreneurs as they seek to launch new ventures (Bylund & McCaffrey, 2017), and were likely to be particularly salient for would-be entrepreneurs attempting to launch new food and restaurant businesses amid the COVID-19 pandemic. In other words, the level of environmental and economic uncertainty was arguably of the utmost importance for nascent restaurant entrepreneurs dealing with COVID-19 because they were forced to consider the boundaries and regulations within which they were expected to operate. Thus, the clarity of purpose in policy communication coming from government leaders likely had a profound impact upon the way prospective entrepreneurs approached launching their ventures.

For example, one would be hard-pressed to open a fine-dining restaurant exclusively for dine-in customers in states with strict regulations about social distancing, masking, and food-handling policies. Instead, nascent entrepreneurs were more inclined to relegate their operations to less traditional food services like food trucks, takeout, or outdoor dining. Within the context of UET, governor discretion likely enabled individuals to recognize and predict the stability and favorability of political and social institutions and craft business strategies accordingly (Bylund & McCaffrey, 2017). Entrepreneurs in stable political environments where there was high unity of command were arguably more capable of capitalizing on potential opportunities while simultaneously avoiding threats by exploiting advantages offered by consistent policy expectations (Kuckertz et al., 2020). Formally, we hypothesize:

  • H2 Governor discretion is positively associated with the number of new food and restaurant ventures.

2.4 Integrating party and discretion

While governor discretion by itself can help promote political stability and policy expectations for nascent food and restaurant entrepreneurs, such promotion may be more or less important depending on the political party that is leading the state. As argued above, governors from the Democratic Party, with greater discretion, are likely to initiate policies that are related to increased health measures for the general citizenry. On the other hand, governors from the Republican Party, with greater discretion, are more likely to enact policies that are advantageous for buoying the economy. As such, we expect the following:

  • H3 Governor party moderates the positive relationship between governor discretion and the number of new food and restaurant ventures, such that the relationship is strengthened when states are led by Republican governors.

2.5 The moderating role of unemployment

Although governor party and discretion help provide clarity for policy expectations among would-be entrepreneurs, additional factors in the economic environment are likely to muddy the waters. In the aftermath of the initial COVID-19 economic shutdown, the US unemployment rate increased from 4.4% in March 2020 to 14.4% in April 2020 (U.S. Bureau of Labor Statistics). Many individuals were forced to reconsider their work options during the initial unemployment surge, as traditional wage employment was not available to everyone (Grashuis, 2021). In lieu of impending financial hardship brought about by unemployment, many individuals turned to self-employment as an alternative to enhance their financial circumstances (Fairlie, 2020). However, doing so should be much easier for those living in states led by governors more concerned about economic recovery than about public health (i.e., governors affiliated with the Republican Party). Moreover, due to uncertainty in the political environment, we expect the risks associated with self-employment to be much higher in states where the governor lacks discretion to unilaterally make decisions. Said simply, when unemployment rates are high, more people are likely to consider alternatives to traditional employment, and when policy expectations are easily distinguishable and presumably stable, individuals are less inhibited from pursuing entrepreneurship. More formally, we hypothesize:

  • H4a Unemployment moderates the relationship between governor party and the number of new food and restaurant ventures, such that when unemployment is high, states led by Republican governors will experience more new food and restaurant ventures than states led by Democratic governors.

  • H4b Unemployment moderates the positive relationship between governor discretion and the number of new food and restaurant ventures, such that the relationship will be strengthened when unemployment is high.

3 Methods

3.1 Sample and procedures

We tested our hypotheses using data from the first 7 months of the COVID-19 crisis (March–September 2020). We selected the early period of the pandemic for specific reasons. First, lockdowns were first instituted in the USA in March 2020; therefore, the pandemic’s effect on business activity was likely more pronounced during the early stages of the pandemic. Furthermore, March through September 2020 witnessed the first major surge in both new COVID-19 cases and deaths associated with the disease, resulting in the highest levels of uncertainty within the institutional environment. Second, because state leaders were largely given early latitude to implement lockdown and physical distancing measures within their respective states at the beginning of the pandemic, the early months of COVID-19 seem to be a logical setting in which to examine how governors imposed their policy of choice, thereby affecting entrepreneurial outcomes.

We perform our study within the context of the food and restaurant industry because this industry was among the hardest hit within the US economy. Furthermore, restaurant and food service businesses comprise significant market share across all regions of the country, unlike other, more geographic specific industries that are located within confined areas due to the inherent raw materials or human capital found therein (e.g., coal mining or software development). Data for the variables in this study were obtained using several different archival sources. Specifically, we procured data from (1) State Governors’ Offices, (2) The Book of the States, (3) U.S. Census Bureau, (4) Ballotpedia, (5) U.S. Bureau of Labor Statistics Local Area Unemployment Statistics Program, and (6) Yelp.

3.2 Measures

Number of new food and restaurant ventures

Our dependent variable is sourced from the third quarter 2020 economic report issued by Yelp, Inc. (Mouras et al., 2020). Yelp is a popular data provider hosting crowdsourced reviews about small businesses from around the world and has proven useful for academic research in the past (e.g., Luca, 2011). Throughout 2020, Yelp economists regularly published articles and reports to capture the effects of COVID-19 on small businesses. In our study, we examine the number of new restaurant and food business ventures per 100,000 population between March 1, 2020, and September 30, 2020, as reported by the economists at Yelp.

Governor party

Similar to our conceptualization of governor political partisanship, we measure governor party in accordance with party membership. Specifically, governor party was operationalized as a dichotomous variable, with 1 representing the Republican Party and 0 representing the Democratic Party.

Governor discretion

Governor discretion is significantly influenced by whether or not the controlling interest in the state legislature shares the same political party as the governor. We followed Wang et al. (2021) and operationalized governor discretion as a dichotomous variable, holding a value of 1 if the state legislature was controlled by the same political party as the governor and 0 otherwise.

Unemployment

We operationalize our moderating variable — unemployment — within the context of small businesses. Due to the nature of our dependent variable (new food and restaurant ventures), we identified small business unemployment as the most salient category of unemployment among those in our sample, as the businesses examined in our data were owner operated. As such, following Fairlie (2020), we capture small business unemployment using information from the Current Population Survey (CPS) to estimate state-level metrics. The CPS is a monthly effort conducted by the U.S. Census Bureau and the U.S. Bureau of Labor Statistics to survey between 70,000 and 140,000 people who are representative of the US population (Grashuis, 2021). Unlike other surveys of small business owners that do not reflect recent trends in ownership, the CPS is released monthly and is considered to be ideal for estimating changes in the number of small businesses over time (Fairlie, 2020). We identified all individuals within the CPS who owned — and were principally employed by — their business. Thus, non-manager owners and owners of side businesses were not included in our estimates. Next, we calculated the percentage of self-employed individuals within each state by dividing the total number of self-employed individuals by the total number of working individuals. Finally, we applied the resulting percentage of self-employed individuals sourced from the CPS to the estimates of monthly unemployment obtained from the U.S. Bureau of Labor Statistics’ Local Area Unemployment Statistics (LAUS) program and converted it to an unemployment rate for small businesses (in percent).

3.3 Control variables

We included controls to account for differences in state size and policy characteristics. Indeed, states with larger populations may experience greater risk of viral exposure and thus be more inclined to exact stricter closure policies. Therefore, we included state population, which we measured as the natural logarithm of the state’s 2020 population.Footnote 2 Because some states are more endowed with healthcare resources than others, we also control for the number of COVID-19 cases per capita. This variable accounts for potential strain on healthcare resources as well as increasing external pressure to enact more restrictive mandates. A high number of COVID-19 cases per capita and more restrictive mandates would likely have a negative impact on the number of food and restaurant ventures created. We additionally control for the percentage of the population that are foreign born immigrants. According to some estimates, immigrants are more than twice as likely to start businesses than are their native counterparts (Vandor, 2021; Wiens & Jackson, 2014), and a large concentration of these businesses are in the food service industry (Kerr & Kerr, 2020; Kosten, 2018). Moreover, longitudinal evidence suggests that one-in-five new food service businesses are created by immigrant entrepreneurs (Kerr & Kerr, 2020), thus we control for this factor. Additionally, because gubernatorial (i.e., relating to the governorship) election results often reflect a state’s general political orientation, we control for the state’s total Republican vote percentage. This is a strong indicator of the degree to which a state voted for the Republican governor, which likely impacts the degree to which governors enact policies consistent with their political party.

We also control for state-level factors regarding employment and property rights. First, because employers typically own or otherwise control the assets of their organizations, their ability to ‘do as they please’ with these resources is important. As such, we controlled for economic freedom of the state. The Fraser Institute’s economic freedom index, which has been used to study whether party ideology influences policymaking at the state level (Bjørnskov & Potrafke, 2013), reflects key aspects of the labor market such as minimum wage legislation, tax policy, and tax structure (Ashby & Sobel, 2008). This control variable represents residents’ individual liberties with respect to life and property. High economic freedom indicates that an individual can live their lives as they see fit, so long as they do not infringe upon the rights of others. Furthermore, because starting a business can require sizeable financial capital, we also control for the state’s average disposable income (in thousands of dollars).

We also added controls for individual differences among governors that may affect decision-making. For example, the age of top executives is important in an organizational context for several reasons. First, previous research suggests that younger managers are often associated with organizational growth and increased risk-taking (Child, 1974). Second, older executives tend to not only seek more information but also spend more time integrating and evaluating information than their younger counterparts, therefore taking longer to make decisions (Taylor, 1975). Third, research shows that older executives have a stronger psychological commitment to maintaining the status quo (Stevens et al., 1978). In summary, because older decision-makers are less likely to disrupt the status quo, their regimes are characterized by greater stability. We posit that the increased stability that accompanies governor age will have a positive impact on the number of new restaurant ventures. Thus, we control for governor age (in years).

Similarly, we also control for governor gender (0 = male and 1 = female). Empirical evidence suggests notable differences in risk aversion between women and men (Faccio et al., 2016), especially during situations where the leader is making a decision on behalf of others (Stone et al., 2002). We expect female governors to be more risk averse in this context, and would be more likely to preserve life at the expense of economic livelihood (Sergent & Stajkovic, 2020). Additionally, executive tenure has been shown to influence risk-taking and confidence in decision-making situations (Orens & Reheul, 2013). For example, Hambrick et al. (1993) found that long-tenured executives were less likely to go against commonly accepted practices, whereas younger executives may be less committed psychologically to existing protocols. Therefore, we controlled for governor term length in office and expect longer tenured governors to have a positive relation with the number of new restaurant ventures. Finally, we control for governor race (0 = white and 1 = ethnic minority). Prior research has suggested that minority leaders are more likely to support strong, cautious governance (Smith & Nkomo, 2003). As such, we expect ethnic minority governors to support the preservation of life and reduction of health risks, thereby having a negative impact on the number of new food and restaurant ventures created.

3.4 Analysis and results

Descriptive statistics are found in Table 1 and results for our primary analysis are presented in Table 2. In each model, we include the full list of controls at both the governor and state level, with model 1 representing how our base model is associated with the number of new food and restaurant ventures. Model 2 is specified to include both independent variables (i.e., governor party and governor discretion), thereby enabling us to test H1 and H2. In testing H1, our results indicate that governor party fails to predict venture creation in the food and restaurant industry (β =  − 1.58; p = n.s.). Consistent with our theorizing, we find that governor discretion is positively associated with the number of new food and restaurant ventures (β = 2.51; p < 0.05), thereby supporting H2.

Table 1 Descriptive statistics and correlation matrix
Full size table
Table 2 OLS regression estimates predicting number of new food and restaurant ventures per 100,000 population
Full size table

Next, model 3 presents the interaction between governor party and governor discretion and shows its effect on new food and restaurant ventures. We find that this relationship is not statistically significant (β = 3.88; p = n.s.). As such, we fail to find support for H3. Model 4 introduces our moderator, small business unemployment, enabling us to test H4a and H4b. We find that the interaction between governor party and small business unemployment is indeed associated with the number of food and restaurant ventures created (β =  − 1774.51; p < 0.05). However, our results demonstrate a relationship opposite of what we initially hypothesized. Figure 1 displays a representation of this relationship. Specifically, Fig. 1 shows that compared to states led by Republican governors, states led by Democratic governors experience a greater number of food and restaurant ventures when facing high small business unemployment rates. Interestingly, food and restaurant ventures are more likely to be created in states led by Republican governors when small business unemployment rates are low. We discuss potential explanations for these relationships in greater detail in “Sect. 4.” Finally, we test H4b and find that the interaction between governor discretion and the number of new food and restaurant ventures is positive and statistically significant (β = 1637.63; p < 0.05), providing support for our hypothesis that states led by governors who possess discretion see a higher number of food and restaurant ventures when small business unemployment in the state is higher (see Fig. 2).

Fig. 1
figure 1

Governor party, small business unemployment, and new food and restaurant ventures per 100,000 population. Figure 1 depicts the number of new food and restaurant ventures per capita for states with high and low levels of small business unemployment, as given by Model 4 of Table 2. Low small business unemployment refers to unemployment levels at one standard deviation below the mean, while high small business unemployment refers to unemployment levels at one standard deviation above the mean. Additionally, the X-axis plots whether the state is led by a Republican or Democratic governor. All other variables were held constant at their mean

Full size image
Fig. 2
figure 2

Governor discretion, small business unemployment, and new food and restaurant ventures per 100,000 population. Figure 2 depicts the number of new food and restaurant ventures per capita for states with high and low levels of small business unemployment, as given by Model 4 of Table 2. Low small business unemployment refers to unemployment levels at one standard deviation below the mean, while high small business unemployment refers to unemployment levels at one standard deviation above the mean. Additionally, the X-axis plots whether or not the state’s governor possessed discretion. All other variables were held constant at their mean

Full size image

3.5 Robustness and endogeneity

To ensure that the results of our analyses were robust, we conducted additional testing. Specifically, we used an alternative operationalization of the dependent variable, the number of new food and restaurant ventures. While our initial analysis measured this variable as the number of new ventures per 100,000 population, we additionally used the total raw number of ventures in a state. Because our outcome of interest in this robustness test is a count variable — the raw number of new food and restaurant ventures — we specified this model using a negative binomial regression with robust standard errors. We confirmed that this type of analysis was preferred to a Poisson distribution by testing for overdispersion. Specifically, the large and statistically significant chi-square value that was obtained from a goodness-of-fit test indicated that a Poisson distribution was inappropriate for our analysis and that a negative binomial regression should be applied instead. Results for these models are found in Table 3. We find that our results are consistent in significance to those models reported in Table 2. Therefore, our robustness tests support our initial findings.

Table 3 Negative binomial regression estimates predicting total number of new food and restaurant ventures (robustness test)
Full size table

There is also an ongoing debate in the academic literature on the merits and pitfalls of including control variables in regression equations (see Bernerth & Aguinis, 2016). As such, in an unreported model, we also tested our models using only significant control variables from the full model of our primary analysis (see Table 2 for the list of significant controls). This analysis provided additional support for our initial results. By testing our hypotheses using only significant controls, we show robustness in our findings.

We also recognize that it is possible for endogeneity to bias our results. Because of the complexity of the political environment, there may have been some variables which we failed to account for and unintentionally omitted during the research process. Because omitted variable bias may be “the most commonly encountered problem in social behavioral sciences” (Bascle, 2008: 290), it is particularly important that we address this form of endogeneity and express confidence that our findings are not biased because of it. To address this issue, we tested the impact threshold of a confounding variable (ITCV), which assesses how strongly correlated an omitted variable would have to be in order to overturn the results of our analysis (see Busenbark et al., 2022). Using this endogeneity test, we specifically examine whether omitted variable bias had an impact on the direct effect models of our study (i.e., how both governor party and governor discretion have an impact on food and restaurant venture creation).

We first performed the ITCV test with regard to governor discretion. Our results indicate that for our results to be biased, an omitted variable would have to be correlated at r > 0.29 with both governor discretion and new food and restaurant ventures, and that approximately 20% of our cases (i.e., 10 states) would need to be replaced with cases for which there is an effect of zero. Upon further examination of our correlation table (Table 1), we find that not one of our control variables meets this threshold. To put these values in perspective, governor race — at 0.48 — was the highest correlation between any of our control variables and our dependent variable. However, it was only correlated at 0.12 with governor discretion. As such, the ITCV tests indicate that overturning the results would require an omitted variable with an impact larger than the strongest control variable that was used in the model. We next performed the same test with regard to governor partyFootnote 3 and found that for an omitted variable to overturn our result, it would need to be correlated at r > 0.36 with both governor party and food and restaurant ventures. Furthermore, the ITCV test suggests that more than 62% of our cases (approximately 31 states) would need to be replaced with cases for which there is an effect of zero to bias our results. Overall, our ITCV tests suggest that omitted variable bias is unlikely to have a meaningful impact on our results.

It is also possible that either of our independent variables might be endogenous. In order to assess this possibility, we performed a multi-stage process, as outlined by Semadeni et al. (2014). We first examine whether or not governor discretion is impacted by endogeneity. An important part of the two-stage model is to identify exogenous instruments that are theoretically relevant to — and associated with — governor discretion, but are unrelated to our dependent variable, food and restaurant venture creation. We identified two particularly strong instrumental variables to include in our endogeneity testing: (1) whether or not the current governor’s party aligns with the preceding governor (i.e., governor predecessor party congruence) and (2) the ratio of the state’s voter turnout to its voting age population (i.e., voter turnout).

First, the literature suggests that governor predecessor party congruence should be related to governor discretion because states tend to be consistent in their political ideology over time (Gupta et al., 2018). For example, if a state has a Republican governor, then there is a strong likelihood that the former governor was also Republican and that the state’s legislature has a Republican majority. Alternatively, governor predecessor party congruence is unlikely to be related to new food and restaurant ventures because congruence can occur for either the Republican or Democratic Party. Moreover, it is improbable that the political affiliation of the preceding governor plays a significant role in predicting new food and restaurant ventures in the current period, particularly under the impact of COVID-19.

Second, we argue that the ratio of total voter turnout to total voter population is also theoretically relevant and an important predictor of governor discretion. Specifically, we advocate that total voter turnout to total voter population is likely to harm the incumbent political party (thereby being negatively related to governor discretion). For example, prior work has shown that increases in voter turnout will negatively impact the election races of the incumbent political party, such that votes for the incumbent officials decrease (Martins & Veiga, 2014). Interestingly, Hansford and Gomez (2010) demonstrated that increased voter turnout results in increases for Democratic candidates’ vote shares. Thus, we expect that voter turnout will be strongly associated with governor discretion. Furthermore, we expect voter turnout to be unrelated to new food and restaurant ventures because increased voter turnout can result in political victories for either party. As such — and in line with our theory development — there is a strong likelihood that voter turnout is independent of food and restaurant venture creation.

Using the two-stage model, the first stage indicated that both instrumental variables were highly predictive of governor discretion but did not predict food and restaurant venture creation when they were accounted for in the second stage, making them strong instruments. In an unreported model, we were able to determine that governor discretion is an exogenous variable and its inclusion does not bias our results, as it is still significant in the model after accounting for the instruments listed above (β = 3.11; p < 0.10). We next attempted to identify strong, exogenous instruments for governor party by evaluating whether predecessor governor political affiliation (Gupta et al., 2018), vote ratio and voter turnout (Kim et al., 2013), or state-level political alignment (e.g., proportion of voters in the firm’s home state who voted Republican in the most recent US presidential election) influenced governors’ political party. Based on these tests, we concur with prior work in the political science literature which proposes that governor party is an exogenous variable and currently lacks an appropriate instrument (Potrafke, 2018).

4 Discussion

The COVID-19 pandemic continues to impact everyday life across the United States. At its onset, each state’s governor enacted policies to help citizens manage the turbulence and uncertainty that the pandemic brought upon them. Governors, and their near daily press conferences, set the tone for their respective states’ environment via COVID-19 policies and post-COVID-19 economic recovery plans. While prior work examines the composition of state leadership and decision-making (e.g., Neelon et al., 2021; Sergent & Stajkovic, 2020; Wang et al., 2021) and COVID-19’s impact on small business survival (e.g., Grashuis, 2021), the phenomenon of new business ventures — a salient indicator of economic recovery — remains largely unaddressed.

For scholars, our research builds on UET by introducing a novel environmental context (e.g., state-level ecosystems) and dependent variable (i.e., new venture creation). We provide an empirical assessment of the extent to which both governor party and discretion are associated with new venture creation. Specifically, we find that states led by governors with greater discretion experienced more new food and restaurant venture activity than those where political influence was divided between the two parties. We further dive into the phenomenon by evaluating its boundary conditions, finding that a state’s small business unemployment rate moderates the relationship between governor discretion and new food and restaurant ventures. We find that the positive relationships that exist between governors with discretion and new food and restaurant ventures created becomes strengthened when small business unemployment increases. Notably, we did not find any significant interaction between governor political party and governor discretion, suggesting that the traits and characteristics of a united leadership team (regardless of political affiliation) is more important than the isolated political orientation of the senior leader.

Perhaps most interestingly, we did not find evidence to support H1, which hypothesized that states led by Republican governors would be associated with a higher level of new venture creation in the food and restaurant industry. However, we believe that UET is still an applicable theory within the context of governors, as our analysis demonstrated that other governor characteristics — namely gender and age — were associated with this outcome. For example, our findings support those of Sergent and Stajkovic (2020), who reported that female governors were associated with stricter COVID-19 policies and lower death counts. Specifically, we found a negative relationship between female governors and new venture creation, suggesting that strict policies implemented by states with female governors may have inadvertedently stymied new economic activity.

4.1 Implications

Our research has implications for research, practice, and policy. For example, with few exceptions (e.g., Belghitar et al., 2022; Block et al., 2022; Kalenkoski & Wulff Pabilonia, 2022), much of the extant research at the crossroads of COVID-19 and entrepreneurship has focused on theory building and commentaries, not constructing empirical evidence. We believe this to be an important gap that our study fills, as we build upon prior empirical work in the area (Fairlie & Fossen, 2022). Moreover, scholars suggest that a unity of command structure is a double-edged sword (Finkelstein & D’aveni, 1994; Krause et al., 2014), but have generally failed to account for crisis-type situations. Our results suggest that during precarious times — including a global pandemic — the centralization of power may produce positive outcomes. Specifically, we argue that as governors gain the requisite discretion to make decisions without experiencing significant pushback from the state’s legislature, it signals to the public a clear message of what to expect regarding the political or regulatory environment. Therefore, governors may wield considerable influence over economic activity, especially during times of crisis, and the effect may be strengthened with proper alignment among the various branches of government. This alignment has very strong implications for nascent and experienced entrepreneurs, as well as state and federal economic development agencies (e.g., startup assistance programs, incubators, and other entrepreneurship centers), such that the alignment of political leaders serves as an indicator of the environment regarding the optimal timing for venture creation decisions.

Accordingly, we make notable contributions to the entrepreneurship literature by appealing to the unity of command perspective and examining conditions of the economic environment that may encourage individuals to start businesses. Specifically, nascent entrepreneurs likely exhibit confidence when the political environment is stable because they know better what will be required of them as they pursue their new ventures. We submit that knowledge of political, institutional, and social expectations enables would-be entrepreneurs to navigate the process of forming a new business more easily. Our results further suggest that, during times of high unemployment, having a consistent approach to economic policy (as manifested by governor discretion) may serve as an important catalyst to economic recovery, as unity of command indicates a calmer, more stable regulatory environment in which to operate. Thus, during times of economic instability (i.e., high unemployment rates), individuals may take cues from political leadership before transitioning to self-employment. Unity of political command therefore may be a preferred approach toward navigating economic uncertainty. Our work extends unity of command to the political arena during times of crisis and complements other work examining alternatives to unemployment during the COVID-19 pandemic (e.g., Grashuis, 2021). Therefore, our findings not only provide strong evidence regarding which environmental conditions are conducive for fostering new business activity but also have strong implications for policymakers. For example, our study provides a potential roadmap of barriers for policymakers to consider when crafting or implementing economic development policies.

Lastly, we complement existing work using UET by extending it to the fields of political science and entrepreneurship. Although we did not find support for our hypothesis that governor party was directly associated with new venture openings, we did find evidence that other personal characteristics, including age and gender, were indeed associated with this relationship. Additionally, we found that political orientation of the senior leader interacted with the level of small business unemployment when predicting new venture openings. We initially hypothesized that Republican governors would be associated with higher new venture creation during periods of high small business unemployment, but our results demonstrated the opposite. Indeed, the level of small business unemployment does not appear to meaningfully impact new food and restaurant openings in Democratic states, but it does impact openings in Republican-led states. Republican states have more new venture creation in environments with low-to-moderate small business unemployment, and also have significantly less unemployment overall. Both findings support our arguments that Republican governors are more friendly to businesses. Yet, we found that openings in Republican-led states fall precipitiously during periods of high small business unemployment. Such findings imply that socially focused policies heralded by Democratic governors may prove useful for bolstering businesses during worst-case scenarios (e.g., very high small business unemployment), but may negatively impact economic outcomes in more benign environments (e.g., low to moderate small business unemployment).

In summary, our cumulative findings regarding age, gender, and political orientation of states’ senior leaders suggest that UET’s main premises are relevant for top leaders beyond the C-suite, and that the theory’s central tenets are also applicable and generalizable to state governors. In doing so, we follow the call made by Wang et al. (2021) to use UET as a theoretical framework to better understand how top leaders — regardless of the setting in which they are studied — can influence important outcomes. Additionally, we identify boundary conditions (e.g., small business unemployment) that provide greater context to these outcomes.

4.2 Limitations

Similar to other studies, our research is not without its limitations. For example, while we capture the events associated with the first several months of the pandemic, causality is assumed rather than empirically tested (Antonakis et al., 2010). We recommend that future scholars employ robust longitudinal designs to determine whether or not a causal relationship exists between governor characteristics and new ventures (see Ployhart & Vandenberg, 2010). We also advocate that a finer-grained investigation is integral to recognize the actual mechanisms through which governor party and discretion impact new venture creation. Moreover, we study venture creation within the context of the food and restaurant industry. We encourage scholars to examine the generalizability of our findings and see if they hold for other types of ventures. Finally, political affiliation is just one of many governor attributes. Future research can observe how additional characteristics (e.g., race, gender, or personality traits) impact entrepreneurial or economic outcomes at the state level.

5 Conclusion

Our study utilizes the US governor context to extend the upper echelons literature, and also integrates UET within the entrepreneurship literature. We examine how governor party and discretion can impact new venture openings in terms of a state’s food and restaurant ventures. Although we did not find evidence that political partisanship was directly associated with new food and restaurant openings, we find that states led by governors who have discretion (i.e., when governor and state legislature political parties align) experience a greater number of ventures created. We additionally observe significant interactions between our focal variables and small business unemployment. Specifically, we found that Democratic governors in states with high small business unemployment were associated with a greater number of new venture openings. Regardless of political party, we found that small business unemployment moderates the relationship between governor discretion and small business openings, such that the number of new food and restaurant ventures was higher when unemployment was higher, and governors had discretion. Overall, our work evidences the value of managerial discretion, especially in times of crisis.