Margin Trading and Cryptocurrency Investment Among U.S. Investors: Evidence from the National Financial Capability Study

Abstract

This study examined the relationship between margin trading and cryptocurrency investment using data from the 2018 and 2021 waves of the National Financial Capability Study (NFCS) Investor Survey. Guided by behavioral finance theory, which suggests that cognitive biases may influence risk-taking, the study explored whether margin loan use and margin calls are associated with higher cryptocurrency participation. Margin loans are inherently risky, as they must be repaid regardless of investment outcomes, and margin calls are triggered when an investor’s equity falls below a required threshold. The results showed a positive and statistically significant association between margin activity and cryptocurrency investment. Specifically, individuals with a margin loan were 17 percentage points more likely to invest in cryptocurrency, while those who have experienced a margin call were 23 percentage points more likely. Given the extreme volatility of cryptocurrencies, these results highlight the increased risks investors face when using leverage in speculative markets. The analysis is based on cross-sectional data from U.S. investors; therefore, the findings should be interpreted as correlational rather than causal.

1. Introduction

Recently, the invention of blockchain technology and the introduction of cryptocurrency in the financial market have brought forth a new global era. Since the birth of Bitcoin in 2008, blockchain and its application in cryptocurrency have gained a lot of popularity (Hileman & Rauchs, 2017). Over the last decade, the cryptocurrency market has experienced tremendous growth (Xi et al., 2020; Zhao & Zhang, 2021). For instance, in 2017, the price of Bitcoin increased dramatically, with a compound annual growth rate of 1300% (Hackethal et al., 2022). With over 10,000 cryptocurrencies available today, including Bitcoin, Ethereum, Ripple, Tether, Dogecoin, and Litecoin, the combined market value of cryptocurrencies was estimated at USD 2.79 trillion as of 2023 (Hossain, 2023).

However, the rapid growth of cryptocurrencies has also been accompanied by extreme volatility, often linked to the lack of regulatory oversight for this market (Abu Bakar & Rosbi, 2017). The substantial volatility and potential for high returns have attracted many individual investors who often treat cryptocurrencies as a new asset class (Ji et al., 2019). As a result, cryptocurrencies tend to attract risk-seeking investors.

One example of risk-seeking behavior is margin trading, where investors borrow money to buy financial assets. Margin use not only intensifies potential gains but also increases potential losses, especially in highly volatile markets like cryptocurrencies. The combination of crypto investing and margin borrowing may therefore represent a particularly aggressive investment strategy.

Despite the growing popularity and investment in cryptocurrencies, relatively few studies have examined the behavioral and financial factors that influence cryptocurrency investment. This study aimed to address this gap by examining the relationship between margin loans and cryptocurrency investment, using data from the 2018 and 2021 waves of the National Financial Capability Study (NFCS) Investor Survey.

Behavioral finance theory suggests that cognitive biases, such as overconfidence and the illusion of control, can lead to irrational investment decisions (Barberis & Thaler, 2003). By examining the use of margin in cryptocurrency investments, this study investigated whether these biases are more obvious among cryptocurrency investors, providing insights into their decision-making processes.

Furthermore, understanding how investors manage risk when using margins to invest in volatile assets like cryptocurrencies can contribute to the broader literature on risk management and leverage in financial markets. Margin loans, which require repayment regardless of asset performance, carry several risks, including the potential for losses exceeding initial investments and forced liquidation of securities (Financial Industry Regulatory Authority [FINRA]).

By examining the relationship between cryptocurrency investment and engaging in margin trading, the study extends the existing literature on risk-taking behavior and financial decision-making under uncertainty. The findings have significant implications for investors, financial advisors, and regulators, providing insights that could enhance investment strategies, improve risk management practices, and inform regulatory policies to ensure market stability and investor protection.

2. Literature Review

Cryptocurrency investment decisions are shaped by a range of economic, technological, behavioral, and contextual factors. Macroeconomic conditions and market trends influence how investors assess risk and potential returns, while access to digital tools and infrastructure supports broader participation in decentralized financial markets. Individual traits such as financial literacy, risk tolerance, and behavioral biases play a significant role in shaping investment decisions. Additionally, demographic and sociocultural factors contribute to differences in investor behavior, and periods of geopolitical uncertainty often lead to increased volatility and herd-like reactions. In this complex environment, margin loan usage serves as a clear indicator of risk-taking behavior, offering valuable insight into the motivations behind speculative investments in cryptocurrencies.

2.1. Economic and Market Factors

Cryptocurrency is very popular as an investment, but despite its popularity, there has been limited research on the factors affecting cryptocurrency investment. Gupta et al. (2021) found that social influence, financial literacy, facilitating conditions, performance expectancy, perceived trust, perceived usefulness, and social support positively affect the intention behind cryptocurrency investment. Prior research on crypto assets shows that, using standard demand and supply factors, it is difficult to explain the price volatility of these assets (Goczek & Skliarov, 2019; Kapar & Olmo, 2021; Qiao et al., 2020). These findings suggest the dominance of speculative trades in crypto assets. Even though cryptocurrency is highly volatile, people still invest in it, anticipating the potential for high returns. Alongside broader economic and market factors, technological accessibility and user competence are important for enabling participation in cryptocurrency markets.

2.2. Technological Readiness

Technological readiness plays an important role in shaping individuals’ ability and willingness to engage in cryptocurrency investment. It reflects individuals’ access to digital tools and their ability to effectively use emerging financial technologies. Gupta et al. (2021) highlighted that facilitating conditions such as access to supportive infrastructure and technological resources play a key role in shaping individuals’ intention to invest in cryptocurrencies. Similarly, Gunawan and Sangka (2025) showed that both financial and digital literacy significantly affect investment decisions, suggesting that technological competence is essential for participating in decentralized financial markets. Although access to technology is necessary, it does not fully explain investment behavior, which is also influenced by cognitive and psychological factors.

2.3. Behavioral and Psychological Influences

To better understand the factors influencing cryptocurrency investment, recent studies have increasingly focused on the psychological aspects of investor behavior. Makarchuk et al. (2023) and Almeida and Gonçalves (2023) emphasized that behavioral psychology, including biases such as overconfidence and herd behavior, impacts investment decisions in cryptocurrency markets. Similarly, Bland et al. (2024) highlighted how perceived enjoyment and risk perception shape cryptocurrency investment behavior, while Qi et al. (2025), using data from the National Financial Capability Study (NFCS), explored how investor confidence and advisory sources influence investment decisions. These findings align with the current paper’s focus on behavioral influences and margin-related risk-taking.

Previous literature investigating the relationship between perceived risk tolerance and cryptocurrency investment found a positive relationship (Zhao & Zhang, 2021). The current study used actual risk-taking behaviors, such as engaging with margin loans and margin calls, to examine their relationship with cryptocurrency investment.

Behavioral tendencies such as overconfidence and risk-seeking often influence individuals to engage in high-risk financial activities, including leveraged investments in volatile markets (Barber & Odean, 2002).

2.4. Margin Lending and Volatility Exposure

Cryptocurrencies have gained a reputation for providing significant returns relatively quickly (Grew, 2023). Many investors are attracted to the possibility of significant gains and believe that the possible returns offset the risks associated with volatility (Grew, 2023). Taking a margin loan and investing in cryptocurrency can be a double risk. Margin loans allow investors to borrow money to increase their potential investment returns. While leverage can help earn higher returns when investments perform well, it also increases the risk of losses (Barber et al., 2020). If the value of the cryptocurrencies decreases significantly, the investor may be required to repay the loan, even if the value of their investment has fallen below the borrowed amount. Since cryptocurrency is highly volatile and the possibility of a sudden drop is high, it may be challenging to liquidate investments quickly and cover the margin loan obligations.

The risks of liquidation and overleveraging are particularly concerning in the cryptocurrency market, where unregulated trading platforms and continuous, around-the-clock volatility increase potential losses. These concerns align with the findings from Tehrani et al. (2021), who examined credit risk in leveraged portfolios, and Halse (2010), who identified margin lending as a high-risk investment strategy.

The relationship between margin loans and asset volatility has been explored in previous studies. For instance, Ricke (2004) showed that the availability of margin loans could cause the development of a stock market bubble by stimulating investors to pay more for a stock than its fundamental value. Similarly, Fortune (2001) found a significant positive relationship between the volume of outstanding margin loans and subsequent stock price levels and volatility. These insights highlight the potential risks of margin lending, particularly when associated with highly volatile assets like cryptocurrencies. Although margin lending has been examined in the context of traditional financial markets (e.g., Fortune, 2001; Ricke, 2004), its impact on cryptocurrency investment remains unexplored in nationally representative individual-level data.

Despite being classified as a highly risky financial asset, there has been no research on the association between margin lending and cryptocurrency investment. A margin loan is considered high-risk lending (Halse, 2010) as it allows investors to borrow against securities or purchase risky assets like stocks. While studies like Tehrani et al. (2021) have used machine learning to predict credit risk in margin lending, they primarily focused on stock markets rather than cryptocurrency investments. This gap in the literature highlights the need for further exploration.

This study used margin loan usage as a proxy for risky investor behavior to examine its relationship with cryptocurrency investment. Margin loans are particularly suitable for capturing leveraged financial behavior as they are directly tied to market activity and reflect an investor’s intent to increase both potential gains and losses. Unlike personal or consumer loans, margin borrowing is inherently speculative. Previous research finds that margin investors show higher levels of overconfidence compared to cash investors (Barber et al., 2020).

2.5. Sociocultural and Demographic Factors

Sociocultural and demographic factors significantly influence cryptocurrency investment behavior. Studies have shown that younger, educated, and tech-savvy individuals are more likely to invest in cryptocurrencies (Auer & Tercero-Lucas, 2022). Although investment willingness generally decreases with age, this decline is more pronounced among men, while women tend to exhibit more stable investment behavior over time (Honold & Oh, 2025). Krause (2024) further notes that Millennials and Gen Z are particularly inclined toward cryptocurrency due to their familiarity with digital technologies and openness to non-traditional financial systems. Additionally, investors who use social media as a source of investment information are more likely to invest in cryptocurrencies and demonstrate continued interest in doing so in the future (Kim & Fan, 2025).

2.6. Geopolitical Context

Global political and economic uncertainty plays an important role in influencing cryptocurrency investment behavior. Wanidwaranan et al. (2025) found that geopolitical risk substantially increases herd behavior among crypto investors, especially during periods of crisis such as the COVID-19 pandemic and the Russia–Ukraine war. In such periods of uncertainty, investors are more likely to engage in fear-driven, imitative trading. Similarly, Yi Fang et al. (2024) found that extreme geopolitical events lead to sharp increases in cryptocurrency market volatility, suggesting that higher geopolitical tension not only increases investment activity but also increases overall market risk.

Although various contextual and demographic factors provide important insights, this study focused specifically on the behavioral dimension of risk-taking, as measured by margin loan usage.

2.7. Contribution of the Current Study

Although a few studies, such as Strych (2022), examined how margin trading and short selling impact Bitcoin market efficiency in retail investor-driven settings, empirical research on this topic remains limited. To date, no study has systematically examined the relationship between individual-level margin behavior and cryptocurrency investment. This study addresses that gap by utilizing the 2018 and 2021 waves of the National Financial Capability Study (NFCS) to examine whether real-world risk-taking behaviors, specifically, margin trading, are associated with cryptocurrency investment.

3. Theoretical Framework

The theoretical foundation for the current study is built on behavioral finance theory, which emphasizes how cognitive biases and emotional influences often lead to irrational financial behaviors. In the context of cryptocurrency investment, where volatility and speculative potential are high, behavioral biases play a crucial role in shaping investors’ decision-making processes. Although the study did not directly assess specific behavioral biases such as overconfidence or the illusion of control, it examined observable risk-taking behaviors, like margin trading, that may indicate these cognitive tendencies. Behavioral finance theory suggests that investors who trade on margin in volatile markets may do so because they expect high returns or underestimate risk; these perceptions were not directly measured in this study.

One of the most relevant biases in this context is overconfidence (Barber & Odean, 2002), where investors tend to overestimate their ability to predict price movements. This perception can result in riskier strategies, such as margin trading, especially in speculative markets like cryptocurrencies. Similarly, the illusion of control (Langer, 1975) can lead investors to overestimate their ability to influence uncertain market outcomes. This perception can increase the likelihood of taking leveraged positions, as investors may feel overly confident in their ability to manage risks.

Additionally, herding behavior is common in cryptocurrency markets (Papadamou et al., 2021; Shrotryia & Kalra, 2022), where investors tend to follow others, driven by fear of missing out (FOMO) or the desire to copy successful strategies. By investigating the relationship between cryptocurrency investment and margin trading, this study aimed to explore how risk-seeking behavior aligns with the principles of behavioral finance theory.

3.1. Hypotheses

The relationship between purchasing securities on margin and cryptocurrency investment is expected to be positive. Buying securities on margin demonstrates risky financial behavior. Cryptocurrency is a risky financial asset because of its high volatility. Prior research found that risk tolerance and perceived risk play important roles in investing in risky financial assets (Cho & Lee, 2006; Li & Qian, 2018).

Similarly, the relationship between experiencing a margin call and cryptocurrency investment is expected to be positive. A margin call is triggered when the investor’s equity falls below the required maintenance level relative to the market value of securities. Getting a margin call is a sign that the investor has invested in highly volatile assets.

Based on these considerations, the study proposes the following hypotheses:

H1: 

Individuals who have taken a margin loan are more likely to invest in cryptocurrencies.

H2: 

Individuals who have received a margin call are more likely to invest in cryptocurrencies.

3.2. Data

This study used the state-by-state and investor survey data from the 2018 and 2021 waves of the National Financial Capability Study (NFCS). The NFCS is a large-scale, multi-year project funded by the Financial Industry Regulatory Authority (FINRA) Investor Education Foundation. It is designed to evaluate the financial capability of U.S. adults and provides nationally representative data on financial knowledge, behavior, and decision-making. For the current study, we combined the 2018 and 2021 datasets to perform a pooled cross-sectional study to maximize the number of observations for our dependent variable, cryptocurrency investment. The NFCS state-by-state survey is a large-scale, multi-year project that measures Americans’ financial capability. The NFCS provides information on Americans’ financial capability, drawing from a sample of over 25,000 respondents.

To obtain more in-depth insight into investment decisions, the FINRA Foundation commissioned a separate follow-up survey of investors as part of the 2018 and 2021 NFCS. This Investor Survey focuses on individuals with investments outside of retirement accounts and includes in-depth questions on risk-taking, margin use, and cryptocurrency investment. It is particularly useful for analyzing investor behavior related to speculative assets. The Investor Survey explores relationships with investment brokers and advisors, understanding and perceptions of fees charged for investment services, usage of investment information sources, attitudes towards investing in general, and investor knowledge. The NFCS study aims to benchmark key indicators of financial capability and evaluate how these indicators vary with underlying demographic, behavioral, attitudinal, and financial literacy characteristics. In each wave, the study deepened the exploration of topics that are highly relevant today.

The 2018 investor survey was administered to 2003 investors who completed the 2018 NFCS state-by-state survey. The 2021 investor survey was administered to 2824 investors who completed the 2021 NFCS state-by-state survey. After combining the 2018 and 2021 datasets, the total number of observations was 4827. These two waves were chosen because they are the most recent NFCS Investor Surveys that provide comprehensive data on both cryptocurrency investment and margin activity. At the time of this study, no newer waves with similar information were available. All respondents in the investor survey have investments outside of retirement accounts. Weights are provided to make the sample representative of the United States population. Since the 2021 data were collected during the COVID-19 pandemic, investor behavior may have been affected by market uncertainty and volatility. Although the analysis does not explicitly control for COVID-related effects, a year dummy variable for 2021 is included to control for systemic differences between the 2018 and 2021 survey waves. However, several observations from the sample were recorded as missing. Responses such as “Don’t know” and “Prefer not to say” are omitted from the dependent and explanatory variables.

The dependent variable used in this study is the decision to invest in cryptocurrencies. This variable was measured based on responses to the question, “Have you invested in cryptocurrencies, either directly or through a fund that invests in cryptocurrencies?” Respondents who stated “Don’t know” or “Prefer not to say” were excluded. The value for the dependent variable is 1 if the respondents answered “yes,” and 0 if the respondents answered “no,” with “no” being the reference category.

The explanatory variables used in this study included margin loan and margin call. The margin loan variable was measured based on responses to the following question: “Have you made any securities purchases on margin?” The margin loan variable is a dummy variable. The value for this variable is 1 if the respondents answer “yes,” and 0 if the respondents answer “no,” with “no” being the reference category. Respondents who stated “Don’t know” or “Prefer not to say” were excluded. The second explanatory variable is a margin call. The margin call variable was measured based on responses to the following question after responding ‘yes’ to the margin loan question: “Have you ever had a margin call?” The margin call variable is a dummy variable. The value for this variable is 1 if the respondents answered “yes” and 0 if the respondents answered “no”, with “no” being the reference category. Respondents who stated “Don’t know” or “Prefer not to say” were also excluded.

The study also included demographic and economic variables as control variables. The model included gender, ethnicity, marital status, age, homeownership, education, annual income, employment status, and investments. Gender was a dichotomous variable with “Female” as the reference category. Ethnicity was reported as either “White Alone (Non-Hispanic)” or “Non-White.” Ethnicity was included as a control variable because differences in financial behavior, risk perception, and access to services across racial and ethnic groups may affect cryptocurrency investment (Hudson et al., 2018; Elu & Williams, 2023). Marital status is a dichotomous variable with “Not married” as the reference category. The respondent’s age was a continuous variable that ranged from 18 to 96. Homeownership was a dichotomous variable with “Not owning a home” as the reference category.

The respondents report their level of education as less than high school, high school regular, high school GED, some college, associate degree, bachelor’s degree, of postgraduate degree. Each response option was represented in the model by a dummy variable, with “postgraduate degree” as the reference category. The respondents also report their annual income in ten different ranges: “Less than USD 15,000,” “USD 15,000 to USD 25,000,” “USD 25,000 to USD 35,000,” “USD 35,000 to USD 50,000,” “USD 50,000 to USD 75,000,” “USD 75,000 to USD 100,000,” “USD 100,000 to USD 150,000,” “USD 150,000 to USD 200,000,” “USD 200,000 to USD 300,000,” and “More than USD 300,000.” The study created nine different income dummies, with “less than USD 15,000” as the reference category.

Employment status was assessed using seven dummy variables: self-employed, work full time, work part time, full-time student, permanently sick or disabled, unemployed, and retired. The reference category was homemaker. Six dummy variables were created for investment in non-retirement accounts, which represented the following categories: “Less than USD 25,000,” “USD 25,000 to USD 50,000,” “USD 50,000 to USD 100,000,” “USD 250,000 to USD 500,000,” “USD 500,000 to USD 1 million,” and “More than USD 1 million,” using “USD 100,000 to USD 250,000” as the reference category. “Don’t know” and “Prefer not to say” responses in any demographic or economic variables were excluded from the analysis. A summary of all the key variables, along with their definitions and coding structures, is provided in Appendix A for reference.

Table 1. Summary statistics (whole sample).

	Mean	Std. Err.
Cryptocurrency Investments	0.1344	0.0049
Margin Loan	0.0870	0.0040
Margin Call	0.0453	0.0029
Male	0.5898	0.0071
White	0.8346	0.0053
Married	0.6515	0.0068
Age	35.6987	0.4141
Homeownership	0.8479	0.0052
Education Level
Less than Highschool	0.0024	0.0007
Highschool Regular	0.0988	0.0043
Highschool GED	0.0282	0.0024
Some College	0.1808	0.0055
Associate Degree	0.0972	0.0043
Bachelor’s Degree	0.3569	0.0069
Postgraduate Degree	0.2355	0.0061
Income Level
Less than USD 15,000	0.0178	0.0019
USD 15,000 to USD 25,000	0.0393	0.0028
USD 25,000 to USD 35,000	0.0576	0.0034
USD 35,000 to USD 50,000	0.1096	0.0045
USD 50,000 to USD 75,000	0.2099	0.0059
USD 75,000 to USD 100,000	0.1920	0.0057
USD 100,000 to USD 150,000	0.2181	0.0059
USD 150,000 to USD 200,000	0.1094	0.0045
USD 200,000 to USD 300,000	0.0331	0.0026
More than USD 300,000	0.0131	0.0016
Work Status
Homemaker	0.0296	0.0024
Self-Employed	0.0824	0.0039
Work Full Time	0.3665	0.0069
Work Part Time	0.0694	0.0036
Full-Time Student	0.0083	0.0013
Permanently Sick or Disabled	0.0106	0.0015
Unemployed	0.0217	0.0021
Retired	0.4114	0.0071
Investment in Non-Retirement Accounts
Less than USD 25,000	0.2080	0.0058
USD 25,000 to USD 50,000	0.0804	0.0039
USD 50,000 to USD 100,000	0.1295	0.0048
USD 100,000 to USD 250,000	0.1699	0.0054
USD 250,000 to USD 500,000	0.1499	0.0051
USD 500,000 to USD 1 million	0.1067	0.0044
More than USD 1M	0.1033	0.0044
The year 2021	0.5850	0.0071

presents the descriptive statistics for the dependent and explanatory variables for the whole sample. Approximately 13 percent of the respondents held cryptocurrency investments, about 9 percent used margin loans, and about 5 percent received margin calls. On average, 59 percent of the respondents were male, 83 percent were white, and 65 percent were married. Over 60 percent of the respondents had either bachelor’s or postgraduate degrees. About 56 percent had an income of more than USD 75,000, but over 64 percent had an investment of USD 250,000 or less. The average age of the respondents was 36 years old. Approximately 41 percent of the respondents were retired, 36 percent worked full time, and 85 percent owned their homes.

Table 2. Summary statistics for the participants who said yes or no to investments in cryptocurrencies.

	Yes		No
	Mean	Std. Err.	Mean	Std. Err.
Margin Loan	0.3020	0.0180	0.0536	0.0035
Margin Call	0.2219	0.0163	0.0180	0.0021
Male	0.7273	0.0175	0.5685	0.0077
White	0.7488	0.0170	0.8480	0.0056
Married	0.5948	0.0193	0.6604	0.0073
Age	33.6995	0.8376	36.0093	0.4603
Homeownership	0.7288	0.0175	0.8664	0.0053
Education Level
Less than Highschool	0.0015	0.0015	0.0026	0.0008
Highschool Regular	0.1017	0.0119	0.0984	0.0046
Highschool GED	0.0308	0.0068	0.0278	0.0025
Some College	0.2111	0.0160	0.1762	0.0059
Associate Degree	0.1186	0.0127	0.0938	0.0045
Bachelor’s Degree	0.3467	0.0187	0.3585	0.0074
Postgraduate Degree	0.1895	0.0154	0.2427	0.0066
Income Level
Less than USD 15,000	0.0354	0.0073	0.0151	0.0019
USD 15,000 to USD 25,000	0.0524	0.0088	0.0373	0.0029
USD 25,000 to USD 35,000	0.0801	0.0107	0.0541	0.0035
USD 35,000 to USD 50,000	0.1171	0.0126	0.1084	0.0048
USD 50,000 to USD 75,000	0.1849	0.0153	0.2137	0.0063
USD 75,000 to USD 100,000	0.1710	0.0148	0.1953	0.0061
USD 100,000 to USD 150,000	0.2096	0.0160	0.2195	0.0064
USD 150,000 to USD 200,000	0.0847	0.0109	0.1132	0.0049
USD 200,000 to USD 300,000	0.0493	0.0085	0.0306	0.0027
More than USD 300,000	0.0154	0.0048	0.0127	0.0017
Work Status
Homemaker	0.0231	0.0059	0.0306	0.0027
Self-Employed	0.1109	0.0123	0.0780	0.0042
Work Full Time	0.6302	0.0190	0.3255	0.0073
Work Part Time	0.0647	0.0097	0.0701	0.0040
Full-Time student	0.0185	0.0053	0.0067	0.0013
Permanently Sick or Disabled	0.0169	0.0051	0.0096	0.0015
Unemployed	0.0401	0.0077	0.0189	0.0021
Retired	0.0955	0.0115	0.4605	0.0077
Investment in Non-Retirement Accounts
Less than USD 25,000	0.3744	0.0190	0.1821	0.0060
USD 25,000 to USD 50,000	0.0955	0.0115	0.0780	0.0042
USD 50,000 to USD 100,000	0.1402	0.0136	0.1278	0.0052
USD 100,000 to USD 250,000	0.1649	0.0146	0.1707	0.0058
USD 250,000 to USD 500,000	0.1109	0.0123	0.1561	0.0056
USD 500,000 to USD 1 million	0.0539	0.0089	0.1149	0.0049
More than USD 1M	0.0447	0.0081	0.1125	0.0049
The year 2021	0.7581	0.0168	0.5582	0.0077

Data from the 2018 and 2021 waves of the National Financial Capability Study (NFCS) combines the state-by-state survey and the investor survey data. Both surveys were commissioned by the FINRA Investor Education Foundation and were conducted by Applied Research and Consulting LLC (ARC).

presents the descriptive statistics for those who said “Yes—C(1)” and “No—C(0)” to owning or holding cryptocurrency investments. Table 2 and Figure 1 show that among the respondents who said “yes” to owning cryptocurrencies, 30% used a margin loan; among those who said “no”, only 5% used a margin loan.

Similarly, among the respondents who said “yes” to owning cryptocurrencies and using a margin loan, about 22% received a margin call; among those who said “no” to owning cryptocurrencies, only 2% received a margin call.

Table 2 also provides the descriptive statistics for the other demographic variables for the respondents who responded “Yes—C(1)” or “No—C(0)” to having cryptocurrency investments. Among the respondents who said “yes” to owning cryptocurrencies, about 73% were male; among those who said “no” to owning cryptocurrencies, only 57% were male.

When comparing respondents who worked full time, over 63% said “yes” to having cryptocurrency investments, and only about 33% said “no” to having cryptocurrency investments.

3.3. Model

The study ran two probit models for the margin loan and margin call variables to examine the relationship between margin and cryptocurrency investment. The dependent variable in both models was whether the respondents invested in cryptocurrencies directly or through a fund that invests in cryptocurrencies. The control variables were also the same across the two models. The probit model is appropriate because the dependent variable is dichotomous. The probit model follows a cumulative normal distribution and has a latent variable containing a linear prediction. It is important to note that, due to the cross-sectional nature of the data, the probit models estimate associations rather than causal relationships. The following estimated probit models were used.

Model 1:

$$\mathrm{Crypto}{{\mathrm{investment}}_{\mathrm{i}}}^{\ast }={\mathsf{\beta }}_0+{\mathsf{\beta }}_1{\mathrm{Margin}\mathrm{loan}}_{\mathrm{i}}+{}_{\mathsf{\gamma }}\mathrm{X}_{\mathrm{i}}+{\mathrm{v}}_{\mathrm{i}}$$

$${\mathrm{Crypto}\mathrm{investment}}_{\mathrm{i}}=1\mathrm{if}\mathrm{Crypto}{{\mathrm{investment}}_{\mathrm{i}}}^{\ast }>0$$

$${\mathrm{Crypto}\mathrm{investment}}_{\mathrm{i}}=0\mathrm{if}\mathrm{Crypto}{{\mathrm{investment}}_{\mathrm{i}}}^{\ast }\le 0$$

Model 2:

$$\mathrm{Crypto}{{\mathrm{investment}}_{\mathrm{i}}}^{\ast }={\mathsf{\beta }}_0+{\mathsf{\beta }}_1{\mathrm{Margin}\mathrm{call}}_{\mathrm{i}}+{}_{\mathsf{\gamma }}\mathrm{X}_{\mathrm{i}}+{\mathrm{v}}_{\mathrm{i}}$$

$${\mathrm{Crypto}\mathrm{investment}}_{\mathrm{i}}=1\mathrm{if}\mathrm{Crypto}{{\mathrm{investment}}_{\mathrm{i}}}^{\ast }>0$$

$${\mathrm{Crypto}\mathrm{investment}}_{\mathrm{i}}=0\mathrm{if}\mathrm{Crypto}{{\mathrm{investment}}_{\mathrm{i}}}^{\ast }\le 0$$

where $\mathrm{Crypto}{{\mathrm{investment}}_{\mathrm{i}}}^{\ast }$ in both models is the unobserved variable indicating respondent _i’s cryptocurrency investment. ${\mathrm{Crypto}\mathrm{investment}}_{\mathrm{i}}$ in both models is the observed variable that takes a value of 1 if the respondent had a cryptocurrency investment and 0 otherwise. β₀ in both models is the intercept. β₁ in model 1 is the association between respondents having margin loans and cryptocurrency investment. ${\mathrm{Margin}\mathrm{loan}}_{\mathrm{i}}$ in model 1 indicates whether the respondents made any security purchases on margin. β₁ in model 2 is the association between respondents having margin calls and cryptocurrency investment. ${\mathrm{Margin}\mathrm{call}}_{\mathrm{i}}$ in model 2 indicates whether the respondents received any margin calls. The matrix X_i in both models contains all the other explanatory variables related to the cryptocurrency investment decision. These explanatory variables include gender, ethnicity, marital status, age, homeownership, level of education, income, work status, and investment in non-retirement accounts. $\mathsf{\gamma }$ is a vector of the corresponding slope parameters for gender, ethnicity, marital status, age, homeownership, level of education, income, work status, and investment in non-retirement accounts. v_i in both models is the error term that is assumed to follow a standard normal distribution.

4. Results

Table 3. Relationship between margin loan and cryptocurrency investment: probit model.

	Marginal Effect	Std. Err.	p Value	95% Conf. Interval
Margin Loan	0.1691 ***	0.0122	0.0000	0.1452	0.1930
Male	0.0578 ***	0.0097	0.0000	0.0387	0.0769
White	−0.0122	0.0112	0.2760	−0.0343	0.0098
Married	0.0091	0.0103	0.3810	−0.0112	0.0293
Age	−0.0042 ***	0.0004	0.0000	−0.0050	−0.0035
Homeownership	−0.0090	0.0126	0.4760	−0.0336	0.0157
Education Level (Versus Postgraduate Degree)
Less than High School	−0.0568	0.0930	0.5420	−0.2391	0.1256
High School Regular	0.0220	0.0176	0.2090	−0.0124	0.0564
High School GED	0.0217	0.0276	0.4320	−0.0324	0.0757
Some College	0.0435 ***	0.0141	0.0020	0.0159	0.0712
Associate Degree	0.0305 *	0.0161	0.0590	−0.0011	0.0621
Bachelor’s Degree	0.0032	0.0119	0.7860	−0.0200	0.0265
Income Level (Versus Less than USD 15,000)
USD 15,000 to USD 25,000	−0.0013	0.0333	0.9690	−0.0666	0.0640
USD 25,000 to USD 35,000	0.0167	0.0307	0.5860	−0.0434	0.0769
USD 35,000 to USD 50,000	0.0051	0.0293	0.8620	−0.0524	0.0626
USD 50,000 to USD 75,000	−0.0177	0.0280	0.5280	−0.0726	0.0373
USD 75,000 to USD 100,000	−0.0204	0.0289	0.4810	−0.0770	0.0362
USD 100,000 to USD 150,000	−0.0114	0.0292	0.6960	−0.0685	0.0458
USD 150,000 to USD 200,000	−0.0243	0.0311	0.4350	−0.0852	0.0367
USD 200,000 to USD 300,000	0.0202	0.0363	0.5780	−0.0510	0.0914
More than USD 300,000	0.0007	0.0462	0.9880	−0.0899	0.0913
Work Status (Versus Homemaker)
Self-Employed	0.0185	0.0296	0.5310	−0.0394	0.0765
Work Full Time	0.0383	0.0265	0.1480	−0.0136	0.0903
Work Part Time	0.0011	0.0303	0.9700	−0.0582	0.0605
Fulltime Student	0.0132	0.0524	0.8010	−0.0896	0.1160
Permanently Sick or Disabled	0.0482	0.0440	0.2730	−0.0380	0.1345
Unemployed	0.0279	0.0363	0.4430	−0.0433	0.0990
Retired	−0.0486 *	0.0277	0.0790	−0.1029	0.0057
Investment in Non-Retirement Accounts (Versus Investment Value USD 100,000-USD 250,000
Less than USD 25,000	0.0571 ***	0.0130	0.0000	0.0315	0.0826
USD 25,000 to USD 50,000	0.0311 *	0.0174	0.0740	−0.0031	0.0652
USD 50,000 to USD 100,000	0.0019	0.0147	0.8990	−0.0270	0.0308
USD 250,000 to USD 500,000	0.0039	0.0147	0.7920	−0.0250	0.0327
USD 500,000 to USD 1 Million	−0.0352 *	0.0188	0.0610	−0.0721	0.0017
More than USD 1 Million	−0.0171	0.0210	0.4150	−0.0583	0.0240
Year 2021 (Versus 2018)	0.2979 ***	0.0207	0.0000	0.2573	0.3386
N	4827
Pseudo R2	0.2021

Notes: This analysis used data from the FINRA Foundation 2018 and 2021 NFCS state-by-state and investor survey datasets. Marginal effect values are shown alongside the standard errors. Survey weights were applied. *** indicates significance at the 1% level; * indicates significance at the 10% level.

and

Table 4. Relationship between margin call and cryptocurrency investment: probit model.

	Marginal Effect	Std. Err.	p Value	95% Conf. Interval
Margin Call	0.2316 ***	0.0158	0.0000	0.2006	0.2626
Male	0.0576 ***	0.0097	0.0000	0.0387	0.0766
White	−0.0161	0.0110	0.1440	−0.0377	0.0055
Married	0.0109	0.0102	0.2840	−0.0090	0.0308
Age	−0.0042 ***	0.0004	0.0000	−0.0050	−0.0034
Homeownership	−0.0138	0.0123	0.2600	−0.0378	0.0102
Education Level (Versus Postgraduate Degree)
Less than High School	−0.0920	0.0981	0.3480	−0.2844	0.1003
High School Regular	0.0200	0.0174	0.2500	−0.0141	0.0540
High School GED	0.0277	0.0274	0.3130	−0.0261	0.0814
Some College	0.0423 ***	0.0140	0.0020	0.0149	0.0697
Associate Degree	0.0337 **	0.0159	0.0340	0.0025	0.0650
Bachelor’s Degree	0.0038	0.0118	0.7480	−0.0193	0.0269
Income Level (Versus Less than USD 15,000)
USD 15,000 to USD 25,000	0.0009	0.0336	0.9800	−0.0649	0.0667
USD 25,000 to USD 35,000	0.0239	0.0310	0.4410	−0.0369	0.0847
USD 35,000 to USD 50,000	0.0085	0.0300	0.7770	−0.0502	0.0672
USD 50,000 to USD 75,000	−0.0126	0.0287	0.6610	−0.0689	0.0437
USD 75,000 to USD 100,000	−0.0195	0.0295	0.5090	−0.0772	0.0383
USD 100,000 to USD 150,000	−0.0058	0.0298	0.8450	−0.0642	0.0525
USD 150,000 to USD 200,000	−0.0237	0.0316	0.4530	−0.0857	0.0382
USD 200,000 to USD 300,000	0.0263	0.0366	0.4730	−0.0455	0.0981
More than USD 300,000	0.0034	0.0467	0.9420	−0.0882	0.0949
Work Status (Versus Homemaker)
Self-Employed	0.0223	0.0293	0.4470	−0.0351	0.0796
Work Full Time	0.0455 *	0.0262	0.0820	−0.0058	0.0969
Work Part Time	0.0084	0.0300	0.7790	−0.0504	0.0673
Full-time Student	0.0143	0.0516	0.7810	−0.0868	0.1154
Permanently Sick or Disabled	0.0563	0.0425	0.1850	−0.0270	0.1395
Unemployed	0.0288	0.0362	0.4270	−0.0422	0.0998
Retired	−0.0431	0.0274	0.1150	−0.0968	0.0106
Investment in Non-Retirement Accounts (Versus Investment Value USD 100,000-USD 250,000
Less than USD 25,000	0.0521 ***	0.0128	0.0000	0.0270	0.0772
USD 25,000 to USD 50,000	0.0251	0.0176	0.1530	−0.0093	0.0595
USD 50,000 to USD 100,000	0.0025	0.0147	0.8660	−0.0264	0.0314
USD 250,000 to USD 500,000	0.0024	0.0146	0.8720	−0.0262	0.0309
USD 500,000 to USD 1 Million	−0.0359 *	0.0190	0.0590	−0.0731	0.0013
More than USD 1 Million	−0.0195	0.0206	0.3440	−0.0598	0.0208
Year 2021 (Versus 2018)	0.2899 ***	0.0209	0.0000	0.2491	0.3308
N	4827
Pseudo R2	0.2143

Notes: This analysis used data from the FINRA Foundation 2018 and 2021 NFCS state-by-state and investor survey datasets. Marginal effect values are shown alongside the standard errors. Survey weights were applied. *** indicates significance at the 1% level; ** indicates significance at the 5% level; * indicates significance at the 10% level.

show the marginal effects and standard errors for the two separate probit models. To verify that multicollinearity is not misrepresenting the estimates, we calculated the Variance Inflation Factors (VIFs) for all the independent variables (Appendix B). Most VIF values were well below the threshold of 10. A few income-related dummy variables slightly exceeded this value, which was expected because they are mutually exclusive categories. Overall, the diagnostics showed no serious multicollinearity concerns. Table 3 shows the marginal effects and standard errors for probit model 1 (relation between margin and cryptocurrency ownership). Consistent with the hypothesis, Table 3 shows that the relationship between margin loans and cryptocurrency investment was positive and statistically significant. Specifically, investors with a margin loan were 17 percentage points more likely to invest in cryptocurrency compared to those without a margin loan.

Similarly, Table 4 shows the marginal effects and standard errors for probit model 2 (relation between margin call and cryptocurrency ownership). The results showed a positive relationship between margin calls and cryptocurrency investment. Individuals who experienced a margin call were 23 percentage points more likely to invest in cryptocurrency than those who did not receive a margin call.

Regarding the demographic and economic variables, both models showed consistent results. Table 3 and Table 4 show that the relationship between being male and cryptocurrency investment was positive and statistically significant. Across both models, the results for respondents’ age were consistent. They showed that age was negatively related to cryptocurrency investment. In terms of education, compared to those with a postgraduate degree, individuals with some college education and an associate degree were more likely to invest in cryptocurrencies.

In both models, the results for individuals’ total value of investments in non-retirement accounts showed that having less than USD 25,000 in non-was is positively related to cryptocurrency investment. Across both models, having a USD 500,000 to USD 1 million investment in a non-retirement account was associated negatively with cryptocurrency investment.

Finally, the year 2021 was positively associated with cryptocurrency investment compared to 2018, indicating a higher likelihood of investment in 2021.

Sub-Analysis: Interaction Between Males with a Margin Loan and Males with a Margin Call

The study found that males are more likely to invest in cryptocurrency than females. An individual with a margin loan or margin call has a higher probability of investing in cryptocurrency than an individual without a margin loan or margin call. To build on this, we further investigated the outcomes when interacting gender (male) with margin loans and margin calls, specifically regarding their relationship with cryptocurrency investments.

Individuals with margin loans usually have a higher risk tolerance and potentially have more aggressive investment strategies. As such, investigating males who engage in these behaviors can provide insights into how leverage influences cryptocurrency investment decisions. Examining the interaction between males with margin loans and margin calls in the context of cryptocurrency investments is important because it helps to discover gender-specific risk behaviors, and the effect of leverage on investment decisions, and provides valuable insights for regulatory policies and risk management strategies.

Table 5. Summary statistics (interaction between male and margin loan and between male and margin call).

	Crypto Investment (Yes)	Crypto Investment (No)
Margin Loan and Male	44.48%	55.52%
No Margin loan and Male	13.30%	86.70%
Margin Loan and Female	52.07%	47.93%
No Margin Loan and Female	6.13%	93.87%
Male	16.58%	83.42%
Female	8.94%	91.06%
Margin Loan	46.67%	53.33%
No Margin Loan	10.28%	89.72%
Margin Call and Male	63.69%	36.30%
No Margin Call and Male	13.83%	86.17%
Margin Call and Female	70.96%	29.04%
No Margin Call and Female	6.93%	93.07%
Male	16.58%	83.42%
Female	8.94%	91.06%
Margin Call	65.75%	34.25%
No Margin Call	10.96%	89.04%

and Figure 2 present detailed descriptive statistics for the interaction terms. Interestingly, when we interacted the variables “male” with “margin loan” and “margin call,” the results reversed, showing a negative association with cryptocurrency investment.

Table 6. Relationship between margin loan and cryptocurrency investment (interaction between male and margin loan): probit model.

	Marginal Effect	Std. Err.	p Value	95% Conf. Interval
Male*Margin Loan	−0.1073 ***	0.0277	0.0000	−0.1617	−0.0529
Margin Loan	0.2710 ***	0.0231	0.0000	0.2258	0.3163
Male	0.0747 ***	0.0109	0.0000	0.0534	0.0961
White	−0.0148	0.0115	0.1960	−0.0373	0.0077
Married	−0.0011	0.0106	0.9170	−0.0219	0.0197
Homeownership	−0.0272 **	0.0124	0.0290	−0.0516	−0.0028
Education Level (Versus Postgraduate Degree)
Less than High School	−0.0695	0.0965	0.4710	−0.2586	0.1196
High School Regular	0.0184	0.0177	0.2980	−0.0162	0.0530
High School GED	0.0196	0.0277	0.4780	−0.0346	0.0739
Some College	0.0408 ***	0.0141	0.0040	0.0132	0.0684
Associate Degree	0.0279 *	0.0161	0.0830	−0.0037	0.0595
Bachelor’s degree	0.0011	0.0118	0.9250	−0.0221	0.0243
Income Level (Versus Less than USD 15,000)
USD 15,000 to USD 25,000	−0.0208	0.0331	0.5300	−0.0856	0.0441
USD 25,000 to USD 35,000	0.0026	0.0307	0.9310	−0.0575	0.0628
USD 35,000 to USD 50,000	−0.0084	0.0292	0.7750	−0.0657	0.0489
USD 50,000 to USD 75,000	−0.0344	0.0281	0.2210	−0.0895	0.0207
USD 75,000 to USD 100,000	−0.0336	0.0290	0.2460	−0.0905	0.0232
USD 100,000 to USD 150,000	−0.0230	0.0294	0.4350	−0.0806	0.0347
USD 150,000 to USD 200,000	−0.0327	0.0316	0.3010	−0.0946	0.0293
USD 200,000 to USD 300,000	0.0016	0.0364	0.9650	−0.0697	0.0728
More than USD 300,000	−0.0150	0.0491	0.7600	−0.1113	0.0813
Work Status (Versus Homemaker)
Self-Employed	0.0127	0.0306	0.6790	−0.0473	0.0727
Work Full Time	0.0514 *	0.0277	0.0630	−0.0029	0.1057
Work Part Time	−0.0087	0.0312	0.7800	−0.0698	0.0524
Full-Time Student	0.0538	0.0502	0.2840	−0.0446	0.1522
Permanently Sick or Disabled	0.0271	0.0454	0.5510	−0.0619	0.1161
Unemployed	0.0282	0.0372	0.4480	−0.0447	0.1012
Retired	−0.1070 ***	0.0286	0.0000	−0.1630	−0.0509
Investment in Non-Retirement Accounts (Versus Investment Value USD 100,000-USD 250,000
Less than USD 25,000	0.0766 ***	0.0132	0.0000	0.0508	0.1024
USD 25,000 to USD 50,000	0.0357 **	0.0181	0.0480	0.0003	0.0711
USD 50,000 to USD 100,000	0.0102	0.0151	0.4990	−0.0194	0.0397
USD 250,000 to USD 500,000	0.0032	0.0155	0.8360	−0.0272	0.0336
USD 500,000 to USD 1 Million	−0.0372 *	0.0195	0.0560	−0.0753	0.0010
More than USD 1 Million	−0.0347	0.0222	0.1190	−0.0783	0.0089
Year 2021 (Versus 2018)	0.1101 ***	0.0099	0.0000	0.0907	0.1294
N	4827
Pseudo R2	0.2061

Notes: This analysis used data from the FINRA Foundation 2018 and 2021 NFCS state-by-state and investor survey datasets. Marginal effect values are shown alongside the standard errors. Survey weights were applied. *** indicates significance at the 1% level; ** indicates significance at the 5% level; * indicates significance at the 10% level.

shows that males with a margin loan had a 0.11 lower probability of investing in cryptocurrency than would be expected based on the separate effects of being male and having a margin loan. Similarly,

Table 7. Relation between margin call and cryptocurrency investment (interaction between male and margin call): probit model.

	Marginal Effect	Std. Err.	P Value	95% Conf. Interval
Male*Margin Call	−0.0859 **	0.0386	0.0260	−0.1615	−0.0103
Margin Call	0.3239 ***	0.0328	0.0000	0.2597	0.3882
Male	0.0640 ***	0.0103	0.0000	0.0438	0.0843
White	−0.0205 *	0.0113	0.0690	−0.0427	0.0016
Married	0.0022	0.0105	0.8380	−0.0184	0.0227
Homeownership	−0.0310 **	0.0122	0.0110	−0.0549	−0.0071
Education Level (Versus Postgraduate Degree)
Less than High School	−0.1020	0.1017	0.3160	−0.3013	0.0972
High School Regular	0.0181	0.0174	0.3000	−0.0161	0.0522
High School GED	0.0264	0.0275	0.3370	−0.0275	0.0804
Some College	0.0412 ***	0.0140	0.0030	0.0138	0.0686
Associate Degree	0.0330 **	0.0159	0.0380	0.0018	0.0642
Bachelor’s Degree	0.0024	0.0118	0.8360	−0.0206	0.0255
Income Level (Versus Less than USD 15,000)
USD 15,000 to USD 25,000	−0.0188	0.0336	0.5770	−0.0847	0.0472
USD 25,000 to USD 35,000	0.0110	0.0314	0.7250	−0.0505	0.0725
USD 35,000 to USD 50,000	−0.0074	0.0303	0.8060	−0.0667	0.0519
USD 50,000 to USD 75,000	−0.0293	0.0292	0.3160	−0.0866	0.0280
USD 75,000 to USD 100,000	−0.0335	0.0300	0.2640	−0.0923	0.0253
USD 100,000 to USD 150,000	−0.0185	0.0304	0.5430	−0.0780	0.0411
USD 150,000 to USD 200,000	−0.0336	0.0324	0.3000	−0.0972	0.0299
USD 200,000 to USD 300,000	0.0053	0.0370	0.8870	−0.0672	0.0778
More than USD 300,000	−0.0079	0.0498	0.8730	−0.1055	0.0896
Work Status (Versus Homemaker)
Self-Employed	0.0199	0.0299	0.5060	−0.0387	0.0786
Work Full Time	0.0627 **	0.0271	0.0210	0.0097	0.1157
Work Part Time	0.0019	0.0307	0.9510	−0.0583	0.0621
Full-Time Student	0.0642	0.0481	0.1820	−0.0301	0.1584
Permanently Sick or Disabled	0.0390	0.0439	0.3740	−0.0471	0.1251
Unemployed	0.0330	0.0366	0.3680	−0.0389	0.1048
Retired	−0.0968 ***	0.0279	0.0010	−0.1515	−0.0422
Investment in Non-Retirement Accounts (Versus Investment Value USD 100,000–USD 250,000
Less than USD 25,000	0.0714 ***	0.0130	0.0000	0.0460	0.0968
USD 25,000 to USD 50,000	0.0292	0.0183	0.1110	−0.0067	0.0651
USD 50,000 to USD 100,000	0.0111	0.0151	0.4630	−0.0185	0.0407
USD 250,000 to USD 500,000	0.0016	0.0153	0.9150	−0.0284	0.0317
USD 500,000 to USD 1 Million	−0.0388 **	0.0195	0.0460	−0.0769	−0.0006
More than USD 1 Million	−0.0374 *	0.0218	0.0870	−0.0802	0.0054
Year 2021 (Versus 2018)	0.1002 ***	0.0098	0.0000	0.0810	0.1193
N	4827
Pseudo R2	0.2157

Notes: This analysis used data from the FINRA Foundation 2018 and 2021 NFCS state-by-state and investor survey datasets. Marginal effect values are shown alongside the standard errors. Survey weights were applied. *** indicates significance at the 1% level; ** indicates significance at the 5% level; * indicates significance at the 10% level.

shows that males with a margin call had a 0.09 lower probability of investing in cryptocurrency than would be expected based on the separate effects of being male and having a margin loan.

There could be several explanations behind these interesting results. Males who have margin loans might experience financial stress, which could lead to risk-averse behaviors. In order to reduce their overall stress and potential financial burden, males might avoid additional risky investments like cryptocurrencies. Males who face margin calls might rebalance their portfolios by incorporating more traditional and less volatile assets. This rebalancing could lead to a decreased likelihood of holding cryptocurrencies.

Therefore, future research could further investigate why males with a margin loan, or a margin call have a lower probability of investing in cryptocurrencies.

5. Discussion

The results of this study support the hypothesis that both margin loans and margin calls are positively related to cryptocurrency ownership. These findings suggest that investors who engage in risky financial behaviors, such as using margin loans, are more likely to own or invest in cryptocurrencies. The results are consistent with those of Cho and Lee (2006) and Li and Qian (2018), who found that risk tolerance and perceived risk play important roles in investing in high-risk financial assets.

The findings also showed that individuals with higher levels of non-retirement investments, particularly those with USD 500,000 to USD 1 million in investments, were less likely to invest in cryptocurrencies than those with lower investment amounts. This may suggest that individuals with less wealth view cryptocurrency as an opportunity to achieve substantial returns quickly, especially if they face limited access to traditional wealth-building opportunities. In contrast, wealthier individuals may adopt more conservative and diversified investment strategies, viewing cryptocurrency as overly speculative.

Similarly, education level was also associated with cryptocurrency investment behavior. Compared to those with a postgraduate degree, individuals with some college education and an associate degree were more likely to invest in cryptocurrencies. This pattern may suggest differences in financial goals and risk tolerance. Individuals with mid-level education might be more interested in alternative investments and may view cryptocurrency as a faster way to grow their money.

From a behavioral finance perspective, engaging in margin trading may indicate an increased level of risk tolerance, potentially driven by cognitive biases such as overconfidence and the illusion of control. These behaviors are consistent with recent studies emphasizing the role of psychological traits in shaping investment decisions in emerging asset classes like cryptocurrencies.

The results of this study raise important concerns because investors are demonstrating risky behavior by participating in the highly risky cryptocurrency market where there is currently no organized regulation. Margin trading not only amplifies profits but also amplifies losses. If the investment value decreases significantly, the investors could find themselves in trouble since they must repay the borrowed amount plus the interest. Cryptocurrencies are highly volatile assets; because of this, the margin trading of the cryptocurrency comes with additional risk.

Additionally, the sub-analysis found that, while margin loans and margin calls were generally associated with higher cryptocurrency investment, this pattern reversed among male investors. Males with margin exposure were less likely to invest in cryptocurrencies, possibly due to financial stress or a preference for more stable assets. These findings emphasize the importance of recognizing gender differences in risk-taking behavior within speculative investment contexts.

Implications

The implications of these results are significant for several stakeholders, including government agencies, policymakers, financial planners, and investors themselves. Government agencies and policymakers can formulate new policies to protect investors. A highly unregulated crypto market and margin trading are potentially dangerous for investors. By introducing more quality ETFs into the crypto market, authorities can provide diversification benefits to investors to reduce the risks.

Financial planners can play an essential role in advising clients on cryptocurrency investments and margin trading. They can inform clients about the risks associated with cryptocurrency investments, especially when combined with margin trading. Financial planners can conduct comprehensive risk assessments to determine clients’ risk tolerance and suitability for such investments. They can also recommend diversification strategies to lessen the risks of cryptocurrency investments using margin trading. This may involve advising clients to allocate only a certain portion of their portfolio to cryptocurrencies and diversify the remaining investments across different asset classes.

In addition, planners can actively monitor clients’ margin positions and provide ongoing advice regarding risk management. They can set specific guidelines and thresholds to help clients make informed decisions regarding margin calls and potential liquidation. Most importantly, financial planners should prioritize client education, ensuring that individuals understand the compounding risks of investing in volatile assets with borrowed funds. By understanding the risks and implementing appropriate strategies, financial planners can assist clients navigate this complex and high-risk investment environment more effectively.

6. Limitations

In addition to the findings and contributions of the study, it is essential to recognize the study’s limitations. This section aims to discuss the two main limitations of this study.

The first limitation is that the data used in this study were pooled from a cross-section, meaning that they were collected at specific points in time (2018 and 2021). Despite controlling for income and other factors, it is essential to acknowledge that the results may be contextually bound to the years in which the data were collected. Cryptocurrency markets are known for their high volatility and quick changes, and the market dynamics may have changed considerably between 2018 and 2021 and afterwards. Therefore, this study’s findings should be interpreted carefully and may not fully capture the current state or future trends in the relationship between margin loans, margin calls, and cryptocurrency ownership.

Given the cross-sectional nature of the data, the results indicate correlations rather than causation. That is, the observed associations do not imply that margin trading behavior causes cryptocurrency investment or vice versa.

Future research could consider longitudinal data or more recent periods to provide a more up-to-date understanding of this relationship.

The second important limitation is the need for more information regarding the actual percentage of cryptocurrency investment in the investors’ overall portfolio. This information would be vital in assessing the level of risk exposure associated with cryptocurrency investments. Without knowing the proportion of a portfolio allocated to cryptocurrencies, it is difficult to fully evaluate the impact and significance of margin loans and margin calls on cryptocurrency ownership. Future studies could explore methods to obtain this information, such as direct surveys or utilizing data from financial institutions or trading platforms, to better understand the risk exposure and diversification strategies that investors employ in the cryptocurrency market.

7. Conclusions

This study examined the relationship between margin loan usage and cryptocurrency investment, adding to the growing body of research on behavioral finance and risk-taking in rising financial markets. Using data from the 2018 and 2021 NFCS Investor Surveys, the study found that individuals who participate in margin trading or receive margin calls are more likely to invest in cryptocurrencies, indicating a clear relationship between leveraged financial behavior and speculative asset ownership.

These findings are consistent with behavioral finance theory, particularly the roles of overconfidence, illusion of control, and risk-seeking behavior in influencing investment decisions. The finding that males involved in margin trading are less likely to invest in cryptocurrencies highlights the complexity of investor psychology and suggests that risk preferences and responses to financial stress may differ across demographic groups.

The study’s findings have important implications for regulators, financial planners, and investors. Regulators should carefully consider the risks related to margin trading in the context of highly volatile and largely unregulated cryptocurrency markets.

Financial planners should educate clients about the increased risks of using leverage to invest in cryptocurrencies and guide them in implementing diversified, risk-managed strategies.

Overall, this study contributes to the literature by examining the association between margin-related behaviors and cryptocurrency investment, providing insights into how these behaviors may be correlated with financial decision-making patterns in high-risk settings.