Differences Between Latino-Owned Businesses and White-,Black-,or Asian-Owned Businesses: Evidence From Census Microdata

Abstract

Growth of the U.S. Latino population translates into policy interest of how business owner, firm, and local characteristics may be different for Latinos. To explore ethnicity and business ownership, this study merges restricted-access data from 11 million businesses. Multinomial logistic regression estimates how characteristics associate with the probability of the business being Latino-owned relative to White-owned, Black-owned, or Asian-owned. There are differences in the source and amount of start-up funds, gender, and the sector of the business. The differences depend on the group to which Latinos are being compared; for example, manufacturing firms are less likely to be Latino owned than White owned, but more likely to be Latino owned than Black owned. An exception is college education and rurality; Latino owners are consistently less likely to be college educated and more likely to locate in rural areas than the other ethnic minorities. The results should be helpful to groups attempting to improve Latino business outcomes.

Keywords

entrepreneurship small business self-employment Latino ethnic

Interest in minority entrepreneurship and business development programs around the United States stems from desire for social equity and research indicating entrepreneurship can provide a road out of poverty or to employment, given discrimination or other barriers present in the formal labor market (Light, 1979, 1980). To foster minority business, federal, state, and local government programs provide set-asides or extra consideration in the scoring process for loans or grants to minorities, women, and high-poverty areas. Government-funded service providers are asked to take special efforts to reach underserved populations, which are often minorities. Proponents of minority outreach programs emphasize the social mobility that may result for those who face barriers to entry into the formal labor market. Although the emphasis of today’s programs is on groups currently perceived as disadvantaged, numerous historical examples from the United States demonstrate the role of self-employment or business ownership in economic success minority groups in the United States, such as Italians, Japanese, Jews, Koreans, Greeks, and Vietnamese (Bonacich & Modell, 1980; Federman, Harrington, & Krynski, 2006; Light, 1980).

Another consideration in minority entrepreneurship is the growth of immigration to Western countries post-World War II and the growth in the rate of small and medium-size enterprises (Light & Rosenstein, 1995). Small businesses are important to the U.S. economy, accounting for approximately 47.8% of employment in the United States (U.S. Small Business Administration, 2017). Recently in the United States, the increased importance of entrepreneurship is occurring in the context of a rapidly growing Latino population, which some have found to be linked to increased local economic growth (Coates & Gindling, 2013). While Latino immigration to the United States has recently slowed, the age mix and fertility of Latinos means that the population will continue to grow faster than the majority population for years (Gonzalez-Barrrera, 2016; Krogstad, 2014). Furthermore, in the 5 years between 2007 and 2012, the number of U.S. businesses owned by Hispanics grew from 2.3 million to 3.3 million, compared with the total number of all U.S. firms, which increased by only 2% during the same period, from 27.1 million to 27.6 million (Bernstein, 2016).¹ Receipt growth behaved similarly, increasing 35.1% to $473.6 billion for Hispanic-owned businesses compared with the 11.7% growth (to $33.5 trillion) for all firms (Bernstein, 2016). Hence, successful Latino-owned businesses (LOBs) are important to the overall growth of the U.S. economy; thus, the focus of this study is on Latinos in general, not just recent immigrants. Successful incorporation leading to minority income parity may help stave off feelings of discrimination that can contribute to social unrest. Business ownership may be one route to achieving those aims. For example, (Munoz & Spain, 2015) conducted case studies of successful Latino entrepreneurs and the majority disavowed minority status as a barrier in their business.

To explore differences across Latinos and other groups, we use LOB as our reference group, using the relative position in comparison with each ethnicity to investigate a more comprehensive picture of outcomes associated with racial and ethnic inequality. The environment that facilitates entrepreneurship, in general, and LOB, in particular, is dependent on local, business, and personal factors as well as activities of local economic development decision makers and practitioners to promote entrepreneurship and business development. For example, Davila, Mendez, and Mora (2003) found that English-only legislation, a proxy for social stratification attitudes, was associated with lower levels of Hispanic homeownership. Homeownership is often a source of capital for business start-ups, so such attitudes may hamper growth. In an econometric study of immigrant gateway metropolitan areas, Wang (2010) found significant effects based on the type of gateway (recent and historic immigration patterns), implying that contextual effects greatly influence entrepreneurship or self-employment outcomes. Munoz and Spain (2015) documented that few of their interviewees took advantage of (or were aware of) programs aimed at helping minority businesses, which might imply a need for changes in those programs.

Small sample size and lack of microdata on ethnic business ownership meant that past research findings were not easily generalizable. This is especially true for relatively sparsely populated rural areas and specific ethnicities such as Latinos. Furthermore, studies using microdata (Lofstrom & Bates, 2009) while valuable, face limitations in their examination of comprehensive demographic and geographic differences with respect to the success of diverse business owners in the United States.

The efficacy of government and NGO efforts to facilitate LOB are dependent on decision makers and practitioners understanding how LOB differ from businesses owned by other ethnicities. Understanding characteristics of LOB may inform appropriate best practice guidelines and training for economic development practitioners that will lead to supporting LOB activities more efficiently and effectively. Comparative studies on businesses in the United States by race, ethnicity, gender, and regional factors gain significant value when put into context. For example, it is not necessarily surprising that LOB are more likely to enter low-barrier industries (Bates, Lofstrom, & Servon, 2011) since Latinos have lower education levels, on average (Pew Research Institute, 2006), and high-barrier industries typically require higher education levels. This study, however, examines the question of whether LOB are more associated with low-barrier industries even while controlling for education level. Such a finding may provide evidence about a larger social or economic problem leading to Latino self-employment in low-barrier industries than is possible with the descriptive statistics often used in prior work. Similarly, the finding that a proportion of LOB in low-barrier industries is relatively high (Puryear et al., 2008; Robles & Cordero-Guzman, 2007) does not become particularly important unless this concentration of LOB into low-barrier industries remains after controlling for numerous other factors.

Most studies seeking to demonstrate context rely on reports by government agencies, such as the U.S. Census Bureau, the U.S. Small Business Administration, or the Minority Business Development Agency. These reports often compare static means and counts and highlight differences by race or other characteristics (e.g., Lichtenstein, 2007; U.S. Census Bureau, 2010, 2016b). A limitation of static means comparisons is that they do not control for other firm and business owner characteristics when examining relative associations. Failure to control for these characteristics in ethnic business comparisons may drastically change results or interpretations. For example, previous reports indicate that LOB are more associated with construction, wholesale trade, and retail trade sectors (U.S. Census Bureau, 2010), when not controlling for business owner, business, and local factors. Much research seeks to answer why this (and other) static ethnic business differences exist, but this approach may lead to the development of the wrong questions (i.e., why are LOB more associated with construction, wholesale trade, and retail trade sectors?). Our data allow for many controls and the results show a more complicated picture, allowing for future work to ask more detailed research questions (i.e., why are minority-owned businesses less associated with the construction industry than White-owned businesses, even when controlling for owner, business, and location characteristics?). Such questions allow future research to examine more systemic (though also specific) disparities in business ownership that remain after controlling for compounding factors, rather than disparities in one aspect of business ownership (e.g., finance). We elaborate on the complicated picture of ethnic business ownership disparities below, but clearly drawing research questions or policy prescriptions from previous descriptive work alone may be misguided, as the associations may disappear or reverse once we control for other factors.

This study extends past research by examining the relative associations of business development factors by comparing other owner ethnicities to Latinos. We draw on the literature to develop testable hypotheses about LOB. We then describe our data, which addresses the limitations of prior works by way of Federal Statistical Research Data Center (FSRDC) access to over 11 million observations on businesses in the United States, allowing us to tease out subtle differences that might be invisible with other methods. We use the FSRDC data to identify which business owner, business, and location factors are statistically associated with LOB relative to businesses owned by Asians, Blacks, and Whites. The analysis uses a multinomial logit to explore relative associations, ceteris paribus. Hence, this investigation allows the application of past findings to develop a deeper understanding of which factors are most associated with LOB compared with other ethnicities, while controlling for other personal, business, and location variables. As such, and with the size and comprehensive nature of the data set and model, the results contribute both toward the practical and scientific understanding of how personal, business, and local factors relate with ethnic business ownership.

The model results show that LOBs are different from those owned by other ethnicities in several dimensions. While many of the differences vary depending on the comparison group, Latino business owners are less likely to be college educated than all other types of business owners. This may imply that in the future many LOB could peak at a lower level than might otherwise be expected, if college training is important for actualizing business potential. Additionally, ceteris paribus, our results show that LOB are more likely to locate in rural areas. We elaborate on this finding and possible implications in the conclusions.

Literature Review

Many of the determinants of business ownership probably transcend race or ethnicity, such as personality traits that may make a person desire more flexible (though often longer) work hours, involuntary changes in employment status (e.g., health status, plant closure, or firm restructuring), or perceived higher monetary gains through ownership. For decades, however, researchers have investigated why minorities have different business formation, survival, and growth rates (Hisrich & Brush, 1986). Although the effect of business and owner characteristics on small businesses is a frequent topic of study, few authors investigate business and business-owner characteristics associated with LOB in the United States. Fairlie and Robb (Fairlie, 2005; Fairlie & Robb, 2007, 2008; Robb & Fairlie, 2007) conducted studies on gender and minority differences in business ownership using FSRDC access, but those studies have not focused on LOB. Shinnar, Cardon, Eisenman, Zuiker, and Lee (2009) compare the motivations and management styles of immigrant and U.S.-born, Mexican-owned businesses using the 2005 National Minority Business Owner Survey (n = 156). Their findings indicated that about one third of Mexican-owned businesses perceived the labor market as discriminatory, and that this discrimination prompted them to become self-employed. Thus, these findings provide some support for disadvantage theory (Light, 1979), which argues minorities are often pushed (rather than pulled) into entrepreneurship because of chronic unemployment, low wages, and labor market discrimination.

Numerous personal, business, and location variables can serve as direct measures or proxies for mechanisms that theory predicts to significantly affect business formation, survival, and growth. Because ethnicity correlates differently with some of these mechanisms, outcome differences may arise. Some examples of mechanisms related to Latino ethnicity that may affect business outcomes include being multilingual, which improves the brain’s executive function (Diamond, 2010); being multinational, which may imply better connections to home country (supply chain or customer base) or exposure to different business models that can occupy new niches in the United States (Rauch, 2001); or ethnicity-based U.S. networking effects, which may allow for higher nonmonetary labor benefits, demand, and business expertise to be shared among members (Greve & Salaff, 2003). On the negative side, rather than increase demand, coethnic clientele may actually impose a financial penalty and LOB may be disproportionately likely to have coethnic clientele (Shinnar, Aguilera, & Lyons, 2011). Coethnic labor markets may also impose a penalty, with successful Latino business owners when interviewed by Munoz and Spain (2015) indicating some unhappy experiences hiring family members. Pressure to hire unqualified staff from within the family may be more intense in situations where the extended family relationships are strong and could result in smaller business size, producing more businesses per capita within the group. Conversely, Shinnar, Cho, and Rogoff (2013) found financial performance improved for Mexican-owned businesses in the National Minority Business Owner Survey (n = 513) as family involvement increased.

While the preceding paragraph provides insights into the possible mechanisms giving rise to differences, these effects are difficult to measure directly; thus, this study examines how business ownership factors relate to various ethnicities and thereby the extent to which racial and ethnic discrepancies exist in self-employment, even when controlling for other business owner, firm, and location factors. Our empirical approach may inform questions asked in future qualitative works. Based on the aforementioned backdrop, one could summarize the decision to start a business as a function of the umbrella concepts of opportunity sets, opportunity costs, and risk tolerance.

In this context, we consider opportunity sets to be the range of options available for work (including different labor market opportunities and different entrepreneurial opportunities), opportunity costs as the loss of potential gain from other work alternatives when one work alternative is chosen, and risk tolerance as the extent to which an individual discounts expected income based on variability in that income. For example, to the extent that they are migrants, Latinos are probably risk tolerant compared with the general population. Supportive family networks could also help LOB tolerate risk, but on the other hand, opportunity costs may be smaller due to language or educational issues, labor market discrimination, and so on. Similarly, if a LOB cannot get a bank loan, this will limit its opportunity set. This limited opportunity set may, for example, limit the ability to enter a high-barrier industry (and perhaps reduce the probability that the LOB is an “opportunity entrepreneur”).

This section uses prior research into the interaction of ethnicity and various business owner, business, and location factors on the formation, survival, and growth of businesses to develop hypotheses that are testable within our data set. The hypotheses inform our choice of models and stem from previous findings that did not have access to the extensive control variables available in our rich data set. Furthermore, as noted in the introduction, our work transcends the descriptive statistics available to the public from government fact sheets that provide statistics on minority business ownership (e.g., U.S. Census Bureau, 2010). While these fact sheets are a useful starting point for minority business research, without controlling for compounding factors related to business owner characteristics, fact sheets alone may miss important relationships. We thus list the following four major findings related to LOB as hypotheses, drawn from government fact sheets and other LOB research that used less detailed data. We then test these hypotheses risk tolerance to examine the extent to which these findings persist while controlling for compounding factors of business owner, business, and location characteristics. The robustness to these previous findings (and current hypotheses) to the extensive controls may be indicative of various unexplored explanations for ethnic disparities in self-employment and open new avenues for research. We draw from the numerous FSRDC variables available to look more specifically at how LOB opportunity costs, opportunity set, and risk preferences may influence business outcomes.

Opportunity Set Hypothesis 1 (OSH1): Latino Business Owners Are More Likely Than Business Owners of Other Ethnicities to Finance Their Businesses With Personal Savings

LOB opportunity sets may be limited by an inability to get a bank loan, but also increased by higher family and community ties, allowing for familial borrowing. Several studies find that Latinos primarily finance their businesses with personal savings and informal loans from friends or family and money lenders (Granier, 2006; Haynes, Onochie, & Lee, 2008; Raijman & Tienda, 2000). Similarly, data from the 2005 National Minority Business Owner Survey suggest that in comparison with Korean Americans, Mexican American business owners borrow more from family, friends, suppliers, and credit cards and have a lower proportion of bank loan debt (Puryear et al., 2008). While the research shows this to be the case descriptively, it is important to understand if this ethnic variation remains even when controlling for other factors that are likely to have an impact on capital source, such as industry and education. Variation in a financial source that is robust to extensive controls for confounding factors (e.g., education) would imply an unobserved cause of ethnic disparities in entrepreneurship, opening the door to future research into other unobserved causes more difficult to observe and quantify, such as ethnic networks, credit history development, and discrimination in lending.

Opportunity Set Hypothesis 2 (OSH2): Latino Business Owners Are Less Likely Than Businesses of Other Ethnicities to Finance Their Businesses With Bank Loans

While this might seem to be a natural outcome of the opportunity set that LOB may disproportionately face noted in Hypothesis 1, financing might also come from other sources such as venture capital, credit card debt, government sources, or family and friends; thus, empirical verification of Hypothesis 1 does not imply Hypothesis 2. Studies on the source of funding Latinos use for start-ups conclude they use fewer formal funds compared with Whites with lower income on which to base the loan-limiting opportunity sets (Blanchard, Zhao, & Yinger, 2008; Cavalluzzo & Wolken, 2005; Granier, 2006; Puryear et al., 2008). These studies support the findings of Blanchard et al. (2008), which show opportunity sets also limited by discrimination that Latinos are quoted higher interest rates on bank loans, which may discourage LOB from seeking formal funds for the businesses. A finding here that otherwise similar LOB remain less likely to use bank loans than other ethnicities would provide support to the finding that, once controlling for risk factors, owner race/ethnicity is important (Bates & Robb, 2016), which may indicate discrimination based on if ethnicity influences the ability of LOB to get a bank loan and thereby succeed.

Opportunity Cost Hypothesis (OCH): Latino Business Owners Are Less Likely to Have a College Degree

If a business owner does not have a college degree, that indicates (all else equal) a lower opportunity cost of starting a business. Past research also indicates an important role of education in the formation, survival, and growth of business ownership (Lechmann & Schnabel, 2014; Millán, Congregado, & Román, 2012; Oberschachtsiek, 2012). As a result, past research has often ascribed the disparate success of races in business ownership to education levels. Fairlie and Robb (2008), for example, pointed out that about half of Asian American business owners have college degrees, one third of White business owners have college degrees, and only one quarter of Black business owners have the same level of education. Some of these results, however, do not control for industry or other personal factors. This study includes ethnicity, which others exploring the relationship between education and business ownership have not considered, as well as controls for other factors such as sector. If disparities in education remain, even when controlling for compounding factors such as industry, then our results may indicate that education does indeed provide an important role itself in disparate success of races in business ownership.

Risk Tolerance Hypothesis (RTH): Latino Businesses Are Overrepresented in Low-Barrier Industries Such as the Services, Construction, Wholesale Trade, and Retail Trade Sectors, and Underrepresented in High-Barrier Industries Such as Manufacturing, Professional Services, and Finance or Insurance

Several authors have examined why Latinos tend to concentrate in low-barrier industries perceived as relatively vulnerable, such as the services (Puryear et al., 2008; Robles & Cordero-Guzman, 2007), construction, wholesale trade, and retail trade sectors (U.S. Census Bureau, 2010). It may be related to higher risk tolerance among LOB or it may be related to their opportunity set, which limits their ability to borrow more for a higher barrier industry. High-barrier industries are those requiring advanced educational credentials or large amounts of start-up capital. Industries such as professional services, finance, or insurance are examples of high-barrier industries, while low-barrier industries include some food services and construction. Bates et al. (2011) have found that businesses within the low-technology sectors or low-barrier industries face a greater likelihood of going out of business compared with businesses within the high-technology sectors or the high-barrier industries. One plausible theoretical mechanism for this finding is the level of innovation. Despite difficulties in measuring innovation directly, recent work finds empirical associations between innovation and firm growth, especially innovating small firms (Audretsch, Coad, & Segarra, 2014). Past studies also examine the impact of legal form on growth, with incorporation or limited liability forms having a positive impact on firm growth (Almus, 2002; Davidsson, Kirchhoff, Hatemi-J, & Gustavsson, 2002). Innovation may be associated with sectors that are less accessible to Latinos due to missing technical expertise, lack of financial resources, or discrimination.

In addition to variables aimed at testing the hypotheses, the model includes additional control variables such as owner gender, age, and location characteristics and to provide insights not previously found in the literature. For example, whether lower rates of business ownership among women results from opportunity set or risk tolerance is subject to some debate. Studies into Latina business owners have long indicated the need for larger data sets such as the one in use here (Shim & Eastlick, 1998). Previous studies have found that gender has an effect both on business formation (Coleman & Robb, 2009; Fairlie & Robb, 2009) and survival (Georgellis, Sessions, & Tsitsianis, 2007; Haapanen & Tervo, 2009). There is limited research into the validity of these results with respect to Latinas. This study addresses this limitation of previous research.

Probably due to data limitations, few studies have attempted to examine the impact of local and geographic characteristics. Despite the limited prior studies, this study includes local and geographic controls for those factors found to be significant in the past. For example, Millán, Congregado, Román, van Praag, and van Stel (2014) have found that the share of a population that is highly educated has a positive impact on business survival due to both demand and supply-side factors. Fertala (2007) has found that population density has a positive relationship with survival rates. Lo and Teixeira (2015) found that barriers are different in communities with small immigrant populations due to lack of the “institutionally complete” enclaves found in larger metropolitan areas.

Another local factor examined in numerous studies is the impact of the unemployment rate. The two competing forces here are the prosperity-pull hypothesis and the recession-push hypothesis, which occur as a result of opportunity entrepreneurship and necessity entrepreneurship, respectively (Millán et al., 2012). It is also possible that, given both effects occur simultaneously, neither outweighs the other. Past results on the impact of the unemployment rate on self-employment have been mixed with support for the recession-push effect (Georgellis et al., 2007; Lin, Picot, & Compton, 2000; van Praag, 2003) and the prosperity-pull effect (Carrasco, 1999; Fertala, 2007; Haapanen & Tervo, 2009; Taylor, 1999).

Method

To assess the hypotheses, this study frames the estimation in terms of the probability of a business being Latino-owned as opposed to a business being Asian-owned, Black-owned, or White-owned. Specifically, the estimation uses a multinomial logit model to estimate how various factors associate with LOB, Asian businesses, and businesses owned by Blacks and by Whites. The racial/ethnic groups in the data are categorical.² The model assesses how these different owner groups relate to personal traits and demographic variables, business and industry characteristics, and location factors. The multinomial logit model format is as follows (Greene, 2012):

\Pr (y = k) = \frac{\exp (\propto + β^{'} X_{i j})}{\sum_{j = 1}^{3} \exp (\propto + β^{'} X_{i j})} k = 1, 2, 3

The categorical dependent variable takes on four levels (LOB, Asian-owned businesses, Black-owned businesses, and White-owned businesses). The vector X denotes the set of characteristics associated with the businesses. Often the omitted group in similar analyses is White individuals and the relative position of Latino individuals must be inferred. As Latino business owners are the focus of this research, they are the base group in this analysis and hence all ethnic associations are relative to Latino business owners. The explanatory variables that represent personal traits refer to the traits of the owner or majority shareholder of the firm, including gender, education, and age.³

Beyond the state fixed effects, which may help control for the effect of unobserved state-level policy, agglomeration, and other variables, the model uses standard county-level control variables to account for local conditions: the percentage of people who have various education levels and age groups as a proxy for local human capital, rurality, and natural amenities. Local conditions may influence the extent of local push/pull factors, which could bias the results. For example, high local unemployment rates may “push” individuals into self-employment (i.e., lowering opportunity costs) in relatively low-barrier industries because they do not have high levels of resources required for high-barrier industries (i.e., because of their opportunity set). If this “pushing” varies by ethnicity or unemployment correlates with populations of different ethnicities, then failure to control for local unemployment rates may bias the likelihood of involvement in various industries by different ethnicities. It may also bias various business owners’ characteristics associated with high- or low-barrier industries (source of capital, amount of capital, education, experience, etc.) and thereby risk tolerance. The model also includes controls for the percentage of the population that is Black, Asian, and Hispanic because county-level demographic composition is important to control for as it reflects the relative structural inequalities of education, race or ethnicity, age, and single-family status, which significantly impact local well-being (Lobao, Zhou, Partridge, & Betz, 2016; Voss, Long, Hammer, & Friedman, 2006), and may influence the extent to which local factors influence push and pull factors. Furthermore, the extent of the local coethnic involvement in the business (which influences a business owner’s opportunity set) has been shown to influence business success (Shinnar et al., 2013), thus controlling for local populations of various ethnicities may act as a proxy for coethnic involvement while also directly controlling for the potential for coethnic demand. Similarly, we include the unemployment rate and the labor force participation rate as a measure of the aggregate strength of the local labor market, which may also indicate the extent of push and pull factors (Lobao et al., 2016; Partridge & Rickman, 2006). The model includes indicators for the rurality of a county to account for how urban–rural location influences agglomeration economies, which has also been shown to be important (Partridge & Rickman, 2008). Although many of the aforementioned county-level variables are included as controls and their coefficient is left to the appendix, others are included to show how local conditions relate to each ethnicity. Table 1 lists all the included business owner, business, and location variables.

Table 1.

Variables List by Type.

Business owner variables	Business variables	Location (county-level) variables
• Gender • Education • Age • Weekly hours • Primary function • Primary income indicator	• Start-up capital amount • Start-up capital source • Expansion capital source • 2-digit NAICS sector • Customer types • Multiple establishment indicator • Business organization type • Firm payroll and employment • Establishment payroll and employment • Same industry employment in county • Same industry employment in state	• Percentage of population with various education levels • Black, White, Asian, and Hispanic population • Black, White, Asian, and Hispanic population squared • Population of various ages • Per capita income • Mean household size • Percentage of households headed by single female • Unemployment rate • Rurality • Natural amenities • State indicators • Year indicators

Note. NAICS = North American Industry Classification System.

Data

This study merges three restricted access Census Bureau data sets by individual firm and establishment level to investigate the factors associated with LOB. The three databases are the Integrated Longitudinal Business Database (ILBD, $n =$ approximately 18 million establishments annually), the Longitudinal Business Database (LBD, $n =$ approximately 8 million establishments annually), and the Survey of Business Owners (SBO, $n =$ approximately 2.3 million establishments in 2007). Factors associated with LOB ownership include characteristics of business and industry, characteristics of business owner, and characteristics of location. The standard economic model predicts that these factors are associated with the firm’s production process. Detailed business owner data are available from the SBO: race, ethnicity, gender, education level, and access to capital. Source and amount of start-up capital is also available from the SBO, while other business characteristics such as firm age, size, and industry can be constructed using information available from the merged ILBD/LBD. Using multiple years of the ILBD/LBD allows for additional variation in the location data than is possible in a cross section created from a single SBO.

The LBD has firm and establishment identifiers, making the linking of the LBD with the SBO feasible. The ILBD consists of administrative records for all nonemployer business units and is needed because the SBO has information about firms that have no paid employees and the LBD has information only about establishments that have paid employees. The final merge uses the 2002 and 2007 SBO data and the 2002-2007 ILBD/LBD data. The final sample is composed of about 4% (~473,900 establishments) LOB, 5% (~592,400 establishments) Asian-owned businesses, 3% (~355,400 establishments) Black-owned businesses, and 88% (~10,425,800 establishments) White-owned businesses. The majority of the observations are of White-owned businesses, but at n = ~11,847,500,⁴ sample size is not a concern. The FSRDC did not approve publication of the means and standard deviations of the firm-level regression variables.

Although establishment location is available from these merged data sets, location-specific characteristics such as agglomeration, racial makeup of location, amenities, and market size are not available in these data and are drawn from the publicly available data of the Bureau of the Census, the Bureau of Economic Analysis, and the U.S. Department of Agriculture (USDA). Specifically, these regional factors include demographic information of the county in which each business is located including population shares of various races/ethnicities, the age distribution, local unemployment rates, natural amenities, and the rurality of U.S. counties. The counties’ natural amenities follow the rankings data developed by the USDA Economic Research Service (ERS) and rurality follows the contemporaneous USDA Urban–Rural Continuum Codes.

Results and Discussion

The results of the multinomial logit analysis of the probability of a business being Latino-owned as opposed to a business being Asian-owned, Black-owned, or White-owned is given in Tables 2.1 to 2.5.⁵ For ease of interpretation, Tables 2.1 to 2.5 display the results of the multinomial logit estimations as odds ratios (or relative risk ratios; i.e., as exponentiated multinomial logit coefficients). For the sake of completeness, Table A1 in the appendix contains the odds ratios on control variables in the estimation that are not included in Tables 2.1 to 2.5. Multicollinearity is not a concern given that most of the location variables remain significant at the 99% confidence level.⁶

Table 2.1.

Multinomial Logit Results (Relative Odds Ratio): Business Owner Variables.

Business owner variables	Asian American	Black American	White American
Female	0.92*** (0.02)	1.24*** (0.02)	0.76*** (0.01)
Education level (base: <high school)
High school	1.66*** (0.05)	2.26*** (0.08)	3.31*** (0.07)
Some college	XX^a (XX)	3.69*** (0.14)	4.77*** (0.11)
Trade school	1.29*** (0.06)	3.50*** (0.16)	3.76*** (0.11)
Associate’s degree	2.15*** (0.09)	XX (XX)	4.64*** (0.14)
Bachelor’s degree	4.67*** (0.15)	3.78*** (0.15)	8.06*** (0.19)
Graduate degree	5.91*** (0.21)	3.34*** (0.14)	7.55*** (0.20)
Owner age (base: <25)
25 to 34	XX (XX)	XX (XX)	0.11*** (0.00)
35 to 44	0.26*** (0.01)	0.37*** (0.02)	0.13*** (0.00)
45 to 54	0.29*** (0.01)	0.49*** (0.03)	0.19*** (0.01)
55 to 64	0.30*** (0.01)	0.62*** (0.03)	0.27*** (0.01)
≥65	0.24*** (0.01)	0.72*** (0.04)	0.37*** (0.01)
Weekly hours (base: <20)
20 to 39	0.78*** (0.02)	XX (XX)	0.78*** (0.02)
40	0.98 (0.03)	0.81*** (0.03)	0.67*** (0.02)
41 to 59	0.88*** (0.03)	0.89*** (0.03)	0.78*** (0.02)
≥60	1.04 (0.03)	1.27*** (0.04)	0.72*** (0.02)
Business is primary income	1.17*** (0.03)	0.59*** (0.01)	0.96** (0.02)
Primary function (multiple response)
Finance	1.05*** (0.02)	0.92*** (0.02)	1.48*** (0.02)
Provide good/service	0.79*** (0.01)	XX (XX)	0.94*** (0.01)
Manage	1.10*** (0.02)	1.17*** (0.03)	1.12*** (0.02)

Note. Robust standard errors are in parentheses.

Note that “XX” may appear rather than a coefficient and standard error in the tables using limited access data if that coefficient and standard error have been suppressed on release of the results because the Census Bureau finds that coefficient and standard error to risk an inappropriate disclosure of individual response.

p < .1. **p < .05. ***p < .01.

Table 2.2.

Multinomial Logit Results (Relative Odds Ratio): Business Finance Variables.

Business variables	Asian American	Black American	White American
Start-up capital (base: <$5,000)
$5,000-$10,000	0.79*** (0.03)	0.94* (0.03)	0.76*** (0.02)
$10,000-$25,000	1.15*** (0.03)	0.82*** (0.03)	0.82*** (0.02)
$25,000-$50,000	1.32*** (0.05)	0.79*** (0.04)	0.87*** (0.02)
$50,000-$100,000	XX^a (XX)	0.73*** (0.04)	0.96 (0.03)
$100,000-$250,000	1.83*** (0.07)	0.75*** (0.04)	1.07** (0.04)
$250,000-$1 million	1.85*** (0.10)	0.74*** (0.05)	1.23*** (0.06)
>$1 million	2.13*** (0.18)	XX (XX)	1.45*** (0.11)
Start-up capital source (base: other)
Savings	1.08*** (0.02)	0.93*** (0.02)	1.00 (0.02)
Credit	0.91*** (0.03)	1.11*** (0.03)	0.85*** (0.02)
Government loan	XX (XX)	1.32*** (0.09)	0.89** (0.05)
Bank loan	1.29*** (0.04)	0.89*** (0.03)	1.19*** (0.03)
Venture capital	1.33*** (0.10)	1.06 (0.10)	1.32*** (0.08)
None needed	0.82*** (0.02)	0.88*** (0.03)	0.92*** (0.02)
Expansion capital source (base: other)
Savings	1.11*** (0.02)	1.06** (0.03)	0.98 (0.02)
Credit	0.74*** (0.02)	0.94** (0.03)	1.09*** (0.02)
Government loan	0.81*** (0.06)	1.14 (0.10)	0.78*** (0.05)
Bank loan	0.89*** (0.03)	0.73*** (0.03)	1.12*** (0.02)
Venture capital	0.94 (0.10)	XX (XX)	0.92 (0.08)
None needed	0.98 (0.02)	0.83*** (0.02)	1.20*** (0.02)

Note. Robust standard errors are in parentheses.

p < .1. **p < .05. ***p < .01.

Table 2.3.

Multinomial Logit Results (Relative Odds Ratio): Business Sector Variables.

Business sector variables:	Asian American	Black American	White American
NAICS industry (base: other)
11: Agriculture, Forestry, Fishing, Hunting	0.66* (0.14)	0.59** (0.12)	1.87*** (0.30)
21: Mining, Quarrying, Oil, and Gas Extraction	0.50*** (0.13)	0.56** (0.16)	3.79*** (0.71)
22: Utilities	0.98 (0.34)	1.27 (0.41)	2.08*** (0.45)
23: Construction	0.37*** (0.07)	0.58*** (0.10)	1.43** (0.20)
31-33: Manufacturing	1.08 (0.19)	0.42*** (0.08)	1.57*** (0.22)
42: Wholesale Trade	1.96*** (0.35)	0.40*** (0.07)	1.64*** (0.23)
44-45: Retail Trade	1.98*** (0.34)	0.62*** (0.11)	1.37** (0.19)
48-49: Transportation and Warehousing	0.74* (0.13)	XX^a (XX)	0.86 (0.12)
51: Information	0.92 (0.16)	1.17 (0.22)	1.95*** (0.28)
52: Finance and Insurance	0.80 (0.14)	0.89 (0.16)	1.76*** (0.25)
53: Real Estate, Rental, Leasing	1.25 (0.22)	0.73* (0.13)	XX (XX)
54: Professional, Scientific, and Technical Services	0.98 (0.17)	0.85 (0.15)	1.66*** (0.23)
55: Management of Companies and Enterprises	1.65** (0.40)	1.14 (0.30)	1.67** (0.33)
56: Admin. Support, Waste Management and Remediation Services	0.47*** (0.08)	0.83 (0.15)	0.88 (0.12)
61: Educational Services	0.97 (0.18)	1.00 (0.19)	1.37** (0.20)
62: Health Care and Social Assistance	1.09 (0.19)	1.41* (0.26)	0.96 (0.14)
71: Arts, Entertainment, & Recreation	0.62*** (0.11)	1.46** (0.26)	2.08*** (0.30)
72: Accommodation & Food Services	3.07*** (0.54)	0.44*** (0.08)	0.71** (0.10)
81: Other Services	XX (XX)	1.01 (0.18)	1.14 (0.16)

Note. NAICS = North American Industry Classification System. Robust standard errors are in parentheses.

p < .1. **p < .05. ***p < .01.

Table 2.4.

Multinomial Logit Results (Relative Odds Ratio): Business and Customer Variables.

Business and customer variables	Asian American	Black American	White American
Multiple establishments	0.68*** (0.03)	1.12** (0.05)	2.00*** (0.07)
>10% Sale (multiple response)
Federal	1.09* (0.05)	1.07 (0.06)	0.66*** (0.03)
State or local government	0.67*** (0.03)	1.40*** (0.05)	0.89*** (0.02)
Other businesses	0.90*** (0.01)	1.11*** (0.01)	0.89*** (0.01)
Individuals	1.02* (0.01)	0.90*** (0.01)	1.05*** (0.01)

Note. Robust standard errors are in parentheses.

p < .1. **p < .05. ***p < .01.

Table 2.5.

Multinomial Logit Results (Relative Odds Ratio): County Variables.

Location (county-level) variables	Asian American	Black American	White American
% In county with (base: <high school)
High school	1.07*** (0.00)	1.10*** (0.01)	1.10*** (0.00)
Some college	1.04*** (0.00)	1.07*** (0.01)	1.02*** (0.00)
Associate’s degree	1.05*** (0.01)	1.05*** (0.01)	1.06*** (0.01)
Bachelor’s degree	1.04*** (0.00)	1.07*** (0.01)	1.05*** (0.00)
Graduate degree	1.01** (0.01)	1.01** (0.01)	1.00 (0.00)
Per capita income	1.00*** (0.00)	1.00*** (0.00)	1.00*** (0.00)
Mean household size	0.60*** (0.04)	0.66*** (0.05)	0.22*** (0.01)
Unemployment rate	1.58*** (0.08)	2.03 (1.91)	1.62 (1.03)
Labor force participation rate	1.53* (0.35)	0.68 (0.17)	3.39*** (0.57)
% Households headed by single female	1.02*** (0.00)	1.11*** (0.01)	0.96*** (0.00)
Natural amenities	0.84*** (0.01)	0.75*** (0.01)	0.86*** (0.01)
Rural–urban code (base: 1)
2: Metro area, 0.25 to 1 million population	1.01 (0.03)	1.08** (0.04)	1.13*** (0.03)
3: Metro area, <250,000	0.91** (0.04)	1.05 (0.05)	1.28*** (0.04)
4: Adjacent to metro, ≥20,000	0.89** (0.05)	1.11 (0.07)	1.36*** (0.06)
5: Not adjacent to metro, ≥20,000	0.89 (0.07)	1.04 (0.10)	1.58*** (0.10)
6: Adjacent to metro, 2,500-19,999	0.89* (0.06)	1.22*** (0.08)	1.76*** (0.09)
7: Not adjacent to metro, 2,500-19,999	0.75*** (0.06)	0.83** (0.07)	1.54*** (0.09)
8: Adjacent to a metro, <2,500	0.74* (0.13)	1.59*** (0.25)	1.94*** (0.25)
9: Not adjacent to a metro, <2,500	0.37*** (0.06)	1.11 (0.17)	1.92*** (0.21)

Note. Robust standard errors are in parentheses.

p < .1. **p < .05. ***p < .01.

Though some of the coefficients on the control variables are not statistically significant, most are statistically significant, but equal to 1.00. Equality to 1.00 implies that (ceteris paribus) the represented ethnic group is near equal to Hispanic Americans in its relationship with the control variable in question. The employment and payroll variables relate similarly.⁷ Or, for example, the indicator variable for whether a business is female owned implies that female-owned businesses are 0.92 times as likely to be Asian owned, 1.24 times more likely to be Black owned, and 0.76 times more likely to be White owned. XX designates values retained by FSRDC to maintain business owner confidentiality. For readability, this section divides the coefficients into different tables and discussions despite all coefficients resulting from the same multinomial regression.

The aforementioned controls also produce some patterns of interest beyond the literature-based hypotheses. For example, when controlling for our included variables, Latino business owners are more likely to be associated with any of the included age groups. Thus, Hispanic American business owners are less associated with the (omitted) less than 25-year-old category. This may run counter to expectations given that 2015 estimates based on the 2010 Census show Hispanic American median age to be more than 12 years younger than the U.S. general population (U.S. Census Bureau, 2016a). Alternatively, this may reflect the relative financial barriers facing younger LOBs, which depend more on friends, family, and personal savings—factors that are unobserved and likely correlated with age. Similarly, other ethnicities may be more likely to have individuals with financial means to cosign a loan.

Hispanic American business owners’ relative association with weekly hours worked is more complex. Hispanic American business owners are more likely to work 20 to 59 hours per week at their business; they are also more likely than White Americans to work 60 or more hours per week, but less likely than Asian and Black American business owners to work 60 or more hours per week. Hispanic American business owners are more likely to be associated with having the business as their primary source of income than both Black and White Americans, but less likely than Asian Americans.

The 2007 SBO divides start-up capital into value ranges. The omitted category is less than $5,000. Hispanic Americans are more likely to be associated with the $5,000 to $10,000 range than the other ethnicities. Hispanic Americans are more likely to be associated with the ranges from $10,000 to $100,000 than White and Black Americans, but less likely than Asian Americans (although the $50,000 to $100,000 range is suppressed for Asian Americans). Hispanic Americans are less associated with all the ranges over $100,000 than Asian and White Americans, but more associated with those (nonsuppressed) ranges than Black Americans.

Table 2.3 contains the 2-digit North American Industry Classification System (NAICS) industry indicator variables. The only industry where LOB have stronger association than White-owned businesses is NAICS 72 (accommodation and food service). LOB are also more associated than Black-owned businesses with NAICS 72, but Asian-owned businesses are over three times more likely to be associated with NAICS 72. Otherwise, Table 2.3 paints a complex picture of how various ethnicities are associated with different industries. LOB are more associated with NAICS 11 (agriculture, forestry, fishing, and hunting), NAICS 21 (mining, quarrying, oil and gas extraction), and NAICS 23 (construction) than Asian- and Black-owned businesses. Conversely, only Black-owned businesses are less associated with manufacturing, wholesale trade, and retail trade.

Table 2.4 indicates that only Asian-owned businesses are less associated with owning firms with multiple establishments. The table also indicates that White-owned businesses are less likely than LOB and Black-owned businesses to sell goods or services to the federal or state/local government. This may be indicative of successful implementation of minority business purchase requirements by some government agencies.

Except for the state fixed effect and state industrial employment, the location variables are at the county level. The percentage of the population with various education levels shows that Hispanic American business owners are less associated with every included county education-level variable. Hence, Hispanic American business owners tend to be in counties with a relatively high percentage of individuals having less than a high school degree. The coefficients on other county-level variables reveal that LOB are more associated with areas in which there are high amenities than any other ethnicity. The USDA ERS constructs the amenities scale by combining six measures of climate, topography, and water area. The ERS natural amenities index favors warm regions, which have relatively high populations of Latinos (and thereby LOB), so this association may simply be the result of historical immigration trends, despite the controls for states and ethnic/racial population sizes. Furthermore, the coefficients on the rural–urban continuum codes indicate that LOB are more associated with rural counties than Asian American business owners, but less associated with rural counties than White Americans.

Table 2.5 shows that Asian American business owners are far more likely than Hispanic American business owners to be in counties with higher unemployment rates, while Whites and Blacks are not statistically different from Hispanic Americans in their association with the unemployment rate. This finding could imply that the somewhat stereotypical networks of Asians who provide convenience store services in inner city neighborhoods (Derickson & Ross, 2015) or who manage rural hotels (Dhingra, 2016) are expressions of a more widespread ability of these groups, possibly by way of family-franchise methods, to find opportunity in low-income regions.

Table 3 summarizes the results in Tables 2.1 to 2.6 with respect to the hypotheses presented earlier. Most prior studies focused on the differences between Hispanics and the majority population. Looking down the last column of Table 3, most of the hypotheses are supported with respect to Whites, but a more complex pattern emerges when the comparison is with other minority groups.

Table 2.6.

Multinomial Logit Results: Constant, Fixed Effects, and Goodness of Fit.

	Asian American	Black American	White American
Constant	0.10*** (0.05)	0.01*** (0.00)	100.23*** (36.10)
State FE	YES	YES	YES
Year FE	YES	YES	YES

Note. FE = fixed effects. Pseudo R² = .224; Log pseudo-likelihood = −28890434; n = ~11,847,500. Robust standard errors are in parentheses.

p < .1. **p < .05. ***p < .01.

Table 3.

Summary of Results With Respect Hypotheses: Probability of Other Group Relative to Latino-Owned Business Characteristics.

	Hypothesized probability relative to LOB	Found probability
	Hypothesized probability relative to LOB	Asians	Blacks	Whites
OSH1: Personal finance start-up	<	>	<	ns
OSH2: Bank loan start-up	>	>	<	>
OCH: College degree	>	>	>	>
RTH: Low-barrier sectors
Construction	<	<	<	>
Retail	<	>	<	>
Accomadation and Food Service	<	>	<	<
High-barrier sectors
Manufacturing	>	ns	<	>
Professional, Scientific, and Technical Services	>	ns	ns	>
Management of Companies	>	>	ns	>

Note. OSH1 = Opportunity Set Hypothesis 1; OSH2 = Opportunity Set Hypothesis 2; OCH = Opportunity Cost Hypothesis; RTH = Risk Tolerance Hypothesis; LOB = Latino-owned business. Symbols: > (<) means the target group is more (less) probable than Hispanics (statistically significant and not 1.00), while ns means the variable was not significant.

Conclusions

Our findings provide a more complete picture of the effect and importance of each business owner, business, and location factor under consideration than was previously available. For example, data from the 2005 National Minority Business Owners Survey suggest that, in comparison with Korean Americans, Mexican American business owners borrow more from family, friends, suppliers, and credit cards and have a lower proportion of bank loan debt (Haynes et al., 2008). This study’s results imply this relative association between Korean and Mexican Americans is more general, and that Asian American business owners are more likely than Latin Americans to take on bank loan debt. This finding implies support for economic development intermediaries and public policy to help improve LOB and Black-owned business access to bank loans. Perhaps surprisingly though, once we control for owner and business characteristics, only Black business owners are less associated with using personal savings, while White business owners do not differ significantly in their association and Asian business owners are more associated with using personal savings. Thus, the decision to use personal savings, at least, may be tied more toward low-barrier industry choice, but given the bank loan association’s finding, LOB may lack personal savings compared with other ethnicities and still be driven to family, friends, alternative financial services, and credit cards as a result.

Some of the findings are untested in the literature. For example, the finding that (ceteris paribus) LOB are more likely to be in rural areas than other ethnic minorities. This finding supports the idea that rural development organizations, in particular, may find benefits in increasing their Latino-focused business support efforts, such as Spanish language business development or translation services. It is possible the finding arises from Latinos taking over formally occupied niches (e.g., small town storefront businesses) that have become unattractive to other groups due to population decline or other structural changes that reduce profitability. Alternatively, it may be that Latinos are more likely to face limited employment options (smaller opportunity sets) in rural areas, resultantly forcing them into necessity rather than opportunity (entrepreneurship). That is, the reservation wage for entering self-employment may be lower for Latinos in rural areas. Although the model controls for the unemployment rate, it does not control for ethnicity-specific unemployment. It is the case that Latino unemployment rates are disproportionately higher in rural areas than the average unemployment rate. For example, in 2000, the difference in the unemployment rate for Latinos between nonrural and rural areas was 0.856% (8.308% and 9.164%), while the difference for the average unemployment rate was only 0.234% (5.701% and 5.935%).⁸ As a result, it may be that rural Latinos may be occupying low-profitability business niches abandoned by other ethnicities.

Previous findings show that, in general (when not controlling for business owner, business, and location factors), Latinos are more likely to work in sectors thought to be relatively vulnerable, such as low-barrier services (Puryear et al., 2008; Robles & Cordero-Guzman, 2007) like construction, wholesale trade, and retail trade sectors (U.S. Census Bureau, 2010). High-barrier industries are those requiring advanced educational degrees or large amounts of start-up capital. Low-barrier industries require less start-up capital and include some food services, some types of retail, and construction.⁹ In fact, the estimates presented here show that, when controlling for owner characteristics, there is a much more complicated picture. LOB may be more associated with wholesale and retail trade than Black-owned businesses, but are less associated with those sectors than Asian- and White-owned businesses. Similarly, LOB are more associated with the construction industry than other minorities in this study, but still less associated than White-owned businesses. Thus, the results here only partially support the conjecture of other authors that LOB lack financial capital to enter high-barrier industries and are often associated with low-barrier industries. Indeed, given that the model also controls for start-up capital source and amount, this research demonstrates that financial capital is not the sole explanation for this ethnic disparity. Other possible explanations for the remaining disparities include labor market discrimination, missing business or technical expertise (not sufficiently captured by the formal education control variables), and higher family and community ties, facilitating familial borrowing (not sufficiently captured by the financial capital source indicator control variables and local population control variables). These explanations would limit the opportunity sets of prospective Latino entrepreneurs and change opportunity costs.

The findings here show that LOB are more associated with the smaller start-up capital amounts than White-owned businesses. Although other ethnicities’ start-up capital amounts vary relative to LOB, LOB’s relation to White-owned business start-up capital amounts is consistent with LOB being associated with lower levels of capital. Thus, the results support and extend the finding of Bates et al. (2011) who found that LOB are a large percentage of businesses in the low-capital-barrier sectors and a small percentage of businesses in the high-capital-barrier sectors compared with White-owned businesses.

In summary, our results paint a rather complex picture of ethnic business differences in a way not possible with smaller and less comprehensive data sets and econometric techniques. Policy implications of our finding is that one-size-fits-all minority-government purchase preferences or additional considerations in federal grant or loan applications may not be sufficient for addressing ethnic disparities present in American business ownership. Especially, given the finding that LOB may be less likely than other minorities to benefit from sales to governments agencies, when controlling for other factors, more nuanced approaches tailored to each group’s needs may be more appropriate. Some possible changes in how local economic development practitioners engage Hispanics might be the following:

Programs that help owners compensate for lack of formal education in their business planning may be more effective than other types of interventions. For example, local economic development practitioners might work to form networks to pair more advanced entrepreneurs with coethnics at earlier stages of business growth. Learning from mentors who have succeeded from the same (or similar) cultural base may help overcome lack of formal training.

Loan guarantee programs with Hispanic youth focus to offset lower use of bank loans among younger Hispanics.

Rural incubator programs to help Hispanics take up abandoned niches (e.g., bakeries, small hardware stores) that could help make declining rural downtowns more attractive.

The results also highlight the need for further investigations beyond the 2-digit NAICS code level to understand the degree of this association, and if the aggregation to the two-digit level hides any exceptions. Another area for future research is to examine how the factors enumerated in this study interact. For example, it may be that if one only examines female business owners, the associations change significantly. Furthermore, future research should investigate associations with finer delineations of ethnicity and industry. As Valdez (2011) demonstrated, further delineations of ethnicity are important because there are often an array of experiences and outcomes among a diverse group of coethnics. For example, future work could apply this study’s model to an examination of relative associations among various Latino countries of origin and 3-digit NAICS codes. Taking the investigation into finer levels of geography could also help generalize the disparities. Dayanim (2011), for example, found neighborhood-level inequality hidden by higher geographic aggregation in city contracts aimed at reducing the disparity in set-aside programs.

These areas for future research highlight how this research provides new insights into disparities that exist in an important and growing part of the U.S. economy. Moving forward, federal, state, and local policy makers interested in economic development can use these results to not only guide their future investigations into disparities among ethnic groups in business ownership but also increase the accuracy and thereby effect of their economic development programs. While the focus of this study is on Hispanic Americans, the results reveal many unexpected differences between Hispanics and other ethnic groups, and provide a more comprehensive picture beyond binomial discussions of racial and ethnic inequality. More in-depth study of how one group overcomes barriers in an area could provide insights for future policies.

Footnotes

Appendix

Table A1.

Multinomial Logit Results (Omitted Relative Risk Ratio Coefficients).

Business and county variables	Asian American	Black American	White American
Business type: Co-op	4.68*** (1.08)	0.89 (0.24)	0.82 (0.18)
Estate	1.03 (0.51)	XX (XX)	XX (XX)
Nonprofit	2.18** (0.85)	8.20*** (2.01)	1.31 (0.28)
Public	0.97 (0.06)	0.97 (0.06)	0.56*** (0.03)
Husband and wife	XX (XX)	0.90*** (0.02)	0.95*** (0.02)
Establishment employment	0.99*** (0.00)	1.00* (0.00)	1.00*** (0.00)
Establishment payroll	1.00** (0.00)	1.00*** (0.00)	1.00*** (0.00)
Firm payroll	1.00* (0.00)	1.00*** (0.00)	1.00*** (0.00)
Firm employment	1.00*** (0.00)	1.00*** (0.00)	1.00*** (0.00)
County-level variables
Population	1.00*** (0.00)	1.00*** (0.00)	1.00*** (0.00)
Population squared	1.00*** (0.00)	1.00*** (0.00)	1.00*** (0.00)
Population density	1.00*** (0.00)	1.00** (0.00)	1.00*** (0.00)
25-34 Population	1.00*** (0.00)	1.00*** (0.00)	1.00*** (0.00)
35-44 Population	1.00 (0.00)	1.00** (0.00)	1.00 (0.00)
45-54 Population	1.00*** (0.00)	1.00*** (0.00)	1.00*** (0.00)
55-64 Population	1.00*** (0.00)	1.00*** (0.00)	1.00*** (0.00)
>64 Population	1.00*** (0.00)	1.00*** (0.00)	1.00*** (0.00)
Female population squared	1.00*** (0.00)	1.00*** (0.00)	1.00*** (0.00)
Black population	1.00*** (0.00)	1.00*** (0.00)	1.00*** (0.00)
Black population squared	1.00 (0.00)	1.00*** (0.00)	1.00*** (0.00)
Asian population	1.00*** (0.00)	1.00*** (0.00)	1.00*** (0.00)
Asian population squared	1.00*** (0.00)	1.00*** (0.00)	1.00*** (0.00)
Hispanic population	1.00*** (0.00)	1.00*** (0.00)	1.00*** (0.00)
Hispanic population squared	1.00*** (0.00)	1.00*** (0.00)	1.00*** (0.00)
Build permits	1.00 (0.00)	1.00*** (0.00)	1.00*** (0.00)
Industry employment	1.00 (0.00)	1.00 (0.00)	1.00*** (0.00)
State industry employment	1.00*** (0.00)	1.00** (0.00)	1.00 (0.00)
State variables
Arkansas	0.60*** (0.09)	0.49*** (0.07)	0.67*** (0.09)
California	0.57*** (0.07)	XX (XX)	0.35*** (0.04)
Colorado	0.29*** (0.03)	0.10*** (0.01)	0.23*** (0.02)
Connecticut	0.49*** (0.06)	0.17*** (0.02)	0.38*** (0.04)
Delaware	0.85 (0.15)	0.31*** (0.06)	0.39*** (0.06)
District of Columbia	0.44*** (0.06)	0.28*** (0.04)	0.21*** (0.03)
Florida	0.27*** (0.03)	0.12*** (0.01)	0.15*** (0.01)
Georgia	0.74*** (0.08)	0.52*** (0.05)	0.45*** (0.04)
Idaho	XX (XX)	0.05*** (0.01)	0.49*** (0.06)
Illinois	0.40*** (0.05)	XX (XX)	0.29*** (0.03)
Indiana	XX (XX)	XX (XX)	XX (XX)
Iowa	0.37*** (0.06)	0.07*** (0.01)	0.37*** (0.05)
Kansas	0.32*** (0.04)	0.14*** (0.02)	0.28*** (0.03)
Kentucky	1.13 (0.17)	0.56*** (0.08)	1.18 (0.15)
Louisiana	0.65*** (0.08)	0.44*** (0.05)	0.56*** (0.06)
Maine	XX (XX)	XX (XX)	XX (XX)
Maryland	0.68*** (0.08)	0.41*** (0.04)	0.37*** (0.04)
Massachusetts	0.59*** (0.07)	0.14*** (0.02)	0.47*** (0.05)
Michigan	XX (XX)	0.23*** (0.03)	0.53*** (0.05)
Minnesota	0.52*** (0.07)	XX (XX)	0.60*** (0.07)
Missouri	0.81 (0.11)	0.41*** (0.05)	0.80* (0.09)
Montana	0.32*** (0.08)	0.05*** (0.02)	0.58*** (0.10)
Nebraska	0.29*** (0.05)	0.09*** (0.02)	0.31*** (0.04)
Nevada	XX (XX)	XX (XX)	0.16*** (0.02)
New Hampshire	0.60*** (0.12)	XX (XX)	0.57*** (0.10)
New Jersey	0.50*** (0.06)	0.10*** (0.01)	0.21*** (0.02)
New Mexico	0.09*** (0.01)	0.02*** (0.00)	0.07*** (0.01)
New York	0.42*** (0.05)	0.12*** (0.01)	0.33*** (0.03)
North Carolina	0.71*** (0.08)	0.71*** (0.08)	0.58*** (0.06)
North Dakota	XX (XX)	XX (XX)	XX (XX)
Ohio	0.76** (0.09)	0.28*** (0.03)	0.63*** (0.07)
Oklahoma	0.56*** (0.07)	0.19*** (0.02)	0.43*** (0.05)
Oregon	0.80* (0.10)	0.08*** (0.01)	0.48*** (0.05)
Pennsylvania	0.67*** (0.08)	0.18*** (0.02)	0.44*** (0.05)
Rhode Island	0.54*** (0.09)	0.19*** (0.04)	0.70*** (0.10)
South Carolina	0.79* (0.10)	0.99 (0.13)	0.75** (0.09)
South Dakota	XX (XX)	XX (XX)	XX (XX)
Tennessee	0.79* (0.10)	0.41*** (0.05)	0.75*** (0.08)
Texas	0.28*** (0.03)	0.10*** (0.01)	0.18*** (0.02)
Utah	XX (XX)	XX (XX)	0.60*** (0.07)
Vermont	XX (XX)	XX (XX)	XX (XX)
Virginia	0.70*** (0.08)	0.43*** (0.05)	0.38*** (0.04)
Washington	0.60*** (0.07)	0.11*** (0.01)	0.39*** (0.04)
West Virginia	1.18 (0.25)	0.22*** (0.05)	0.69* (0.13)
Wisconsin	0.39*** (0.05)	XX (XX)	0.40*** (0.05)
Wyoming	0.23*** (0.05)	XX (XX)	0.22*** (0.04)
Year variable
2003	0.92*** (0.01)	1.12*** (0.01)	0.88*** (0.01)
2004	0.87*** (0.01)	1.25*** (0.02)	0.79*** (0.01)
2005	0.84*** (0.02)	1.40*** (0.03)	0.71*** (0.01)
2006	0.80*** (0.02)	1.43*** (0.04)	0.64*** (0.01)
2007	0.85*** (0.02)	1.42*** (0.04)	0.66*** (0.01)

Note. Robust standard errors in parentheses.

p < .1. **p < .05. ***p < .01.

Acknowledgements

Thanks to Anil Rupasingha, Myriam Quispe-Agnoli, Julie L. Hotchkiss, Melissa Banzhaf, Margaret C. Leventsin, J. Clint Carter, Mark Fossett, and Bethany Desalvo for their assistance with the FSRDC process. Support for this research at the Michigan and Texas FSRDC from the USDA-supported North Central Regional Center for Rural Development, the Department of Agricultural, Food, and Resource Economics at Michigan State University, the Interuniversity Consortium for Political and Social Research, and the TXRDC Consortium is also gratefully acknowledged.

Authors’ Notes

Any opinions and conclusions expressed herein are those of the author and do not necessarily represent the views of the U.S. Census Bureau. All results have been reviewed to ensure that no confidential information is disclosed.

Declaration of Conflicting Interests

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

Funding

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

Notes

ORCID iD

Craig Wesley Carpenter

Author Biographies

Craig Wesley Carpenter is an assistant professor and extension specialist in community economics and business development, in the Department of Agricultural Economics at Texas A&M University. His research expertise includes the interaction of race, ethnicity, entrepreneurship, and economic growth.

Scott Loveridge is a professor in the Department of Agricultural, Food, and Resource Economics at Michigan State University. His research focuses on social conditions associated with economic growth and the development of subnational regions.

References

Almus

(2002). What characterizes a fast-growing firm? Applied Economics, 34, 1497-1508. doi:10.1080/00036840110105010

Audretsch

D. B.

Coad

Segarra

(2014). Firm growth and innovation. Small Business Economics, 43, 743-749. doi:10.1007/s11187-014-9560-x

Bates

Lofstrom

Servon

L. J.

(2011). Why have lending programs targeting disadvantaged small business borrowers achieved so little success in the United States? Economic Development Quarterly, 25, 255-266. doi:10.1177/0891242411409206

Bates

Robb

A. M.

(2016). Impacts of owner race and geographic context on access to small-business financing. Economic Development Quarterly, 30, 159-170. doi:10.1177/0891242415620484

Bernstein

(2016). Hispanic-owned businesses on the upswing. Retrieved from https://www.census.gov/newsroom/blogs/random-samplings/2016/12/hispanic-owned_busin.html

Blanchard

Zhao

Yinger

(2008). Do lenders discriminate against minority and woman entrepreneurs? Journal of Urban Economics, 63, 467-497. doi:10.1016/j.jue.2007.03.001

Bonacich

Modell

(1980). The economic basis of ethnic solidarity in the Japanese American community. Berkeley: University of California Press.

Carrasco

(1999). Transitions to and from self-employment in Spain: An empirical analysis. Oxford Bulletin of Economics and Statistics, 61, 315-341. doi:10.1111/1468-0084.00132

Cavalluzzo

Wolken

(2005). Small business loan turndowns, personal wealth, and discrimination. Journal of Business, 78, 2153-2178. doi:10.1086/497045

10.

Coates

Gindling

T. H.

(2013). Is Hispanic population dispersion into rural counties contributing to local economic growth? Contemporary Economic Policy, 31, 649-668. doi:10.1111/j.1465-7287.2012.00334.x

11.

Coleman

Robb

(2009). A comparison of new firm financing by gender: Evidence from the Kauffman Firm Survey data. Small Business Economics, 33, 397-411. doi:10.1007/s11187-009-9205-7

12.

Davidsson

Kirchhoff

Hatemi-J

Gustavsson

(2002). Empirical analysis of business growth factors using Swedish data. Journal of Small Business Management, 40, 332-349. doi:10.1111/1540-627X.00061

13.

Davila

Mendez

Mora

M. T.

(2003). Are Hispanic immigrants in English-only states at a homeownership disadvantage? Evidence from the 1980 and 1990 U.S. Censuses. Growth and Change, 34, 40-63. doi:10.1111/1468-2257.00198

14.

Dayanim

S. L.

(2011). Do minority-owned businesses face a spatial barrier? Measuring neighborhood-level economic activity differences in Philadelphia. Growth and Change, 42, 397-419. doi:10.1111/j.1468-2257.2011.00558.x

15.

Derickson

K. D.

Ross

R. J. S.

(2015). Asia comes to main street and may learn to speak Spanish: Globalization in a poor neighborhood in Worcester. Journal of World-Systems Research, 13, 179-197. doi:10.5195/jwsr.2007.356

16.

Dhingra

(2016). Just getting a job is not enough: How Indian Americans navigate the workplace. In Zhou

Ocampo

A. C.

(Eds.), Contemporary Asian America: A multidisciplinary (3rd ed., pp. 217-233). New York: NYU Press.

17.

Diamond

(2010). The benefits of multilingualism. Science, 330, 332-333.

18.

Fairlie

R. W.

(2005). Entrepreneurship among disadvantaged groups: An analysis of the dynamics of self-employment by gender, race, and education. In Parker

S. C.

Acs

Z. J.

Audretsch

D. B.

Fairlie

R. W.

(Eds.), Handbook of entrepreneurship (pp. 437-475). Dordrecht, Netherlands: Kluwer Academic.

19.

Fairlie

R. W.

Robb

A. M.

(2007). Why are Black-owned businesses less successful than White-owned businesses? The role of families, inheritances, and business human capital. Journal of Labor Economics, 25, 289-323. doi:10.1086/510763

20.

Fairlie

R. W.

Robb

A. M.

(2008). Race and entrepreneurial success: Black-, Asian-, and White-owned businesses in the United States. Cambridge: MIT Press.

21.

Fairlie

R. W.

Robb

A. M.

(2009). Gender differences in business performance: Evidence from the Characteristics of Business Owners survey. Small Business Economics, 33, 375-395. doi:10.1007/s11187-009-9207-5

22.

Federman

M. N.

Harrington

D. E.

Krynski

K. J.

(2006). Vietnamese manicurists: Are immigrants displacing natives or finding new nails to polish? ILR Review, 59, 302-318.

23.

Fertala

(2007). The shadow of death: Do regional differences matter for firm survival across native and immigrant entrepreneurs? Empirica, 35, 59-80. doi:10.1007/s10663-007-9051-2

24.

Georgellis

Sessions

Tsitsianis

(2007). Pecuniary and non-pecuniary aspects of self-employment survival. Quarterly Review of Economics and Finance, 47, 94-112. doi:10.1016/j.qref.2006.03.002

25.

Gonzalez-Barrrera

(2016, April 14). Apprehensions of Mexican migrants at U.S. borders reach near-historic low. Pew Research Center. Retrieved from http://www.pewresearch.org/fact-tank/2016/04/14/mexico-us-border-apprehensions/

26.

Granier

(2006). Hispanic entrepreneurship: A comparison of two Georgia communities. Retrieved from http://stip.gatech.edu/wp-content/uploads/2010/05/Granier-Hispanic-Entrepreneurship-Georgia-2006.pdf

27.

Greene

W. H.

(2012). Econometric analysis (7th ed.). Boston, MA: Prentice Hall.

28.

Greve

Salaff

J. W.

(2003). Social networks and entrepreneurship. Entrepreneurship Theory and Practice, 28, 1-22. doi:10.1111/1540-8520.00029

29.

Haapanen

Tervo

(2009). Self-employment duration in urban and rural locations. Applied Economics, 41, 2449-2461. doi:10.1080/00036840802360278

30.

Haynes

G. W.

Onochie

J. I.

Lee

(2008). Influence of family’s social relationships on the debt structure of Mexican-American and Korean-American small business. Journal of Developmental Entrepreneurship, 13, 343-361. doi:10.1142/S1084946708001022

31.

Hisrich

R. D.

Brush

(1986). Characteristics of the minority entrepreneur. Journal of Small Business Management, 24(4), 1-8.

32.

Krogstad

J. M.

(2014, December 16). With fewer new arrivals, Census lowers Hispanic population projections. Pew Research Center. Retrieved from http://www.pewresearch.org/fact-tank/2014/12/16/with-fewer-new-arrivals-census-lowers-hispanic-population-projections-2/

33.

Lechmann

D. S. J.

Schnabel

(2014). Are the self-employed really jacks-of-all-trades? Testing the assumptions and implications of Lazear’s theory of entrepreneurship with German data. Small Business Economics, 42, 59-76. doi:10.1007/s11187-012-9464-6

34.

Lichtenstein

(2007). Demographic characteristics of business owners (Issue Brief No. 2). Retrieved from https://www.sba.gov/sites/default/files/Issue%20Brief%202,%20Business%20Owner%20Demographics.pdf

35.

Light

(1979). Disadvantaged minorities in self-employment. International Journal of Comparative Sociology, 20, 31-45.

36.

Light

(1980). Asian enterprise in America: Chinese, Japanese, and Koreans in small business. In Cummings

(Ed.), Self-help in urban America: Patterns of minority business enterprise (pp. 33-57). New York, NY: National University Publications.

37.

Light

Rosenstein

C. N.

(1995). Race, ethnicity, and entrepreneurship in urban America. New York, NY: Aldine de Gruyter.

38.

Lin

Picot

Compton

(2000). The entry and exit dynamics of self-employment in Canada. Small Business Economics, 15, 105-125. doi:10.1023/A:1008150516764

39.

Teixeira

(2015). Immigrants doing business in a mid-sized Canadian city: Challenges, opportunities, and local strategies in Kelowna, British Columbia. Growth and Change, 46, 631-653. doi:10.1111/grow.12103

40.

Lobao

Zhou

Partridge

Betz

(2016). Poverty, place, and coal employment across Appalachia and the United States in a new economic era. Rural Sociology, 81, 343-386. doi:10.1111/ruso.12098

41.

Lofstrom

Bates

(2009). Latina entrepreneurship. Small Business Economics, 33, 427-439. doi:10.1007/s11187-009-9203-9

42.

Millán

J. M.

Congregado

Román

(2012). Determinants of self-employment survival in Europe. Small Business Economics, 38, 231-258. doi:10.1007/s11187-010-9260-0

43.

Millán

J. M.

Congregado

Román

van Praag

van Stel

(2014). The value of an educated population for an individual’s entrepreneurship success. Journal of Business Venturing, 29, 612-632. doi:10.1016/j.jbusvent.2013.09.003

44.

Munoz

J. M.

Spain

M. I.

(2015). Hispanic-Latino entrepreneurship: Viewpoints of practitioners. New York, NY: Business Expert Press.

45.

Oberschachtsiek

(2012). The experience of the founder and self-employment duration: A comparative advantage approach. Small Business Economics, 39, 1-17. doi:10.1007/s11187-010-9288-1

46.

Partridge

M. D.

Rickman

D. S.

(2006). The geography of American poverty: Is there a need for place-based policies? Kalamazoo, MI: W.E. Upjohn Institute. doi:10.17848/9781429454872

47.

Partridge

M. D.

Rickman

D. S.

(2008). Distance from urban agglomeration economies and rural poverty. Journal of Regional Science, 48, 285-310. doi:10.1111/j.1467-9787.2008.00552.x

48.

Pew Research Institute. (2006). Hispanic trends. Cubans in the United States: Fact sheet. Retrieved from http://www.pewhispanic.org/2006/08/25/cubans-in-the-united-states/

49.

Puryear

A. N.

Rogoff

E. G.

Lee

M.-S.

Heck

R. K. Z.

Grossman

E. B.

Haynes

G. W.

Onochie

(2008). Sampling minority business owners and their families: The understudied entrepreneurial experience. Journal of Small Business Management, 46, 422-455. doi:10.1111/j.1540-627X.2008.00251.x

50.

Raijman

Tienda

(2000). Immigrants’ pathways to business ownership: A comparative ethnic perspective. International Migration Review, 34, 682-706. doi:10.2307/2675941

51.

Rauch

J. E.

(2001). Business and social networks in international trade. Journal of Economic Literature, 39, 1177-1203.

52.

Robb

A. M.

Fairlie

R. W.

(2007). Access to financial capital among U.S. businesses: The case of African American firms. ANNALS of the American Academy of Political and Social Science, 613(1), 47-72. doi:10.1177/0002716207303578

53.

Robles

B. J.

Cordero-Guzman

(2007). Latino self-employment and entrepreneurship in the United States: An overview of the literature and data sources. ANNALS of the American Academy of Political and Social Science, 613(1), 18-31. doi:10.1177/0002716207303541

54.

Shim

Eastlick

M. A.

(1998). Characteristics of Hispanic female business owners: An exploratory study. Journal of Small Business Management, 3(36), 18-34.

55.

Shinnar

R. S.

Aguilera

M. B.

Lyons

T. S.

(2011). Co-ethnic markets: Financial penalty or opportunity? International Business Review, 20, 646-658. doi:10.1016/J.IBUSREV.2011.02.014

56.

Shinnar

R. S.

Cardon

M. S.

Eisenman

Zuiker

V. S.

Lee

M.-S.

(2009). Immigrant and US-born Mexican-owned businesses: Motivations and management. Journal of Developmental Entrepreneurship, 14, 273-295.

57.

Shinnar

R. S.

Cho

Rogoff

E. G.

(2013). Outcomes of family involvement in minority owned family businesses. Journal of Family Business Strategy, 4, 22-33. doi:10.1016/J.JFBS.2012.12.001

58.

Taylor

M. P.

(1999). Survival of the fittest? An analysis of self-employment duration in Britain. Economic Journal, 109, 140-155. doi:10.1111/1468-0297.00422

59.

U.S. Census Bureau. (2010). Survey of Business Owners: Hispanic-owned firms: 2007. Retrieved from https://www.census.gov/library/publications/2007/econ/2007-sbo-hispanic.html

60.

U.S. Census Bureau. (2016a). Annual estimates of the resident population by sex, age, race, and Hispanic origin for the United States and states: April 1, 2010 to July 1, 2016. Retrieved from https://factfinder.census.gov/faces/tableservices/jsf/pages/productview.xhtml?src=bkmk

61.

U.S. Census Bureau. (2016b). QuckFacts United States. Retrieved from https://www.census.gov/quickfacts/table/PST045215/00#headnote-js-b

62.

U.S. Small Business Administration. (2017). Office of Advocacy. Small Business Profile. https://www.sba.gov/sites/default/files/advocacy/All_States.pdf

63.

Valdez

(2011). The new entrepreneurs: How race, class, and gender shape American enterprise. Stanford, CA: Stanford University Press.

64.

van Praag

C. M

. (2003). Business survival and success of young small business owners. Small Business Economics, 21, 1-17. doi:10.1023/A:1024453200297

65.

Voss

P. R.

Long

D. D.

Hammer

R. B.

Friedman

(2006). County child poverty rates in the US: A spatial regression approach. Population Research and Policy Review, 25, 369-391. doi:10.1007/s11113-006-9007-4

66.

Wang

(2010). Immigration and ethnic entrepreneurship: A comparative study in the United States. Growth and Change, 41, 430-458. doi:10.1111/j.1468-2257.2010.00532.x