Abstract
Chinese migrants face numerous socioeconomic disadvantages. Little is understood about their vulnerability to consumer financial fraud, which can impede their integration into urban China and hamper the national priority of continued urbanization. This study utilizes nationally representative data from 2015 China Household Finance Survey to investigate migrants’ risk factors for consumer financial fraud across two stages: fraud exposure and victimization. Results reveal a distinct pattern of fraud risks based on migrant status and stages of fraud, partially supporting theories of disadvantaged consumer and routine activity. After adjusting for other factors, urban migrants exhibit a 28.2% higher likelihood of encountering fraud than urban residents, with rural migrants having a 57.3% higher probability of overall victimization. Recent movers face a higher risk. Market and digital engagement emerge as key factors influencing fraud exposure. These results underscore the necessity for targeted policy interventions aimed at fostering inclusive urbanization and promoting rural-urban integration.
Plain Language Summary
This study looks at the extent to which migrants are vulnerable to consumer financial fraud, which hinders their integration into urban China and affects national development goals. Migrants face many challenges largely because China's hukou system, or household registration system, restricts their access to education, jobs, and public services. This research shows they are also more prone to consumer fraud than local residents. Using nationally representative data from 2015, the study employs a two-stage approach to separately examine fraud exposure and victimization. The results show that rural migrants are the most likely to fall victim to fraud, while urban migrants are more likely to encounter fraud attempts than urban residents. Factors like mobility and engagement with the market and digital services increase the risk of fraud. Overall, the study sheds light on the specific risks that migrants face in China and identifies the institutional and socioeconomic factors contributing to their vulnerability. These insights can help inform policies aimed at reducing fraud risk and improving the well-being of migrants in urban China.
Introduction
Consumer financial fraud refers to unethical and illegal activities in the marketplace that deceive people for financial gain (Deevy, Lucich, & Beals, 2012; Morgan, 2021). It is a significant concern, with far more victims affected than street crimes (Smith & Budd, 2009). Even though it does not involve physical violence, consumer fraud harms people’s well-being and trust in the economic system (Anderson, 2019). Victims not only suffer financial losses but also face health issues, legal troubles, and damage to their social lives (Brenner et al., 2020). Therefore, preventing and addressing fraud is crucial to safeguard individuals and society as a whole.
Consumer financial fraud is a significant issue with staggering impacts. Fan and Yu (2021a) estimated that in 2014 China alone saw an annual loss of 108 billion USD and 14 million households affected. This problem seems to be growing in China (Yu, 2021), which may hinder the national goal of making China a consumption-driven society (Browne et al., 2016). However, not all groups face the same fraud risk. There is a lot of research on fraud victimization among older adults and ethnic minorities (e.g., Burnes et al., 2017; DeLiema et al., 2018; Hall et al., 2016), but little attention has been paid to migrants.
Migrants in China are a crucial group to study as the country has experienced rapid economic growth, urbanization, and a significant increase in migration. The number of migrants in China has risen from 121 million in 2000 to 247 million in 2015 (Liang, 2016). The Chinese government has prioritized rural-urban integration by making it a core economic policy and a national priority, requiring local and provincial governments to implement hukou reform, ease migration, and integrate rural migrants into cities (CPC & State Council, 2014; Fang et al., 2016). Consequently, it’s important to understand the challenges that migrants face in urban China.
Migrants in China often face obstacles and limitations in accessing urban education, housing, and social services. Even though migrants are highly selected, they have had worse socioeconomic and health outcomes and limited upward mobility compared to urban local residents (e.g. Fu et al., 2018; Liu, 2005; Wu & Treiman, 2004; Yang, 2013). These government policy-related disadvantages are well-studied, but less is known about their vulnerability to consumer financial fraud. Migrants may be more vulnerable to fraud because migration is often associated with the lack of familial support and social resources in the destination (Bhugra, 2004; Wang & Tian, 2014). Migrants in China also face stigmatization and exclusion in the marketplace and society (Chen & Hoy, 2011; Cheng et al., 2014; Wang & Tian, 2014). Meanwhile, migrants tend to be younger, more risk-tolerant, and more engaged with the marketplace than local residents (Marlowe & Atiles, 2005). These factors may affect migration-specific fraud risks.
The changing demographics in China has made it more complicated to understand the consumer fraud vulnerability of migrants. Historically, the majority of migrants were rural-to-urban workers, making up 87% of the overall migrant population in 2010 (National Bureau of Statistics of China, 2012). However, there has been an increase in urban-to-urban migration recently, a trend expected to continue (Liang, 2016; Liang et al., 2014; Shen, 2020). Urban migrants (urban hukou migrants), despite having urban hukou, are registered outside their current place of residence, which can limit their access to public services in their destination city (Cheng et al., 2014). However, little is known about the differences in fraud risk between urban and rural migrants (rural hukou migrants).
Finally, there is a discrepancy between findings from fraud-prevalence studies and laboratory experiments on consumer fraud (Anderson, 2016; Fan & Yu, 2021a). Experimental studies often show that older adults and the less educated are more likely to fall victim to fraud. However, observational studies reveal that younger adults and more educated individuals are more likely to be fraud victims (Anderson, 2019; DeLiema, 2018; Holtfreter et al., 2005; Lee & Soberon-Ferrer, 1997). The reason for this inconsistency might be that migrants are different from local residents in their levels of market and digital world engagement, which have not been examined in the literature. Additionally, separating fraud exposure from victimization as two steps may provide further insight into the discrepancy (Deevy et al., 2012).
This study employs a two-stage approach, which separates fraud exposure from fraud victimization after exposure, using nationally representative data to examine the risk factors for consumer fraud among urban and rural migrants, using urban local residents as a reference. The study aims to answer three interrelated research questions: (1) Are migrants at a higher risk for consumer financial fraud than local urban residents? (2) Are rural migrants more likely to be victims of fraud than urban migrants? (3) To what extent can these risk differentials be explained by variations in demographic, human capital, market engagement, and psychological characteristics? This study also attempts to fill three gaps in the existing literature on consumer fraud, namely: the risks specific to migration, the distinctions between rural and urban migrants, and the significance of market/digital world engagement.
In the following sections, we review the literature on migration and consumer financial fraud, present our conceptual framework and hypotheses, and discuss the data and methodology used in this study. Then, we report the differences in fraud exposure and victimization by migrant groups, conduct a multivariate analysis of factors associated with vulnerability to consumer fraud, and provide a counterfactual simulation. We draw conclusions and provide policy recommendations based on our findings in the end.
Literature Review
Hukou and Migration Status in China
Since 1978, China has experienced continuous economic growth and rapid urbanization. The economic reform has allowed for more freedom of movement, while growing regional disparities have incentivized people to migrate to more developed areas (Liang et al., 2014; Shen, 2020; Sun & Fan, 2011; Wu et al., 2019).
As a part of economic reform, the governments have gradually relaxed the hukou (household registration) system, where a person’s place of birth determines their hukou. Prior to the hukou reform, it was almost impossible to transfer one’s hukou to another location or change it from rural (agricultural) to urban (non-agricultural) (Chan, 2019; Liu, 2005). Hukou is inextricably linked to a person’s access to social services in their place of residence (Chan & Zhang, 1999). “It is a major source of injustice and inequality, perhaps the most crucial foundation of China’s social and spatial stratification…” (Chan & Buckingham, 2008, p. 583). It is also a barrier to rural-urban integration and continued urbanization (Afridi et al., 2015; Liu, 2005; Song, 2014).
Migration status is closely linked to the hukou registration system in China (Solinger, 1999; Wang & Zuo, 1999). The hukou status has two key aspects (Chan & Zhang, 1999). The first aspect denotes the original location of hukou registration. Migrants are those who have moved away from the city or village where their hukou was initially registered. Historically, changing hukou status was nearly impossible unless an individual had an officially sanctioned reason, such as attending college, joining the military, or taking a government job (Chan, 2019; Chan & Zhang, 1999). Even after the recent hukou reform, it is still difficult for migrants, particularly rural migrants, to permanently move their hukou to their new city of residence.
The second aspect is hukou trait—whether it is registered as urban (non-agricultural) or rural (agricultural). This system was initially used during the socialist era to ration food and resources, as rural residents were not entitled to access social benefits and welfare programs that were exclusively available to urban residents. As a result of China’s economic reform, many residents have moved away from their city or town of official residence for at least 6 months and are classified as migrants in the official statistics. The hukou location was often linked to the hukou trait, as nearly all residents who live in rural counties, towns, and villages have rural hukou.
Migration has a crucial role in China’s economic development, contributing to 21% of annual GDP growth since economic reform began in 1978 (Fang & Wang, 1999). Migrants constitute a significant proportion of the Chinese labor force, accounting for 35% of China’s total workforce of 770 million in 2015, which is 15 percentage points higher than their share of the total population (National Bureau of Statistics of China, 2017). Despite the rapid increase in the number of urban migrants (National Bureau of Statistics of China, 2017, 2021), there is limited understanding of the challenges they encounter in their destination cities compared to rural migrants.
Nearly all internal migration in China is toward urban areas, which offer more economic opportunities and amenities than rural areas. Two groups of migrants exist: (1) rural migrants, who hold a rural (agricultural) hukou and have moved away from their hukou origin, and currently reside in an urban area, and (2) urban migrants, who hold urban (non-agricultural) hukou and have moved to another city. Meanwhile, there are two groups of local residents: (1) urban local residents who hold an urban hukou, and (2) rural local residents who hold a rural hukou. In this study, the sample is divided into four mutually exclusive groups based on their hukou and migrant status: urban local residents (urban hukou and non-migrants), urban migrants (urban hukou and migrants), rural migrants (rural hukou and migrants, and rural local residents (rural hukou and non-migrants). Those who have received an urban-rural integrated hukou are placed into the rural hukou category, as most of them previously had rural hukou. Urban local residents are used as a reference group in this study, as they are the most numerous and well-off group and often serve as an aspiration for rural residents and migrants.
Consumer Financial Fraud and Risk Factors
The literature on consumer fraud has largely focused on western countries, with the definition and scope of consumer fraud evolving over time (e.g., Anderson, 2019; Baker et al., 2005; Hill & Sharma, 2020; Morgan, 2021). Representative surveys conducted by government agencies and official organizations are often used to determine the prevalence of consumer financial fraud (Anderson, 2016; Deevy & Beals, 2013). Deevy, Lucich, and Beals (2012) distinguished consumer fraud from financial abuse and exploitation as a discrete domain of study. Unlike financial abuse and exploitation which is carried out by family members or persons of trust, consumer financial fraud is committed by strangers for financial gains (Burnes et al., 2017). Studies that cover more types of fraud tend to show a higher prevalence of consumer financial fraud (Anderson, 2004, 2019).
A growing body of literature has examined the general pattern of consumer fraud (e.g., Anderson, 2016; DeLiema, 2018; Holtfreter et al., 2005; Raval, 2021). Recent studies have investigated consumer vulnerability, the associated factors (e.g., Anderson, 2016; Hill & Sharma, 2020) and specific at-risk groups, such as older adults (e.g., DeLiema et al., 2018; Ross et al., 2014). Despite these advancements, our theoretical understanding of consumer fraud remains limited.
New research has examined the disadvantaged consumer theory and the routine activity theory. The theory of disadvantaged consumers suggests that individuals who are deprived of resources and hold lower positions in society are more susceptible to becoming fraud victims. (Raval, 2021). Disadvantaged consumers are typically identified as individuals who are older, belong to minority groups, have lower levels of education, income, and assets. If this theory holds true, rural migrants, who are considered a disadvantaged group, should be more susceptible to becoming fraud victims than urban local residents.
According to the routine activity theory (Felson & Cohen, 1980), fraud risks increase with the convergence of motivated perpetrators, suitable targets, and the absence of capable guardians. Migrants may be more vulnerable to consumer fraud due to their increased involvement in market transactions and online activities, making them more visible targets for potential fraudsters. Furthermore, their unfamiliarity with the area and lack of proximity to friends and family members may make them more easily identifiable to fraud perpetrators than local residents.
Most of the conceptual frameworks discussed in the literature concern the consumers’ decision-making skills once targeted by perpetrators. Hill and Sharma (2020) highlighted three levels of risks—individual, interpersonal, and structural. Demographic, financial, psychological, and social variables can all influence consumer fraud vulnerability. For example, migrants may be more vulnerable due to social isolation and the lack of social support. Furthermore, rural migrants who moved from socio-centric cultures to egocentric cities tend to feel more alienated than urban migrants (Bhugra, 2004). This cognitive and social vulnerability might increase rural migrants’ risk of consumer fraud. In contrast, urban migrants are more familiar with city life than rural migrants. Education is considered an important determinant of cognitive vulnerability because knowledge and skills gained through formal schooling are useful resources when coping with marketplace complexities (McGhee, 1983). As such, lower levels of education are expected to be associated with higher levels of victimization risk.
In addition, consumers in poor physical health may also be more vulnerable due to their desire to get help and because of the link between physical and mental health (Ohrnberger et al., 2017). Migrants tend to be younger and healthier than urban local residents, which may have a protective effect against consumer fraud.
Recently, Fan and Yu (2021a) highlight the need to distinguish consumer fraud exposure from victimization when examining consumer fraud risk factors. Older adults are less likely to be exposed to fraud than younger adults in the first place, but once exposed, they are more likely to become victims. These two opposing effects led to an overall higher risk of consumer financial fraud victimization by older consumers. Xu et al. (2022) validate the two-stage approach and highlight the significance of incorporating individual factors into models that predict fraud risks. Despite the progress made in recent years, our theoretical understanding of fraud risks remains limited (Baker et al., 2005; Dove, 2018).
In summary, the empirical evidence on the risk factors of consumer financial fraud is far from conclusive (Deevy et al., 2012). One example is about the role of education, which may have a non-linear effect on fraud risks. Those with the lowest (less than high school) and highest (graduate degrees) education were the least likely to be victims (Anderson, 2016), while those with some college or more being at the highest risk (DeLiema, 2015).
Research Gaps
This study also addresses three gaps in the current research on consumer fraud: (1) fraud risks that are specific to migrants; (2) differences in fraud risks differences between urban and rural migrants; (3) the impact of market and digital world engagement on fraud risks.
The first gap concerns migration-specific risks and how they differ from those of local residents. Marlowe and Atiles (2005) pointed out that Latino immigrants may be more vulnerable to consumer fraud due to their limited knowledge of the local marketplace. Similarly, rural migrants may encounter significant challenges in urban China due to government policies and a lack of integration (J. J. Wang & Tian, 2014). However, these studies are qualitative and do not compare fraud rates between different groups. Anderson (2019) found that Hispanic consumers reported greater instances of fraud victimization than non-Hispanic whites in three consecutive Federal Trade Commission surveys. However, these studies do not separate migrants from non-migrants. No empirical studies based on our knowledge have specifically investigated whether the fraud risks for migrants are linked to their migration-specific factors or other socioeconomic and demographic characteristics.
The second gap is the lack of research on the differences in fraud risk between urban and rural migrants. Migration is self-selective, as migrants choose to relocate to take advantage of the opportunities available at their destination. However, migrants in China often face restrictions in access to and control over resources in their destination areas. While many studies have shown that skilled migrants to the US tend to do better in the housing and labor markets than less skilled migrants and, in some cases, than native-born residents (e.g., Painter & Yu, 2014), migrants in China typically fare worse in the housing sector than local residents (Li & Hung, 2018; Yu, 2021). Urban migrants in China tend to be more educated and resourceful than rural migrants. Conversely, rural migrants are typically younger in age and have lower income and fewer assets than urban migrants (Chen & Liu, 2018). The observed variations in these factors may help to explain the differences in fraud rates between migrant groups and between migrants and local residents.
The third gap is the lack of research on the role of market and digital world engagement in consumer fraud. In experimental studies and qualitative examinations of consumer fraud victimization, researchers see what personal characteristics are correlated with a person being susceptible to consumer fraud (e.g., Anderson, 2016; Fischer et al., 2013). Anderson (2016) shows that consumers with higher levels of education and financial literacy are less likely to be victims of fraud in the experimental study. Shao et al. (2019) argue that these experimental studies often scarify external validity. In other words, when we study the overall consumer fraud risk, a consumer will have to participate in marketplace activities to be exposed to fraud and become a fraud target. Those who are more exposed to fraud should be more likely to be victimized by fraudsters (Van Wyk & Mason, 2001). Several recent COVID-19 studies show that exposure is a crucial step in the virus infection, as individuals who have higher levels of exposure to the virus are more likely to become infected (e.g., Hong et al., 2021). While past research has conceptualized consumer fraud victimization as a two-staged approach (Deevy et al., 2012), only recently have researchers started to separate exposure from victimization in empirical studies (e.g., Fan & Yu, 2021b; Xu et al., 2022). It remains unclear whether market engagement has affected fraud exposure and victimization differently, particularly for migrants who may be more active in the marketplace than local residents.
Conceptual Framework
General Expectations
As China transitions toward a market economy, we anticipate that the prevalence of consumer fraud and associated risk factors among Chinese consumers will align with those described in existing literature. To explore this further, we’ll begin by developing a conceptual framework that distinguishes between fraud exposure and fraud victimization following exposure. Specifically, we'll examine the differences in fraud risk between migrants and local residents.
We then build on the existing literature and propose that migrant status can affect both fraud exposure and conditional fraud victimization (fraud victimization after exposure) through market and digital world engagement, financial resources, human capital factors, and psychological characteristics (Figure 1).
The conceptual diagram.
Fraud Exposure
The more a consumer is engaged in the market and the digital world, the more likely his/her personal and contact information is available to fraudsters, leading to a higher likelihood of fraud exposure. Migrants are likely more engaged with the market than local residents as they have less informal ways to acquire goods and services in their migration destinations due to a lack of local social support networks. In addition, migrants may be more likely to utilize digital technology to stay in touch with family and friends back home. As such, migrants are more likely to be targeted by fraudsters than locals due to their higher level of market and digital world engagement. In addition, fraudsters are more likely to target consumers with more financial resources due to a potentially higher gain. Urban migrants have a similar level of financial resources as urban locals, but rural migrants and locals tend to have lower levels of income and wealth than urban local residents. As such, rural migrants and rural locals may have less exposure to fraud compared to both urban locals and urban migrants.
Fraud Victimization After Exposure
Once exposed to fraud, consumers with higher human capital and certain psychological characteristics may be more likely to become fraud victims than otherwise similar consumers. Human capital has many components, including cognitive ability, education, financial knowledge, health, life experiences, and social support. Compared to locals, migrants have better health but lower life experiences and social support levels in migration destinations (Chen & Hoy, 2011; Wang & Tian, 2014). In addition, rural migrants and rural locals have lower levels of education compared to urban migrants and urban locals. As such, with the exception of a health advantage due to younger average age, migrants have multiple disadvantages and a generally lower level of human capital, leading to a higher risk of fraud victimization once exposed. In addition, psychological factors may influence a consumer’s susceptibility to fraud victimization once exposed. For example, consumers who are more risk-tolerant may take on high-risk, high-potential return opportunities presented by potentially fraudulent offers. Migrants tend to be more ambitious and risk-tolerant than local residents (Borjas, 1991; Chiswick, 2000).
Hypotheses
Based on the conceptual framework and our measurements, we form several hypotheses for our research questions. For research questions of (1) whether migrants are at a higher risk for consumer financial fraud than local residents and (2) whether rural migrants are more likely to be fraud victims than urban migrants, we hypothesize that (H1) urban migrants have a higher risk of fraud exposure and fraud victimization after exposure than urban local residents, while (H2) rural migrants have a similar risk of fraud exposure but higher risk of fraud victimization after exposure than urban local. For research question (3) of how variations in market and digital world engagement, financial resources, human capital, and psychological characteristics can explain these group risk differentials, we hypothesize that (H3) these factors can partially explain the risk differentials between migrants and locals. Specifically, we expect that variables indicating a higher level of market and digital world engagement and more financial resources to be associated with a higher fraud exposure risk, and variables indicating a higher level of human capital and certain psychological characteristics to be associated with a lower risk of fraud victimization after exposure.
Data and Sample
Data Source
To test our hypotheses, we used the 2015 China Household Finance Survey microdata (CHFS). The CHFS was collected by the Survey and Research Center for China Household Finance at Southwestern University of Finance and Economics in China (Gan et al., 2013). It is a nationwide and comprehensive biennial survey for household finance in China, providing detailed information on various aspects of household finances. This includes data on household assets, liabilities, income, expenditure, demographics, intergenerational transfers, employment, and subjective well-being. In 2015, the survey covered consumer fraud and included households from 1,439 communities in 363 counties across 29 mainland Chinese provinces. The data were representative at both the national and provincial level.
Our sample included 25,292 households with a reference person aged between 16 and 60. We exclude those who are over 60, since the mobility variable (whether someone moved in the past 2 years) is only available for those who are aged between 16 and 60. In addition, older adults and retirees have different migration intentions and mobility patterns (Shen, 2020). For each household, the person who was the most knowledgeable about their household finance was interviewed. For convenience, we referred to this person as the household reference person. This study was deemed exempt by our university IRB as the data are publicly available.
Variables and Measurements
The 2015 CHFS had a special section on consumer financial fraud over the past year, with questions relating to both fraud exposure and victimization. Below are the operational definitions of the variables we used in this study.
Fraud Exposure
For consumer fraud exposure, the survey asked if the household encountered any of the following forms of consumer fraud and scams over the past year—telemarketing, text messages, smartphone apps such as Q.Q. or WeChat, other phishing schemes, from acquaintances, or other forms. If a respondent answered “yes” to any of these forms, the respondent is considered to have been exposed to consumer fraud in the past year.
Fraud Victimhood
If a respondent answered “yes” to the fraud exposure question, the respondent was then asked if money was lost to the fraud. If yes, then the respondent was classified as a consumer fraud victim.
Migrant Status
Four exclusive categories are created to reflect the respondent’s migration and hukou status: urban migrants, rural migrants, urban local residents, rural local residents. We focus on migrants in this study, using local residents as the reference group.
Market and digital world engagement: We include four variables to control for market engagement. First, we use the percent food budget spent away from home as a proxy for market engagement. The higher the percentage, the more the respondent is expected to be engaged with the market. In addition, we have two variables that capture digital world engagement: whether the respondent shopped online last month and whether the respondent owns a smartphone. We expect that shopping online and owning a smartphone are associated with more engagement with the market. Moreover, we use whether the respondent has chronic health conditions. The survey asked if the respondent had any of the listed chronic diseases. If the respondent answered yes, then s/he is classified as having at least one chronic health condition. Consumers in poor health are more likely to engage in specialized markets of health care and healthcare products.
Moreover, multiple demographic, human capital, and geographical variables are used as proxies for market and digital world engagement, including age, education, and region. Younger and better-educated consumers are more likely to have higher levels of digital world engagement than older and less-educated consumers.
Financial resources: Financial resources are measured using household income (logged), asset (logged), and debt (logged).
Human capital: Human capital is measured by multiple variables, including education level, financial knowledge, the recent mover status, and social support. The financial knowledge variable is created from three simple objective personal finance questions: (1) interest calculation, (2) inflation calculation, and (3) risk related to stock and mutual fund investments. Each correct answer receives two points, while each incorrect answer receives zero points. A response of “don’t know” receives one point. The financial knowledge score is the sum of the scores on these three questions. The highest possible point is six, while the lowest is zero.
We include the recent mover status or whether the respondent moved in the past 2 years. Recent movers may be at a higher risk of consumer fraud for several reasons. When people move to a new location, they may not be familiar with the local scams and may be more vulnerable to fraudulent activities. In addition, when people move, they often have to update their contact information with various organizations, and if they miss any important notifications, they may be at risk of falling victim to fraud. Furthermore, when people move, they may be more likely to share their personal information, such as their address and banking details, with various organizations, and this information could be vulnerable to theft or misuse. Scammers may also use the fact that people are in the midst of moving to try and trick them into giving up personal information or money.
For social support, the survey asked whom the respondent would ask for help in times of personal difficulties or in cases of conflict with other community members. The respondent was given several options, such as community cadres, families and friends, governments, social organizations, work units, or “resolve by myself.” A lack of social support is coded if the respondent chose “resolve by myself” as the only choice.
Psychological characteristics: Two psychological variables are used: risk tolerance level and trust in professionals. For risk tolerance, the survey asked what type of risk-return investment the respondent would make if s/he wins a lottery of RMB500,000 (about $70,000). A set of six questions ranging from “high-risk and high-return” to “no-risk and no return” were asked. Based on the answer, we create a set of risk tolerance dummy variables for each respondent. For the trust variable, the survey asked if the respondent trusted professionals such as doctors, scientists, teachers, lawyers, and government officials. We create the trust variable based on the number of professional groups the respondent trusted, up to 4 points for four groups or more.
Additional variable: We also control for region (east, mid, northeast, or west) in which the respondent is located.
Method
To study the consumer financial fraud risks of migrants relative to that of urban local residents, we use descriptive analyses to estimate means first, weighted to be population-representative. Next, we conduct logistic regressions to investigate migrants’ risk factors for consumer fraud victimization. We estimate models with the log-odds ratios of three dependent variables: (1) fraud exposure, (2) conditional fraud victimization, and (3) overall fraud victimization. Lastly, we conduct logistic regression analyses separately within three subsamples with urban local residents as the reference group. The first one includes urban migrants and urban local residents only. The second one includes rural migrants and urban local residents only. The third one includes rural local residents and urban local residents only.
Results
Summary Statistics
Table 1 first presents one-year prevalence estimates of consumer financial fraud exposure and victimization, together with descriptive statistics for all variables used in this study. The estimated 1-year prevalence of consumer financial fraud exposure is high at 59.9% for Chinese households with a householder whose age was between 16 and 60 in 2014 to 2015. 3.4% of all households lost money to consumer financial fraud during a 1-year period in 2014 to 2015. In contrast, migrants have reported a higher level of fraud exposure than the average, so have urban local residents. Migrants have reported a higher incidence of fraud loss. In contrast, rural local residents have a much lower level of fraud exposure and a higher level of conditional victimization than urban local residents.
Prevalence of Consumer Financial Fraud Exposure and Victimization: Total and by Migrant Status.
Note. The sample is weight adjusted.
Table 2 shows migrants are quite different from urban local residents: (1) Migrants are much younger. The average age of the reference person is 45.1, while migrants are about 7 years younger on average. (2) Women are slightly more represented in migrant households than in local households. (3) Migrants are polarized with their educational attainment, with urban migrants being the most educated group of all. In contrast, rural local residents are the least educated. (4) One’s level of financial knowledge seems to correspond with their educational attainment.
Weighted Descriptive Statistics for the Full Sample and Subsamples for (1) Urban Local Residents, (2) Urban Migrants, (3) Rural Migrants, (4) Rural Local Residents.
(5) As expected, migrants are more mobile than local residents. 14.7% and 15.8% of urban and rural migrants have moved in the last 2 years respectively, while only 6.9% of urban local residents moved in the same time period. (6) Migrants are much more engaged with the market than local residents, as urban and rural migrants spent 35.2% and 27.7% of the food budget out of home respectively. Migrants are more likely to shop online than urban local residents and have higher rates of smartphone ownership. (7) The reported risk toleration of migrants is significantly higher than that of local residents. (8) Migrants are less trustful of professionals and less likely to have chronic conditions. (9) With respect to regional distribution, migrants are overrepresented in the East—the most productive region of China.
There are also important differences between urban and rural migrants, as urban migrants tend to have a higher level of education, income, wealth, and debt. Meanwhile, rural local residents have the lowest levels of education, financial knowledge, and risk tolerance.
In sum, migrants are quite different from urban local residents, as they are on average more likely to be exposed to fraud, more likely to be fraud victims, and much younger. They are also more likely to live in the East and have a higher self-reported tolerance for risk. Moreover, they tend to be more mobile, healthy, and engaged with the market than urban local residents, but less trustful of professionals. Meanwhile, urban migrants are more educated, having a higher level of income, wealth, debt, and financial knowledge than rural migrants. The variances in fraud exposure and victimization between migrants and local residents may be attributed to differences in covariates such as demographic, health, and sociopsychological factors. It is essential to account for these variables in a multivariate framework.
Regression Results
Table 3 presents logistic regression results with the logs odds of three different dependent variables: (1) fraud exposure (exposed = 1, not exposed = 0, full sample), (2) conditional fraud victimization (victimized = 1, not victimized = 0, only those exposed to fraud in the sample) and (3) overall fraud victimization (victimized = 1, not victimized = 0, full sample). For each dependent variable, we estimate three sets of models: (I) migrant status only, (II) migrant status plus the controls for demographics and education, (III) additional controls for human capital, financial resource, market and digital world engagement, psychological characteristics, and geographic locations. We chose the reference group as those who are urban local residents, aged 45 to 54, female, completed high school but no college, stayers (those who did not move in the last 2 years), having not shopped online last month, having no smartphones, having no chronic conditions, having a low level of risk tolerance, having social support, and living in East China.
Weighted Logistic Regression Results for Consumer Financial Fraud Exposure, Conditional Victimization, and Ooverall Victimization.
Note. Standard error is available upon request.
p < .05, **p < .01, ***p < .001.
For fraud exposure, summary statistics show that migrants seem more likely to be targeted by fraudsters. However, results from Model I show that urban migrants are statistically more likely to be exposed to fraud than urban local residents by 42.2%. Once they are exposed to fraud, rural migrants are much more likely to be victimized. Rural migrants are 69.9% more likely to be fraud victims than urban local residents. In contrast, rural local residents have a much lower rate of fraud exposure and a higher rate of conditional victimization.
To what extent can the disparities be explained by differences in demographics and education—two sets of largely unchangeable variables? Rural migrants are less educated and younger than urban local residents. Results from Model II show that the inclusion of these variables partially explains the high fraud exposure of urban migrants and the high overall victimization rate of rural migrants. Meanwhile, men and the less educated have a lower exposure rate to fraud.
Model III controls for more covariates including human capital, household economic conditions, market and digital world engagement, health vulnerability and psychological vulnerability. Results show that, after controlling all the covariates, migrants are still more exposed to fraud than urban local residents. Rural migrants are much more likely than urban local residents to be fraud victims. Recent movers are more likely to be fraud victims than stayers. Consistent with past research, income and assets are positively associated with fraud exposure. Household debt is positively linked to fraud victimization. Market and digital world engagement is positively associated with fraud exposure, but not statistically associated with fraud victimization. The exception is shopping online, which elevates both fraud exposure and victimization. Educational attainment and risk tolerance are positively linked to fraud exposure, while financial knowledge is positively associated with both fraud exposure and victimization. Social trust has a protective effect against fraud exposure. Finally, lack of social support increases the probability of fraud exposure.
To test the robustness of our result, we show the regression results for three subsamples in Table 4. Urban local residents are used as a reference in the three comparisons. They are urban migrants, rural migrants, and rural local residents respectively.
Weighted Logistic Regression Results for Consumer Financial Fraud Exposure, Conditional Victimization, and Overall Victimization—Urban Local Residents as the Reference Group.
Note. Standard error is available upon request.
p < .05, **p < .01, ***p < .001.
After controlling for the covariates, the results mirror what we found in the full model. Urban migrants are 30.7% more likely to be exposed to fraud relative to urban local residents. Overall, rural migrants are 45.2% more likely to be fraud victims. Rural local residents have a very low exposure to fraud. However, they are as likely to be fraud victims as urban local residents, because they have a higher probability of conditional victimization. Recent movers are much more likely to be fraud victims. Most market engagement variables are positively linked to fraud exposure, but not victimization.
Simulations
Building on the regression results reported in Table 3, we rely on Blinder–Oaxaca non-linear decomposition techniques (Blinder, 1973; Oaxaca, 1973), using the multivariate regression coefficients and the characteristics of urban local residents, to simulate two counterfactual scenarios. Figure 2 shows three sets of bars, which represent the simulation results of fraud exposure, conditional fraud victimization, and overall fraud victimization respectively. Urban local residents are used as the reference group. The light grey bars on top report the actual differences in rates. The bars in the middle show the differences after controlling for demographics and education, while the bars on the bottom reveal the simulated differences after controlling for all covariates.
Predicted rates relative to those of urban local residents.
The goal is to quantify their relative importance of the covariates in explaining the differences in fraud exposure and victimization. Through this process, we address two key questions:
What would the fraud exposure and victimization rates be if other groups had the same demographics and education of urban local residents (i.e., being older and having a high level of education)? The answer is that fraud exposure would change little. While fraud victimization rates would increase for urban migrants and decrease for rural migrants. That is to say if urban and rural migrants had the same demographics and level of education as urban local residents, urban migrants would have even higher rates of victimization. Meanwhile, rural migrants would be lower in the victimization rate.
What would the fraud exposure and victimization rates be if migrants had the same characteristics of urban local residents on all accounts? Urban migrants would have similar rates of victimization as their actual rates, rural migrants would have lower rates of victimization than the actual rates. In any case, migrants are more likely to be fraud victims than urban local residents. Rural local residents are largely protected from fraud by their low exposure to fraud, in part explained by their very low levels of market engagement.
Discussion and Implications
The discussion of this study’s results must be prefaced with three caveats. First, our consumer fraud questions are limited to definitions available in the CHFS survey. Second, our study is observational in nature, examining the prevalence of self-reported consumer fraud exposure and victimization over a 1-year period. As such, the results are likely different from those of laboratory experiments and from studies using different surveys. Third, our study is based on cross-sectional data; we cannot draw firm conclusions on causal relationships. Despite these limitations, it is still a valuable effort to use nationally representative Chinese data to examine migrants’ consumer fraud risks in urban China.
The first two aims of this study are to examine whether migrants are at a higher risk for consumer financial fraud than local residents in China and whether there are differences between urban and rural migrants. Our findings partially support H1, in that compared to urban local residents, urban migrants have a higher level of risk of fraud exposure (76.8% vs. 70.1%, p = .05). However, while the victimization risk after exposure is higher for urban migrants than for urban locals, the difference is not statistically significant (6.4% vs. 4.5%). Our findings support H2 in that rural migrants have a similar level of fraud exposure risk (69.6% vs. 70.1%, p = .05) but a higher level of victimization risk after exposure (7.6% vs. 4.5%, p = .05). Converting to odds ratios, this means urban migrants have 42.2% higher odds for fraud exposure than urban locals, while rural migrants have 69.9% higher odds of fraud victimization after exposure.
The third aim of this study is to investigate the extent to which the fraud risk differentials can be explained by variations in market and digital engagement, financial resources, human capital, and psychological characteristics.
We used (1) the percentage of the food budget spent away from home, (2) whether shopped online last month, (3) whether owning a smartphone, and (4) whether having chronic health conditions as proxies for market and digital world engagement. For fraud exposure risks, our findings show that market engagement is associated with a higher level of fraud exposure risk. Those with chronic health conditions are likely to seek health care and health care products, exposing themselves to a market that sells many credence goods with fraud potentials (Wheatley & Spink, 2013). Moreover, multiple demographic, human capital, and psychological characteristics variables are found to be significantly associated with fraud exposure risk. Specifically, people with lower education are less likely to be internet savvy and thus less likely to engage in the digital marketplace and social media, and are thus less likely to be exposed to fraudsters. Those with less social support are more likely to turn to the digital world for products and services search and for virtual connections. Those with better financial knowledge and higher levels of risk tolerance are more likely to engage in investment activities, exposing themselves to investment market fraud possibilities. Consumers who place more trust in professionals face lower fraud exposure risk, likely indicating that these consumers have specific market behavior that reduces their risk of encountering fraudsters. For financial resources, our findings show that income and assets are associated with a higher risk of fraud exposure, while the amount of debt is not. This makes sense because when income and assets are held constant, the debt amount should not make a consumer a more or less profitable target.
Controlling for these variables reduces the fraud exposure risk differential between urban migrants and urban locals but do not explain away all the differences. Adjusting for these process variables, urban migrants now have 28.2% higher odds of fraud exposure than urban locals, down from 42.2%. On the other hand, controlling for these same factors increases the fraud exposure risk differential between rural migrants and urban locals. Adjusting for these process variables, rural migrants have 22.2% higher odds of fraud exposure than urban locals, a substantial increase from no significant statistical differences between these two groups without controls. Note that rural migrants are expected to have a higher market and digital world engagement but lower financial resources than urban locals, two opposing forces that may cancel out each other. The finding that rural migrants have a higher fraud exposure risk after controlling for these process variables indicates that the forces that increase their risks are stronger than the forces that decrease their risks. Overall, our models are unable to explain the remaining 28.2% higher odds of fraud exposure risk for urban migrants and 22.2% higher odds of fraud exposure risk for rural migrants compared to urban locals.
For fraud victimization risk after exposure, as expected, having low education and being a recent mover, two variables indicating lower general and destination-specific human capital are associated with a higher fraud victimization risk after exposure. It is interesting to note that having better financial knowledge, not responding to the risk question, and having a higher amount of debt are associated with a higher fraud victimization risk after exposure. Because the financial knowledge questions in this survey ask basic financial literacy information, it is possible that those with some knowledge may be overconfident about their knowledge level, leading to higher susceptibility to fraud victimization after exposure. Not responding to the risk question likely indicates a lack of understanding of the concept of risk and thus becomes a proxy for low human capital. Having a higher amount of debt may be associated with certain psychological characteristics not controlled in our models. One such psychological characteristic is time preference, in that debtors are more likely to be present-oriented and more impulsive than non-debtors (Webley & Nyhus, 2001). Impulsivity and present orientation may increase a consumer's risk of fraud victimization as these consumers are more likely to act quickly without thinking through the risks involved in potentially fraudulent offers.
The fraud victimization risk differential between urban migrants and urban locals is reduced after controlling for the covariates but remains statistically insignificant. On the other hand, rural migrants now have 47.0% higher odds of fraud victimization risk after exposure than urban locals, down from 69.9%. Again, our process variables capture some sources of this risk differential but not all.
Overall, our research addresses the three gaps in the literature on consumer fraud. First, our findings highlight migration-specific risks—the special risks faced by recent movers, who are at a significantly higher risk of consumer fraud during the victimization stage. Migrants who moved recently are facing double jeopardy, as they will be more vulnerable to consumer fraud in both the exposure and victimization stages. Second, the problem of consumer fraud is much more significant among migrants, both urban and rural. Urban migrants are at higher risk during the exposure stage, while rural migrants are more vulnerable during the victimization stage after exposure. Our two-stage model focuses on individual characteristics and behaviors, providing partial explanations for these risk differences. Third, our study emphasizes the importance of market and digital engagement in understanding consumer fraud risks, as they substantially increase fraud exposure. Lastly, our findings align with the disadvantaged consumer theory and routine activity theory, both of which suggest that migrants, especially those from rural areas, are more vulnerable to consumer fraud. These theories highlight the increased risk that migrants face due to their socio-economic status and the lack of social support networks in their new environments.
There are two potential sources for unexplained risk differentials: imperfect measures of the concepts in our model, and institutional constraints that are inherent in the hukou system, which may manifest themselves beyond the realm of socioeconomic struggles for migrants. As China marches toward an urbanized society and integrates more migrants into cities, policymakers may want to provide information and education to migrants on how to recognize and avoid fraud through community-based interventions and support networks. While demographic characteristics and behaviors are difficult to modify, institutional constraints can be addressed through changes in public policy. Targeted policy intervention may be necessary for more inclusive urbanization and rural-urban integration.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Zhou Yu is supported in part by the Asia Center and University Research Committee (URC) Faculty Research Grant at the University of Utah.
