Consumer Experience of Mistreatment and Fraud in Financial Services: Implications from an Integrative Consumer Vulnerability Framework

Abstract

This study utilizes the National Financial Well-Being Survey (NFWS) from the Consumer Financial Protection Bureau (CFPB) to investigate the profiles of American consumers who experience mistreatment or a type of fraud in financial services (compromised accounts). An integrative consumer vulnerability framework was used as the theoretical framework to examine how disadvantaged consumer characteristics and vulnerable consumer characteristics are associated with mistreatment and compromised accounts. Consumers in vulnerable states, due to low financial capability, cognitive decline, material hardships, financial shocks, and more exposure to various financial services, were more likely to report experiencing mistreatment and having their financial accounts compromised. Consumers from higher socio-economic status were more likely to have been victims of mistreatment and compromised accounts in financial services. These findings offer implications for consumer financial education and protection.

In 2021, more than 2.8 million fraud complaints were made to state law enforcement agencies and federal agencies, such as Consumer Financial Protection Bureau (CFPB), Better Business Bureau (BBB), and Federal Trade Commission (FTC) (Federal Trade Commission, 2022) in the USA. Among those reports, one-fourth of the complaints indicated financial loss estimated at $5.8 billion (FTC, 2022). When looking at complaints regarding financial products and services, approximately 994,000 consumer complaints were submitted to the CFPB in 2021, which is a substantial increase in volume from 2020 (542,300 complaints) and 2019 (352,400 complaints) (CFPB, 2020, 2021, 2022). The top five complaints before and during the pandemic were the same—credit or consumer reporting (44% in 2019; 59% in 2020; 71% in 2021), debt collection (21% in 2019; 15% in 2020; 12% in 2021), credit cards (8% in 2019; 7% in 2020; 4% in 2021), mortgages (8% in 2019; 5% in 2020; 3% in 2021), and checking or saving accounts (8% in 2019; 6% in 2020; 4% in 2021) (CFPB, 2020, 2021, 2022). However, dramatic changes in the volume were observed in credit or consumer reporting (107% increase between 2019 and 2020; 122% increase between 2020 and 2021), prepaid cards (106% increase between 2019 and 2020; 17% decrease between 2020 and 2021), and money transfer, money services, and virtual currencies (56% increase between 2019 and 2020; 63% increase between 2020 and 2021) (CFPB, 2020, 2021, 2022).

Despite regulatory efforts to protect consumers in the financial market, consumer complaints tend to increase with the appearance of complicated financial products and services and aggressive marketing practices (Jackson, 2015). Victims not only suffer from financial and emotional difficulties due to the problems they experience, but they also spend time and effort to redress the problems (DeLiema et al., 2017; FTC, 2022; Ganzini et al., 1990; Jackson, 2015). Half of financial fraud victims experienced severe stress, 44% experienced anxiety, and 35% experienced depression (FINRA, 2015).

The purposes of this study are to (1) describe the profiles of victims of mistreatment and a specific type of fraud in the financial services market and (2) discuss implications for consumer policy to reduce consumer problems in the financial services market. It is important to note that the measure of fraud in this study is a limited definition of fraud and deals only with unauthorized access to certain accounts. We will refer to this as “compromised accounts” throughout the paper. This study expands previous research on consumer dissatisfaction and financial fraud in several ways. First, consumer mistreatment or dissatisfaction has been explored in marketing and retailing, but there has been limited research specifically in financial services. Research indicates that the demographic profiles of those who complain about financial services differ from those who complain in other industries, highlighting the importance of understanding the differentiating factors in the financial services industry (Raval, 2020b). Next, a majority of consumer fraud research focuses specifically on the elderly population (e.g., Lichtenberg et al., 2013). By using National Financial Well-Being Survey (NFWS) data from the CFPB, this study uses information on two types of consumer problems in financial services, consumer mistreatment, and compromised accounts, from a nationally representative sample.

Using data from NFWS allows this study to identify the characteristics of the victims, not the complainers. Complaints are one channel for consumers to voice dissatisfaction. There are some reports from government agencies analyzing consumer complaints in financial services, but the focus is on complaint type and complaint resolution rather than who experienced problems. Not all consumers who experience problems in the marketplace report complaints (Garrett & Tourmanoff, 2010; Goodman & Newman, 2003). Wall (2007) suggests that victims of fraud may be hesitant to report because of their embarrassment, ignorance of the reporting process, or just simply because they want to get on with their lives. Warland et al. (1975) warned that complaints data cannot be regarded as a true measure of consumer dissatisfaction. Evidence suggests that consumers who experience problems in the marketplace and consumers who complain are not one in the same (Grønhaug & Gilly, 1991; Hogarth et al., 2001; Raval, 2020a). This is important to understand because if information is gleaned only from those who complain, then we will have an incomplete picture of who is suffering from mistreatment and fraud.

This paper uses an integrative approach of consumer vulnerability. The integrative approach embraces both class-based disadvantaged consumer perspectives and state-based vulnerable consumer perspectives (Commuri & Ekici, 2008). This approach not only allows us to examine whether the typical socio-demographic characteristics used to determine the disadvantaged consumers are viable in identifying financial services victims but also allows us to newly investigate more market-specific consumer vulnerability characteristics. The findings from this study will contribute a more comprehensive perspective on consumer dissatisfaction and fraud research and provide more specific implications to consumer educators and consumer policy agencies.

Literature Review

Consumer Mistreatment/Dissatisfaction

Research on consumer dissatisfaction started with the consumerism movement in the 1970s with a focus on consumer complaint behavior (Day, 1977; Singh, 1990; Warland et al., 1975). According to Day’s (1977) consumer complaining behavior framework, consumer satisfaction or dissatisfaction occurs after consumption, in the evaluation stage. Characteristics of the products (i.e., complexity) and circumstances of purchase and use (i.e., marketing practices) have an effect on the evaluation stage. Typical socio-demographic variables and some unexplored individual factors such as depth of experience, personal involvement, and propensity to be critical were also useful in the evaluation stage. Those who are not satisfied decide whether to act and how and where to take action (i.e., complain).

Using data from specific survey questions on credit card use, problems, and satisfaction, Hogarth et al. (2001) found that among credit card users, those with credit card problems were more likely to be single, younger, highly educated, earn more income, and have more cards compared to those without problems. Furthermore, they found that those who complained to third parties were more likely to be single, minority, younger, less educated, and earn less income compared to those who took private actions or voiced their complaints directly to sellers. The complainers were also found to be less knowledgeable and to have unfavorable attitudes towards financial institutions.

Other research has utilized the data of consumer complaints to third-party agencies, such as the CFPB, the Better Business Bureau (BBB), or the Federal Reserve System (FRS) to examine consumer problems. Garrett and Toumanoff (2010) analyzed consumer complaints filed to the BBB and matched them with the socio-demographic characteristics of zip codes. They found that that disadvantaged consumers (i.e., older consumers with lower income, less formal education, and minority status) were less likely to complain. Raval (2021) found that the complaints to the CFPB were predominantly from Black and college educated areas compared to complaints filed to the Federal Trade Commission (FTC) or the BBB. Ayres et al. (2014) used the data from consumer complaints to the CFPB and matched them with zip code level information. They found that certain financial institutions were slower in responding to consumer complaints and that responses varied by type of financial product involved, type of issue, and certain geographic regions with more vulnerable consumers receiving slower responses. Hogarth and English (2002) used the data from consumer complaints to the Federal Reserve to provide the profiles of consumers who complained.

Fraud Experience

While different terminologies such as financial fraud, financial abuse, and financial exploitation have been used interchangeably, Jackson (2015) proposed making clarifications and distinguishing between financial exploitation and financial fraud. The author clarified that financial fraud involves deception, encompassing a range of crimes such as investment fraud. Financial exploitation is a broader concept which may or may not involve fraud. The target is also different; financial exploitation mainly targets older adults, although it has also been reported with children as the victims (Betz-Hamilton, 2020), while individuals of any age can be the targets of financial fraud. The distinction between financial exploitation and financial fraud may vary from researcher to researcher.

A pilot study collaboration between the Stanford Center on Longevity and FINRA (DeLiema et al., 2017) categorized financial fraud into seven categories: (1) investment fraud, (2) consumer products and services fraud, (3) employment opportunity fraud, (4) prize and lottery fraud, (5) phantom debt collection fraud, (6) charity fraud, and (7) relationship and trust fraud. Among the 2000 online panel of US adults, half reported victimization in one or more of those categories in the past year. The study found that victims are generally younger and more diverse in race and ethnicity compared non-victims.

A report by the Bureau of Economics and the Federal Trade Commission (FTC) (Anderson, 2019) identified 19 different types of fraud. The most reported types of fraud in the survey were fraudulent weight-loss products, fraudulent computer repair, being falsely told you owe money to the government, and unauthorized billing for buyers’ club memberships and unauthorized billing on a cell phone bill. Approximately 15% of the survey population reported being the victim of one or more of the frauds asked about in the survey. Middle-aged consumers (ages 35–54) had the highest rate of fraud victimization. Education was correlated with fraud; those holding at least a bachelor’s degree were less likely to report being the victims of fraud. Women had a higher rate of reported fraud victimization than men, and those with a higher risk tolerance had a higher risk of being victimized. The report also found that people who had experienced a serious life event in the past two years, such as divorce or the death of a loved one, were more likely to report being a victim of fraud. In addition, consumers with more debt were also more likely to be victims of fraud.

Raval (2021) utilized the data sets of victims from 23 legal cases of fraudulent activities, such as payday loan applications, student debt relief, and business opportunity scams, and matched the location information of victims with the demographic information of the community at the zip code level. The author found that victimization rates were higher among the communities with more Black and elderly populations and with higher median income.

A majority of the research on fraud has focused on the elderly, and most profiling work has been done on lottery and investment fraud victims. Using the data from the Health and Retirement Study (HRS), where the participants are at least 50 years old, Lichtenberg et al. (2013) found that older adults who are relatively younger, better educated, in financial need, and with psychosocial difficulties were more likely to have experienced financial fraud. Focusing on race, Han et al. (2021) found that older black adults are less susceptible to scams compared to older white adults who had similar characteristics.

Previous studies have also found an association between human and social capital and fraud victimization. Lee and Soberon-Ferrer (1997) found that the elderly, the less educated, and the poor are more vulnerable to consumer fraud due to lower levels of cognitive ability and social interaction. Marlowe and Atiles (2005) examined the consumer fraud experience among Latino immigrants qualitatively and found that low-income immigrants are at risk of being victims to consumer fraud because they lack English proficiency and experience with the local marketplace. Deliema et al. (2019) found that those who were isolated from others and did not have anyone to discuss a scam offer, those who were financially insecure, and those with less financial knowledge were more likely to be victims of financial scams. However, other studies on investment-specific fraud found the opposite—that higher financial literacy is associated with higher rates of fraud victimization (AARP, 2007; Kieffer & Mottola, 2017). Fraud is a complex topic, and the individuals who are targeted differ based on the type of fraud. This makes it difficult to simply use demographics to identify vulnerable populations.

Theoretical Framework

To investigate the characteristics of those who have experienced problems with financial services, this study utilized an integrative consumer vulnerability framework, which incorporates a disadvantaged consumer framework with a vulnerable consumer framework (Baker et al., 2005; Commuri & Ekici, 2008; Garrett & Toumanoff, 2010; Raval, 2020a). Historically, a disadvantaged consumer framework was used to identify the consumers most likely to experience vulnerability in the marketplace. The original model of a disadvantaged consumer was based on the notion that certain demographic characteristics, such as low income, less education, minority status, and old age render a consumer disadvantaged (Garrett & Toumanoff, 2010).

The disadvantaged consumer framework has since been augmented by a vulnerable consumer framework. Baker et al. (2005) defined consumer vulnerability as “a state of powerlessness that arises from an imbalance in marketplace interactions or from the consumption of marketing messages and products” (p. 134). Unlike a disadvantaged consumer framework which assumes that consumers in some categories are always vulnerable because of certain characteristics, a vulnerable consumer framework assumes that everyone has the potential to be vulnerable depending on the situation, such as when a person is powerless or not in control. In this framework, vulnerability arises from multiple factors including individual characteristics (e.g., cognitive capacity), individual states (e.g., life transitions), and external conditions (e.g., discrimination). Vladeck (2013) also noted that the most important predictor of vulnerability to fraud was not a demographic factor, such as age, education level, race, or gender, but whether “individuals have more debt than they can comfortably handle” (p. 367). According to Vladeck, individuals can be in debt due to things outside of their control, like an unexpected negative life shock or an economic downturn, and this may cloud their judgment and make them vulnerable to fraud (Vladeck, 2013).

Rather than abandoning the disadvantaged consumer perspective, Commuri and Ekici (2008) proposed an integrative view of consumer vulnerability, which incorporates a systematic, class-based disadvantaged consumer component along with a transient, state-based vulnerable consumer component. They argued that a class-based perspective is still useful because a certain demographic group can be vulnerable not because of their characteristics but because of the exploitative practices targeted to that group. Indeed, certain types of frauds such as pyramid schemes and Ponzi scams are known to target specific ethnic, racial, religious, professional, or age groups (Bosley & Knorr, 2018).

Following the integrative consumer vulnerability framework, this study examines class-based disadvantaged consumer characteristics and state-based vulnerable consumer characteristics in the empirical models on consumer mistreatment and compromised accounts in financial services. Based on previous research focusing on socio-demographic characteristics of consumers (Clifton et al., 2014; Garrett & Toumanoff, 2010), this study classifies consumers as disadvantaged if they are elderly, female, single, a racial or ethnic minority, less educated, have lower income, or are living in rural area or county with high poverty. A review of the literature on consumer vulnerability guided this study to regard vulnerable consumers as those who lack financial capabilities (e.g., financial skills and knowledge); those who are in difficult situations to make sound decisions (e.g., declines in cognitive ability, experience of material hardships); had negative shocks in their lives; and those who are more exposed to the risk in the market (e.g., owning several financial products and services) (Anderloni et al., 2012; DeLiema et al., 2020; Financial Conduct Authority, 2015; Goodman & Newman, 2003; Grønhaug & Gilly, 1991; Hogarth et al., 2001; Lee & Soberon-Ferrer, 1997; Lichtenberg et al., 2013; O’Connor et al., 2019; Ofgem, 2013; Vladeck, 2013). Disadvantaged or vulnerable consumers are expected to be more likely to experience problems with financial services. The hypotheses tested in this study are as follows:

• H1a: Disadvantaged consumers are more likely to have experienced mistreatment in financial services.

• H1b: Vulnerable consumers are more likely to have experienced mistreatment in financial services.

• H2a: Disadvantaged consumers are more likely to have experienced compromised financial accounts.

• H2b: Vulnerable consumers are more likely to have experienced compromised financial accounts.

Methods

Data

This study uses National Financial Well-Being Survey (NFWS) data from the Consumer Financial Protection Bureau (CFPB). The CFPB was created in 2011 to provide a single point of accountability for enforcing federal consumer financial laws and protecting consumers in the financial marketplace. The NFWS was launched to collect data on the factors that support consumer financial well-being of Americans. Between October 2016 and December 2016, 6394 adult respondents completed the survey on financial well-being and variables that are hypothesized to influence financial well-being, such as individual and household characteristics, financial experiences and skills, and financial attitudes and behaviors (CFPB, 2017). The sample of NFWS includes a both general population sample, which represents adults in the USA across key population groups (age, race and ethnicity, and poverty level), and an oversample of the those aged 62 and older (CFPB, 2017).

The sample size of this study was reduced to 6,205 for the mistreatment model (Model 1) and to 5,738 for the compromised accounts model (Model 2). In both models, the sample size was reduced because 119 respondents refused to answer about employment, 25 did not answer about material hardships, and 69 did not answer about cognitive decline. In the mistreatment model, 50 respondents refused to answer the question about mistreatment. Because each of these variables accounted for less than 2% of the total sample, we proceeded with listwise deletion. The sample size for the compromised accounts model was further reduced because 512 reported that they did not know whether they had experienced this type of fraud. These samples were excluded from the analysis, though a robustness test was run on the excluded sample which is presented in the discussion section. This study used a weight for descriptive analysis to adjust the sample to be representative of the US population because the sample was not selected with equal probability in the survey (CFPB, 2017).

Dependent Variables

The dependent variables of this study are two separate variables of mistreatment and fraud (compromised accounts) in financial services. Respondents were asked “How often have you had experiences with financial services where you did not feel respected or where you felt mistreated?” The response options were 1 = never, 2 = rarely, 3 = sometimes, and 4 = often. Respondents were also asked “In the past 5 years, has someone without your permission used or attempted to use an existing account of yours, such as a credit or debit card, checking, savings, telephone, online, or insurance account?” The compromised accounts variable was created as a binary variable, with yes = 1 and no = 0.

Independent Variables

Independent variables are categorized into two groups, which are disadvantaged consumer factors and vulnerable consumer factors. Disadvantaged consumer factors include socio-demographic variables, and vulnerable consumer factors include financial capability variables, situational variables, and exposure variables.

Disadvantaged Consumer Factors

Following the literature on disadvantaged consumers, this study included socio-demographic variables, such as generation (age), gender, marital status, race/ethnicity, education level, employment status, and household income. Based on the information on respondents’ age at the survey period in fall 2016, the NFWS created four categories of generations for user convenience: Millennial (age between18 and 35), Gen X (age between 36 and 51), Boomer (age between 52 and 70), and Pre-Boomer (age 71 and older). Gender is a binary variable with female as 1 and male as 0. In regards with race and ethnicity, respondents were categorized as non-Hispanic White, non-Hispanic Black, Hispanic, and non-Hispanic others. Marital status is categorized as married or cohabiting, never married, and others (widowed, divorced, and separated). Education level has five categories: less than high school, high school degree/GED, some college/associate degree, bachelor’s degree, and higher. Employment status is categorized as currently working, not working, and retired. Responses to household income in the NFWS survey were originally reported with detailed categories but combined into low income (less than $40,000), middle income (between $40,000 and $99,999), and high income ($100,000 or more).

Vulnerable Consumer Factors

Financial Capability Variables

Financial capability variables include financial skill and financial knowledge. A financial skill scale was constructed by the CFPB using 10 items such as “I know how to make complex financial decisions” and “I know where to find the advice I need to make decisions involving money.” The details on how the financial skill scale score was developed are reported in the CFPB’s (2018a) guide on this topic. The score ranges from 0 to 100. Financial knowledge was measured by the number of correct answers to Lusardi and Mitchell (2007)’s three financial knowledge questions on compound interest, inflation, and diversification, ranging from 0 to 3.

Situational Variables

Situational variables include the respondents’ experiences in the past 12 months which may put the respondents in a vulnerable position. The first situational variable is cognitive ability. Respondents were asked if they experienced confusion or memory loss that is happening more often or has gotten worse during the past 12 months. A binary variable of decline in cognitive ability was constructed from the response. Respondents were also asked to indicate if they experienced various types of material hardships in the past 12 months: (1) worried whether food would run out, (2) food did not last or they did not have money to get more, (3) could not afford a place to live, (4) needed to see a doctor or go to the hospital but did not go because they could not afford it, (5) stopped taking a medication or took less than directed due to the costs, and (6) utilities were shut off due to non-payment. The material hardship variable ranges from 0 to 6, by summing the number of experiences respondents reported. Lastly, respondents were asked to indicate if they or any members of their household experienced a variety of shocks. The original survey questions consist of 11 items: (1) lost a job, (2) had work hours and/or pay reduced or financial difficulty in business, (3) received a foreclosure notice, (4) had a major car or home repair, (5) had a health emergency, (6) got a divorce or separation, (7) added a child to the household, (8) experienced the death of primary breadwinner, (9) received a large sum of money beyond normal income, (10) had a child start daycare or college, and (11) provided unexpected financial support to a family member or friend. This study constructed a continuous variable of shock experience by only using 10 items and constructed a separate binary variable using the excluded item (i.e., receiving a large sum of money beyond normal income) since it was the only item viewed as a positive financial experience.

Exposure Variables

This group of variables includes two separate variables on the use of financial products and services. The first variable is a measure of the number of financial products or services the respondent currently owns. They include products or services such as (1) checking or savings account at a bank or credit union, (2) life insurance, (3) health insurance, (4) retirement account, (5) pension, (6) non-retirement investments, (7) education savings account, and (8) student/education loan. The resulting variable is the sum of all the responses selected by the respondent. The second variable is a measure of the number of non-traditional financial products or services the respondent has used in the past 12 months. These products include (1) payday loan or cash advance loan, (2) pawn loan or auto title loan, (3) a reloadable card that is not linked with a checking or savings account, (4) a place other than a bank or credit union to give or send money to relatives or friends outside the USA, and (5) a place other than a bank or credit union to cash a check or purchase a money order. The resulting variable is the sum of all the responses selected by the respondent.

Analysis

A generalized probit model was used to estimate the mistreatment experience model. Since the response options for mistreatment experience are ordinal (i.e., never, rarely, sometimes, and often) and a large proportion (40%) of the sample never experienced mistreatment, we considered two-stage models in which we estimate the probability of mistreatment experience first and then estimate the frequency of mistreatment experience among those who reported mistreatment. However, Cragg’s tobit was not appropriate because the dependent variable was not continuous, and an ordered logit/probit approach was not appropriate because the Brant test (Brant, 1990) revealed several variables violate the parallel line assumption (e.g., education, employment, number of material hardship, and number of traditional financial products). This means that the estimated coefficients vary across the different levels of the dependent variable. A generalized ordered probit model provides information on how each of the variables associates to mistreatment experience differently depending on the frequency. A generalized probit also produces more parsimonious and interpretable results than those fitted by a non-ordinal method, such as a probit regression (Williams, 2006). To test the hypotheses on fraud experience, a binary probit regression model was used because the response variable is dichotomous.

Results

Sample Description

The sample description is reported in Table 1, with and without sample weights. The weighted sample of the NFWS is representative of the US population in demographic characteristics such as age, sex, education, and race/ethnicity (CFPB, 2017). Therefore, the demographic characteristics of our samples were quite comparable to those of the general US population. For example, according to the US census, the educational attainment of the population 18 years and over in 2016 was as follows: 11.7% with less than high school, 29.0% with high school graduate, 28.7% with some college, 19.5% with bachelor’s degree, and 11.2% with graduate degree, which are very close to our sample characteristics.

Table 1.

Descriptive statistics

	Sample for mistreatment model (N = 6,205)		Sample for fraud model (N = 5,738)
	Unweighted	Weighted	Unweighted	Weighted
Mistreatment experience
Never	0.400	0.403	-	-
Rarely	0.417	0.403	-	-
Sometimes	0.157	0.163	-	-
Often	0.027	0.031	-	-
Fraud experience	-	-	0.289	0.270
Disadvantaged consumer variables
Generation
Pre-Boomer (age ≥ 71)	0.176	0.123	0.182	0.127
Boomer (age between 52 and 70)	0.355	0.310	0.362	0.317
Gen-X (age between 36 and 51)	0.222	0.249	0.221	0.249
Millennial (age between 18 and 35)	0.247	0.318	0.235	0.307
Female	0.476	0.516	0.476	0.515
Marital status
Couple	0.657	0.621	0.666	0.630
Separated, divorced, widowed	0.164	0.157	0.164	0.157
Never married	0.179	0.222	0.170	0.213
Race/ethnicity
White, Non-Hispanic	0.709	0.649	0.719	0.661
Black, Non-Hispanic	0.106	0.117	0.099	0.111
Other, Non-Hispanic	0.053	0.081	0.052	0.078
Hispanic	0.133	0.154	0.130	0.151
Education
Less than high school	0.065	0.115	0.006	0.110
High school	0.252	0.288	0.250	0.286
Some college	0.303	0.287	0.301	0.285
Bachelor’s degree	0.207	0.197	0.209	0.202
Graduate school	0.173	0.113	0.179	0.118
Employment
Work for employer	0.533	0.573	0.533	0.575
Currently not working	0.174	0.218	0.164	0.207
Retired	0.293	0.209	0.304	0.217
Household income
Low (income < $40 k)	0.284	0.318	0.271	0.305
Middle ($40 k ≤ income < $100 k)	0.405	0.370	0.412	0.377
High ($100 k ≤ income)	0.312	0.312	0.318	0.318
Vulnerable consumer variables
Financial knowledge score (0 ~ 3)	2.522	2.444	2.545	2.469
Financial skill score (0 ~ 100)	50.801	49.882	51.254	50.378
Cognitive ability decline	0.104	0.107	0.099	0.101
# of material hardships (0 ~ 6)	0.786	0.950	0.699	0.852
# of financial shocks (0 ~ 10)	0.721	0.751	0.699	0.726
Received a large sum of money	0.065	0.065	0.066	0.065
# of traditional financial products (0 ~ 8)	3.555	3.274	3.635	3.357
# of alternative financial services (0 ~ 5)	0.239	0.279	0.221	0.257

In the sample for the mistreatment model, 40% report they were never respected or mistreated when using financial services, 40% rarely felt this way, 16% said sometimes they felt this way, and 3% said they often felt this way. In the sample for the compromised accounts model, about 27% reported that they experienced this type of fraud in the past five years.

The two samples were quite similar in disadvantaged consumer characteristics. Millennials (ages 18–35) and Boomers (ages 52–70) were the highest proportion in both samples each accounting for about 30% of the sample, followed by Gen-Xers (36–51) (25%) and Pre-Boomers (age 71 and older) (12 ~ 13%). Couples and Whites accounted for approximately two-thirds of the sample in marital status categories and race/ethnicity categories, respectively. About 30% of the sample were high school graduates, and another 30% had some college experiences, followed by bachelor’s degree (20%), graduate degree (11 ~ 12%), and not graduating high school (11 ~ 12%). More than half of the respondents in each sample were currently working, while about 20% were not working, and another 20% were retired. Slightly less than one-third of each sample had household income less than $40,000 and more than $100,000, respectively.

The two samples also showed similar characteristics in vulnerable consumer characteristics. Both samples had high financial knowledge (about 2.5 from the ranges between 0 and 3) and scored about the middle (about 50 from the ranges between 0 and 100) for financial skill, on average. About 10% of each sample experienced confusion or memory loss more frequently or more severely during the past 12 months. On average, respondents experienced slightly less than one type of material hardship and less than one financial shock in the past 12 months. Respondents in each sample reported that they use more than three traditional financial products among eight and less than one alternative financial service among five, on average.

Multivariate Results

Model 1: Mistreatment Experience

The full results of a generalized ordered probit regression of Model 1 are shown in Table 2. As seen in the table, different levels of coefficients were reported for some of the variables because they violated the parallel line assumptions, indicating they associate to mistreatment differently across the level of frequency. These variables include education level, employment status, financial skill score, number of material hardship, and number of traditional financial products.

Table 2.

Generalized ordered probit regression results on mistreatment experience

		Coef	S.E	95% C.I
				Lower	Upper
Disadvantaged consumer variables
Generation	Boomer (ages 52–70)	0.185***	0.049	0.089	0.281
(Ref: Pre-Boomer)	Gen-X (ages 36–51)	0.109†	0.062	− 0.012	0.231
	Millennials (ages 18–35)	0.134*	0.064	0.009	0.260
Gender (Ref: male)	Female	− 0.126***	0.030	− 0.184	− 0.067
Marital status	Separated, divorced, widowed	0.067	0.042	− 0.015	0.149
(Ref: couple)	Never married	− 0.091*	0.043	− 0.175	− 0.007
Race/ethnicity	Black	− 0.011	0.050	− 0.109	0.087
(Ref: White)	Hispanic	− 0.113*	0.046	− 0.203	− 0.023
	Other	0.029	0.064	− 0.096	0.155
Education	High school [1]	0.120	0.075	− 0.026	0.266
(Ref: < high school)	High school [2]	− 0.121	0.083	− 0.284	0.042
	High school [3]	− 0.276*	0.129	− 0.529	− 0.024
	Some college [1]	0.278***	0.076	0.129	0.426
	Some college [2]	0.047	0.083	− 0.116	0.209
	Some college [3]	− 0.219†	0.128	− 0.470	0.032
	Bachelor’s degree [1]	0.401***	0.083	0.239	0.563
	Bachelor’s degree [2]	0.025	0.092	− 0.156	0.205
	Bachelor’s degree [3]	− 0.340*	0.152	− 0.639	− 0.041
	Graduate school [1]	0.407***	0.086	0.238	0.576
	Graduate school [2]	0.189*	0.095	0.002	0.377
	Graduate school [3]	− 0.175	0.158	− 0.485	0.134
Employment	Not working [1]	− 0.168**	0.049	− 0.264	− 0.072
(Ref: Working)	Not working [2]	0.015	0.053	− 0.090	0.120
	Not working [3]	0.080	0.087	− 0.090	0.250
	Retired	− 0.071	0.046	− 0.161	0.019
Household income	Middle (40 ~ 100 k)	0.094*	0.040	0.016	0.172
(Ref: low(< $40 k))	High (≥ 100 k)	0.049	0.048	− 0.044	0.143
Vulnerable consumer variables
Financial knowledge score		0.029	0.022	− 0.014	0.073
Financial skill score [1]		− 0.009***	0.001	− 0.012	− 0.006
Financial skill score [2]		− 0.007***	0.002	− 0.010	− 0.004
Financial skill score [3]		0.000	0.003	− 0.005	0.006
Cognitive ability decline		0.151**	0.048	0.057	0.244
# of material hardships [1]		0.147***	0.014	0.119	0.175
# of material hardships [2]		0.207***	0.014	0.180	0.233
# of material hardships [3]		0.183***	0.020	0.144	0.221
# of financial shocks		0.085***	0.016	0.054	0.115
Received a large sum of money	0.041		0.058	-0.072	0.155
# of traditional financial products [1]	0.006		0.012	− 0.017	0.030
# of traditional financial products [2]	− 0.036	**	0.014	− 0.064	− 0.009
# of traditional financial products [3]	− 0.030		0.025	− 0.080	0.020
# of alternative financial services	0.155***	0.030	0.097	0.214
Constant [1]		0.099	0.128	− 0.152	0.349
Constant [2]		− 0.951***	0.141	− 1.227	− 0.675
Constant [3]		− 2.230***	0.207	− 2.637	− 1.823
Pseudo R2				0.066

[1] indicates a threshold between “Never” and “Rarely”; [2] indicates a threshold between “Rarely” and “Sometimes”; [3] indicates a threshold between “Sometimes” and “Often”

^† < . p < .1

^*p < .05

^**p < .01

^***p < .001

Some of the disadvantaged consumer factors consistently showed a significant relationship with the frequency of mistreatment experience but in the opposite direction that was expected from previous research. Boomers (ages 52–70), Gen-Xers (ages 36–51) (p < 0.1), and Millennials (ages 18–35) experienced mistreatment more often, compared to Pre-Boomers (age 71 and older). Females and Hispanics felt disrespected or mistreated when using financial services less frequently compared to males and white respondents. Those who never married experienced mistreatment less frequently than coupled individuals, and those with middle income (between $40,000 and $100,000) experienced disrespect or mistreatment more often compared to those with low income (less than $40,000).

Education level and employment status were significant in the mistreatment model, but the relationship varies depending on frequency. More educated individuals were more likely to have experienced mistreatment (rarely, sometimes, and often) rather than “never.” However, being more educated was related to having experienced mistreatment less frequently than “often.” For example, compared to those who had not graduated from high school, those with an associate degree or some college experience, those with bachelor’s degree, and those with graduate or professional degree were 10.4, 14.8, and 15.0 percentage points less likely to have “never” experienced mistreatment. However, those with bachelor’s degree were 2.0 percentage points less likely to have experienced mistreatment “often” compared to those without high school degree (see Table 3). Not working was negatively related to being mistreated only at the threshold between “never” and “rarely.”

Table 3.

Average marginal effects of selected variables from generalized ordered probit regression results on mistreatment experience

Variables	Predict	Average marginal effects
Education (Ref: < high school)
High school	1 (None)	− 0.045
	2 (Rarely)	0.072**
	3 (Sometimes)	− 0.009
	4 (Often)	− 0.017†
Some college	1 (None)	− 0.104***
	2 (Rarely)	0.093***
	3 (Sometimes)	0.025
	4 (Often)	− 0.014†
Bachelor’s degree	1 (None)	− 0.148***
	2 (Rarely)	0.142***
	3 (Sometimes)	0.026
	4 (Often)	− 0.020*
Graduate school	1 (None)	− 0.150***
	2 (Rarely)	0.103***
	3 (Sometimes)	0.059**
	4 (Often)	− 0.012
Employment (Ref: working)
Not working	1 (None)	0.062**
	2 (Rarely)	− 0.065***
	3 (Sometimes)	− 0.001
	4 (Often)	0.005
Financial skill score	1 (None)	0.003***
	2 (Rarely)	− 0.002**
	3 (Sometimes)	− 0.002***
	4 (Often)	0.000
# of material hardships	1 (None)	− 0.054***
	2 (Rarely)	0.005
	3 (Sometimes)	0.039***
	4 (Often)	0.010***
# of traditional financial products	1 (None)	− 0.002
	2 (Rarely)	0.011*
	3 (Sometimes)	− 0.007*
	4 (Often)	− 0.002

Average marginal effects are presented for variables which violate the parallel line assumption in Table 2. Full results are available upon request

^† < . p < .1

^*p < .05

^**p < .01

^***p < .001

In contrast, most of the vulnerable consumer variables were significant with the frequency of mistreatment in the same direction expected from the hypotheses. Those who experienced a decline in cognitive ability, those who suffered from more financial shocks, and those who were using more alternative financial services were more likely to have experienced mistreatment compared to their counterparts. Financial skill was negatively related with the frequency of mistreatment at the thresholds between “none” and “rarely” and between “rarely” and “sometimes,” but it was not significant at the threshold between “sometimes” and “often.” The number of material hardships experienced in the past 12 months was positively related to the frequency of mistreatment, and the relationships were stronger at the higher thresholds of frequency level. The number of traditional financial products used was negatively related to mistreatment only at the threshold between “rarely” and “sometimes.” Financial knowledge was not a significant factor in the mistreatment model.

To examine which type of vulnerability factors is significant in the mistreatment model, we included individual items of material hardship, financial shocks, traditional financial products, and alternative financial services, instead of the composite measures of them. Table 5 in the Appendix presents the results from analysis using individual measurements. Among the six material hardship items, two food-related items and two medical-related items were positively related with experiencing mistreatment more frequently. Not being able to afford a place to live was positively related but only at the threshold between “rarely” and “sometimes.” Among ten financial shocks, having work hours or pay reduced, having a major car or home repair, a health emergency, and providing unexpected financial support to a family member or friend were positively related with the frequency of mistreatment. However, losing a job was negatively related to mistreatment only at the threshold between “never” and “rarely.” The use of traditional financial products showed limited association to mistreatment experience. Ownership of life insurance or retirement accounts was both negatively associated with mistreatment at a certain threshold. Except for non-bank international transfers, using alternative financial services was positively related to the frequency of mistreatment.

Table 5.

Generalized ordered probit regression results on mistreatment experience: individual measurements

		Coef	S.E	95% C.I
				Lower	Upper
Disadvantaged consumer variables
Generation	Boomer (ages 52–70)	0.188***	0.049	0.092	0.285
(Ref: Pre-Boomer)	Gen-X (ages 36–51)	0.120†	0.063	− 0.003	0.244
	Millennials (ages 18–35)	0.138*	0.067	0.007	0.269
Gender (Ref: male)	Female	− 0.138***	0.030	− 0.196	− 0.079
Marital status	Separated, divorced, widowed	0.078†	0.043	− 0.005	0.162
(Ref: couple)	Never married	− 0.093*	0.044	− 0.179	− 0.008
Race/ethnicity	Black [1]	− 0.051	0.058	− 0.164	0.063
(Ref: White)	Black [2]	0.094	0.063	− 0.029	0.217
	Black [3]	− 0.059	0.111	− 0.276	0.159
	Hispanic	− 0.103*	0.046	− 0.194	− 0.012
	Other	0.039	0.064	− 0.087	0.165
Education	High school [1]	0.143†	0.075	− 0.004	0.290
(Ref: < high school)	High school [2]	− 0.133	0.083	− 0.296	0.031
	High school [3]	− 0.265*	0.130	− 0.520	− 0.011
	Some college [1]	0.299***	0.076	0.150	0.448
	Some college [2]	0.012	0.083	− 0.151	0.175
	Some college [3]	− 0.229†	0.128	− 0.480	0.022
	Bachelor’s degree [1]	0.433***	0.083	0.270	0.596
	Bachelor’s degree [2]	− 0.022	0.092	− 0.202	0.159
	Bachelor’s degree [3]	− 0.324*	0.152	− 0.621	− 0.027
	Graduate school [1]	0.443***	0.087	0.274	0.613
	Graduate school [2]	0.139	0.095	− 0.048	0.326
	Graduate school [3]	− 0.146	0.155	− 0.450	0.159
Employment	Not working [1]	− 0.142**	0.050	− 0.240	− 0.044
(Ref: working)	Not working [2]	0.004	0.055	− 0.104	0.111
	Not working [3]	0.011	0.090	− 0.166	0.188
	Retired [1]	− 0.022	0.051	− 0.123	0.078
	Retired [2]	− 0.141*	0.059	− 0.256	− 0.026
	Retired [3]	− 0.276*	0.112	− 0.496	− 0.057
Household income	Middle (40 ~ 100 k)	0.098*	0.040	0.019	0.177
(Ref: low (< $40 k))	High (≥ 100 k)	0.059	0.049	− 0.037	0.154
Vulnerable consumer variables
Financial knowledge score		0.021	0.023	− 0.023	0.066
Financial skill score [1]		− 0.009***	0.001	− 0.012	− 0.006
Financial skill score [2]		− 0.007***	0.002	− 0.011	− 0.004
Financial skill score [3]		0.001	0.003	− 0.004	0.007
Cognitive ability decline		0.145**	0.048	0.051	0.239
Material hardships
Worry about food		0.164**	0.063	0.041	0.288
No money for food		0.192**	0.067	0.061	0.322
No money for a place to live [1]		− 0.016	0.077	− 0.167	0.135
No money for a place to live [2]		0.206**	0.074	0.060	0.352
No money for a place to live [3]		0.089	0.101	− 0.110	0.288
No money to see a doctor		0.184***	0.053	0.080	0.287
Stopped taking a medication		0.256***	0.053	0.152	0.361
Utility shut-off		0.129†	0.074	− 0.016	0.273
Financial shocks
Lost a job [1]		− 0.185*	0.072	− 0.326	− 0.043
Lost a job [2]		0.085	0.075	− 0.062	0.232
Lost a job [3]		0.068	0.118	− 0.163	0.299
Reduced work/pay		0.119*	0.059	0.003	0.234
Received a foreclosure notice		0.128	0.160	− 0.185	0.441
Car/home repair		0.093	0.036	0.022	0.163
Health emergency		0.124**	0.042	0.041	0.207
Divorce/separation		− 0.012	0.107	− 0.222	0.198
Added a child [1]		0.043	0.097	− 0.147	0.233
Added a child [2]		0.194†	0.101	− 0.003	0.391
Added a child [3]		− 0.303	0.209	− 0.712	0.106
Death of breadwinner		− 0.034	0.150	− 0.328	0.260
Child starting daycare/college		− 0.026	0.076	− 0.175	0.124
Provided financial support		0.142**	0.043	0.058	0.226
Received a large sum of money		0.044	0.058	− 0.070	0.159
Traditional financial products
Checking/savings account		− 0.014	0.046	− 0.104	0.076
Life insurance [1]		0.015	0.037	− 0.057	0.088
Life insurance [2]		− 0.102*	0.043	− 0.186	− 0.017
Life insurance [3]		− 0.040	0.082	− 0.200	0.120
Health insurance		0.017	0.038	− 0.056	0.091
Retirement account [1]		− 0.002	0.041	− 0.084	0.079
Retirement account [2]		− 0.076	0.048	− 0.169	0.018
Retirement account [3]		− 0.226**	0.087	− 0.397	− 0.056
Pension		− 0.043	0.036	− 0.113	0.026
Non-retirement investments		0.035	0.036	− 0.035	0.106
Education savings account		− 0.055	0.062	− 0.177	0.066
Student loan		0.034	0.045	− 0.054	0.122
Alternative financial services
Payday loan/cash advance loan		0.263**	0.089	0.089	0.438
Pawn loan/auto title loan		0.190†	0.109	− 0.024	0.403
Reloadable card		0.098†	0.052	− 0.005	0.200
Non-bank international transfer		0.114	0.072	− 0.026	0.255
Non-bank money order		0.180**	0.060	0.063	0.297
Constant [1]		0.081	0.131	− 0.175	0.337
Constant [2]		−0.918***	0.142	− 1.196	− 0.640
Constant [3]		− 2.182***	0.203	− 2.580	− 1.785
Pseudo R2				0.071

[1] indicates a threshold between “Never” and “Rarely”; [2] indicates a threshold between “Rarely” and “Sometimes”; [3] indicates a threshold between “Sometimes” and “Often”

^† < . p < .1

^*p < .05

^**p < .01

^***p < .001

Model 2: Compromised Accounts

The full probit regression results for Model 2 are in Table 4.The findings for compromised accounts were mixed. Among the disadvantaged consumer characteristics, results for generation were consistent with the hypothesis. Millennials (ages 18–35) were less likely to have experienced this type of fraud compared to Pre-Boomers (ages 71 and older). However, most other disadvantaged consumer characteristics were in the opposite direction than expected. Individuals in a relationship, with more education, and currently working were more likely to have experienced this type of fraud compared to those who were single, less educated, and not working. For example, compared to those without a high school diploma, respondents with more education were more likely to have compromised accounts (high school diploma, 7.6%; associate degree or some college experience, 12.5%; bachelor’s degree, 18.1%; graduate or professional degree, 20.3%). Also, those with high income ($100,000 or more) were 4.5 percentage points more likely to have experienced this type of fraud compared to those with low household income (less than $40,000).

Table 4.

Probit regression results on compromised accounts experience

		Coef	S.E	95% C.I		AME
				Lower	Upper
Disadvantaged consumer variables
Generation	Boomer (ages 52–70)	− 0.039	0.059	− 0.155	0.078	− 0.013
(Ref: Pre-Boomer)	Gen-X (ages 36–51)	− 0.040	0.076	− 0.189	0.109	− 0.013
	Millennials (ages 18–35)	− 0.164*	0.080	− 0.320	− 0.009	− 0.052
Gender (Ref: male)	Female	0.024	0.037	− 0.049	0.097	0.008
Marital status	Separated, divorced, widowed	− 0.129*	0.053	− 0.234	− 0.025	− 0.041
(Ref: couple)	Never married	− 0.135*	0.056	− 0.245	− 0.024	− 0.043
Race/ethnicity	Black	− 0.070	0.066	− 0.199	0.059	− 0.022
(Ref: White)	Hispanic	0.063	0.059	− 0.053	0.179	0.021
	Other	− 0.112	0.084	− 0.275	0.052	− 0.035
Education	High school	0.289**	0.097	0.099	0.479	0.076
(Ref: < high school)	Some college	0.447***	0.097	0.257	0.638	0.125
	Bachelor’s degree	0.612***	0.103	0.409	0.814	0.181
	Graduate school	0.672***	0.106	0.464	0.880	0.203
Employment	Not working	− 0.128*	0.057	− 0.239	− 0.017	− 0.041
(Ref: working)	Retired	− 0.050	0.057	− 0.161	0.062	− 0.016
Household income	Middle (40 ~ 100 k)	0.030	0.052	− 0.071	0.131	0.009
(Ref: low (< $40 k))	High (≥ 100 k)	0.139*	0.061	0.021	0.258	0.045
Vulnerable consumer variables
Financial knowledge score		0.110***	0.030	0.052	0.168	0.036
Financial skill score		− 0.003*	0.002	− 0.006	0.000	− 0.001
Cognitive ability decline		0.244***	0.060	0.126	0.363	0.083
# of material hardships		0.049**	0.015	0.018	0.079	0.016
# of financial shocks		0.132***	0.019	0.094	0.170	0.043
Received a large sum of money		0.128†	0.070	− 0.011	0.266	0.042
# of traditional financial products		0.041**	0.013	0.015	0.067	0.013
# of alternative financial services		0.100*	0.039	0.023	0.177	0.032
Constant		− 1.411***	0.159	− 1.724	− 1.099
Pseudo R2				0.051

^† < . p < .1

^*p < .05

^**p < .01

^***p < .001

All vulnerable consumer characteristics were significantly related with this type of fraud experience in the expected direction, except financial knowledge. Surprisingly, financial knowledge was positively related with compromised accounts. Other characteristics, such as lower financial skill, decline in cognitive ability, number of material hardships and financial shocks, and number of traditional and alternative financial products and services used, were all positively related with this type of fraud experience.

The results from an additional analysis using individual measurements of material hardship, financial shocks, traditional financial products, and alternative financial services are presented in Table 6 in the Appendix. Having stopped taking medication due to the cost was positively related to having accounts compromised. Similar to the mistreatment experience, having work hours or pay reduced, a health emergency, and providing unexpected financial support to a family member or friend were positively related to this type of fraud experience. Among traditional financial products, having a retirement account was positively related to compromised financial accounts. Among alternative financial services, using payday loans, reloadable cards, and non-bank international transfers were positively related to having financial accounts compromised (p < 0.1).

Table 6.

Probit regression results on compromised accounts experience: individual measurements

		Coef	S.E	95% C.I
				Lower	Upper
Disadvantaged consumer variables
Generation	Boomer (ages 52–70)	− 0.048	0.060	− 0.166	0.069
(Ref: Pre-Boomer)	Gen-X (ages 36–51)	− 0.019	0.078	− 0.171	0.133
	Millennials (ages 18–35)	− 0.127	0.083	− 0.290	0.036
Gender (Ref: male)	Female	0.017	0.038	− 0.056	0.091
Marital status	Separated, divorced, widowed	− 0.104†	0.054	− 0.211	0.003
(Ref: couple)	Never married	− 0.146*	0.057	− 0.259	− 0.034
Race/ethnicity	Black	− 0.051	0.067	− 0.182	0.080
(Ref: White)	Hispanic	0.066	0.060	− 0.051	0.184
	Other	− 0.109	0.084	− 0.273	0.056
Education	High school	0.301**	0.098	0.108	0.494
(Ref: < high school)	Some college	0.457***	0.099	0.263	0.650
	Bachelor’s degree	0.618***	0.105	0.413	0.824
	Graduate school	0.683***	0.108	0.472	0.895
Employment	Not working	− 0.099†	0.058	− 0.213	0.014
(Ref: Working)	Retired	− 0.042	0.059	− 0.157	0.074
Household income	Middle (40 ~ 100 k)	0.023	0.053	− 0.080	0.126
(Ref: low (< $40 k))	High (≥ 100 k)	0.135*	0.062	0.014	0.257
Vulnerable consumer variables
Financial knowledge score		0.100**	0.030	0.041	0.159
Financial skill score		− 0.003*	0.002	− 0.007	0.000
Cognitive ability decline		0.237***	0.061	0.117	0.356
Material hardships
Worry about food		0.015	0.086	− 0.154	0.185
No money for food		0.116	0.092	− 0.064	0.295
No money for a place to live		− 0.150	0.094	− 0.334	0.034
No money to see a doctor		− 0.001	0.070	− 0.138	0.137
Stopped taking a medication		0.203**	0.070	0.065	0.340
Utility shut-off		0.105	0.103	− 0.097	0.308
Financial shocks
Lost a job		− 0.102	0.084	− 0.267	0.062
Reduced work/pay		0.231**	0.077	0.081	0.381
Received a foreclosure notice		− 0.009	0.214	− 0.428	0.410
Car/home repair		0.206***	0.044	0.119	0.293
Health emergency		0.154**	0.052	0.051	0.257
Divorce/separation		− 0.167	0.145	− 0.451	0.117
Added a child		− 0.035	0.106	− 0.243	0.172
Death of breadwinner		0.046	0.191	− 0.328	0.420
Child starting daycare/college		0.062	0.094	− 0.122	0.245
Provided financial support		0.185***	0.053	0.081	0.288
Received a large sum of money		0.132†	0.071	− 0.007	0.271
Traditional financial products
Checking/savings account		0.055	0.062	− 0.068	0.177
Life insurance		0.008	0.041	− 0.073	0.088
Health insurance		0.004	0.048	− 0.091	0.099
Retirement account		0.158**	0.046	0.067	0.248
Pension		0.022	0.044	− 0.063	0.107
Non-retirement investments		0.035	0.043	− 0.050	0.120
Education savings account		− 0.056	0.073	− 0.200	0.087
Student loan		0.040	0.057	− 0.071	0.151
Alternative financial services
Payday loan/cash advance loan		0.196†	0.117	− 0.034	0.425
Pawn loan/auto title loan		0.027	0.156	− 0.279	0.333
Reloadable card		0.114†	0.068	− 0.020	0.247
Non-bank international transfer		0.162†	0.094	− 0.022	0.346
Non-bank money order		0.055	0.081	− 0.104	0.215
Constant		− 1.438***	0.166	− 1.764	− 1.112
Pseudo R2				0.058

^† < . p < .1

^*p < .05

^**p < .01

^***p < .001

Discussion

Several important findings emerge from this study. The first is the large number of consumers who have experienced mistreatment and compromised financial accounts. About 60% of the sample reported that they have experienced being disrespected or mistreated with financial services at least once, and about 30% of the sample reported that they have had financial accounts compromised in the past five years. The survey conducted by the Stanford Center on Longevity and FINRA Investor Education Foundation (DeLiema et al., 2017) found 50% of respondents reported financial fraud, which is higher than the rate in this study, and it is probably due to more expansive measure of financial fraud used in that survey. Similarly, the FTC survey on mass-market consumer fraud (Anderson, 2019) found that just under 16% of the sample reported being victims of fraud. The measure of fraud in the NFWS survey is limited to unauthorized access of certain accounts. This aligns with two of the top five reported frauds in the FTC study, which is likely why the percentage reporting fraud in the NFWS is slightly higher. While there is no comparable recent data on consumer mistreatment, the complex attributes of financial services can be a contributing factor to the widespread consumer problems with financial services, as this makes the consumer experience more complicated. For example, Day (1977) discussed some aspects that contribute to complexity in the marketplace, such as the use of a product or service over a considerable period, being very complex with many different features, and consisting of professional judgments or advice, and these are precisely the attributes of present-day financial services. The prevalence of consumer problems in financial services calls attention to both preventive and protective approaches to increase consumer financial capability and reduce the cost of taking actions to redress problems in the market.

We found strong support for the vulnerable consumer framework but not for the disadvantaged consumer framework in both the mistreatment model and compromised accounts model. Therefore, Hypotheses 1a and 2a were not supported, and Hypotheses 1b and 2b were mostly supported. These findings support the supposition put forward by Baker et al. (2005) that vulnerable consumers cannot be identified by demographic characteristics and that any consumer can be vulnerable due to inequities in marketplace interactions.

In general, those who were in vulnerable states were more likely to have experienced mistreatment more frequently and to have experienced compromised accounts. The significant vulnerable characteristics were being less financially capable (lower financial skill), in difficult situations (experiencing decline in cognitive ability, more material hardships, and more financial shocks), and more exposed to risk (using more alternative financial services). This finding is consistent with previous research which found that those under financial strain, having recently experienced a negative life event, or exposed to offenders and risky behaviors are more susceptible to scams and investment frauds (Anderson, 2019; DeLiema et al., 2019, 2020). According to the survey results from 600 self-reported fraud victims, financial fraud victims tended to blame themselves for being defrauded and feel ashamed and guilty (FINRA, 2015). The agents and counselors at the consumer agencies and organizations ought to communicate that everyone has the potential to be vulnerable to reduce the shame or stigma associated with being mistreated or defrauded (Baker et al., 2005).

The significance of the financial capability variables, situation variables, and exposure variables provides valuable insight for consumer education, policy, and research. Financial literacy and education are often touted as a solution to the problems consumers face, but the findings from this study show that general understanding of financial concepts failed to alleviate experiences of this form of fraud or mistreatment. In fact, those who are more financially knowledgeable had a higher probability of having their financial accounts compromised, which is consistent with previous studies finding the positive association between financial literacy and investment fraud victimization (AARP, 2007; Kieffer & Mottola, 2017). However, the fact that financial skill shows a consistent relationship with positive outcomes (being less likely to be mistreated and defrauded) provides insights to what an effective intervention looks like. Financial skill, which is defined as knowing how to find, process, and use relevant financial information by the CFPB (2018b), can be an effective tool to prevent consumers from financial fraud and actively respond to the problems in financial services.

Experiencing material hardships and financial shocks may put consumers at risk of victimization. An increased number of material hardships and financial shocks were associated with an increased likelihood of mistreatment and having financial accounts compromised. Foregoing healthcare and having a health emergency are examples of material hardships and financial shocks that were significantly related to both mistreatment and having compromised accounts. In the USA, healthcare is not universally accessible. Healthcare costs in the USA often fall on consumers, at least some portion, and these costs can be expensive and unaffordable for many. It seems like those who are vulnerable in healthcare might be vulnerable in financial services, and extra caution from policymakers and practitioners is required for those who are experiencing transitions in their lives. To alleviate these types of problems, our first recommendation is to implement universal healthcare, however, unlikely that recommendation is to be implemented in the USA. In lieu of that, minor protections could help, like making financial advice and assistance available to individuals and families who are struggling or limiting the amount of medical debt an individual or family can be held responsible for. States who have not expanded Medicaid should consider doing so to allow more low-income people access to affordable healthcare.

The finding that increased use of financial products is associated with both mistreatment and compromised accounts supports the notion that experiencing problems increases with transaction frequency (Goodman & Newman, 2003; Grønhaug & Gilly, 1991). Considering the growing concerns regarding alternative financial services (Bhutta, 2014; Melzer, 2011), the relationship between alternative financial services and consumer problems ought to be explored further. Since both individual factors such as financial knowledge (e.g., Robb et al., 2015) and environmental factors such as access to those services and regulation of the services (e.g., Friedline & Kepple, 2017) play a role in the use of alternative financial services, existing laws and regulations need to evolve to keep up with the ever-changing financial services environment with an eye on consumer protection.

As briefly mentioned above, our findings do not support the disadvantaged consumer framework, indicating that consumer problems may arise more from activities in the marketplace rather than from demographic characteristics. In general, consumers experiencing mistreatment or compromised accounts appear to be from higher socio-economic groups (i.e., those who are in a relationship, more educated, working, with higher household income). This is contrary to the traditional “disadvantaged consumer” profile but is consistent with previous findings on victims of consumer fraud. For example, Lichtenberg et al. (2013) found the victims of fraud were highly educated, and Hogarth et al. (2001) found that those who experienced credit card problems were highly educated and high-income individuals. Our findings may be due to the characteristics of consumer financial markets that those in higher socio-economic groups are more active in the market so that they are more exposed to the potential problems. Ganzini et al. (1990) compared the victims of white-collar crime, such as financial fraud, to violent crime and pointed out that “the term ‘white collar’ not only describes the deceivers but the deceived” (p. 61).

Some caution is needed in the interpretation of the disadvantaged consumer characteristics on mistreatment. Education was positively correlated with the frequency of mistreatment at the lowest threshold but negatively correlated with it at the highest threshold. This implies that higher socio-economic status may allow them to be aware of mistreatment but experience it less frequently. Indeed, early work (Warland et al., 1975) on consumer dissatisfaction and complaint behavior noted that consumers who are more socially active and have higher socio-economic status are upset and take action, but those who are less concerned about consumer interests are not even upset with the way they were treated in the marketplace. There is a possibility that those from lower socio-economic status groups were unaware that they were poorly treated or were used to such treatment. In our robustness tests on “Don’t know” responses to the compromised account question, we found that those with disadvantaged (i.e., minority in race/ethnicity group, less educated) and vulnerable (i.e., lower financial knowledge and financial skill scores, having more financial products) consumer characteristics were more likely not to fully understand their exposure to this type of fraud or to identify it (results are available from authors upon request). This implies that the profiles of consumers who complain can be very different from those of the pool of victims (Raval, 2020a). Policymakers and researchers who are using consumer complaint data or survey data to craft policies should consider the fact that there may be victims who do not raise their voices. We ought to also consider new ways to collect information on those who have been mistreated or been victims of fraud that circumvents the reporting requirement, such as gathering general information after a known occurrence to better understand the profiles of the victims to understand who else could be affected. Of course, the issue is how to identify instances that have not been reported. Perhaps non-traditional data sources such as big data (i.e., Google search data) or social media can be used to identify people who experience these issues but do not report them through traditional means.

Lastly, the profiles of mistreated consumers and defrauded consumers were somewhat distinct. Older consumers (i.e., Pre-Boomers, ages 71 and older) were less likely to have experienced mistreatment in financial services but more likely to have experienced fraud compared to younger consumers (i.e., Millennials, ages 18–35). This means that a consumer may be vulnerable to one problem but not the other. Consumers experiencing fraud appear to be from higher socio-economic groups. However, some of the higher socio-economic characteristics were only related to becoming aware of the mistreatment with financial services but not related to the frequency of the experience. Therefore, consumer vulnerability needs to be measured in a specific way, and the profiles of the victims need to be identified in specific areas, not in general.

The current study has some limitations and suggestions for future studies. First, since the data used in the current study were from a one-time cross-sectional survey and the survey questions referred to different time periods, the results can be interpreted as correlational rather than causal. For example, while the survey asked about fraud experience “in the past 5 years” and situational variables (i.e., cognitive ability decline, material hardship, and financial shocks) “in the past 12 months,” financial capability variables were measured at the time that the survey was conducted. Second, the survey questions on mistreatment and fraud experience can be interpreted differently by respondents. Feeling not respected or mistreated is subjective and the fraud experience asked in the survey (“has someone without your permission used or attempted to use an existing account of yours”) is narrowly defined. There is a possibility that the experience of mistreatment can be either overestimated or underestimated and that the experience of fraud is underestimated in the current study.

Future studies can address those issues by utilizing panel data, by asking more detailed questions about both mistreatment and fraud, or by using more non-traditional data sets. If the NFWS is launched again in the future, we suggest more robust questions about different types of fraud such as, not limited to, identity theft, securities fraud, and insurance fraud. The COVID-19 pandemic has put certain consumers in more vulnerable states and shined a light into how vulnerable some consumers really are. For example, due to the accelerated shift to online from offline during the pandemic, individuals have been exposed to cyber financial fraud, and the cybercriminals targeted emotionally vulnerable individuals such as those with COVID-19-related anxiety (Ma & McKinnon, 2022). Both the research community and the policy community need to be sensitive and responsive to shifts and changes in financial markets that can create disparate impacts among consumers.

Conclusion

The findings of this study offer valuable implications for consumer education and consumer policy. First, more individual and external factors need to be considered to identify vulnerable consumers. Considering that the characteristics of victims from different problems in the marketplace vary, a class-based disadvantaged consumer approach may produce type 1 and type 2 errors, failing to allocate intervention efforts efficiently. Like the victims of fraud in this study, those who have been considered as capable consumers can be vulnerable to other issues. This becomes more important as financial services evolve, and technology becomes more sophisticated, and adoption increases. For example, with fast-growing mobile financial services, young adults, who are considered as vulnerable in traditional settings due to their inexperience, may not be vulnerable in this context due to their familiarity with mobile technologies, and the opposite may be true for older adults. Therefore, a domain-specific consumer vulnerability approach needs to be taken to accurately identify vulnerable consumers in specific areas.

Second, both preventive and protective approaches are needed to increase consumer financial capability and reduce the cost of taking action to redress problems in the market. More oversight and scrutiny are required on the part of lawmakers to ensure that financial products and services are neither unnecessarily complex nor predatory and that consumers are not disadvantaged or mistreated in the financial services industry. Given the disparate nature of financial regulation in the USA, a coordinated interagency focus on all forms of consumer fraud would go a long way in addressing fraud more comprehensively and working towards a coordinated effort to reducing fraud. More practical interventions to increase consumer financial skill and caution for non-traditional financial products, such as money transfers and virtual currencies, need to be made available to consumers to reduce the risk of becoming victims in the financial marketplace.

Appendix

Tables 5, 6

Data Availability

The data that authors used is from the National Financial Well-Being Survey, which is available to download at: https://www.consumerfinance.gov/data-research/financial-well-being-survey-data/.

Declarations

Ethics Approval

This study did not require ethics approval as no human subjects were used.

Conflict of Interest

The authors declare no competing interests.