Credit Choices in Rural Egypt: A Comparative Study of Formal and Informal Borrowing

Abstract

Access to finance is essential for fostering financial inclusion, improving household economic well-being, and stimulating economic growth. However, if not prudently managed, it can become a double-edged sword, increasing the risk of over-indebtedness, particularly among low-income households. This paper investigates the borrowing behavior of rural households in Egypt, exploring whether it is motivated by the optimization of intertemporal consumption or reflects deeper financial vulnerabilities. The study enhances our understanding of rural households’ financial behavior in Egypt and contributes to the literature by introducing perceived general self-efficacy as a key behavioral factor. The paper employs a quantitative methodology using a probit analysis of the Egypt Labor Market Panel Survey to explore the factors affecting the demand for formal loans, informal borrowing, and Rotating Saving and Credit Associations (RoSCAs). The results show that informal credit plays a dominant role in meeting rural households’ financial needs. A significant positive relationship between formal and informal credit suggests they are complementary. Elderly, married, less educated, and poorer individuals are more likely to seek both forms of credit, with employment stability being a key differentiator. Self-efficacy also has a significant positive effect. No significant regional differences are observed, except in the case of informal borrowing, with rural households in Upper Egypt showing less reliance, suggesting that social image may influence financial behavior in this region. The results suggest that demand for credit is driven by economic and financial vulnerability of rural households. The paper highlights key policy implications. First, to enhance participation in formal credit market, credit policies should offer more affordable, tailored credit relevant to starting a business rather than financing consumption, part of which is conspicuous. Second, the low self-efficacy among the rural poor suggests a need for policies that combine credit access with financial literacy and debt management support to prevent over-indebtedness.

1. Introduction

Access to credit is essential for enabling households to manage their resources effectively and enhance their overall economic well-being, allowing them to smooth consumption, save and invest for the future, and cope with unforeseen financial shocks. Recognizing its importance for poverty reduction and economic growth, developing economies, including Egypt, position access to finance as one of their top priorities (Rashdan and Eissa 2020; Abdel Hamid 2021). While the potential benefits of credit access are evident, they are contingent upon effective debt management practices. Poorly managed debt can lead to over-indebtedness, posing risks not only to individual households but also to the broader financial system (Dharmadasa and Gunatilake 2023; Vandone 2009). In this context, it is crucial to examine which population segments are more likely to participate in credit markets and whether their borrowing behavior is driven by the goal of optimizing intertemporal consumption or indicates deeper financial and economic vulnerabilities (Vandone 2009).

Egypt, mirroring trends in other developing countries, has made significant strides toward enhancing financial inclusion, particularly for rural populations. Despite these efforts, less than 25% of Egypt’s rural population is financially included, as indicated by bank account ownership (World Bank 2021). However, formal and informal borrowing among the poorest 40% has increased significantly since 2017, reaching 58% in 2021, with rural areas exhibiting the highest proportion of borrowers (World Bank 2021). In light of growing concerns that increasing access to credit may lead to serious challenges when accompanied by imprudent debt utilization, this paper investigates the use of formal and informal borrowing in rural Egypt across different regions and wealth quintiles. It also examines individuals’ demographic, socio-economic characteristics, and personal attributes—particularly self-efficacy—as key determinants of participation in formal and informal credit markets.

1.1. Theoretical Background

The determinants of household borrowing are discussed within two main economic theories; the Life Cycle Hypothesis (LCH) developed by Modigliani and Brumberg (1954) and Ando and Modigliani (1963), and the Permanent Income Hypothesis (PIH) proposed by Friedman (1957). According to the LCH and the PIH, individuals aim to maximize an inter-temporal utility function by smoothing their streams of consumption over their lifetime. Within this framework, consumption choices are subject to an intertemporal budget constraint and individuals determine their consumption, and consequently their saving and borrowing decisions, based on their wealth, current disposable income, and future income expectations (Sablik 2016). Accordingly, individuals seek to maintain the same level of consumption throughout their lifetimes by taking on debt or liquidating assets early or late in life when their income is low, and by saving during their prime earning years when their income is high (Deaton 2005).

While the LCH and PIH remain central to analyzing household consumption, saving, and borrowing decisions, and have been supported empirically by some researchers (e.g., Bernanke 1984; Hall and Mishkin 1980), others have found conflicting evidence (e.g., Stephens 2002; Kohara and Horioka 2006). This opposing empirical evidence suggested the importance of considering other factors that influence household demand for credit including liquidity constraints. It also paved the way for notable contributions from behavioral economics, which demonstrate the significant influence of psychological factors on household borrowing behavior (Thaler and Shefrin 1988; Beshears et al. 2018). Moreover, with the rising trend of household indebtedness, empirical studies have examined the underlying motives for borrowing, revealing that social comparison and the pursuit of status, consistent with Veblen’s theory of conspicuous consumption (1899), significantly shape borrowing decisions (Banuri and Nguyen 2020; Lee and Mori 2021).

According to the intertemporal choice models of the LCH and PIH, the demand for credit is influenced by the demographic and socioeconomic characteristics of individuals such as level of education, family size, type of employment, income, and wealth, and institutional factors related to the local credit market specifically, the existence and quality of information sharing mechanisms among financial institutions (i.e., credit bureaus), the effectiveness of the legal system, and the presence of an informal credit market (Vandone 2009; Chen and Chivakul 2008).

In recognition of the significance of behavioral factors and personal attributes on household borrowing behavior, the analysis included examining the effect of perceived general self-efficacy on the demand for credit. According to Bandura (1977), self-efficacy is defined as the belief in one’s ability to succeed in specific situations and accomplish particular tasks. Individuals with a strong sense of self-efficacy are more likely to accept challenges, have self-confidence in their ability to succeed and cope with adversity, and embrace an optimistic attitude toward the future (Lown et al. 2015; Ismail et al. 2017; Chong et al. 2021; Bari et al. 2020). In line with Bandura’s definition, individuals with higher levels of self-efficacy are expected to demand credit, driven by their confidence in their ability to navigate the financial landscape and effectively manage debt.

1.2. Empirical Evidence

The determinants of credit demand have been widely examined in the empirical literature, with socioeconomic and demographic characteristics showing varying effects across different contexts (Kofarmata et al. 2016). For example, while some studies report a negative relationship between age and credit demand (Asiamah et al. 2021; Cheng and Ahmed 2014; Chandio et al. 2021), others find either a positive effect (Rashdan and Eissa 2020; Fahmy and Ghoneim 2023; Chen and Chivakul 2008) or no significant impact (Dharmadasa and Gunatilake 2023). Similar mixed results are observed in rural areas. In some cases, older farmers are more likely to use credit, while in others, younger farmers, who are economically active and keen to invest, show higher credit demand (Kofarmata et al. 2016). With regards to household size, several studies have found a positive effect (Dharmadasa and Gunatilake 2023; Vandone 2009; Mohieldin and Wright 2000; Mpuga 2010; Chandio et al. 2021; Chen and Chivakul 2008). However, having elderly members in the household has been shown to decrease loan demand (Dharmadasa and Gunatilake 2023). As for marital status, while some researchers (Mohieldin and Wright 2000; Vandone 2009; Asiamah et al. 2021) report that married individuals are more likely to seek credit, Dharmadasa and Gunatilake (2023) report no significant effect. Similarly, studies indicate that males tend to have a higher likelihood of demanding credit (Mohieldin and Wright 2000; Nguyen et al. 2021; Kofarmata et al. 2016; Mpuga 2010), yet others (Dharmadasa and Gunatilake 2023; Chandio et al. 2021; Rashdan and Eissa 2020; Fahmy and Ghoneim 2023; Chen and Chivakul 2008) suggest that sex is insignificant. Additionally, Sayed and Shusha (2019) emphasize the significant influence of religion on credit use decisions.

Regarding the impact of education, studies generally indicate that it increases the likelihood of using credit (Vandone 2009; Asiamah et al. 2021; Dharmadasa and Gunatilake 2023; Chandio et al. 2021; Nguyen et al. 2021; Moahid and Maharjan 2020; Qin et al. 2018; Cull et al. 2019; Rashdan and Eissa 2020). However, Chen and Chivakul (2008) found a negative relationship. A positive effect of current and expected income on credit demand has also been documented (Vandone 2009; Chandio et al. 2021; Mohieldin and Wright 2000; Fahmy and Ghoneim 2023). While Vandone (2009) reports a significant negative effect of net wealth on credit use, Chen and Chivakul (2008) observe a hump-shaped pattern, suggesting that the negative relationship only emerges after individuals reach a certain threshold of net wealth. Mixed findings are reported for asset ownership; Mohieldin and Wright (2000) identify a positive effect, while Nguyen et al. (2021) suggest that asset ownership negatively impacts credit demand. Studies also show a positive effect of larger landholdings (Chandio et al. 2021; Nguyen et al. 2021).

Employment stability, reflecting uncertainty in current and future income prospects, is another significant factor influencing credit demand, with the unemployed being less likely to participate in the credit market (Vandone 2009; Rashdan and Eissa 2020). Vandone (2009) also highlights notable differences between employed, self-employed, and retirees, showing that the employed are more likely to use credit compared to the self-employed, while retirees have a lower probability due to the effect of age and the limited future income. In contrast, Chen and Chivakul (2008) report no significant difference between the employed and unemployed, but inactive individuals, such as housewives, students, and people with disabilities, are less likely to seek credit than the employed. Additionally, households experiencing shocks are more likely to demand credit (Kofarmata et al. 2016; Dharmadasa and Gunatilake 2023), and the desire to keep up with others’ socioeconomic status, particularly among lower-income groups, is positively associated with credit demand (Dharmadasa and Gunatilake 2023; Banuri and Nguyen 2020).

When examining the effect of the existence of informal credit markets, the literature on developing countries highlights the duality of credit markets, which is a well-documented feature of the rural financial sector (Cull et al. 2019; Barslund and Tarp 2008; Mohieldin and Wright 2000). Researchers have examined the interaction between formal and informal credit markets, yielding varying results and interpretations (Nguyen et al. 2021; Barslund and Tarp 2008). In this context, several studies underscore the significant role of the informal credit market, either emphasizing its complementarity with the formal credit market (Mohieldin and Wright 2000; Abdel Zaher 2019; Wabwire 2019; Gyeltshen 2012; Barslund and Tarp 2008) or noting its limited or imperfect substitutability (Nguyen et al. 2021). Furthermore, Vandone (2009) suggests that the existence of an informal credit market may negatively affect borrowers’ attitudes toward repayment, creating incentives for strategic insolvency.

Concerning the effect of self-efficacy on financial management behavior, it has been examined in various dimensions within the literature. Some studies explored the relationship between self-efficacy and debt and saving behavior (e.g., Lim et al. 2014; Lown et al. 2015; Ismail et al. 2017). Others examined its role as a mediator that influences financial management behavior through its effect on financial literacy and cognitive abilities (e.g., Lone and Bhat 2024; Chong et al. 2021; Bari et al. 2020; Tang 2021). Additionally, some studies investigated the relationship between self-efficacy and risk appetite and the choice of financial services (e.g., Pavani and Alagwadi 2023; Liu and Zhang 2021; Farrell et al. 2016). Similarly, the effect of self-efficacy has been examined across various participant categories, such as young and older adults, women in different professions, and college students.

Regarding the effect of self-efficacy on debt and saving behavior, higher levels of self-efficacy are found to be associated with lower levels of debt, more savings, and less financial distress and anxiety (Lown et al. 2015; Chong et al. 2021; Tang 2021). Self-efficacy also significantly differentiates between prudent credit users and credit abusers. Unsuccessful credit users have lower levels of self-efficacy and link money with power and prestige. They also exhibit external locus of control and have low risk tolerance levels. Also, individuals with a stronger sense of self-efficacy are more likely to resolve their debt issues and take precautionary actions to prevent financial distress (Lown et al. 2015).

In examining the impact of self-efficacy on financial management behavior, studies show that the more confidence people have in using financial knowledge to manage money, the better their financial management behavior (Farrell et al. 2016). Financial self-efficacy is also found to mediate the relationship between financial literacy and financial management behavior (Lown et al. 2015; Chong et al. 2021), financial well-being (Lone and Bhat 2024), and risky credit behavior of college students (Liu and Zhang 2021). The mediating role of self-efficacy on financial behavior has also been examined via its effect on cognitive abilities by Tang (2021). In his study on older adults, Tang (2021) found that self-efficacy is a secondary source of cognitive influence. He concluded that lower cognitive abilities reduce older adults’ belief in their ability to control their lives, significantly decreasing their financial management efficiency.

The effect of self-efficacy on the choice of financial services and risk-taking behavior has also been examined in several studies (e.g., Pavani and Alagwadi 2023; Llewellyn et al. 2008). In this context, individuals with higher levels of self-efficacy, driven by a strong belief in their abilities, are more likely to take risks, influencing their choice of financial products. For example, in their study on Australian women, Farrell et al. (2016) showed that women with high levels of self-efficacy are more likely to use investment and saving products than women with low self-efficacy levels who are more inclined to use debt-related products. Murray and Dulebohn (2008) also highlighted the difference in people’s perceptions of investment in retirement plans according to their sense of self-efficacy. People with low levels of self-efficacy perceived it as a threat, unlike those with high ones who considered it as an opportunity. Financial self-efficacy also predicts financial market participation and wealth accumulation (Tang 2021; Lone and Bhat 2024).

Despite the growing research interest in examining the demand side determinants of financial inclusion in Egypt (Rashdan and Eissa 2020; Sayed and Shusha 2019; Fahmy and Ghoneim 2023), studies on rural household demand for financial services are limited.1 Adding to the existing literature, this paper attempts to provide a better understanding of rural household borrowing behavior, examining who are the population segments most likely to use credit and whether its usage is driven by economic and financial vulnerabilities. It also incorporates a behavioral perspective examining the effect of perceived general self-efficacy on the demand for formal loans, informal borrowing, and participation in RoSCAs. Additionally, it explores the impact of the informal credit market on the demand for formal credit. Using the latest Egypt Labor Market Panel Survey (ELMPS) national representative dataset in 2018, this paper examines the LCH and PIH within the context of rural Egypt, and in comparison to the literature, by testing the following hypotheses:

H1: 

Wealth has a positive relationship with demand for credit.

H2: 

Uncertainty of income has a negative relationship with demand for credit.

H3: 

Existence of informal market has a negative relationship with demand for credit.

H4: 

Self-efficacy has a positive relationship with demand for credit.

Following the introduction, this paper is organized as follows: Section 2 discusses the data and methods. Section 3 presents the main findings on rural households’ formal and informal credit use and the results of the probit analysis. Section 4 covers the discussion of the results, identifying factors that influence an individual’s likelihood of applying for a formal loan, engaging in informal borrowing, or participating in RoSCAs, and examining the relationship between the demand for formal and informal credit. Section 5 concludes the paper.

2. Materials and Methods

The quantitative analysis is based on the micro-level data of the latest round of the Egypt Labor Market Panel Survey (ELMPS), collected by the Central Agency for Public Mobilization and Statistics (CAPMAS) in 2018. Using a two-stage stratified random sampling design, the dataset is representative of Egyptian households and covers a total sample size of 15,746 households (61,231 individuals).2 The ELMPS 2018 provides detailed information on individual and household demographic and socioeconomic characteristics such as health, education, employment, job characteristics, housing, marriage, fertility, migration, information technology, perceptions, exposure to shocks, and saving and borrowing.3 For our research purposes, examining the demand for formal and informal credit in rural areas, the analysis is applied to the rural sample aged 15 years and above who responded to the questions on saving and borrowing behavior and included 24,465 individuals (9693 households). A brief description of the sample is provided in

Table 1. Sample profile of survey participants (N = 24,465).

Sample Categories	Frequency	Percentage
Gender
Male	11,913	49%
Female	12,358	51%
Age
15–19	3489	14.5%
20–29	5756	24%
30–39	5637	23%
40–49	3733	15.3%
50–59	2661	11%
60–64	1023	4.2%
65+	1973	8%
Marital Status
Ever Married	18,083	75%
Never Married	6126	25%
Education
Illiterate	7572	31%
Read and Write	1239	5%
Less than Intermediate	5125	21%
Intermediate	7624	31.5%
Above Intermediate	431	2%
University	2220	9.5%
Employment Status
Employed	13,424	60%
Unemployed	729	3%
Out of Labor Force	8423	37%
Region
Rural Upper Egypt	10,556	43.5%
Rural Lower Egypt	13,716	56.5%

Source: Calculated by the author.

below.

Using descriptive analysis, we first analyzed households’ use of formal and informal financial mechanisms in rural Egypt. This was followed by a probit regression to examine the determinants of demand for formal loans, informal borrowing, and participation in RoSCAs. A probit analysis was employed following the approach of previous studies on credit demand in other countries (e.g., Asiamah et al. 2021; Barslund and Tarp 2008) and rural Egypt (Mohieldin and Wright 2000). A dummy variable indicating that an individual applied for a loan from a formal financial institution was used as a proxy to capture the demand for formal credit. Similarly, two dummy variables were used to reflect the use of informal credit; first, informal borrowing was defined as borrowing from other individuals, and second, participation in one or more RoSCAs. To indicate individuals’ demographic and socioeconomic characteristics, we included several variables, specifically age, sex, marital status, and education. Health status, stability, and type of employment were also included to reflect the individual’s potential ability to work and his/her future income-earning prospects. Household characteristics, including household size, region, and financial position proxied by wealth quintiles and ownership of assets, including dwelling and agricultural land, were incorporated in the analysis. Besides reflecting the household financial position, agricultural land is commonly used as loan collateral in rural Egypt. Due to their high level of vulnerability (Helmy and Roushdy 2019), facing emergencies is a major financial need for rural households. Accordingly, the analysis included whether the household experienced an emergency in the year before the survey, and health insurance coverage, which may reduce the need for borrowing to cover health emergencies.

To reflect the influence of behavioral factors, we constructed a measure of perceived general self-efficacy using Schwarzer and Jerusalem’s (1995) Generalized Self-Efficacy Scale (GSES), which assesses ten statements on a scale from 1 to 4, where 1 = “Not at all true”, 2 = “Hardly true”, 3 = “Moderately True”, 4 = “Exactly True” (Cronbach’s alpha = 0.93).4 Each statement pertains to successful coping and indicates an internal, stable attribution of success (e.g., “I can always manage to solve difficult problems if I try hard enough”, “It is easy for me to stick to my aims and accomplish my goals”, “I am confident that I could deal efficiently with unexpected events”). The score was calculated for individual i by summing the scores of the ten statements. The score ranges from 10 to 40, with higher values indicating stronger self-efficacy (See

Table 2. Definitions of key variables.

Variables	Definitions
Formal loan	Dummy variable = 1 if applied for a loan from a formal financial institution
Informal borrowing	Dummy variable = 1 if borrowed money from others
RoSCAs	Dummy variable = 1 if a member in one or more RoSCAS
Age	Continuous variable measuring individuals’ age in years
Sex	Dummy variable = 1 if female
Marital Status	Dummy variable = 1 if ever married, i.e., married, widowed, divorced, or contractually married
Education	Categorical variable reflecting different education levels: illiterate, read and write, <Intermediate, Intermediate, >Intermediate, University education
Employment Stability	Categorical variable measuring employment stability in a primary job three months before the survey: permanent, temporary, seasonal, casual, unemployed, out of labor force
Health Status	Dummy variable = 1 if an individual has chronic illness or disability
Household Size	Continuous variable measuring the number of individuals in the household
Region	Dummy variable indicating household geographical location, = 1 if rural Upper Egypt
Wealth Quintile	Categorical variable reflecting the wealth status of rural households where 1 is the poorest 20% and 5 is the richest 20%. The wealth quintiles are calculated based on a wealth score computed based on the number of rooms, total housing area, building material, ownership of assets and durables such as fridge, freezer, dish washer, TV, satellite, radio, air conditioner, microwave, cooker, fan, heater, camera, car, bicycle, scooter, computer, cell phone, wireless router.
Ownership of agricultural land	Dummy variable = 1 if a member in the household owns agricultural land
Ownership of dwelling	Categorical variable indicating the type of dwelling ownership: owned, rented, fringe benefit, granted, or put a hand
Insurance coverage	Dummy variable = 1 if an individual is covered by any type of health insurance
Emergencies	Dummy variable = 1 if an individual was subject to any type of emergency in the 12 months before the survey
Self-efficacy	An individual’s perceived general self-efficacy measured using Schwarzer and Jerusalem’s (1995) Generalized Self-Efficacy Scale (GSES), assessing ten statements on a scale from 1 to 4. A score is calculated for every individual by summing the scores of the ten statements and ranges from 10 to 40, wherein a higher score indicates stronger self-efficacy.

) for a list of our variables and their definitions.

Using the maximum likelihood estimation method, we modeled the independent probability that an individual applies for formal loans (Pr (Formal Credit) = 1), borrows informally (Pr (Informal Borrowing) = 1), and participates in RoSCAs (Pr (RoSCAs) = 1), employing a binary probit model specification. Since the dependent variables were binary, i.e., the decision to borrow formally, informally, or join RoSCAs, the models were estimated using non-linear regression methods like probit regression (Wooldridge 2009). By maximizing the log likelihood function, assuming that disturbances follow a standard normal distribution, the probit model yields efficient and consistent estimates.

The probit model is based on a latent variable framework, where the true likelihood (Y_mi*) that an individual uses a form of credit is not directly observable. Instead, it is estimated as a probability (Y_mi) ranging between 0 and 1, where i represents the individual and m represents the type of credit.

$${\mathrm{Y}}_{\mathrm{m}\mathrm{i}}^{\mathrm{*}}={\mathsf{\beta }}_{°}+{\mathsf{\beta }}_1{\mathrm{X}}_{\mathrm{m}\mathrm{i}}+{\mathsf{\beta }}_2{\mathrm{H}}_{\mathrm{m}\mathrm{i}}+{\mathsf{\beta }}_3{\mathrm{I}}_{\mathrm{m}\mathrm{i}}+{\mathsf{\beta }}_4{\mathrm{E}}_{\mathrm{m}\mathrm{i}}+{\mathsf{\beta }}_5{\mathrm{G}\mathrm{S}\mathrm{E}}_{\mathrm{m}\mathrm{i}}+{\mathsf{\epsilon }}_{\mathrm{m}\mathrm{i}}$$

(1)

$${\mathrm{Y}}_{\mathrm{m}\mathrm{i}}=\left\{\begin{array}{c}1,\mathrm{if}{\mathrm{Y}}_{\mathrm{m}\mathrm{i}}^{\mathrm{*}}>0\\ 0,\mathrm{o}\mathrm{t}\mathrm{h}\mathrm{e}\mathrm{r}\mathrm{w}\mathrm{i}\mathrm{s}\mathrm{e}\end{array}\right.$$

(2)

$$\mathrm{m}=1,.....,\mathrm{M}$$

where Y_mi denotes the probability outcomes for the demand for each type of credit: formal loans, informal borrowing, and RoSCAs. X_mi denotes the vector of individual demographic and socioeconomic characteristics, and H_mi denotes the vector of household characteristics. I_mi denotes health insurance coverage, E_mi denotes emergencies, and GSE_mi denotes perceived general self-efficacy. B₁ to B₅ represents the estimated coefficients. A dummy variable is added to the equation to examine the relation between the demand for formal and informal credit to indicate if the individual concerned uses informal borrowing or participates in RoSCAs while using formal credit.

The probit models’ coefficients and average marginal effects5 were estimated as shown in Section 3 below. Additionally, several measures of goodness of fit and model specification tests were applied. The tests showed that the models were well specified and correctly predicted 97% of observed data. The matrix of correlations among independent variables was also estimated to check against multicollinearity. The results are available in Appendix A.2.

Research Limitations

While the literature on financial management behavior underscores the importance of perceived general self-efficacy, we acknowledge that employing a domain-specific measure focusing on financial self-efficacy would likely yield more accurate results. However, the ELMPS 2018 dataset, the most recent wave available at the time the paper was written, only includes data on general self-efficacy, which limited our ability to use a more specific measure.

3. Results

In this section, we present the findings of the data analysis in the following structure. First, we provide a detailed analysis of rural household participation in formal and informal credit markets, examining variations across regions, wealth quintiles, and demographic and socio-economic characteristics. Second, we analyze the relationship between perceived general self-efficacy and demographic and socioeconomic characteristics. Finally, we outline the results of the probit regression, focusing on the determinants of participation in formal and informal borrowing, as well as RoSCAs.

3.1. Rural Households’ Use of Formal and Informal Credit: Wealth Quintile and Regional Variations

Mapping the rural financial landscape in Egypt from a demand-side perspective reveals a range of formal and informal financial mechanisms. While microcredit and post office savings accounts are the most commonly used formal financial services, the informal financial market in Egypt encompasses several distinct mechanisms. These include interest-free, occasional lending and borrowing, typically practiced among relatives and friends, based on reciprocal personal relationships. Another prevalent form is interlinked credit, where two or more mutually dependent transactions are simultaneously agreed upon in a bundled deal. A common example is the trader–farmer relationship, in which a trader extends credit to a farmer based on the farmer’s commitment to sell crops. This arrangement, relying on trust and ongoing relationships, substitutes for collateral and reduces transaction costs and default risk. Rotating Savings and Credit Associations (RoSCAs) are a widely used informal financial instrument based on collective arrangements and are considered crucial means for saving and a source of credit. In a RoSCA, a group of individuals deposits funds with a designated leader, and the pooled savings are then lent to group members with no interest rate at the end of every cycle. Finally, informal lending from specialized moneylenders and pawnbrokers, who leverage their in-depth knowledge of borrowers, is relatively expensive but quite rare in Egypt (Mohieldin and Wright 2000).

Rural households are more active in the informal financial sector. However, variations exist across regions and wealth quintiles. For instance, rural households in Lower Egypt are more financially active and show higher engagement with the formal sector than those in Upper Egypt. This may be partially due to the higher economic activity in Lower Egypt and its closer proximity to the capital. Additionally, the type of formal and informal financial services used differs by wealth quintile. Households in the poorest wealth quintiles rely more on formal and informal borrowing than those in the richest quintiles. Notable differences also emerge in loan sizes, reasons for borrowing, and the formal financial institutions approached by the poorest compared to the richest rural households, as detailed below.

3.1.1. The Use of Formal Loans in Rural Egypt

Due to the low levels of engagement with the formal financial sector, the use of formal loans in rural Egypt is notably limited. Only 2.3% (n = 558) of the total rural sample borrow loans from formal financial institutions. Although it varies significantly across wealth quintiles, the average loan size is EGP 25,230. The average loan size for the poorest 20% is EGP 9525 compared to EGP 68,350 for the richest 20%. The reasons for obtaining formal loans also vary across wealth quintiles. The data show that the poorest 20% primarily borrow to cover medical emergencies (27%) and pay off existing debt (26.5%), followed by buying a house (20.7%), financing marriage (14.3%), and funding an agricultural business (14.3%). Conversely, the richest 20% borrow mainly to buy a house (23%), finance marriage (18.8%), and finance a non-farm enterprise (12%). Financing marriage is a common reason for borrowing across all wealth quintiles, averaging 22%. Contrary to variations across wealth quintiles, the reasons for applying for a formal loan are almost the same across rural Upper and Lower Egypt. However, a higher percentage of rural households applied for formal loans in Lower Egypt.

Regarding the sources of formal loans, 45% of the sample borrows from public banks, 18.3% from private banks, and 18.2% from NGOs and charitable organizations. This distribution differs for the poorest 20% of borrowers, who primarily rely on NGOs and charitable organizations (34%), followed by public banks (32%). In contrast, 61% of the richest 20% rely on public banks for access to finance. This discrepancy may be due to the relatively easy access to loans from NGOs, which require fewer collaterals and less documentation, in addition to the established relationship between the poor and NGOs, perceived as community supporters, thus making the poor feel more secure borrowing from them. Regionally, rural households in Lower Egypt exhibit different preferences for formal financial institutions than those in Upper Egypt. As households in Lower Egypt favor public banks and government institutions, those in Upper Egypt primarily rely on private banks, companies, and NGOs6.

When asked whether they pay interest or fees on formal loans, 10% of respondents answered “none” or “do not know”. Among those who report paying interest and/or fees, only 12.4% report paying both, 72% report paying interest, and 6% report paying fees. This highlights a lack of clear understanding of the loan terms and conditions, including fees and interest. It is also possible that the varied responses are due to the different titles people use to refer to the additional amount paid on the loan principal. For example, the interest paid on subsidized loans provided by NGOs, due to their small amount, is usually referred to as additional fees to cover the cost of loan issuance and management. This terminology is more acceptable to the rural population, as interest is considered usury and is prohibited by the Islamic Shariah. Regarding loan terms, on average borrowers paid interest that reached five times the initial loan principal, with an average monthly installment of EGP 1063. Notably, only 3.5% of the rural individuals who applied for formal loans were rejected because of incomplete documentation or insufficient collateral.

3.1.2. Use of Informal Borrowing and RoSCAs in Rural Egypt

Compared to the use of formal loans, a larger proportion of the sample engages with the informal financial sector. Informal borrowing, RoSCAs, and purchasing on credit and in installments, are the key informal financial mechanisms rural households use. Specifically, 5.6% (n = 1398) of the sample borrowed from informal sources, primarily relatives (59%), friends (24%), and neighbors (11%). Local money lenders play a minimal role in the rural informal financial sector, with only 2.7% of informal borrowers utilizing them for finance.

Unlike formal loans, informal borrowing is often used to meet regular expenses, such as daily needs (19%) and medical emergencies (25%) with minor differences in usage across wealth quintiles. The average size of informal borrowing is approximately half that of formal loans, averaging EGP 11,484. For the poorest 20% of borrowers, the average amount of informal borrowing is EGP 6200, compared to EGP 19,432 for the wealthiest 20%. Notably, aside from local money lenders, 97% of informal loans are interest- and fee-free. In addition to formal loans and informal borrowing, buying on installment plans is another source of debt for rural individuals. Specifically, 3.5% (n = 835) of the sample purchased goods in installments, mainly electric and home appliances, and clothes.

Of the total sample, 4.23% (n = 1021) participated in an average of two RoSCAs. RoSCAs are characterized by their flexibility in terms of the periodicity of payment, which could range from a daily to a yearly basis, with monthly contributions being the most common. RoSCAs are formed among friends, work colleagues, and family members as other informal means of finance that rely on trust and relationships. The average contribution per round in a RoSCA is EGP 615, though this amount varies across wealth quintiles, averaging EGP 340 for the poorest 20% compared to EGP 930 for the richest 20%.

Figure 1 illustrates the percentage of rural households across different wealth deciles who use formal loans, informal borrowing, and RoSCAs. Rural households are more dependent on informal financial mechanisms, with the poorest households being the most reliant on formal and informal credit and the least reliant on RoSCAs.

As shown in

Table 3. Number of individuals using formal loans and informal borrowing in rural Egypt.

Informal Borrowing	Formal Loans		Total
	Yes	No
Yes	177	1175	1352
No	375	22,093	22,468
Total	552	23,269	23,821

Source: Calculated by the authors using the ELMPS 2018 dataset.

and

Table 4. Number of individuals using formal loans and RoSCAs in rural Egypt.

Formal Loans	RoSCAs Membership		Total
	Yes	No
Yes	128	423	551
No	893	22,674	23,567
Total	1021	23,097	24,118

Source: Calculated by the authors using the ELMPS 2018 dataset.

, a segment of the rural sample engages with the formal and informal financial sectors, though the majority interact with only one. Among informal borrowers, 13% also withdraw formal loans, and of those who buy in installment plans, 11% borrow formal loans. Similarly, 12.5% of RoSCAs members apply for formal loans. It is worth noting that most rural individuals who use formal and informal borrowing are males, permanent laborers, married, and nearly 50% are aged between 30 and 45, with an almost even distribution across wealth quintiles.

Figure 2 illustrates the overlap in the financial needs addressed by formal loans, informal borrowing, and RoSCAs. The predominant reliance on the informal financial sector indicates a strong preference for these means, suggesting that formal loans are often considered a last resort, utilized only after informal options are exhausted. The significantly larger average size of formal loans suggests that rural households may resort to formal loans when they need substantial funds that cannot be secured through informal financial means. Additionally, the relative importance of formal loans increases compared to RoSCAs and informal borrowing in addressing the need to finance agriculture and non-farm enterprises. For these reasons, formal loans seem to complement informal credit.

3.2. Rural Households’ Use of Formal and Informal Credit: Demographic and Socioeconomic Variations

Analyzing rural households’ participation in formal and informal credit markets reveals similar patterns of engagement across the different demographic and socioeconomic characteristics. Except for RoSCAs, where the gap between males and females is minimal, males represent the majority of formal and informal borrowers. Similarly, married individuals aged between 30 to 59 represent the highest percentage of participants in formal and informal borrowing and RoSCAs. Individuals with intermediate education and the illiterate together account for more than half of formal credit users, informal borrowers, and RoSCA members. A notable difference is observed between individuals with health insurance coverage and those without, with the latter representing the larger proportion of formal and informal borrowers. Emergency appears to be a major reason for seeking finance. As shown by the data, around 30% of the individuals who borrow formally or informally or join RoSCAs have been subject to an emergency. Contrary to expectations, agricultural landowners are less reliant on formal and informal credit compared to those who do not own agricultural land.

Table 5. Percentage of individuals using formal credit, informal borrowing, and RoSCAs by socio-economic and demographic characteristics in rural Egypt.

	Formal Loans	Informal Borrowing	RoSCAs
Sex
Male	72.6%	65.4%	51%
Female	27.4%	34.6%	49%
Marital Status
Ever Married	97.2%	95.5%	91%
Never Married	2.8%	4.5%	9%
Age
15–19	0.18%	0.9%	4%
20–29	13.5%	14%	15%
30–39	28.5%	34.5%	33%
40–49	21.8%	25.6%	24%
50–59	27.5%	15%	16%
60 and above	8.5%	10%	8%
Education
Illiterate	25%	30%	21%
Read and Write	10%	9%	6.6%
Less than Intermediate	15%	16.5%	16%
Intermediate	39.5%	35%	37.8%
Above Intermediate	2.5%	2%	4%
University	8%	7.5%	14.6%
Employment Stability
Permanent	69%	57.8%	68%
Temporary	4%	8%	9%
Seasonal	0.7%	2.5%	0.5%
Casual	10.5%	14.5%	7%
Unemployed	2.8%	3.4%	1.5%
Out of labor force	13%	13.8%	14%
Household Size
1–4	43.5%	45%	42%
5–8	54.5%	53%	57%
9 and above	2%	2%	1%
Health Status
Chronic Illness/Disability	42%	46.4%	41%
No Chronic Illness/Disability	58%	53.6%	59%
Health Insurance Coverage
Covered by Health Insurance	36%	23.4%	39%
Not Covered by Health Insurance	64%	76.6%	61%
Emergency
Subject to Emergency	31%	37%	25%
Not Subject to Emergency	69%	63%	75%
Ownership of Agricultural Land
None of family members own land	88%	87.6%	85%
A family member own land	12%	12.3%	15%

Source: Calculated by the authors using the ELMPS 2018 dataset.

presents the distribution of participants in formal and informal credit markets by demographic and socioeconomic characteristics.

As illustrated in

Table 6. Average size of formal and informal loans by socioeconomic and demographic characteristics in rural Egypt.

	Average Loan Size (EGP)
	Formal Loans	Informal Borrowing
Sex
Male	37,668	12,906
Female	10,858	6136
Marital Status
Ever Married	30,636	10,894
Never Married	16,533	3758
Health Status
Chronic Illness/Disability	16,079	11,709
No Chronic Illness/Disability	40,502	9657
Health Insurance Coverage
Covered by Health Insurance	30,669	17,394
Not Covered by Health Insurance	30,189	8783
Emergency
Subject to Emergency	13,645	9273
Not Subject to Emergency	37,887	11,395
Ownership of Agricultural Land
None of family members own land	29,686	9767
A family member own land	35,349	16,937

Source: Calculated by the authors using the ELMPS 2018 dataset.

, there is a significant disparity in average loan sizes between the formal and informal credit markets, with loan amounts in the formal sector being substantially larger than those in the informal sector.

3.3. General Self-Efficacy, Individual Characteristics, and Financial Behavior in Rural Egypt

Examining the relationship between perceived general self-efficacy and rural individuals’ demographic and socioeconomic characteristics, we found that older, male, married, and more educated individuals have a stronger sense of self-efficacy. This aligns with some studies that show there is a correlation between demographic characteristics, specifically sex and general self-efficacy, where males are found to have a stronger sense of self-efficacy (Lown et al. 2015). Similarly, rural individuals in higher wealth quintiles, with stable employment, and who own assets such as agricultural land and dwellings have higher levels of self-efficacy. In contrast, rural individuals with lower levels of self-efficacy are relatively higher in rural Upper Egypt, belong to larger households, have chronic illness or disability, and have been subject to an emergency.

Controlling for demographic and socioeconomic characteristics, the data reveal a significant positive correlation between perceived general self-efficacy and the likelihood of using formal loans. This aligns with the literature, which shows that individuals with higher levels of self-efficacy are more confident in their financial management abilities and are more likely to utilize various financial services (Lown et al. 2015; Lown 2011). However, while the correlation between self-efficacy and informal borrowing is positive, it is only significant at the 90% confidence level and is insignificant for participation in RoSCAs.

3.4. Probit Analysis of the Demand for Formal and Informal Credit in Rural Egypt

3.4.1. Demand for Formal Loans in Rural Egypt

The results of the probit model estimation, including the estimated coefficients and average marginal effects, for the determinants of demand for formal loans in rural Egypt are presented in

Table 7. Demand for formal loans in rural Egypt.8

Dependent Variable: Demand for Formal Loans
Variables	Model 1		Model 2		Model 3
	Coefficient	Marginals	Coefficient	Marginals	Coefficient	Marginals
Age	0.0661 ***	0.00317 ***	0.062 ***	0.00289 ***	0.0582 ***	0.00271 ***
Age squared	−0.0007 ***	−0.00003 ***	−0.00062 ***	−0.00003 ***	−0.00057 ***	−0.00002 ***
Female	−0.270 ***	−0.01252 ***	−0.239 ***	−0.0108 ***	−0.309 ***	−0.01392 ***
Ever married	0.539 ***	0.01825 ***	0.473 ***	0.01636 ***	0.573 ***	0.01859 ***
Education ***
Illiterate	0.186 *	0.00643 *	0.193 *	0.006615 *	0.206 *	0.007005 *
Read and write	0.426 ***	0.01851 ***	0.409 ***	0.01709 ***	0.420 ***	0.01738 ***
<Intermediate	0.403 ***	0.01711 ***	0.375 ***	0.01519 ***	0.409 ***	0.01681 ***
Intermediate	0.371 ***	0.01527 ***	0.375 ***	0.01517 ***	0.376 ***	0.01496 ***
>Intermediate	0.202	0.00709	0.212	0.00739	0.141	0.00450
Health Status
Chronic illness/disabilities	0.279 ***	0.01475 ***	0.211 ***	0.01053 ***	0.236 ***	0.01191 ***
Health insurance	0.385 ***	0.02182 ***	0.390 ***	0.02146 ***	0.332 ***	0.01779 ***
Household size	0.0236	0.00113	0.0233	0.00109	0.0253	0.00117
Region
Rural Upper Egypt	0.00600	0.00028	0.0431	0.0020249	0.0128	0.00059
Rural wealth quintile **
1st wealth quintile (Poorest 20%)	0.352 ***	0.01687 ***	0.333 ***	0.01556 ***	0.412 ***	0.01877 ***
2nd wealth quintile	0.184 **	0.00756 **	0.182 **	0.00746 **	0.253 ***	0.00996 ***
3rd wealth quintile	0.128	0.00501	0.117	0.00453	0.187 **	0.00694 **
4th wealth quintile	0.237 ***	0.01024 ***	0.213 **	0.00897 **	0.268 ***	0.01070 ***
Ownership dwelling ***
Rented dwelling (old or new law)	0.677 ***	0.05417 ***	0.605 ***	0.04461 ***	0.626 ***	0.04642 ***
Fringe benefit/grant/put a hand	0.0434	0.002034	0.00612	0.00027	0.0442	0.002018
Ownership agricultural land	−0.180 ***	−0.00776 **	−0.173 **	−0.00732 **	−0.176 **	−0.00741 **
Emergency	0.266 ***	0.01438 ***	0.214 ***	0.010961 ***	0.262 ***	0.01364 ***
Self-efficacy	0.0134 ***	0.00064 ***	0.0129 ***	0.000603 ***	0.0141 ***	0.00065 ***
Employment Stability ***
Temporary	−0.387 ***	−0.01633 ***	−0.423 ***	−0.01683 ***	−0.423 ***	−0.01648 ***
Seasonal	−0.623 **	−0.02171	−0.714 ***	−0.02271 ***	−0.579 ***	−0.019939 ***
Casual	−0.0926	−0.00499	−0.124	−0.00627	−0.0859	−0.004402
Unemployed	−0.0693	−0.00380	−0.0612	−0.003264	−0.0192	−0.00104
Out of labor force	−0.317 ***	−0.01417 ***	−0.275 ***	−0.01231 ***	−0.259 ***	−0.01153 ***
Borrow informally	-	-	0.710 ***	0.05318 ***	-	-
RoSCAs	-	-	-	-	0.734 ***	0.05768 ***
Constant	−4.946 ***	-	−4.871 ***	-	−4.960 ***	-
Observations	22,860	-	22,578	-	22,860	-
AIC	258.431		250.681		250.4
McFadden Adjusted R2	0.162		0.194		0.188

*** p < 0.01, ** p < 0.05, * p < 0.1.

. The results show that demographic and socioeconomic characteristics are significant determinants of demand for formal loans in rural Egypt. Older and ever-married individuals are found to be more likely to apply for formal loans. Notably, the probability of applying for formal loans increases significantly starting from aged 35 and keeps increasing for older age groups. Additionally, the results indicate that rural women are 1.2% less likely than men to apply for formal loans. Education is also a significant determinant as lower levels of education are associated with a higher likelihood of applying for formal loans. Similarly, the results show the significance of employment stability wherein permanent laborers are more likely to apply for formal loans than temporary and irregular labor and those unemployed and out of the labor force. Comparing the different types of employment, employers and self-employed individuals are 1.4% and 1.8% more likely than waged employees to apply for formal loans, potentially due to their need for capital to run their owned businesses.7

Regarding the significance of household characteristics, the analysis shows no significant relation between household size and demand for formal loans. Similarly, there is no significant difference between rural Upper and Lower Egypt. On the contrary, the results show a significant effect of household financial position proxied by wealth quintiles. Rural individuals in lower wealth quintiles are more likely to apply for formal loans. Additionally, homeownership and ownership of agricultural land have a significant negative effect, suggesting that individuals who own assets are in a better financial position, and thus have a lower need for finance than those who do not.

Consistent with the data showing that emergencies are a primary reason for using formal loans, the results show that emergencies lead to a 1.4% increase in the likelihood of applying for formal loans. Similarly, rural individuals with chronic illness or disability are more likely to apply for formal loans. Health insurance coverage increases the probability of applying for a formal loan by 2%. Additionally, rural individuals with higher self-efficacy are found to be more likely to apply for formal loans.

When comparing our findings on rural household participation in the formal credit market, we observe that most results contradict the theoretical expectations. Older, less educated, and poorer rural individuals, as well as those with chronic illness or disability, and those who have faced emergencies, are more likely to seek formal credit. A consistent finding relates to employment stability, which may reflect a supply-side factor that provides easier access to credit for those with stable employment compared to their counterparts who are casual or temporary laborers. There is no significant regional difference between rural Upper and Lower Egypt.

3.4.2. Demand for Informal Borrowing and Participation in RoSCAs in Rural Egypt

To explore the difference between the determinants of demand for formal loans and informal financial mechanisms, a probit analysis of the determinants of informal borrowing, and participation in RoSCAs in rural Egypt is undertaken, as shown in

Table 8. Demand for informal borrowing, RoSCAs, and formal loans in rural Egypt: a comparison.9

	Informal Borrowing		Participation in RoSCAs		Demand for Formal Loans
Variables	Model 4		Model 5		Model 2		Model 3
	Coefficient	Marginals	Coefficient	Marginals	Coefficient	Marginals	Coefficient	Marginals
Age	0.0681 ***	0.00668 ***	0.0817 ***	0.00649 ***	0.062 ***	0.00289 ***	0.0582 ***	0.00271 ***
Age Squared	−0.0008 ***	−0.00007 ***	−0.0009 ***	−0.00007 ***	−0.0006 ***	−0.00003 ***	−0.0005 ***	−0.00002 ***
Female	−0.255 ***	−0.02481 ***	0.260 ***	0.02100 ***	−0.239 ***	−0.0108 ***	−0.309 ***	−0.01392 ***
Ever married	0.517 ***	0.03942 ***	0.0506	0.0039175	0.473 ***	0.01636 ***	0.573 ***	0.01859 ***
Education
Illiterate	0.156 **	0.013297 **	−0.116	−0.00805	0.193 *	0.006615 *	0.206 *	0.0070058 *
Read and write	0.298 ***	0.02806 ***	0.159	0.013607	0.409 ***	0.01709 ***	0.420 ***	0.01738 ***
<Intermediate	0.272 ***	0.02510 ***	0.0946	0.007721	0.375 ***	0.01519 ***	0.409 ***	0.01681 ***
Intermediate	0.207 ***	0.01830 ***	0.0851	0.006899	0.375 ***	0.01517 ***	0.376 ***	0.01496 ***
>Intermediate	0.116	0.0095679	0.223	0.020116	0.212	0.00739	0.141	0.0045036
Employment Stability
Temporary	0.0447	0.0050026	0.197 **	0.023106 **	−0.423 ***	−0.0168 ***	−0.423 ***	−0.01648 ***
Seasonal	0.118	0.0138173	−0.472 ***	−0.0336 ***	−0.714 ***	−0.0227 ***	−0.579 ***	−0.01993 ***
Casual	0.0577	0.0065094	−0.0965	−0.00913	−0.124	−0.0062796	−0.0859	−0.0044024
Unemployed	0.00652	0.0007102	−0.400 ***	−0.0301 ***	−0.0612	−0.0032642	−0.0192	−0.0010404
Out of labor force	−0.315 ***	−0.02759 ***	−0.492 ***	−0.0346 ***	−0.275 ***	−0.0123 ***	−0.259 ***	−0.01153 ***
Health Status
Chronic illness/ disabilities	0.441 ***	0.04968 ***	0.338 ***	0.03063 ***	0.211 ***	0.01053 ***	0.236 ***	0.01191 ***
Health insurance	−0.00148	−0.0001454	0.364 ***	0.03346 ***	0.390 ***	0.02146 ***	0.332 ***	0.01779 ***
Household size	−0.00577	−0.0005665	0.00197	0.0001562	0.0233	0.00109	0.0253	0.0011775
Region
Rural Upper Egypt	−0.119 **	−0.01158 **	0.0734	0.0058959	0.0431	0.0020249	0.0128	0.0005984
Rural wealth quintiles
1st wealth quintile (Poorest 20%)	0.0562	0.0114317	−0.237 ***	−0.0200 ***	0.333 ***	0.01556 ***	0.412 ***	0.01877 ***
2nd wealth quintile	0.0562	0.0051825	−0.269 ***	−0.0222 ***	0.182 **	0.007467 **	0.253 ***	0.00996 ***
3rd wealth quintile	0.0560	0.005166	−0.298 ***	−0.0241 ***	0.117	0.0045317	0.187 **	0.0069401 **
4th wealth quintile	0.136 **	0.013253 **	−0.0456	−0.0044388	0.213 **	0.00897 **	0.268 ***	0.01070 ***
Assets ownership
Rented dwelling (old or new law)	0.394 ***	0.04698 ***	0.421 **	0.04446 **	0.605 ***	0.04461 ***	0.626 ***	0.0464 ***
Fringe benefit/grant/ put a hand	0.195 ***	0.02034 ***	0.0331	0.0026158	0.00612	0.0002753	0.0442	0.0020183
Ownership agricultural land	−0.106 *	−0.009937 *	−0.0340	−0.0026497	−0.173 **	−0.00732 **	−0.176 **	− 0.007411 **
Emergency	0.388 ***	0.04374 ***	0.101 *	0.008411 *	0.214 ***	0.01096 ***	0.262 ***	0.013643 ***
Applied for formal loans	0.799 ***	0.12577 ***	0.768 ***	0.10317 ***
Self-efficacy	0.00675 **	0.000662 **	0.00346	0.0002746	0.0129 ***	0.00060 ***	0.0141 ***	0.000657 **
Informal borrowing	-	-	-	-	0.710 ***	0.05319 ***	-	-
RoSCAs	-	-	-	-		-	0.734 ***	0.057681 ***
Constant	−3.849 ***		-3.577 ***	-	−4.871 ***	-	−4.960 ***
Observations	22,578		22,860	-	22,578	-	22,860
AIC	508.4		418.5		250.681		250.4
McFadden Adjusted R2	0.162		0.135		0.194		0.188

*** p < 0.01, ** p < 0.05, * p < 0.1.

. According to the results, demographic and socioeconomic characteristics are still significant determinants of demand for informal borrowing and participation in RoSCAs in rural Egypt. Age continues to be a significant factor that increases the likelihood of informal borrowing and using RoSCAs. Similarly, married and less educated rural individuals are found to be more likely to borrow informally and join RoSCAs. As in the case of formal loans, females are 2.4% less likely to borrow informally but 2% more likely to join RoSCAs, than males. Although significant for informal borrowing and RoSCAs, employment stability has a different effect. In this context, individuals employed in irregular labor are found to be more likely to borrow informally but less likely to join RoSCAs.

Regarding household characteristics, the results show no significant effect of household size on informal borrowing and participation in RoSCAs. However, a significant difference exists between rural Upper and Lower Egypt households concerning informal borrowing. In this regard, households in rural Upper Egypt are significantly less likely to borrow informally. Unlike the case of formal loans, informal borrowing does not significantly differ between households across wealth quintiles. In contrast, households in lower wealth quintiles are 2% less likely to participate in RoSCAs. Ownership of agricultural land significantly increases the likelihood of informal borrowing but has an insignificant positive effect on participation in RoSCAs. Conversely, homeowners are less likely to borrow informally and join RoSCAs.

The analysis also demonstrates the significant positive effect of emergencies and chronic illness or disabilities on the likelihood of borrowing informally and RoSCA participation. Rural individuals who have been subject to an emergency are 4.3% more likely to borrow informally and only 0.8% more likely to join RoSCAs. This highlights that informal borrowing is the most convenient means to address the urgent need for finance compared to RoSCAs, which need relatively more time to be formed, and formal loans that require going through the application procedures. Health insurance coverage does not significantly affect informal borrowing but significantly positively affects joining RoSCAs. Like the demand for formal loans, rural individuals with stronger general self-efficacy are significantly more likely to borrow informally. However, no significant relationship was found between self-efficacy and participation in RoSCAs.

Considering the findings on informal credit market participation, the effects of the majority of demographic and socioeconomic characteristics are similar to those observed in formal credit market participation. Individuals who are poorer, less educated, experiencing chronic illness or disability, and have been subject to emergency are more likely to borrow informally. This further underscores financial need and vulnerability as primary drivers of borrowing behavior. A notable distinction arises regarding participation in RoSCAs, where poorer individuals and irregular laborers are less likely to join. Additionally, a significant regional difference is observed in informal borrowing, with residents of Upper Egypt being less likely to engage in informal borrowing.

4. Discussion

4.1. Comparing Demand for Formal Loans, Informal Borrowing, and RoSCAs in Rural Egypt

Comparing the results of the probit analysis for the demand for formal loans, informal borrowing, and RoSCA participation in rural Egypt reveals that most of the determinants have a consistent effect across them, despite varying in magnitude, with a few notable exceptions. The lack of distinction in many of the determinants of demand for formal and informal credit partially lies in the evident overlap in the financial needs addressed by formal loans, informal borrowing, and RoSCAs. This contradicts studies in other countries that found formal loans are almost entirely used for production and informal borrowing for consumption smoothing (Nguyen et al. 2021; Barslund and Tarp 2008). Additionally, the predominance of informal finance means suggests that formal loans are considered a last resort after exhausting informal financial means.

Concerning demographic and socioeconomic characteristics, driven by the higher financial commitments and responsibility of household expenses, married rural men aged 35 and above are more likely to apply for formal loans and borrow informally. This result aligns with the findings of Asiamah et al. (2021) and Mohieldin and Wright (2000), who found that married rural individuals are more likely to have a formal loan, and contradicts Dharmadasa and Gunatilake (2023) who suggested an insignificant effect of marital status. Regarding the effect of age, our results align with Chen and Chivakul (2008), who found a positive relationship between age and demand for credit; however, they contradict Asiamah et al. (2021), Cheng and Ahmed (2014), and Chandio et al. (2021), who suggested a negative effect, and Rashdan and Eissa (2020), Dharmadasa and Gunatilake (2023), and Fahmy and Ghoneim (2023), who found age is of insignificant influence. Interestingly, as household heads get older, they become responsible for financing their children’s marriage expenses, which is known to be inflated in rural Egypt (Assaad and Krafft 2014). This could be one potential reason for the increasing need for finance as age increases. This interpretation is also consistent with the analysis of the reasons for formal borrowing that show that financing marriage is among its primary reasons. In line with the results of Mohieldin and Wright (2000), due to cultural reasons, rural women are less likely to have a formal loan since they mostly depend on their male relatives to manage their assets. Additionally, according to the rural culture, men are the main breadwinners and women are less likely to work, which could also explain why men are more likely to apply for formal loans and borrow informally. Our results align with the findings of Nguyen et al. (2021), Mpuga (2010), and Kofarmata et al. (2016). In contrast, they differ from those of Dharmadasa and Gunatilake (2023), Rashdan and Eissa (2020), Fahmy and Ghoneim (2023), Chandio et al. (2021), and Chen and Chivakul (2008), who reported an insignificant effect of sex. The fact that most RoSCAs are formed and managed by women, who tend to have a large network of relatives and neighbors, explains why women are more likely than men to join RoSCAs.

A notable difference between the demand for formal loans and RoSCAs and informal borrowing is the effect of employment stability. Rural individuals with regular employment are more likely to demand formal loans and participate in RoSCAs but less likely to borrow informally. This is due to the common use of monthly income as collateral for formal loans, which indicates the borrower’s ability to repay (Mohieldin and Wright 2000). Additionally, unstable and irregular-income individuals will likely be reluctant to take loans due to the difficulty of committing to a fixed schedule of loan installments. Similarly, as rural individuals are keen to preserve their social image and sustain RoSCAs as a financial means, they consider it a binding commitment that should be fulfilled. This explains why irregular laborers will not likely join RoSCAs to avoid the possibility of being unable to pay its regular installments. Our findings are consistent with Vandone (2009), Chandio et al. (2021), Fahmy and Ghoneim (2023), and Rashdan and Eissa (2020), who reported a positive effect of employment stability on participation in formal credit market. However, they contradict Chen and Chivakul (2008), who found no significant difference between employed and unemployed individuals in credit market participation.

As for the level of education, the results show a negative relationship between education and the likelihood of both applying for formal loans and engaging in informal borrowing. However, it is worth noting that the probability of applying for formal loans and informal borrowing increases significantly as education levels rise from illiterate to literate, then starts decreasing with higher education levels. The reason for this is unclear. A potential explanation for the negative effect of education is that rural individuals with higher levels of education could be in a better financial position, and thus have a lower need for finance. Also, more educated individuals are expected to be more capable of understanding loans’ terms and conditions and assessing the resulting debt burden, which could increase their risk aversion towards using formal loans. This contradicts the existing literature that highlights the positive impact of education on demand for formal credit (Nguyen et al. 2021), (Moahid and Maharjan 2020), (Qin et al. 2018), (Cull et al. 2019), (Rashdan and Eissa 2020), (Dharmadasa and Gunatilake 2023), (Chandio et al. 2021), (Asiamah et al. 2021). However, it aligns with Chen and Chivakul (2008), who reported that formal credit market participation decreases with higher education levels, and Barslund and Tarp (2008), who found that an additional year of education reduces the probability of household heads demanding credit from informal sources. As for the use of RoSCAs, no significant difference exists across rural individuals with different levels of education.

Regarding household characteristics, household size has an insignificant positive effect on the demand for formal loans, informal borrowing, and joining RoSCAs. The results on the demand for formal credit contradict Chen and Chivakul (2008) and Dharmadasa and Gunatilake (2023), who suggested a significant positive effect, and Mohieldin and Wright (2000), who documented a positive though diminishing effect of household size on the demand for formal loans. However, it aligns with Mpuga (2010) and Chandio et al. (2021), who reported an insignificant impact of household size on demand for formal credit. It is noteworthy that household composition is an important factor besides household size when assessing how rural households address their financial needs. According to the rural culture, male household members, including sons, are expected to work and fulfill household expenses, while females are not. This affects the household earning potential and income level, and consequently their need for finance and use of financial services.

The financial position of a household also proves to be a significant determinant of the demand for formal loans, informal borrowing, and using RoSCAs, though with different effects. While the poorest 20% of rural individuals are more likely to apply for a formal loan and borrow informally, they are less likely to participate in RoSCAs. This can be explained by the poorest individuals’ need for finance. This disagrees with the formal and informal borrowing results in Mohieldin and Wright (2000) and Chandio et al. (2021), and Fahmy and Ghoneim (2023), Vandone (2009), and Chen and Chivakul (2008) on formal credit, but aligns with Nguyen et al. (2021). It is also worth noting that we observe whether an individual applies for a loan and not having a loan, which means that despite their willingness, they might not be successful in getting it. Another notable observation is that the poorest individuals are willing to apply for a formal loan and not join a RoSCA despite the higher cost of the formal loan and the similar need to commit to paying its regular installments. This indicates the strength of the effect of peer pressure in the rural community and the crucial importance of social image (El Shayeb 2015; Abdel Rahman 2017). In a similar context, despite the insignificant effect of the household region on the demand for formal loans and participation in RoSCAs, rural individuals in Upper Egypt, where social image is of crucial importance, are significantly less likely to borrow informally, potentially indicating their preference not to ask others for money and show their need for finance to preserve their social image.

Concerning the type of dwelling ownership, rural individuals who rent their dwellings are 5.4% more likely to apply for formal loans, and 4.5% more likely to borrow informally and join RoSCAs than those who own their dwellings. A potential explanation is that rural individuals who own their dwellings are probably in a better financial position, and thus less likely to be in need for finance. This is consistent with the results of Nguyen et al. (2021) on formal and informal credit and Barslund and Tarp (2008) on informal borrowing, but contradicts Chen and Chivakul (2008), who reported an insignificant effect of home ownership on formal credit market participation. Similarly, ownership of agricultural land is negatively associated with the demand for formal loans, informal borrowing, and RoSCA participation, even though agricultural land is commonly used as collateral by banks in rural Egypt (Mohieldin and Wright 2000). Several reasons could explain this result. First, similar to home ownership, rural households who own agricultural land are probably in a good financial position, and accordingly less likely to be in need for finance. Second, it is important to note that we observe whether a household member owns agricultural land, which implies that those who do not own agricultural land could still be renting it from others and using its productive claim as collateral for formal loans. In the Egyptian rural context, renting land from others is an important determinant of whether or not an individual can apply for formal loans (Mohieldin and Wright 2000).

The results on the significance of health status and emergencies show the more significant role of informal borrowing compared to formal loans in addressing rural individuals’ urgent liquidity needs. In this regard, rural individuals who have been subject to an emergency are found to be 4.3% more likely to borrow informally compared to a 1.4% higher probability of applying for formal loans. The results also confirm that the need for liquidity to cope with an emergency or cover a shortage in expenses is a major financial need for rural individuals and that informal borrowing is their first preferred choice. This aligns with the existing literature that shows that households who experience shocks are more likely to demand formal and informal credit (Kofarmata et al. 2016; Dharmadasa and Gunatilake 2023). In a similar context, rural individuals who have chronic illness or disability are more than twice as likely to engage in informal borrowing or join RoSCAs compared to applying for formal loans. This aligns with the findings of Chen and Chivakul (2008), who found that economically inactive individuals, including people with disabilities, housewives, and students, are less likely to participate in the formal credit market compared to employed individuals. A notable observation is that rural individuals with illnesses or disabilities are more likely to participate in RoSCAs as a form of precautionary savings due to uncertainties about their future earning capacity. This tendency is reinforced by their lower self-efficacy and confidence in managing future financial situations, prompting them to take precautionary measures. Additionally, health status is a significant factor in the rural context, as 30% of the rural sample are irregular laborers in the informal sector whose income largely depends on their ability to work.

Unexpectedly, health insurance coverage increases the likelihood of applying for formal loans and joining RoSCAs but reduces the likelihood of informal borrowing. This indicates that coverage with health insurance might not necessarily reduce the need for finance to cover health shocks. The varying effects of insurance coverage on formal loans and RoSCAs compared to informal borrowing could potentially be explained by the fact that rural individuals who have access to health insurance are likely to be permanent laborers in stable jobs, and thus more likely to apply for formal loans and join RoSCAs, as previously shown.

As for perceived general self-efficacy, rural individuals with higher levels of self-efficacy are significantly more likely to apply for formal loans and borrow informally. However, a positive but statistically insignificant relationship is found to exist with participation in RoSCAs. Examining this result in light of the available literature highlights the complexity of understanding the effect of self-efficacy on the demand of rural individuals for formal and informal financial services. On one hand, our results align with the literature, suggesting that individuals with higher levels of self-efficacy, due to their confidence in managing financial situations and navigating the financial landscape effectively, are more likely to adopt risk-taking behaviors and seek investment products and loans over risk-free savings accounts (Lown 2011; Lown et al. 2015; Pavani and Alagwadi 2023; Liu and Zhang 2021; Farrell et al. 2016). However, while our findings are consistent with the literature regarding loan usage, further investigation is needed to confirm whether this extends to the demand for investment products.

On the other hand, an alternative interpretation, supported by empirical evidence, is that individuals with higher levels of self-efficacy generally exhibit more prudent financial behavior and are less likely to demand debt related products (Lown et al. 2015; Farrell et al. 2016). Although this interpretation contradicts with our overall results, it agrees with the results obtained for the rural poor, less educated, and individuals with chronic illness or disabilities, who are found to have lower levels of self-efficacy and higher demand for formal loans and informal borrowing. A potential explanation is that the rural poor often feel a lack of control over their financial situation and face significant economic pressures. This leads them to seek formal loans as a last resort when other informal means are exhausted. Consequently, they take formal loans out of necessity rather than confidence in their ability to manage the borrowed funds effectively. Additionally, it is worth noting that the rural poor and less educated individuals tend to have lower levels of financial literacy. This lack of understanding drives them to depend on formal loans without fully comprehending the terms and conditions, particularly the compounded interest rates, in the hope that these loans will address their financial issues.

4.2. The Relationship Between the Demand for Formal Loans, Informal Borrowing, and Use of RoSCAs

The relationship between informal borrowing and the use of RoSCAs and the demand for formal loans is examined in models (2, 3, 4, and 5) as shown in Table 7 and Table 8. Based on the results, a significant positive relationship exists between the demand for formal loans and informal borrowing and RoSCA participation in rural Egypt. Rural individuals who apply for formal loans are 12.5% and 10.3% more likely to borrow informally and join RoSCAs, respectively. This underscores the significant role of the informal financial sector in meeting rural households’ financial needs, with formal loans serving more as a complement rather than a substitute. This finding aligns with the results obtained by Mohieldin and Wright (2000) who concluded that the informal credit market in rural Egypt compensates for the credit constraints in the formal sector. Other studies (e.g., Abdel Zaher 2019; Wabwire 2019; Gyeltshen 2012) have also documented the duality of rural credit markets in other developing countries and emphasized the significant role of the informal financial sector in rural areas (Cull et al. 2019; Barslund and Tarp 2008). It is noteworthy that while our results align with Vandone (2009) in identifying a positive effect of the informal credit market on formal credit participation, we diverge in our interpretation, specifically regarding the creation of incentives for strategic insolvency.

5. Conclusions

This paper examined the determinants of demand for formal loans, informal borrowing, and participation in RoSCAs in rural Egypt, and explored their relationship. Additionally, it investigated the effect of perceived general self-efficacy as a key behavioral factor influencing financial management behavior. The paper concludes that financial vulnerability tends to be the underlying driver of demand for formal and informal credit. Contrary to the intertemporal choice models, our findings reveal that the elderly, less educated, and poorer individuals are more likely to seek credit. Although income uncertainty, as indicated by employment instability, demonstrates a negative relationship with credit demand, consistent with theoretical expectations, the presence of chronic illness and disability positively influences the demand for both formal and informal credit, despite limited future income prospects. This underscores the critical effect of financial vulnerability in shaping rural households borrowing behavior. Our findings also reveal a positive relationship between self-efficacy and formal and informal credit market participation.

In light of the results, several conclusions are worth highlighting. First, the relationship between formal and informal means of finance indicates that they complement rather than substitute each other. The higher activity in the informal sector suggests that informal financial means are the first preference for rural individuals. This highlights the need for further research to explore the reasons behind rural households’ preference for informal means of finance and investigate how formal financial services can be designed to better suit the rural context. Second, understanding the debt behavior of the rural poor is crucial, as they are the most likely to apply for formal loans despite their low levels of self-efficacy, which suggests potential difficulties in effectively managing their debt. With current government efforts to enhance access to finance, and given that the poorest 20% of the rural population are the most likely to borrow formally, credit expansion must be approached cautiously. This is to ensure that increased access to formal credit helps the rural poor improve their economic well-being rather than pushing them toward a debt trap.

Finally, given the potential influence of self-efficacy on behavior change, targeted interventions could be designed to enhance individuals’ self-efficacy, thereby improving financial management practices, promoting better saving behavior, and facilitating more effective use of credit. Additionally, enhancing rural individuals’ financial literacy and self-efficacy, could boost their confidence in using financial knowledge and increase their participation in the formal financial sector. This is an area that merits further investigation in future research. Additionally, incorporating self-efficacy alongside financial characteristics appears to improve the categorization of debtors according to their debt behavior. This could be a potential area for investigation as it might help financial institutions better assess the expected behavior of credit applicants. It is also important to acknowledge the limitations of this paper, which should be considered in future research. Since the analysis is constrained by the data collected in the ELMPS in 2018, supported by the existing literature, the paper had to use the perceived general self-efficacy scale. We recognize that using a domain-specific measure focusing on financial self-efficacy would likely yield more accurate results.