Entrepreneurship and Student Loans: An Analysis of the Association Between Self-Employment and Student Loans

Abstract

This study examines the association between self-employment, otherwise known as entrepreneurship, and student loan debt. Based on data from the 2021 National Financial Capability Study, which were analyzed using a multinomial probit model, our results suggest that households that used student loans to finance the education of one household member are more likely to report that the other member of the household pursued self-employment, with everything else remaining equal. This association between student loans and household self-employment is both economically and statistically significant. Our results highlight the importance of alternative forms of financing for entrepreneurship and expand upon existing knowledge about the unintended effects of student loans. Policymakers should consider these results when analyzing the various intended and unintended benefits of policies affecting the availability of student loans.

1. Introduction

Student loans are a form of debt taken to aid in the pursuit of education. Recently, student loans have faced immense media, government, and academic scrutiny, largely focused on determining whether student loans are a net positive. While research suggests that loans can assist financially constrained students by providing opportunities to further their education and achieve higher income (Korankye et al., 2024), there is concern that the cost of student loan debt outweighs the benefits of completing higher education (Farrington, 2024). The cost of college has dramatically increased, with tuition and fees at public and private 4-year institutions increasing by 141% and 181%, respectively, in the last 20 years (Hanson, 2024). Following the rise in college costs, total student loan balances in the United States ballooned to USD 1.61 trillion in the third quarter of 2024—almost five times the balance reported 20 years ago (Federal Reserve Bank of New York, 2024). In this study, we contribute to the existing literature on the consequences of student loans by exploring the relationship between student loan debt and entrepreneurship. Specifically, we examine whether one member of the household having student loan debt affects the likelihood of one member of said household pursuing self-employment.

Self-employment is risky and often requires significant financing (Ewing Marion Kauffman Foundation, 2020). Budding entrepreneurs often lack the collateral necessary to obtain traditional loans and/or access to capital, which then becomes a barrier to entry and ultimate business success (Al-Fattal, 2024). According to the United States Bureau of Labor Statistics, approximately 20% of small businesses fail in the first year of business (Bureau of Labor Statistics, 2024a). The number of loans granted by financial institutions has declined; a recent report from the Small Business Administration (2025) revealed that only 15% of funds were allocated to startups to open a business—meaning that entrepreneurs must often rely on personal funds and/or credit to finance their businesses (Friedline & West, 2016).

While student loans share similarities with other forms of debt, they have several unique features that influence their relationship with entrepreneurship. First, student loans are widespread and relatively easy to acquire. The funds from student loans are subject to fewer restrictions than many other forms of debt, meaning they can be used to pay for a variety of general living expenses (Litant, 2024). Additionally, prior research on couples has shown that they often intermingle resources, including income and debt (Beblo & Beninger, 2017; Evans & Gray, 2021; Huang et al., 2019; Kan & Laurie, 2014; Kukk & van Raaij, 2022). Members of couples are willing to make sacrifices for their partner or spouse’s career (Van Lange et al., 1997; McKinnish, 2008; Burke & Miller, 2018). The combination of the features of student loans with the willingness of couples to intermingle resources means the proceeds from student loans might be utilized to pay for living expenses while the borrower or their spouse or partner pursues self-employment.

We used data from the 2021 National Financial Capability Study (NFCS) and employed a multinomial probit model to examine the association between student loans and self-employment among household members. These results from the proposed model are both economically and statistically significant, providing insight into how student loans affect household entrepreneurship. Prior studies have examined antecedents of self-employment from an individual difference perspective (considering factors such as risk-taking or impulsive personality, family history of entrepreneurship, and self-efficacy). This research addresses a gap in the literature by exploring household financial intermingling as it relates to student loans and self-employment.

The remainder of the paper is organized as follows: Section 2 reviews the literature on student loans, self-employment, and couples intermingling resources. Section 3 describes the data used in the analysis. Section 4 outlines the model selection process. Section 5 presents empirical results. Section 6 addresses the study’s limitations and highlights the study’s contribution to the literature.

2. Literature Review

2.1. Consequences of Student Loans

Student loans, which are debts acquired to fund post-high school education, can be viewed as both an investment in human capital and a liability on the individual, and often have similarly detrimental effects on life and financial outcomes to other types of debt. Student loans reduce homeownership rates and affect the credit scores of young public-school attendees (Mezza et al., 2020). The constraining nature of debt repayments can also affect job search outcomes; those with student loan debt often take lower-paying jobs (Ji, 2021). Also, while student loans increase access to education, high student loan balances can reduce the likelihood of completing college (Dwyer et al., 2012). Student loans can also have adverse effects on borrowers, reducing life satisfaction (Korankye & Kalenkoski, 2021a), financial well-being (Korankye & Kalenkoski, 2021b), and the likelihood of owning stocks (Korankye & Guillemette, 2020).

However, student loans are unsecured and are relatively easy to acquire compared to other types of debt. Additionally, student loans have fewer spending restrictions than many other types of debt. Borrowers may use the funds from student loans to make payments directly to their education institution; however, these loans can also be used to fund various living expenses, including housing, food, transportation, technology, childcare, and personal costs (Litant, 2024). According to Montalto et al. (2019), 24.7% of college students borrowed the maximum amount available, regardless of the amount needed for school. These findings suggest that student loans can be used as a source of excess funds that do not contribute directly to the education of the individual taking out the loan, and the relative freedom of spending could allow borrowers to employ those funds for non-educational purposes. In the context of self-employment, numerous federal programs have made student loan repayment income-dependent and, in some cases, offer opportunities for deferment, which many people find appealing, especially given the income uncertainty associated with entrepreneurial start-ups (Devaraj & Patel, 2020).

2.2. Self-Employment and Financing

Economic trends and societal shifts over the last few decades have resulted in increased demand for more innovative and self-driven career paths (Al-Fattal, 2024). According to the Internal Revenue Service (2025), self-employed individuals “carry on a trade or business as a sole proprietor or an independent contractor; are a member of a partnership that carries on a trade or business; or are otherwise in business for [them]self.” The entrepreneurial environment is unique in that it is high in uncertainty, but also high in autonomy, allowing people to control their own work and schedules, highlighting their strengths and minimizing their weaknesses (Yu et al., 2021). At its core, entrepreneurship is linked to competitive advantage and innovation (De Carolis & Saparito, 2006). It is a complex concept that includes not only the creation of ideas or discovery of opportunities, but also the appropriate exploitation of these ideas and opportunities (Jiménez et al., 2015). Entrepreneurs often put their finances, personal relationships, careers, and reputations at risk to pursue their business ventures (Van Ness et al., 2020).

Several factors contribute to an individual’s likelihood of self-employment, including a family history of entrepreneurship, social capital, mentorship, entrepreneurial self-efficacy, achievement motivation, risk-taking, impulsivity, and wealth (Ahmed et al., 2021; Al-Fattal, 2024; Wiklund et al., 2017; Yadav & Batra, 2023; Yin & Wu, 2023; Yu et al., 2021). Education also plays an important role in entrepreneurial intentions, as it shapes the minds of future entrepreneurs and builds necessary skills and capabilities required to start and maintain a business (Al-Fattal, 2024). Higher levels of formal education tend to increase self-confidence and exploration of entrepreneurial opportunities and activities (Jiménez et al., 2015). Additionally, students with innovative ideas can become business savvy and learn industry regulations and market conditions as well as build social capital in institutes of higher education (Al-Fattal, 2024).

One of the main challenges faced by entrepreneurs is securing funding for their entrepreneurial ventures due to limited credit and/or access to financial resources (Friedline & West, 2016). Many young entrepreneurs simply lack the necessary collateral or steady income required to procure a traditional loan (Al-Fattal, 2024). Interestingly, students with loans are more likely to be “hybrid” entrepreneurs, simultaneously working another job while running their business in order to mitigate the risk of failure (Neneh, 2020). Lending for small business start-ups from financial institutions has also declined (Friedline & West, 2016), and small business owners have reduced access to capital due to high interest rates, inflation, and economic uncertainty (FDIC, 2024). Twenty years of research in entrepreneurial bank financing have also identified a gender gap: female entrepreneurs experience more business loan rejections and pay higher interest rates on loans compared to male entrepreneurs (Malmström et al., 2024). Entrepreneurial ventures can be risky, with a high potential for failure and without an immediate predictable income; as such, entrepreneurs often turn to their network of family and friends for financial support to start their business (Ewing Marion Kauffman Foundation, 2020).

2.3. Household Financial Intermingling

Financial intermingling is defined as “the use of household assets for the support of the business and/or the use of business assets (other than wage and salary payments) for support of the household” (Yilmazer & Schrank, 2006, p. 729). Previous research has found that self-employed households intermingle their personal and business resources, using consumer loans to fund business activities and business assets to support the household (Kneiding & Kritikos, 2013). In an experiment involving real-life couples in Germany, Beblo and Beninger (2017) found that a substantial share of couples intermingle their incomes. This finding is far from an isolated phenomenon, as couples from various countries exhibit income-pooling behavior (Evans & Gray, 2021). Using survey data, Evans and Gray (2021) determined that 80% of coupled respondents in the United States intermingled all their income, compared to just 12% that did not intermingle any income. Couples also share savings accounts, investments, and debt (Huang et al., 2019; Kan & Laurie, 2014; Kukk & van Raaij, 2022).

G. S. Becker (1965) states that a household acts as an economic unit, as people who are married or are cohabitating benefit from a comparative advantage. Therefore, each individual within the household engages in specialized activities in which they have a comparative advantage. This allows each member of the two-unit partnership to benefit from the intermingling of resources, maximize their wealth, and minimize the potential risks associated with their independent activities in the labor market or in household production.

Members of a couple are often willing to make personal sacrifices to benefit either their partner or their household (Van Lange et al., 1997). This willingness to sacrifice improves relationship outcomes (Van Lange et al., 1997; Wieselquist et al., 1999; Rusbult & Van Lange, 2003). A common concession a member of a couple can make is in their work or career, as one or both members of the couple might restructure or reduce their career commitments to benefit the household, placing limits on their work or choosing to stop working outside of the home entirely (P. E. Becker & Moen, 1999). Individuals might also relocate for the sake of their partner’s career, disrupting their own career and social life in the process (Shihadeh, 1991; Nivalainen, 2004; Copeland & Norell, 2002; McKinnish, 2008; Burke & Miller, 2018). In conclusion, couples often intermingle resources, engage in comparative advantage, and make sacrifices for their partner or for the household as a whole.

As a lack of financial resources is one of the main obstacles for potential entrepreneurs (Friedline & West, 2016; Al-Fattal, 2024), new sources of financing could boost self-employment rates. The wide availability of student loans and the ability to use them for non-educational purposes means that student loans can serve as an alternate source of funds for those looking to start their own business, effectively subsidizing the living costs of potential entrepreneurs as they begin their business and freeing them to take the risk of pursuing self-employment. Since couples often intermingle finances, engage in comparative advantage, and make sacrifices for their partner’s career (G. S. Becker, 1965; Van Lange et al., 1997; Kneiding & Kritikos, 2013; Beblo & Beninger, 2017), one member of the household may take out student loans to fund shared living expenses, such as rent or a mortgage, food, transportation, childcare, etc. This can allow for greater flexibility in the couple’s financial situation, allowing one or both partners to pursue self-employment.

H1. 

Student loans increase the likelihood that one or both members of the household are self-employed.

3. Data

The National Financial Capability Study (NFCS) dataset, published by the FINRA Investor Education Foundation, is a publicly available dataset for researchers exploring financial capability topics across diverse populations. Since its inception in 2009, the survey has been conducted every three years. The NFCS gathers responses from thousands of adults across the United States, considering multiple indicators of financial capability, “including financial behaviors, attitudes, knowledge, and access to financial products and services” (Financial Industry Regulatory Authority, 2025). The NFCS is a cross-sectional survey, not a longitudinal one, so all variables were captured for 2021 only. This study utilizes data from the 2021 NFCS, specifically analyzing households with individuals who are married or cohabiting with a partner and whether student loans are associated with the likelihood of self-employment of either the survey respondent or their spouse or partner.

The NFCS has been administered by the FINRA Investor Education Foundation since its launch in 2009. This large-scale, multiyear project aims to monitor and better understand Americans’ financial situations (Mottola & Kieffer, 2017). The 2021 NFCS survey was conducted between June and October during the COVID-19 pandemic, which was a period of significant financial disruption: 20% of respondents reported job loss, over 25% experienced a substantial and unexpected decline in income, and more than 75% received at least one pandemic-related stimulus payment, which 38% of recipients used to reduce their debt (Lin et al., 2022). Given these factors, this study seeks to examine whether student loan resources are associated with supporting entrepreneurial efforts within households.

In 2020, approximately 16.5 million Americans were self-employed, accounting for 10.4% of the total working population (Bureau of Labor Statistics, 2024b). The agricultural sector remains a significant contributor, representing 7.55% of unincorporated self-employed workers (Bureau of Labor Statistics, 2024b). Diosdado et al. (2024) found that in farming households in the United States, each additional dollar of student loan debt is associated with a decrease in farmland ownership. However, due to data limitations, this analysis was unable to identify the specific industries served by respondents within the broader U.S. economy. This study builds on these findings by exploring the interplay between student loan debt and entrepreneurial activities during a period of heightened economic uncertainty. Our research aims to offer a deeper understanding of how financial resources, such as student loans, are leveraged to foster economic resilience and innovation within American households.

4. Model

This study examines the relationship between student loan debt and household entrepreneurial activities during a period of heightened economic uncertainty. The NCFS Survey asks respondents about their and their partner or spouse’s employment, including self-employment. The dependent variable in this analysis is nominal and represents four categories: (1) no one in the household is self-employed, (2) the spouse or cohabitating partner is self-employed, (3) the individual respondent is self-employed, and (4) both members of the household are self-employed. Given the categorical nature of the dependent variable, it was determined that using an econometric model that assigned numerical significance to individuals within the household would be neither appropriate nor feasible.

To address these challenges, a multinomial probit model was selected as the most suitable econometric approach for analyzing the association between student loan debt and self-employment. This model predicts the probability of an outcome based on an unobserved latent variable using the independent variables specified in the proposed framework. The equation below illustrates the structure of the latent variable model:

$$Y\left\{\begin{array}{l}{Y}_i^{1*}, No one in Houshold is Self Employed\\ Y_i^{2*}, Spouse or partner is Self Empoyed\\ Y_i^{3*}, Respondent is Self Employed\\ Y_i^{4*}, Both Household Members are Self Employed\end{array}\right.$$

and

• $Y_i^{1*}$ = ${\beta }_0$ + ${\beta }_1$Respondent’s Student Loan + ${\beta }_2$Spouse or Partner’s Student Loan + ${\beta }_3$Both Student Loan + ${\beta }_k$Z + ${\epsilon }_1$
• $Y_i^{2*}$ = ${\beta }_0$ + ${\beta }_1$Respondent’s Student Loan + ${\beta }_2$Spouse or Partner’s Student Loan + ${\beta }_3$Both Student Loan + ${\beta }_k$Z + ${\epsilon }_2$
• $Y_i^{3*}$ = ${\beta }_0$ + ${\beta }_1$Respondent’s Student Loan + ${\beta }_2$Spouse or Partner’s Student Loan + ${\beta }_3$Both Student Loan + ${\beta }_k$Z + ${\epsilon }_3$
• $Y_i^{4*}$ = ${\beta }_0$ + ${\beta }_1$Respondent’s Student Loan + ${\beta }_2$Spouse or Partner’s Student Loan + ${\beta }_3$Both Student Loan + ${\beta }_k$Z + ${\epsilon }_4$

where

$$\epsilon ~N(0,\Sigma )$$

The NCFS Survey collects information about household debt, including whether anyone in the household has student loans and whose education was funded with the proceeds of said loans. The variable “Respondent’s Student Loan” takes the value of yes if the household has outstanding student loan debt that was accrued to finance the respondent’s education. The variable “Spouse or Partner’s Student Loan” takes the value of yes if the household has outstanding student loan debt that was used to finance the respondent’s spouse or partner’s education. The variable “Both Student Loan” represents an interaction between “Respondent’s Student Loan” and “Spouse or Partner’s Student Loan”. Due to the NFCS survey’s limitations, the exact uses of the student loan proceeds are unknown. Z represents a matrix of control variables.

Table 1. This table contains a description, frequency, average, and standard deviation for each of the variables included in the multinomial probit.

Variable	Freq	Average	Std. Dev.
Household Self-employment Status
No one is self-employed	13,189	0.8747	0.3311
Partner/Spouse is self-employed	774	0.0513	0.2207
Respondent is self-employed	637	0.0422	0.2011
Both are self-employed	479	0.0318	0.1754
Respondent’s Student Loan
No	12,997	0.8619	0.3499
Yes	2082	0.1380	0.3449
Spouse or Partner’s Student Loan
No	13,600	0.9059	0.2919
Yes	1419	0.0941	0.2919
Both Have Student Loans	604	0.0400	0.1960
Marital Status
Married	12,629	0.8375	0.3689
Unmarried Living with Partner	2450	0.1624	0.3689
Respondent’s Gender
Male	6992	0.4636	0.4987
Female	8087	0.5363	0.4987
Household has Children
Do not have children	3508	0.2326	0.4225
Yes, but not financially dependent	4980	0.3302	0.4703
Yes, financially dependent children	6591	0.4370	0.4960
Respondent’s Level of Education
Did not complete high school	323	0.0214	0.1148
High school graduate	2359	0.1564	0.3633
GED or alternative	943	0.0625	0.2421
Some level of college education, but no degree	3704	0.2456	0.4305
Associate’s degree	1675	0.1110	0.3142
Bachelor’s degree	4112	0.2726	0.4453
Postgraduate degree	1963	0.1301	0.3365
Annual Household Income
Less than USD 15,000	707	0.0468	0.2114
USD 15,000 to USD 25,000	923	0.0612	0.2397
USD 25,000 to USD 35,000	1218	0.0807	0.2725
USD 35,000 to USD 50,000	1994	0.1322	0.3387
USD 50,000 to USD 75,000	3132	0.2077	0.4057
USD 75,000 to USD 100,000	2639	0.1750	0.3800
USD 100,000 to USD 150,000	2797	0.1854	0.3887
USD 150,000 to USD 200,000	1008	0.0668	0.2498
USD 200,000 to USD 300,000	461	0.0305	0.1722
USD 300,000 or more	200	0.01326	0.1144

summarizes the statistics for each of the variables included in the multinomial probit model.

The estimated coefficients from the multinomial probit model do not directly represent changes in probabilities and are not interpreted in the same way as coefficients in OLS models. Therefore, they can be challenging to interpret and are best evaluated by considering the direction of the predicted coefficients rather than their numerical values. The estimated coefficients for the model are presented in

Table 2. This table presents the estimated coefficients from the multinomial probit model on the association between household self-employment status and student loans.

	Household Self-Employment Status
Variable	Spouse or Partner	Respondent	Both
Respondent’s Student Loan (Ref: No)
Yes	0.1553 **	0.1215	−0.0939
	(0.0761)	(0.0844)	(0.0971)
Spouse or Partner’s Student Loan (Ref: No)
Yes	−0.3123	0.2116 **	−0.0565
	(0.1081)	(0.1006)	(0.1224)
Both Student Loans (Ref: No)
Yes	−0.0646	−0.1836	0.0369
	(0.1700)	(0.1719)	(0.2081)
Marital Status (Ref: Married)
Unmarried—Living with Partner	0.2429 ***	0.3202 ***	0.3376 ***
	(0.0651)	(0.0675)	(0.0718)
Respondent’s Gender (Ref: Male)
Female	0.2101 ***	−0.2213 ***	−0.1145 **
	(0.0489)	(0.0512)	(0.0558)
Household has Children (Ref: Do not have children)
Yes, but not financially dependent	0.0812	−0.0448	−0.1252
	(0.0673)	(0.0712)	(0.0742)
Yes, financially dependent children	0.2100 ***	0.1320 **	−0.0531
	(0.0614)	(0.0638)	(0.0676)
Respondent’s Level of Education (Ref: Did not complete High School)
High school graduate	0.0913	−0.1752	0.0814
	(0.1801)	(0.1735)	(0.1859)
GED or alternative graduate	0.0778	−0.0404	0.1055
	(0.1950)	(0.1865)	(0.2020)
Some level of college education, but no degree	0.1480	−0.0631	0.1353
	(0.1786)	(0.1711)	(0.1851)
Associate’s degree	0.1972	−0.1104	−0.0525
	(0.1865)	(0.1809)	(0.2002)
Bachelor’s degree	0.1937	−0.2993 *	0.1342
	(0.1812)	(0.1753)	(0.1890)
Postgraduate degree	0.1676	−0.3149 *	0.0219
	(0.1893)	(0.1857)	(0.2011)
Annual Household Income (Ref: Less than USD 15,000)
USD 15,000 to USD 25,000	0.0555	0.0717	0.0174
	(0.1413)	(0.1497)	(0.1387)
USD 25,000 to USD 35,000	0.0302	−0.2248	−0.2972 **
	(0.1352)	(0.1517)	(0.1413)
USD 35,000 to USD 50,000	−0.0388	0.1971	−0.2924 **
	(0.1287)	(0.1327)	(0.1305)
USD 50,000 to USD 75,000	0.0089	0.0404	−0.4351 ***
	(0.1240)	(0.1311)	(0.1283)
USD 75,000 to USD 100,000	0.0389	0.0468	−0.3443 ***
	(0.1278)	(0.1358)	(0.1323)
USD 100,000 to USD 150,000	−0.0450	0.0716	−0.4122 ***
	(0.1304)	(0.1377)	(0.1357)
USD 150,000 to USD 200,000	−0.0247	0.1844	−0.2887 *
	(0.1512)	(0.1572)	(0.1595)
USD 200,000 to USD 300,000	0.2064	0.3634 **	−0.1289
	(0.1746)	(0.1829)	(0.1898)
USD 300,000 or more	0.6201 ***	1.0326 ***	0.4948 **
	(0.2085)	(0.2067)	(0.2161)

Source: 2021 National Financial Capability Study. Standard errors in parentheses. Statistical significance: * <0.10, ** <0.05, and *** <0.01.

. The marginal effects, which can be interpreted as an increase or decrease in the probability of an outcome resulting from a change in an independent variable, are also calculated. These results are reported in

Table 3. The table presents the marginal effects from the multinomial probit model regarding the association between household self-employment status and student loans.

	Household Self-Employment Status
Variable	No One	Spouse or Partner	Respondent	Both
Respondent’s Student Loan (Ref: No)
Yes	−0.0138	0.0125 *	0.0077	−0.0064
	(0.0096)	(0.0066)	(0.0061)	(0.0045)
Spouse or Partner’s Student Loan (Ref: No)
Yes	−0.0085	−0.0038	0.0164 **	−0.0039
	(0.0122)	(0.0079)	(0.0081)	(0.0059)
Both Student Loans (Ref: No)
Yes	0.0112	−0.0038	−0.0108	0.0034
	(0.0186)	(0.0122)	(0.0090)	(0.0117)
Marital Status (Ref: Married)
Unmarried-Living with Partner	−0.0513 ***	0.01351 ***	0.0195 ***	0.0165 ***
	(0.0087)	(0.0057)	(0.0054)	(0.0047)
Respondent’s Gender (Ref: Male)
Female	0.0034	0.0187 ***	−0.0160 ***	−0.0062 **
	(0.0055)	(0.0036)	(0.0034)	(0.0029)
Household has Children (Ref: Do not have children)
Yes, but not financially dependent	0.0028	0.0068	−0.0026	−0.0071 *
	(0.0072)	(0.0046)	(0.0043)	(0.0040)
Yes, financially dependent children	−0.0180 ***	0.0154 ***	0.0077 *	−0.0051
	(0.0068)	(0.0044)	(0.0041)	(0.0029)
Respondent’s Level of Education (Ref: Did not complete High School)
High school graduate	0.0027	0.0072	−0.0145	0.0046
	(0.0193)	(0.0114)	(0.0140)	(0.0086)
GED or alternative graduate	−0.0055	0.0050	−0.0046	0.0052
	(0.0213)	(0.0124)	(0.0152)	(0.0096)
Some level of college education, but no degree	−0.0095	0.0102	−0.0072	0.0065
	(0.0203)	(0.0121)	(0.0146)	(0.0090)
Associate’s degree	−0.0027	0.0157	−0.0101	−0.0027
	(0.0203)	(0.0121)	(0.0146)	(0.0090)
Bachelor’s degree	−0.0001	0.0157	−0.0233	0.0076
	(0.0195)	(0.0116)	(0.0141)	(0.0088)
Postgraduate degree	0.0071	0.0144	−0.0235	0.0018
	(0.0204)	(0.0123)	(0.0145)	(0.0093)
Annual Household Income (Ref: Less than USD 15,000)
USD 15,000 to USD 25,000	−0.0079	0.0036	0.0038	0.0003
	(0.0168)	(0.0106)	(0.0088)	(0.0107)
USD 25,000 to USD 35,000	0.0227	0.0060	−0.0100	−0.0188 **
	(0.0155)	(0.0102)	(0.0078)	(0.0096)
USD 35,000 to USD 50,000	0.0069	−0.0022	0.0156 *	−0.0202 **
	(0.0148)	(0.0093)	(0.0081)	(0.0091)
USD 50,000 to USD 75,000	0.0182	0.0034	0.0049	−0.0266
	(0.0143)	(0.0091)	(0.0075)	(0.0089)
USD 75,000 to USD 100,000	0.0126	0.0052	0.0046	−0.0225 **
	(0.0148)	(0.0095)	(0.0078)	(0.0093)
USD 100,000 to USD 150,000	0.0192	−0.0010	0.0072	−0.0254 ***
	(0.0150)	(0.0095)	(0.0081)	(0.0093)
USD 150,000 to USD 200,000	0.0067	−0.0011	0.0145	−0.0200 **
	(0.0175)	(0.0110)	(0.0101)	(0.0105)
USD 200,000 to USD 300,000	−0.0276	0.0153	0.0259 *	−0.0136
	(0.0225)	(0.0146)	(0.0138)	(0.0127)
USD 300,000 or more	−0.1601 ***	0.0415 *	0.0927 ***	0.0258
	(0.0360)	(0.0225)	(0.0262)	(0.0216)

Source: 2021 National Financial Capability Study. Standard errors in parentheses. Statistical significance: * <0.10, ** <0.05, and *** <0.01.

; statistically significant effects are indicated by an asterisk.

5. Results

Our analysis revealed that households with student loans are more likely to report that at least one member of said household is self-employed. The key finding of this analysis suggests that the beneficiary of the student loan proceeds (i.e., whether the respondent or spouse is the one with the student loan debt) is a crucial determinant of which member in the household is self-employed. Specifically, when one member of the household has student loan debt, generally, the other member of the household is self-employed. For respondents who used student loans to finance their own education, there was a 1.25% increase in the probability of their partner or spouse being self-employed, while using student loans to fund a partner or spouse’s education increased the probability of the respondent being self-employed by 1.64%. Student loan resources are a source of capital, and our results suggest that capital may help entrepreneurs navigate and possibly circumvent traditional external financing costs associated with entrepreneurial ventures. The following tables contain the data along with the econometric results from the multinomial probit to support the association between entrepreneurship and student loan debt. The remainder of Section 5 describes the results of the analysis in further detail.

5.1. Summary Statistics

Table 1 summarizes the statistics for each of the variables included in the multinomial probit model. The analyzed sample consisted of 15,079 respondents, each representing a household of married or cohabitating partners. Of these households, 9.3% stated that one member of the household was self-employed, 3.2% reported that both members were self-employed, and 87.5% reported that neither member of the household was self-employed. Overall, 19.2% of the households in the sample reported having at least some student loan debt, whereas both members had student loan debt in 4.0% of households. Of the respondents, 53.6% were female. The sample also had a broad distribution based on respondents’ education level, whether the household had financially dependent children, and annual household income.

5.2. Estimated Coefficients from Multinomial Probit and Marginal Effects

The estimated coefficients and marginal effects according to the multinomial probit are reported within Table 2 and Table 3, respectively. The reference outcome variable in Table 2 is no member of the household being self-employed. The marginal effects from the multinomial probit reveal several significant associations between student loan debt and the self-employment status of households who were either married or cohabitating. The decision regarding whose education was financed with those student loan proceeds was confirmed to have statistical significance, suggesting that households were strategic in deciding which party would obtain an education and which party would pursue an entrepreneurial enterprise.

Respondents who took out student loans to finance their own education (“Respondent’s Student Loan” = yes) had a 1.25% greater probability of reporting that their partner or spouse was self-employed. When the household reported using student loan debt to finance a partner or spouse’s education (“Spouse or Partner’s Student Loan” = yes), there was a 1.64% increase in the probability of respondent self-employment (p-value = 0.057), indicating a positive association between financing a partner’s education and personal entrepreneurship. The coefficients for the interaction “Both Student Loan” were statistically insignificant in every specification. This means that for households in which both the respondent and the respondent’s partner or spouse had student loan debt, the likelihood of self-employment was unaffected.

Other interesting findings were obtained across the four models. One such finding is that marital status significantly influences entrepreneurial engagement. Specifically, individuals who cohabit with a partner outside of marriage are more likely to pursue entrepreneurship. The results, which are economically and statistically significant in all models, suggest that unmarried couples may be better positioned to engage in entrepreneurial activities compared to their married counterparts. Unmarried respondents reported greater self-employment rates; specifically, unmarried respondents were 1.35% more likely to have a self-employed partner/spouse, 1.95% more likely to be self-employed themselves, and 1.65% more likely to report both them and their partner/spouse as self-employed. This aligns with the research findings that married individuals tend to be more risk-averse than unmarried individuals (Abdelkerim et al., 2024). This pattern suggests that individuals benefit from certain advantages associated with marriage, such as stability or the intermingling of resources, by choosing to cohabitate while avoiding or limiting the potential risks associated with a failed marriage. By limiting exposure to the negative consequences of marriage dissolution, cohabiting appears to maximize potential entrepreneurial opportunities while mitigating risks.

Gender analysis revealed significant differences in self-employment patterns. Female respondents were 1.87% more likely to have a self-employed spouse or partner, a statistically significant finding. Conversely, female respondents were 1.60% less likely to be self-employed compared to their male counterparts. These results align with the marginal effects of this analysis, which suggest that female respondents report having self-employed spouses and are less likely to report self-employment after financing their education with student loan resources. Having financially dependent children also affects the likelihood of self-employment. Households with financially dependent children were 1.80% less likely to have neither member self-employed.

Overall, we were unable to identify a consistent relationship between annual household income and household self-employment status across all four of the models. The income variables were statistically significant only in the model that examined whether both partners were engaged in entrepreneurial activities. The results suggest that as household income increases, the likelihood of both parties pursuing entrepreneurship decreases. Notably, the model that accounted for both partners’ involvement in self-employment was the only one in which a significant portion of the estimated coefficients supported this observation.

6. Conclusions

This analysis examines whether members of households financed by student loans are more likely to be self-employed. Student loans are widely available and can be used for a wide variety of expenses (Litant, 2024); as such, the economic opportunities associated with access to student loan resources may encourage individuals to pursue entrepreneurial ventures, particularly in households in which said loans are used to finance a spouse or partner’s education. Marginal effects from a multinomial probit model suggest that the provision of student loan financing by one member of the household increases the likelihood that the person’s spouse or partner will be self-employed. Future research should examine how student loan resources are utilized at the household level. Further exploration of this consumption of resources would improve our understanding of how those resources may be used for other household expenditures and/or investments, perhaps creating opportunities for other government agencies to assist with household entrepreneurial endeavors rather than using student loans to facilitate those pursuits.

We contribute to the literature on entrepreneurship by examining a previously unexplored determinant of self-employment: household student loans. We highlight the use of student loans as an alternative source of financing for entrepreneurship, given their flexibility, low interest rates, and minimum qualification conditions. Specifically, we find that when one member of the household takes out student loans, that member’s partner or spouse is more likely to be self-employed. This research fills a gap in the entrepreneurship literature, expanding upon our understanding of intermingling personal and business resources as a possible financing strategy for self-employed households.

Our study informs policymakers about a potential unintended effect of student loans: an increase in household self-employment. Student loans have been subject to immense scrutiny in recent years. The cost of attending college in the US has increased drastically over time, with a resulting increase in the amount of student loan debt (Federal Reserve Bank of New York, 2024; Hanson, 2024). Policymakers are interested in the use of student loans, especially since the U.S. government plays a large role in facilitating and financing higher education (Council on Foreign Relations, 2024). Policies that require closer monitoring of student loans might be appropriate if policymakers are concerned about the procurement of student loans for purposes that do not directly benefit educational attainment. Our study expands the knowledge of the spillover effects of student loan financing, providing useful information for policymakers as they attempt to navigate the benefits and costs of student loans. The intermingling of business and personal resources may represent additional liabilities and financial risks for households and businesses alike (Kneiding & Kritikos, 2013). Future research might expand on our findings by investigating whether using student loans for non-educational purposes affects educational attainment or debt repayment.

Our study focuses on household entrepreneurship and therefore utilizes control variables related to the household. We conducted a link test to check if additional independent variables are needed. The results of the link test indicated the vector of variables included was sufficient. However, an inherent limitation of this study is the potential for omitted variable bias. It is possible that additional individual-level factors affect self-employment patterns in households. Specifically, future studies can examine the effect of behavioral traits, skillsets, and circumstances. These factors, such as financial literacy, risk-tolerance, race or ethnicity, and previous work experience, may also be important determinants of household self-employment. Future studies would benefit from using datasets that provide these variables for both members of the household, as the NFCS provides some variables for only the respondent, not the partner or spouse.

Another limitation of this study was the inability to determine the exact use of student loan proceeds. While student loans can be used for a variety of general living expenses (Litant, 2024), they are not intended to support entrepreneurial business ventures. Future research can explore the specific uses of student loan proceeds and more clearly identify the mechanism through which these loans facilitate household self-employment. Future research might also explore how student debt works in conjunction with other forms of financing to enable one or both members of the household to pursue self-employment. Particularly, other projects can consider whether and how potential entrepreneurs who lack access to traditional forms of business financing use student loan financing due to its relatively widespread availability and comparative flexibility.