Financial Overconfidence and High-Cost Borrowing: The Moderating Effect of Mobile Payments

Abstract

Inadequate financial literacy and overconfidence in financial knowledge, coupled with the use of mobile payments (MPs), may contribute to harmful financial behaviors. While the relationship between financial knowledge confidence and financial behaviors is well documented, there is limited understanding of how financial confidence affects the use of alternative financial services (AFSs), such as payday loans, and how MPs moderate this relationship. This study examines the moderating effect of MPs on the association between financial knowledge confidence and the demand for AFS, utilizing data from U.S. adults surveyed in the 2018 National Financial Capability Study. The results show that individuals who use MPs are significantly more likely to engage with AFSs compared to non-users, with MPs increasing the likelihood of AFS usage by 92% (odds ratio: 1.92). Furthermore, overconfident individuals who use MPs are 94% more likely to rely on AFSs (odds ratio: 1.94). These findings highlight the need for targeted financial education and policymaking to mitigate the risks associated with financial overconfidence and MP usage.

1. Introduction

The rise of FinTech has democratized access to financial services, with mobile payments (MPs) growing 41%, creating a USD 100 billion industry in 2019 in the U.S. [1]. MPs offer benefits like convenience, affordability, and security, but they also influence household spending behaviors [2]. Ref. [3] noted that the convenience and diminished pain associated with mobile transactions can promote increased spending by shifting the focus toward the benefits of the purchases.

AFS providers are leveraging mobile technology for peer-to-peer lending [4]. In addition, widespread financial illiteracy continues to expose households to poor financial decision-making [5]. Research indicates that low financial literacy and overconfidence are significant factors contributing to the use of high-cost AFSs [6,7,8]. While advances in MPs offer substantial benefits to most consumers, they also pose risks of misuse, particularly for individuals who inappropriately assess their financial knowledge (e.g., overconfidence) [1,9]. Although studies have examined the effects of financial confidence and MPs on financial behavior, less is known about how these factors influence AFS usage, especially the role of MPs in moderating the relationship between financial confidence and AFS usage.

Our study adds to the literature by (a) examining the interaction between MPs and financial confidence; (b) adopting a holistic approach that considers both external factors like MPs and internal factors like financial confidence; and (c) integrating bounded rationality and the pain of payment concepts. By doing so, we provide a comprehensive framework to understand the relationship between financial confidence, MPs, and AFS usage. The primary objective of this study is to explore not only the well-established relationship between financial confidence and high-cost borrowing, but also the moderating role of MPs in this relationship. Specifically, we investigate how MPs influence the association between financial confidence and the reliance on high-cost borrowing options, such as AFSs. To achieve this, the study analyzes data from the 2018 National Financial Capability Study (NFCS) using logistic regression.

2. Literature Review

2.1. Alternative Financial Services (AFSs)

AFSs are a range of convenient, easy-to-access, and user-friendly financial services delivered by non-bank providers [10,11]. These services include pawnshop loans, payday loans, check cashing, money transmission, auto title loans, and rent-to-own arrangements [11,12]. Critics of these instruments argue that their use entails forgoing more economical options, suggesting that utilizing these services may not always represent rational decision-making [13]. The decision to borrow through AFSs may require households’ ability to objectively understand financial markets [13].

Disney and Gathergood [14] highlighted that credit customers often underestimated their borrowing costs, and those possessing lower financial literacy showed higher debt-to-income ratios among a representative sample of UK consumers. However, their study was limited to exploring subjective motivations underlying AFS usage. Ref. [15], in their meta-analysis, synthesized the causal effects of financial education on financial knowledge and behavior. They found that downstream financial behaviors (e.g., budgeting and credit card behavior) were significantly influenced by financial knowledge gained through educational programs. However, the translation of financial knowledge into behavior may be impacted by both cognitive and non-cognitive factors, such as attitudes and confidence.

In line with this, Ref. [16] found that borrowers were less likely to enter rent-to-own agreements when a full cost disclosure was made. This suggests that in the absence of explicit cost disclosure, AFS customers may struggle to understand the total cost of their borrowings. This study extends the literature by exploring factors that may obscure the cost of AFSs.

Studies, such as [17], argue that part of the decision to use payday borrowing can be attributed to cognitive biases rather than fully informed decision-making [5,6]. However, there seems to be a gap in understanding the broader psychological dimension, such as financial confidence, in influencing AFS decisions. Research by [18,19] highlights the role of overestimating one’s financial knowledge in relation to financial behaviors. However, their exploration into how this perception affects AFS usage remains underexplored.

2.2. Financial Overconfidence

Recent research emphasizes the significance of household preparedness in making sound financial decisions [20]. Prior studies document the importance of possessing financial knowledge and effectively applying it to financial behavior [21,22,23]. Robb and Woodyard [24] found a notable link between knowledge and financial behaviors. It is evident that both objective and subjective financial knowledge strongly impact financial behavior [6].

In particular, overconfidence in relation to financial knowledge has been associated with an array of undesirable financial behaviors [25,26] found that overconfident male investors were more likely to trade actively, which resulted in reduced net returns among them [25]. Additionally, Ref. [27] noted that overconfident individuals might undertake business ventures with low chances of success. Refs. [28,29] further reported that overconfident households often experience financial stress and are more likely to require debt counseling.

Prior studies have shown the role of financial confidence in financial behaviors [28,30,31]. However, limited research has examined the influence of subjective financial knowledge on high-cost borrowing behaviors. In particular, the relationship between various levels of financial confidence and the utilization of AFSs remains underexplored. This study explores four possible categories of financial confidence (based on subjective and objective financial knowledge) to investigate the moderating effect of MPs on the relationship between a household’s financial confidence and AFS usage, similar to the approach followed by [29].

2.3. Role of Mobile Payments (MPs)

MPs refer to transactions conducted using a mobile device, such as a mobile phone, through methods like text messages, apps, web browsers, or other formats [32]. As MPs become more rampant, it is critical to understand what kind of users they appeal to and whether technological advancements are changing users’ financial behaviors [1].

Existing research indicates that the adoption of mobile payments (MPs) significantly impacts consumer behavior. For example, Ref. [33] found that the adoption of MPs led to an increase in online shopping expenditures among Chinese females, based on an analysis of the 2017 Chinese General Social Survey using an instrumental-variable-based Tobit model. Similarly, Ref. [34] observed that MPs encouraged excessive spending and a greater consumption of luxury goods among young Chinese consumers.

Garrett et al. [35] identified a connection between MP usage and financial issues in the U.S., such as reliance on high-cost debt and the mismanagement of credit cards. Furthermore, Ref. [36], using the 2018 National Financial Capability Study, reported that households utilizing mobile financial services were two to three times more likely to use alternative financial services (AFSs) than those who did not use MPs. In line with this, Ref. [37] found that the adoption of mobile payments was associated with increased credit card activities at a specific bank, both online and offline. Specifically, the total transaction amount and frequency of credit card use at the bank increased by 9.4% and 10.7%, respectively.

Chun and Johnson [38] noted that the increased pain of payment—the negative feelings associated with spending money—made socially excluded individuals less likely to engage in high-risk investments. Moreover, Ref. [3] found that the adoption of mobile payment methods was linked to a decrease in the immediate negative emotions associated with spending, often referred to as the pain of paying, especially for higher-priced items. This alleviation of discomfort made consumers more inclined to spend when using mobile payments instead of cash. Conversely, Ref. [39] highlighted the positive role of financial technology in enhancing financial inclusion, particularly in ASEAN member countries and India, where digital devices have improved access to financial services. Therefore, it is plausible that using mobile payments, which reduces the pain of payment, may be associated with high-risk behaviors such as using AFSs.

Research has established a prominent relationship between objective financial knowledge and the use of mobile payment systems (MPs). For instance, Ref. [40] utilizing data from the 2015 and 2018 National Financial Capability Study, found a significant correlation between financial literacy and the usage of MPs.

While the literature documents the relationship between MP usage, financial knowledge, and financial behaviors, such as the use of AFSs, there remains a limited understanding of how MPs influence the relationship between subjective financial knowledge and AFS usage. Further exploration of this association is needed to uncover the intersection between these factors. In this study, we aim to extend the existing literature by analyzing how MPs interact with varying degrees of confidence in household financial knowledge.

2.4. Theoretical Framework

Bounded rationality has long served as a conceptual framework to demonstrate that individuals, due to cognitive capacities and incomplete information, often make sub-optimal decisions, such as using AFSs for high-cost borrowing [13]. MPs offer timely information and simplify access to various financial services, potentially steering individuals toward both sound and unsound (sub-optimal) choices [9].

Behavioral economics helps explain consumer behavior, particularly through the concept of “hyperbolic discounting”. This concept illustrates why consumers often make impulsive purchases that can result in high-cost borrowing through mobile devices [41,42].

Impulsive buying refers to unplanned and spontaneous actions influenced by a mix of internal and external stimuli [43,44]. Emotional factors such as excitement, stress, or boredom, combined with cognitive limitations, often override rational decision-making processes and lead individuals to focus on immediate gratification [41,42]. Environmental influences like mobile advertising and point-of-purchase strategies, along with social factors such as influencer marketing, can further contribute to this behavior. Additionally, psychological triggers, including the “Fear of Missing Out” (FOMO), are associated with impulsive purchases and the use of services, particularly in time-sensitive or limited-quantity situations [9,41,42,45].

Research has consistently linked lower cognitive ability to heightened impatience and a range of suboptimal choices [43,44]. One key concept, bounded rationality, has been widely utilized to analyze consumer behavior [13,46,47]. Limitations in objective financial knowledge, along with an over-reliance on subjective knowledge, significantly impact households’ suboptimal financial decision-making processes [18,25].

The use of AFSs is well explained within the bounded rationality framework, which indicates that suboptimal financial decisions result from deviations from underlying neoclassical axioms on the limited information processing capabilities [13,47]. For example, ref. [13] used bounded rationality to explain the role of borrowers’ overconfidence in high-cost borrowing. Yet, this framework does not account for how MPs might alleviate household payment pains. Our study bridges this gap by integrating the “pain of payment” concept with bounded rationality, shedding light on MPs’ influence on the link between financial confidence and AFS usage.

The pain of payment is a behavioral economics concept that refers to the painful state of mind or negative emotions consumers experience while purchasing a product or service [48]. Mobile devices can mitigate this pain, enhancing purchasing enjoyment and willingness to buy [38]. The pain of payments perspective has been extensively utilized in mental accounting [49], budgeting [38], and spending behavior [2].

High-cost AFS instruments can be accessed both through physical stores and mobile phones [10]. Multifunctional mobile phones provide immediate and low-pain access to AFSs [3]. This can encourage people with cognitive limitations to seek quick financial solutions to overlook the expenses tied to AFSs and stimulate a household’s propensity to access AFSs [35]. Hence, MP users are more likely to use AFSs due to pain attenuation during the loan term.

Building on these theoretical foundations and previous research, we hypothesize that overconfident consumers, influenced by bounded rationality, may underestimate the true costs of high-cost AFSs. This tendency may be further exacerbated by MPs’ convenience and opacity, which reduce the perceived pain of payment.

H1: 

Those who are overconfident in their financial knowledge will have a higher likelihood of using AFSs than otherwise.

Given the reduced pain of payment associated with MPs, it may be reasonable to expect that those who utilize this form of payment might also be more inclined toward AFSs. This is especially true as reduced pain and ease of access obscure AFS costs.

H2: 

MP users will have a higher likelihood of using AFSs than otherwise.

Finally, integrating the concepts of bounded rationality and pain of payment [50] we posit a synergistic effect on AFS usage.

H3: 

Using MPs will positively moderate (i.e., increase) the impact of financial overconfidence on the use of AFSs.

The above hypotheses establish a link between cognitive limitations in decision-making, as explained by bounded rationality; the behavioral response of a payment mechanism using the pain of payments; and the empirical manifestations of these theories.

3. Methods

3.1. Data and Sample

The dataset used for investigating the moderating effect of MPs on the use of AFSs is the 2018 State-by-State NFCS database, which was collected under the auspices of the FINRA Investor Education Foundation [51]. This study utilizes the 2018 FINRA dataset to emphasize more typical financial behaviors related to AFS usage. Relying on the 2021 data from during the COVID-19 pandemic could have introduced confounding variables. Furthermore, the financial hardships experienced during this time may have influenced responses, reflecting unique challenges rather than standard financial behaviors. Therefore, this research focuses on standard circumstances rather than the temporary disruptions caused by external crises, such as the COVID-19 pandemic.

In 2018, NFCS surveyed 27,091 respondents nationwide. Similar to other studies, our investigation of missing values did not reveal a systematic bias or process that would influence our investigation [52]. Responses marked as “don’t know”, “Prefer not to say”, or left unanswered were excluded from the analysis, resulting in a sample size of 26,017. Given that the resulting missing values were randomly eliminated and the sheer number of observations in the analytical sample, we are confident that missing values do not pose problems for our reported findings. We conducted a preliminary investigation of multicollinearity; we found the correlation between the predictors to be lower than 0.40. Therefore, inefficacy (imprecision) in parameter estimates will not hinder our analysis.

3.2. Measures

3.2.1. Dependent Variable

AFS use (y) is based on the household’s use of high-cost borrowing (including rent-to-own stores, tax refunds, payday loans, pawn shops, and auto title loans) during the last five years. We set y = 1 if the respondent has used AFSs during the last five years; otherwise, y = 0. The data show that almost 27% of respondents used AFSs in the past five years.

3.2.2. Independent Variables

Financial Confidence. Similar to the previous studies, we constructed a categorical variable for gauging financial confidence (x) by combining measures of subjective and objective financial knowledge [53,54]. We created an objective measure of financial knowledge using a series of six personal finance questions related to compound interest rates, inflation, risk, and diversification. This allowed for the identification of those with a higher-than-median (i.e., 3) objective financial knowledge as households with ‘high objective financial knowledge’ and the rest as households with ‘low objective financial knowledge’. Second, we classified those with subjective financial knowledge higher than the median (i.e., 5) as households with ‘high subjective financial’ and the rest as ‘low subjective financial knowledge’.

Consistent with the previous studies [53,54], we categorized financial confidence into four subcategories (x_s (where, s = 1, …, 4)) as follows:

x₁: Appropriate low (low subjective and low objective financial knowledge).

x₂: Appropriate high (high subjective and high objective financial knowledge).

x₃: Overconfident (high subjective and low objective financial knowledge).

x₄: Underconfident (low subjective and high objective financial knowledge).

Furthermore, given that the frequency distribution for objective financial knowledge (−0.155) had a near-symmetrical distribution, while the distribution for subjective financial knowledge displayed slight negative skewness (−0.812), utilizing the median was considered more appropriate than utilizing the mean. In the presence of outliers, the median has been considered a superior measure because it remains relatively unaffected by the direction of the skew, unlike the mean, which might not offer an accurate representation of the dataset [55].

MPs. The responses to the question “How often do you use your mobile phone to pay for a product or service in person at a store, gas station, or restaurant (e.g., by waving/tapping your mobile phone over a sensor at checkout, scanning a barcode or QR code using your mobile phone, or using some other mobile app at checkout)?” measure MPs (m) as a moderator [37]. Responses consisting of “Frequently” and “Sometimes” were set to one (m = 1), while “Never” responses were set to zero (m = 0).

Building on prior research, this study utilizes mobile payments (MPs) to capture transactions conducted via smartphones. Both in-store mobile transactions at checkout and online payments frequently utilize overlapping technologies, particularly through e-wallets such as PayPal, Apple Pay, and Google Pay. This trend reflects a notable increase in the adoption of these payment methods [56]. The enhanced convenience and expedited checkout processes are anticipated to drive an uptick in purchasing activity [56]. Furthermore, Ref. [37] observed that the adoption of mobile payments has contributed to a rise in credit activity at bank branches, both online and offline.

Demographics and Socioeconomic Characteristics. Seven major socioeconomic and demographic characteristics (z_j; where j = 1) of the sample, consisting of gender, ethnicity, marital status, age, education, occupation, and income, were decomposed into their subcomponents and cast as a set of binary (0, 1) variables for further analysis.

3.3. Analytical Model

Figure 1 presents our analytical model of the moderating effect of MPs on the relationship between financial confidence and the use of AFSs. Firstly, in this model, parameters β and γ capture the direct impact of financial confidence (x) and MPs (m) on the use of AFSs (y^*), respectively. Then, parameter δ measures MPs’ moderating (interaction) effect on the relationship between financial confidence and the use of AFSs. Finally, through α, the model allows us to control the influence of the respondents’ socioeconomic characteristics on their use of AFSs. The model was designed to include only the most important variables to explain the relationship, ensuring parsimony [57,58]. This approach minimizes the risk of overfitting while enhancing the model’s generalizability [57,58]. Moreover, parsimonious models are inherently easier to interpret and understand, as they avoid unnecessary complexity. This aligns with the principle of Occam’s Razor, which suggests that the simplest explanation is often the most accurate [57,58].

The above model, including the interaction between MPs and financial confidence (mx), is represented as follows:

$$y_i^{*}={\mathsf{\theta }}_0+{\sum }^{s=4}{\beta }_s{x}_{is}+\gamma m_i+{\sum }_{s=2}^{s=4}{\delta }_s{m}_i{x}_{is}+{\sum }_{j=2}{\alpha }_j{z}_{ij}+{\epsilon }_i$$

(1)

where $x_{is}$ captures our four categories of financial confidence—appropriate low $\left(s=1\right)$, appropriate high $\left(s=2\right)$, overconfident $\left(s=3\right),$ and underconfident $\left(s=4\right)$. In our regression analysis, appropriate low $\left(s=1\right)$ and a base category for the socioeconomic and demographic indicators (J = 1) are used as the reference categories. Using the thresholds for $y_i^{*}$, the observed binary outcome $y$ replaces $y_i^{*}$, as follows:

$$y=\left\{\begin{array}{c} 1 if y^{*}>0, \\ 0 if y^{*}=0,\end{array}\right.$$

(2)

By letting $P_i=Prob\left(y_i=1\right)$ for the ith household, the following logit model is obtained:

$$\log {\displaystyle \frac{P_i}{1-P_i}}={\mathsf{\theta }}_0+{\sum }_{s=2}^{s=4}{\beta }_s{x}_{is}+\gamma m_i+{\sum }_{s=2}^{s=4}{\delta }_s{m}_i{x}_{is}+{\sum }_{j=2}{\alpha }_j{z}_{ij}+u_i$$

(3)

This provides an appropriate framework for estimating and inferring the odds of using AFSs. Given our three hypotheses, H1–H3, a priori, ${\beta }_3>0$, $\gamma >0$, and ${\delta }_3>0$. A positive value for ${\beta }_3$ implies that overconfident households $\left(x_3\right)$ are more likely to use AFSs $\left(y_i^{*}\right)$ than otherwise. Second, we expect the users of MPs ($m$), who benefit from reduced pain of payments, to have a higher likelihood ($\gamma >0$) of using AFSs than those who refrain from using MPs. Finally, we expect that the inclusion of interactions between MPs and financial knowledge categories ($m_i{x}_{is})$ will shed light on the moderating effect of MPs ($m_i$) on the relationship between financial confidence $x_{is}$ and use of AFSs. In particular, we expect MPs ($m$) to play a strong enabling role for the overconfident $\left(x_{i3}\right)$ to use AFSs, $i.e., ({\delta }_3>0).$

4. Results

4.1. Descriptive Statistics

Table 1. Descriptive statistics (%) of key variables: 2018 NFCS (n = 26,017).

Variables (0, 1)	Full Analytical Sample		Did Not Use MPs		Used MPs
	(n = 26,017)		(n = 17,222)		(n = 8795)
	Mean	SD	Mean	SD	Mean	SD
Used Alternative Financial Services (AFSs)	0.27	0.44	0.2	0.39	0.39	0.51
Mobile Payment (MP) Users	0.36	0.48
Financial Knowledge
Very High Financial Knowledge	0.07	0.26	0.08	0.27	0.06	0.24
Very Low Financial Knowledge	0.06	0.25	0.06	0.23	0.07	0.27
Very High Self-Assessed Financial Knowledge	0.16	0.36	0.12	0.32	0.22	0.43
Very Low Self-Assessed Financial Knowledge	0.03	0.16	0.03	0.16	0.03	0.17
Self-assessed Financial Knowledge	5.11	1.38	5.04	1.35	5.26	1.41
Objective Financial Knowledge	3.10	1.65	3.25	1.63	2.87	1.619
Financial knowledge Confidence
Appropriate High	0.21	0.4	0.22	0.41	0.18	0.39
Overconfident	0.21	0.4	0.16	0.36	0.28	0.47
Underconfident	0.21	0.4	0.23	0.41	0.17	0.39
Appropriate Low	0.38	0.48	0.38	0.48	0.37	0.5
Gender
Male	0.49	0.5	0.46	0.49	0.55	0.52
Female	0.51	0.5	0.54	0.49	0.45	0.52
Age
18 to 24	0.11	0.32	0.08	0.27	0.17	0.39
25 to 34	0.18	0.39	0.13	0.33	0.28	0.47
35 to 44	0.16	0.37	0.14	0.34	0.21	0.42
45 to 54	0.17	0.37	0.17	0.37	0.16	0.38
55 to 64	0.18	0.38	0.22	0.41	0.11	0.32
65 and above	0.19	0.39	0.26	0.43	0.07	0.27
Ethnicity
White	0.64	0.48	0.71	0.45	0.52	0.52
Non-White	0.36	0.48	0.29	0.45	0.48	0.52
Marital Status
Married	0.51	0.5	0.53	0.49	0.48	0.52
Single	0.32	0.47	0.28	0.44	0.4	0.51
Separated/Divorced/Widowed	0.02	0.12	0.02	0.12	0.02	0.13
Education
Postgraduate	0.1	0.31	0.1	0.29	0.12	0.33
Bachelor’s Degree	0.18	0.38	0.17	0.36	0.19	0.41
Associate Degree	0.09	0.28	0.08	0.27	0.1	0.31
Some College or no degree	0.19	0.4	0.18	0.38	0.22	0.43
High School Graduate/GED	0.33	0.47	0.35	0.47	0.3	0.47
Did not complete High School	0.09	0.28	0.09	0.28	0.07	0.27
Occupation Stage
Retired	0.22	0.41	0.29	0.44	0.09	0.3
Self-employed	0.05	0.22	0.07	0.24	0.09	0.29
Full-time	0.05	0.22	0.34	0.46	0.52	0.52
Part-time	0.04	0.2	0.09	0.28	0.1	0.3
Homemaker	0.07	0.26	0.07	0.26	0.07	0.26
Full-time Student	0.09	0.29	0.03	0.17	0.06	0.24
Disabled	0.4	0.49	0.06	0.24	0.04	0.19
Unemployed	0.07	0.26	0.05	0.22	0.04	0.21
Income (USD)
Less than 15,000	0.12	0.32	0.12	0.32	0.11	0.33
15,000–25,000	0.11	0.31	0.11	0.31	0.1	0.3
25,000–35,000	0.11	0.31	0.11	0.31	0.11	0.32
35,000–50,000	0.15	0.35	0.15	0.35	0.13	0.35
50,000–75,000	0.19	0.39	0.19	0.39	0.19	0.4
75,000–100,000	0.14	0.35	0.13	0.32	0.16	0.38
100,000–150,000	0.12	0.33	0.12	0.32	0.13	0.35
150,000 and above	0.06	0.24	0.06	0.23	0.07	0.27

provides descriptive statistics for the full analytical sample (n = 26,017) and its breakdown by MPs. It shows that a quarter (26%) of respondents used AFSs for high-cost borrowing. Table 1 also shows that about one-third (36%) of the full sample used MPs. Almost 40% of MP users also use AFSs (Table 1). While 38% of the sample fell into the “Appropriate Low” financial knowledge category, around 21% each were in the “Appropriate High”, “Overconfident”, and “Underconfident” categories. However, a cross-tabulation of financial knowledge confidence categories according to MPs clearly shows a high variation among these categories, where the combination of the “Overconfident” (28%) and “Appropriate Low” (37%) categories included two-thirds of the respondents in the sample.

Socioeconomic and demographic indicators in Table 1 have a relatively uniform distribution by gender and age. Males (49%) and females (51%) were evenly distributed. Age distribution was mostly even, except for the 18–24 age category, which makes up 11%. The majority of the households identified themselves as White (64%) and married (51%). Among MP users, most respondents were males (55%), aged between 25 and 34 years (28%), unmarried (40%), and with some or no college degree (22%). The reported statistics on income show that the most significant number of respondents are in the USD 50,000–USD 75,000 income bracket (19%), and the smallest number is in the USD 150,000 and higher bracket (6%).

4.2. Multivariate Logistic Regression Results

Table 2. Logistic regression estimates (dependent variable—used AFSs): 2018 NFCS (n = 26,017).

	Panel A			Panel B
	Including MPs			Including Moderation
	Coef.	S.E.	Odds	Coef.	S.E.	Odds
Constant	−3.82 ***	0.16	0.02	−3.63 ***	0.16	0.03
Subjective Financial Knowledge	0.14 ***	0.01	1.15	0.07 ***	0.01	1.07
Objective Financial Knowledge	−0.13 ***	0.01	0.88	−0.06 ***	0.01	0.94
Mobile Payments (MPs)	0.65 ***	0.04	1.92	0.62 ***	0.05	1.86
Interaction Term (Financial Knowledge * MPs)
Subjective Financial Knowledge * MPs				0.52 ***	0.06	1.68
Objective Financial Knowledge * MPs				−0.52 ***	0.06	0.59
Bank account Ownership	−0.53 ***	0.06	0.59	−0.53 ***	0.06	0.59
Credit Record (reference: Very Good)
Very Bad	1.78 ***	0.08	5.92	1.76 ***	0.08	5.82
Bad	1.86 ***	0.06	6.39	1.89 ***	0.06	6.61
About average	1.19 ***	0.05	3.28	1.23 ***	0.05	3.42
Good	0.61 ***	0.05	1.85	0.64 ***	0.05	1.89
Gender (reference: Female)
Male	0.32 ***	0.04	1.38	0.29 ***	0.04	1.35
Age (reference: 65 and above)
18–24	1.42 ***	0.10	4.12	1.40 ***	0.09	4.06
25–34	1.42 ***	0.09	4.15	1.39 ***	0.09	4.03
35–44	1.14 ***	0.09	3.14	1.13 ***	0.09	3.11
45–54	0.83 ***	0.09	2.30	0.83 ***	0.08	2.30
55–64	0.41 ***	0.08	1.88	0.40 ***	0.08	1.49
Ethnicity (reference: White)
Non-White	−0.28 ***	0.04	1.50	0.27 ***	0.03	1.31
Marital Status (reference: Married)
Single	−0.29 ***	0.04	0.75	−0.28 ***	0.04	0.75
Separated/Divorced/Widowed	0.10	0.12	1.11	0.09	0.12	1.10
Education (reference: Did not complete High School)
Postgraduate	−0.26 ***	0.08	0.77	−0.26 **	0.08	0.77
Bachelor’s Degree	−0.19 **	0.07	0.83	−0.18 **	0.07	0.83
Associate Degree	−0.03	0.08	0.97	−0.03	0.07	0.97
Some College or No-degree	0.13 *	0.06	1.14	0.11	0.06	1.12
High School Graduate/GED	−0.03	0.06	0.97	−0.04	0.06	0.96
Occupation Stage(reference: Retired)
Self-employed	0.25 ***	0.08	1.29	0.23 *	0.09	1.26
Full-time	0.11	0.08	1.12	0.10 ***	0.07	1.10
Part-time	0.09	0.09	1.09	0.08	0.09	1.08
Homemaker	−0.06	0.09	0.94	−0.06	0.09	0.93
Full-time Student	−0.07	0.11	0.93	−0.07	0.11	0.93
Disabled	0.22 **	0.09	1.25	0.21 *	0.09	1.24
Unemployed	−0.03	0.10	0.97	−0.02	0.10	0.98
Income (USD): (reference: 150,000 and above)
Less than 15,000	0.75 ***	0.11	2.11	0.73 ***	0.11	2.08
15,000–25,000	1.11 ***	0.11	3.05	1.11 ***	0.11	3.04
25,000–35,000	1.04 ***	0.11	2.84	1.03 ***	0.10	2.81
35,000–50,000	0.97 ***	0.10	2.63	0.96 ***	0.10	2.63
50,000–75,000	0.69 ***	0.10	1.99	0.68 ***	0.10	1.98
75,000–100,000	0.87 ***	0.10	2.39	0.83 ***	0.10	2.29
100,000–150,000	0.48 ***	0.11	1.61	0.47 ***	0.10	1.60
Likelihood Ratio Test ($𝒳$2) (p-value)	5934.34 (0.00)			6062.95 (0.00)
df	36			38
McFadden R2	0.19			0.21

Note. *** p < 0.001; ** p < 0.01; * p < 0.05. Reference groups in parentheses.

presents the possibility of further investigating the association of financial confidence categories and their interaction with MPs. Panel A and Panel B use self-assessed and objective financial knowledge as predictors, respectively.

In Panel A, a one-unit increase in subjective financial knowledge increases the odds of using alternative financial services (AFSs) by 0.14. This translates to a 1.15 times higher likelihood of using AFSs or an approximately 15% increase in the odds of AFS usage. This finding indicates that individuals with higher self-assessed financial knowledge are more likely to use AFSs. Conversely, a one-unit increase in objective financial knowledge reduces the log odds of AFS usage by 0.13, which corresponds to 0.88 times the odds of AFS usage. This suggests that a higher objective financial knowledge is associated with a reduced likelihood of using AFSs.

In Panel B, the use of MPs increases the log odds of AFS usage by 0.65, or 1.92 times the odds of AFS usage. This indicates that increased use of MPs is associated with a higher likelihood of using AFSs. Additionally, introducing MPs improves the model’s fit, as evidenced by an increase in McFadden’s R² from 0.75 to 0.76.

The interaction terms in Panel B show that both self-assessed and objective financial knowledge, when combined with MPs, significantly affect AFS usage. In particular, the interaction maintains the same directional signs for financial knowledge predictors, even when MPs are included. McFadden’s R² values, ranging from 19% to 21%, indicate a gradual improvement in explaining variations in AFS usage when MPs and their interactions with financial knowledge indicators are added to the model (p < 0.001).

Table 3. Logistic regression estimates (dependent variable—used AFSs): 2018 NFCS (n = 26,017).

	Panel A			Panel B
	Including MPs			Including Moderation
	Coef.	S.E.	Odds	Coef.	S.E.	Odds
Constant	−3.46	0.14	0.031	−3.37 ***	0.14	0.035
Confidence Category (reference: Appropriate Low)	0.14 ***	0.01	1.15	0.07 ***	0.01	1.07
Appropriate High	−0.26 ***	0.06	0.77	−0.26 ***	0.07	0.77
Overconfident	0.26 ***	0.02	1.29	0.08 *	0.03	1.08
Underconfident	−0.10 ***	0.02	0.91	−0.09 ***	0.02	0.92
Mobile Payments (MPs)	0.64 ***	0.04	1.90	0.50 ***	0.05	1.66
Interaction Terms (reference: Appropriate Low × MPs)
Appropriate High × MPs				0.05	0.10	1.05
Overconfident × MPs				0.66 ***	0.09	1.94
Underconfident × MPs				−0.09	0.10	0.91
Bank account Ownership	−0.51 ***	0.06	0.60	−0.51 ***	0.06	0.60
Credit Record (reference: Very Good)
Very Bad	1.70 ***	0.08	5.48	1.70 ***	0.08	5.48
Bad	1.85 ***	0.05	6.37	1.87 ***	0.05	6.48
About average	1.19 ***	0.05	3.29	1.22 ***	0.05	3.38
Good	0.61 ***	0.05	1.85	0.63 ***	0.05	1.88
Gender (reference: Female)
Male	0.32 ***	0.04	1.38	0.30 ***	0.04	1.36
Age (reference: 65 and above)
18–24	1.40 ***	0.10	4.06	1.37	0.10	3.95
25–34	1.42 ***	0.09	4.12	1.37	0.09	3.95
35–44	1.14 ***	0.09	3.14	1.12	0.09	3.05
45–54	0.83 ***	0.09	2.29	0.81	0.09	2.24
55–64	0.42 ***	0.08	1.52	0.40	0.08	1.49
Ethnicity (reference: White)
Non-White	0.27 ***	0.03	1.31	0.27 ***	0.03	1.31
Marital Status (reference: Married)
Single	−0.30 ***	0.04	0.74	−0.31 ***	0.04	0.74
Separated/Divorced/Widowed	0.10	0.12	1.11	0.09	0.12	1.10
Education (reference: Did not complete High School)
Postgraduate	−0.24 **	0.08	0.78	−0.25 **	0.08	0.78
Bachelor’s Degree	−0.18 **	0.07	0.84	−0.18 **	0.07	0.84
Associate Degree	−0.01 **	0.07	0.99	−0.01	0.07	0.99
Some College or No-degree	0.12	0.06	1.13	0.11	0.06	1.12
High School Graduate/GED	−0.05	0.06	0.95	−0.05	0.06	0.95
Occupation Stage (reference: Retired)
Self-employed	0.27 **	0.09	1.31	0.26 **	0.09	1.29
Full-time	0.11	0.08	1.12	0.10	0.08	1.11
Part-time	0.05	0.09	1.05	0.05	0.09	1.05
Homemaker	−0.06	0.09	0.94	−0.07	0.09	0.93
Full-time Student	−0.04	0.11	0.96	−0.04	0.11	0.96
Disabled	0.21 *	0.09	1.23	0.20 *	0.09	1.22
Unemployed	−0.03	0.10	0.97	−0.04	0.10	0.97
Income (USD): (reference: 150,000 and above)
Less than 15,000	0.70 ***	0.11	2.01	0.70 ***	0.11	2.01
15,000–25,000	1.06 ***	0.11	2.90	1.07 ***	0.11	2.90
25,000–35,000	1.00 ***	0.11	2.72	1.00 ***	0.11	2.73
35,000–50,000	0.92 ***	0.10	2.52	0.93 ***	0.10	2.54
50,000–75,000	0.64 ***	0.10	1.90	0.65 ***	0.10	1.91
75,000–100,000	0.81 ***	0.10	2.25	0.79 ***	0.10	2.19
100,000–150,000	0.43 ***	0.10	1.54	0.43 ***	0.11	1.54
Likelihood Ratio Test ($𝒳$2) (p-value)	6100.94 (0.00)			6173.99 (0.00)
df	37			40
McFadden R2	0.21			0.22
Model Fit	0.75			0.76

Note. *** p < 0.001; ** p < 0.01; * p < 0.05. Reference groups in parentheses.

expands on these findings through multiple logistic regressions, presented in Panels A and B, showing results before (δ_s = 0) and after moderation (δ_s ≠ 0). In Panel A, financial confidence subcategories significantly predict AFS usage at the 5% significance level. Overconfident individuals have a 0.26 increase in the log odds of AFS usage compared to those with “Appropriately Low” confidence, meaning overconfident respondents are 1.29 times more likely to adopt AFSs. Conversely, a one-unit increase in underconfidence decreases the log odds of AFS usage by 0.10, with underconfident individuals being 0.91 times as likely to use AFSs. These results suggest that overconfidence is positively associated with increased AFS usage, while underconfidence is linked to reduced AFS usage. These findings align with previous research [13,53,54].

Additionally, for each one-unit increase in MP usage, the log odds of AFS usage increase by 0.64. This corresponds to individuals using MPs being 1.90 times more likely to use AFSs compared to those who do not, assuming other factors remain constant. However, the model in Panel A assumes no moderating effect of MPs on the relationship between financial confidence and AFS usage.

Panel B highlights the significant positive role of MPs in AFS usage, supporting the second hypothesis (H2). A one-unit increase in MP usage increases the log odds of AFS usage by 0.50, meaning individuals using MPs are 1.66 times as likely to use AFSs. Interaction terms show that overconfident individuals who use MPs experience a 0.66 increase in the log odds of AFS usage compared to “Appropriately Low” confidence individuals who use MPs. This interaction indicates that being overconfident and using MPs increases the odds of AFS usage by 1.94 times. A positive and significant coefficient (p < 0.001) for the interaction between MPs and overconfidence (δ₃ > 0) confirms the third hypothesis (H3).

Finally, the full model in Panel C provides additional insights into demographic factors influencing AFS usage. Being male increases the log odds of AFS usage by 0.35 (odds ratio = 1.36) compared to females (reference group). Younger individuals are more likely to use AFSs than those aged 65 or older (reference group). For example, individuals aged 18–24 are 4.06 times more likely to use AFSs.

Non-White respondents are more likely to use AFSs compared to those who self-identify as White. Additionally, being separated, divorced, or widowed is associated with a 1.10 times higher likelihood of AFS usage relative to married individuals. Higher education, particularly having a bachelor’s degree or higher, is associated with lower AFS usage compared to those without a high school diploma. Occupations such as self-employment (odds ratio = 1.29) or being disabled (odds ratio = 1.22) also increase the likelihood of AFS usage compared to retired individuals. Finally, respondents in all income categories below USD 150,000 are more likely to use AFSs than those earning USD 150,000 or above. These results altogether emphasize the roles of financial confidence, mobile payments, and demographic factors in shaping AFS usage behaviors.

4.3. Discussion and Implications

Given the expansion of rapid cashless transactions through MPs, this study investigated the moderating effect of MPs on the relationship between financial confidence and the use of AFSs as a proxy for high-cost borrowing (Table 2 and Table 3). By addressing this, the study bridges a gap in the existing body of literature. Even after accounting for sociodemographic factors, individuals who are overconfident in their financial knowledge were more likely to use AFSs, aligning with previous research [32,57]

Multiple logistic regression analyses indicated that overconfident respondents had a higher propensity to adopt AFSs. This finding is consistent with the behavioral finance literature, which suggests that cognitive biases, such as overconfidence, can lead to suboptimal financial behaviors [13,59]. This finding may imply that individuals who overestimate their financial literacy may be more prone to risk-taking behaviors, including utilizing AFSs, believing they can effectively manage the associated costs or consequences.

The results also indicated that an increase in underconfidence was associated with a decrease in AFS usage. This aligns with findings from previous research [13,53,54], which suggest that underconfidence tends to encourage more cautious financial behavior, potentially preventing individuals from making irrational financial decisions, such as relying on AFSs.

The findings also indicated that individuals who use MPs are more inclined to utilize alternative financial services (AFSs). This supports existing research, such as that by [36], which has established an association between MP usage and AFS adoption. This outcome also reinforces the “pain of payment” perspective, which suggests that digital transactions conducted via smartphones result in a lower perceived sense of monetary loss compared to traditional cash payments [9,60]. Furthermore, consumers who use mobile payments at checkout, as well as for online transactions, predominantly utilize mobile payment technologies through e-wallets such as PayPal, Apple Pay, and Google Pay. Consequently, improved convenience and checkout experience are expected to drive an increase in the adoption of mobile technologies, leading to a rise in purchases both online and offline [37,56]. To address potential risks, regulators should consider implementing targeted interventions, such as consumer education campaigns and protective measures.

Regarding the interaction terms, the findings suggest that overconfident individuals who use MPs exhibit increased usage of AFSs. This indicates that the combination of restricted rationality due to overconfidence and the diminished pain of payment effect associated with MPs can lead to undesirable financial behaviors, such as greater reliance on AFSs. This highlights the need for tailored financial products and services to address the unique needs of this user segment.

Our study demonstrates a robust and significant moderating effect of MPs on the relationship between overconfidence in financial knowledge and higher AFS usage. The results show that overconfident MP users are at a greater risk of utilizing AFSs than others. This study contributes to the literature by establishing the moderating role of mobile payments and its theoretical grounding, drawing from the concepts of bounded rationality and payment pain. This framework supports the moderated relationship between financial overconfidence, mobile payments (MPs), and alternative financial services (AFSs). By establishing this theoretical foundation, we create a solid basis for understanding these interconnections. Furthermore, these insights are also valuable for financial educators, policymakers, and regulatory bodies, such as the Consumer Financial Protection Bureau, in designing protective measures for overconfident MP users who may be vulnerable to disadvantageous financial decisions stemming from their AFS usage.

Financial educators should adapt education programs to the evolving technological landscape to effectively address consumer vulnerabilities. They could introduce dedicated modules that outline the risks and benefits associated with MPs. Additionally, financial educators can collaborate with FinTech companies to embed financial education content directly into FinTech apps, including notifications and tips that promote healthy financial behaviors.

Policymakers, on the other hand, should focus on implementing policies that ensure consumers are fully informed about the implications of using AFSs through MP platforms. These initiatives could include making transparent disclosures of high interest rates to empower consumers to make informed financial decisions.

Furthermore, this study also contributes to the existing body of literature by providing sociodemographic insights into the users of MPs and AFSs. Our study showed that individuals with postgraduate degrees and bachelor’s degrees were less likely to use AFSs relative to those who did not complete their schooling, signifying the relevance of education on the use of high-cost borrowing decisions [6,12].

Providing financial education to the youth in their early years may better equip them for transitioning to adulthood. Furthermore, separation, divorce, and widowhood increase the likelihood of using AFSs. The finding that those with proper financial knowledge and those holding higher education degrees, i.e., postgraduate degrees, are less likely to use AFSs is crucial for the faculties involved in offering financial education.

A higher likelihood of using AFSs among people with disabilities may be an indication of barriers in the mainstream financial system. Likewise, increased AFS usage among non-White, lower-income, and younger individuals indicates the traditional financial system’s perceived and structural barriers among households with such characteristics. Policymakers and financial institutions should strive to address this disparity to ensure broader financial inclusion, particularly for minorities. While policies to curb AFS usage reduce access to funds, policymakers should expand opportunities for disabled households to access the mainstream banking system. Overall, our findings align with existing studies emphasizing the significant impact of sociodemographic factors (e.g., [12,61]).

4.4. Limitations

This study contributes to the literature by examining the moderating role of MPs in the relationship between financial confidence and consumers’ use of AFSs. However, the findings of this study should be interpreted with caution due to several limitations.

First, the current analysis relied on a cross-sectional dataset (NFCS 2018), which limited the ability to establish causality and restricted the investigation of whether the relationship between MPs and AFS usage remains linear over time or varies at different levels of MP adoption. Additionally, using a 2018 dataset may have restricted our ability to account for more recent developments that could impact the findings and their relevance.

Second, the reliance on survey data, such as the NFCS, introduces limitations inherent to the retrospective nature of the questionnaire. Respondents were asked to recall their use of AFSs over the past few years, which may have increased the risk of recall bias and potential inaccuracies in their responses.

Third, this study’s use of median splits to categorize financial confidence may have oversimplified the variable’s continuous nature. This approach likely overlooked important variations across the spectrum of financial knowledge and confidence, potentially concealing nuanced relationships within the data.

Fourth, while this study provides valuable insights into the relationship between MPs and AFS usage using a nationally representative sample of U.S adults, these findings are not generalizable to other countries. Payment behaviors may be influenced by country-specific factors, including cultural attitudes and financial infrastructure [62]. Lastly, while our study identifies a significant association between MP and AFS engagement, certain forms of AFSs (e.g., payday lending and pawnshops) may incentivize or require the use of MP methods. This could have influenced the adoption of MPs and subsequently contributed to the rapid utilization of AFSs.

5. Future Research

To address these limitations, future studies should retest the hypothesized relationships using longitudinal data to explore the moderating role of MPs and the linear association between financial confidence and AFS usage over time. Additionally, such studies should examine these relationships during and after the COVID-19 pandemic to evaluate potential changes in the dynamics of financial confidence and AFS usage.

Furthermore, given the limitations of using median-based categories, future research should adopt more nuanced approaches to measure financial confidence. Additionally, future studies should focus specifically on the various forms of AFSs and their interaction with MPs. Exploring how MP usage varies with different combinations of financial confidence and types of AFS (e.g., payday loans) could yield valuable insights.

Future studies should incorporate additional demographic and contextual variables to gain a more comprehensive understanding of financial confidence and the use of alternative financial services (AFSs). Factors such as risk tolerance, impulsiveness, behavioral tendencies, and technological adoption could also provide valuable insights into the dynamics of this relationship. Moreover, these investigations should extend beyond the United States to account for cultural preferences and country-specific financial infrastructures.

6. Conclusions

This study examined the relationship between four categories of financial confidence and the usage of AFSs, such as payday loans and pawnshops. Utilizing data from the 2018 National Financial Capability Study and applying logistic regression, the study was grounded in the theoretical frameworks of bounded rationality and the pain of payment concept.

The findings indicated that individuals with a higher subjective financial knowledge were more likely to utilize AFSs. In contrast, an increase in objective financial knowledge was associated with a reduced likelihood of AFS usage. Overconfident individuals exhibited a greater tendency to engage with AFSs compared to those displaying “Appropriate Low” confidence levels. Furthermore, when interacting with MPs, those displaying overconfidence experienced an enhanced likelihood of AFS usage. The combination of overconfidence and MPs significantly intensified the incidence of AFS usage, thus supporting the hypothesis that MPs play a moderating role in the relationship between financial overconfidence and AFS usage. These results are consistent with the existing literature in behavioral finance [59,63], which suggests that overconfidence can lead to biased financial behaviors.

Overconfidence, which reflects a limited capacity for rational decision-making [13], when combined with MPs that alleviate the “pain of payment”, further undermines cognitive control. This exacerbates the relationship between financial overconfidence and the use of AFSs. This study contributes to the existing body of literature by not only highlighting the moderating role of MPs but also by providing a theoretical basis for understanding the well-known relationship between financial overconfidence and AFS usage through the lens of bounded rationality, as well as incorporating the concept of pain of payment to explain the moderating effect of MPs.

These results have valuable implications for policymakers, educators, and financial institutions. Interventions should prioritize improving financial awareness, particularly among individuals prone to financial overconfidence. Programmatic efforts could emphasize the risks associated with AFSs and provide better insights into the usage of MPs. Policymakers could also consider regulations that foster transparency in mobile payment systems to mitigate the use of high-cost borrowing.

Despite its valuable contributions, this study has several limitations. The use of cross-sectional data, particularly from 2018, restricts the ability to draw causal inferences. Additionally, the study employed categorical measures of financial confidence, which may oversimplify the construct. Future research should address these limitations by employing longitudinal data to better capture the dynamics of financial overconfidence and AFS usage. Expanding the analysis to include data from 2021 would allow for a comparative perspective, reflecting pre- and post-COVID-19 scenarios and offering more profound insights into how the pandemic may have influenced these relationships.