1. Introduction

The data analyzed in this research is from 2021, a year marked by the uncertainty caused by the Covid-19 pandemic. The Covid-19 pandemic has severely affected economies, and as a consequence, many households around the world faced severe financial challenges (^{Szustak et al., 2021}; ^{Cziriak, 2022}). In the face of an immediate drop in household income, the financial fragility of families and the difficulties they have in facing an economic emergency are revealed (^{Demertzis et al., 2020}; ^{Cantor & Landry, 2020}; ^{Clark et al., 2021}; ^{Despard, et al., 2020}). According to the Mexican Central Bank the acute turmoil generated by the pandemic stands among the four largest episodes of financial distress experienced by the country (^{Carrillo and Garcia, 2021}). This situation highlights the importance of households to be trained to manage their finances and carry out an adequate financial planning to avoid a crisis (^{Fox & Bartholomae, 2020}; ^{Babiarz and Robb, 2014}; ^{Beshears, et al., 2020}).

While financial literacy literature emphasizes the importance of personal savings (^{Nguyen, et al, 2017}), there is international evidence that many households have insufficient savings in the face of income losses, spending crises, and other financial emergencies (^{Preston, 2022}; ^{Despard, et al., 2020}; ^{Morduch & Schneider, 2017}).

Financial literacy, savings and spending decisions can impact people’s financial situation in the short and long term (^{Lusardi et al., 2021}; ^{OECD, 2020a}). According to Babiarz and Robb (²⁰¹⁴), the low household savings rate has effects not only in the short term, due to households’ inability to face an emergency, but also due to its potential effect of financial insecurity in the long term. Thus, in a situation of income loss, households are more likely to report that they cannot face an unexpected expense and are more likely to resort to both formal and informal sources of credit (^{Cziriak, 2022}; ^{CNBV, 2022}; ^{Bratberg & Monstad, 2015}). Therefore, when facing an emergency situation, adequate planning is of great importance, primarily in low-income households, who have less access to traditional credit and whose tighter budget makes it more difficult to save (^{Nguyen, et al., 2017}; ^{Collins and Gjertson, 2013}).

However, most households do not plan for foreseeable events such as retirement (^{Harahap, et al., 2022}; ^{Ghadwan, et al, 2022}; ^{Lusardi and Mitchell, 2011}) and do not anticipate unexpected events and emergencies (^{Rothwell et al., 2022}), leaving them exposed to financial shocks. Likewise, most households have few or no assets and do not have emergency funds, making them very vulnerable to changes in their income (^{Demertzis et al., 2020}; ^{Hasler and Lusardi, 2019}; ^{Lusardi, 2011}). Not having a savings fund for emergencies is the result of poor planning or a forecast error in households (^{Babiarz and Robb, 2014}). Financial education, which is associated with less financial fragility, can mitigate the negative consequences of income loss by increasing people's capacities to face emergency expenses (^{Cziriak, 2022}; ^{Clark et al., 2021}; ^{Hasler and Lusardi, 2019}; ^{Sherraden, 2010}).

Based on the above, the following questions are formulated: How does financial literacy influence households´ financial fragility? What is the effect that assets owning, savings, financial inclusion and financial education have on financial fragility? What are the sociodemographic variables related to financial fragility?

The objective of this research is to analyze the relationship between Mexicans’ financial fragility and the following variables: financial literacy, savings, assets owning, financial shocks, financial education, and financial inclusion. Likewise, the socioeconomic and demographic determinants of the population’s financial fragility are identified. The study is organized as follows: Section 2 presents the literature review, Section 3 the research methodology and the model description. In Section 4 the results are presented. The discussion of the results is found in Section 5 followed by the conclusions and implications in Section 6.

2. Literature review

Financial fragility and its relationship with financial literacy in young and adult populations is a topic of great interest in the scientific literature due to its impact on individual well-being (^{Baker et al., 2023}; ^{Chhatwani & Mishra, 2021}; ^{Clark et al., 2021}; ^{Nikolaos & Christos, 2020}; ^{Ramli et al., 2022}). Empirical evidence determines a strong negative and statistically significant relationship between financial literacy and financial fragility (^{Cziriak 2022}; ^{Lusardi et al., 2021}; ^{Nikolaos & Christos, 2020}; ^{Lusardi et al., 2011}). According to INEGI (^2022b), financial fragility is defined as the maximum period during in which people aged 18 and over could cover their expenses with their savings, in the event of an unforeseen income loss.

Planning and savings availability is important to deal with any income drop. The results of the International Financial Literacy Survey (INFE) indicate that less than a third (28%) of the adults in the sample have savings to cover their expenses for a week if they lose their main source of income (^{OECD, 2020a}). In developing economies, a common alternative to saving in a financial institution is informal saving (^{Demirguc-Kunt et al., 2017}). Roa and Villegas (²⁰²²), CNBV (²⁰²²) and Lusardi et al. (²⁰¹¹) identified that to face changes in their income or an overspending, people resort to various forms of informal savings. There is also evidence that young people prefer to save informally, such as giving money to a trusted person or family member, keeping money at home, participating in savings groups, or in-kind (^{OECD, 2020b}).

Hilgert, Hogarth, and Beverly (²⁰⁰³) identified a positive relationship between financial knowledge and behavior towards saving. Their results show that the households with the lowest scores on the general measure of financial knowledge are those with low scores on the savings index, as well as the lowest scores on the savings sub-index (savings account, emergency savings, long-term savings). Bhutta et al., (²⁰²¹) identify that financial literacy is an element that helps explain the preparation of families to face unexpected expenses or their income interruption. In their estimation, they find that high levels of financial literacy are positively correlated with having liquid savings.

According to the OECD (^2020a), financial education aims to make people better prepared to manage their money and thereby achieve financial well-being. Wagner (²⁰¹⁹) identified that financial education is positively related to financial literacy. In this regard, Cardona-Montoya et al., (²⁰²²) identified that households with more financial knowledge are better prepared to face adverse events, thereby reducing the probability of financial fragility in the household. Lusardi et al., (²⁰¹¹) find that those who received financial education at school are more likely to cover expenses in an economic emergency, compared to their counterparts. In the results of Anderson, et al., (²⁰¹⁷), answering an additional financial literacy question correctly increases the probability of answering affirmatively to the precautionary savings-holding question by about 8%.

Sherraden's (²⁰¹⁰) conceptual model of financial capabilities states that good financial behavior and the ability to cover expenses with savings in the face of an income loss occurs when individuals have financial knowledge, skills, and access to financial products. Friedline and West (²⁰¹⁶) identify that among young adults, access to financial products (checking account, credit card, bank account and savings account), and having received formal financial education are factors related to the ability to face an economic emergency. Being financially capable was associated with a 176% increase in the probability of meeting expenses in a financial emergency. Being financially included was associated with a 123% increase, compared to their counterparts, and being financially educated was associated with a 40% increase in the probability of meeting their expenses in an economic emergency.

Various authors have identified factors related to financial fragility. Table 1 presents the variables, their relationship (positive or negative) with financial fragility, and their references. Women are more likely to be financially frail than men (^{Cziriak, 2022}; ^{Lusardi et al., 2011}; ^{Hasler et al., 2018}). Financial fragility is more accentuated in young people and in older age groups (^{Cziriak, 2022}; ^{Lusardi et al., 2011}). People with a higher level of education are less likely to be financially fragile (^{Roa and Villegas, 2022}; ^{Hasler et al., 2018}; ^{Cziriak, 2022}; ^{Friedline and West, 2016}; ^{Lusardi et al., 2011}). Significant differences were also found among people in different regions. These were about their ability to handle unexpected spending (^{Lusardi et al., 2011}; ^{Friedline and West, 2016}).

Married people are more capable of coping with a situation of financial fragility, compared to people in a different marital status (^{Cziriak, 2022}; ^{Lusardi et al., 2011}; ^{Friedline and West, 2016}). Those who work full-time or part-time, compared to those who are unemployed, are more likely to cover expenses in an economic emergency (^{Babiarz and Robb, 2014}; ^{Friedline and West, 2016}; ^{Cziriak, 2022}). People with higher income are more likely to cover their expenses in an economic emergency. People with higher levels of financial literacy are less likely to be financially fragile (^{Cziriak 2022}; ^{Lusardi et al., 2021}; ^{Nikolaos & Christos, 2020}; ^{Lusardi et al., 2011}). People who have access to financial services are better prepared to cope with an income crisis (^{Demirgüç-Kunt et al., 2022}; ^{Demirguc-Kunt et al., 2017}; ^{Ramli et al., 2022}; ^{Pomeranz and Kast, 2022}).

People who have received training on how to save, budget or use credit responsibly are more likely to face an economic emergency (^{Cardona-Montoya et al., 2022}; ^{Cziriak, 2022}; ^{Clark et al., 2021}; ^{Lusardi et al., 2011}). People who have formal savings or informal contingency and emergency savings mechanisms are less financially fragile (^{Roa and Villegas, 2022}; ^{Lusardi et al., 2011}). Homeowners may be able to leverage this asset to establish financial stability (^{Cziriak, 2022}; ^{Friedline and West, 2016}). People who experienced a decrease in their income are more vulnerable to economic emergencies (^{Cziriak, 2022}; ^{Lusardi et al., 2011}; ^{Ramli et al., 2022}).

Consistent with Sherraden’s (²⁰¹⁰) model that suggest a relationship between financial literacy, savings, financial inclusion, and financial education with financial fragility, the following hypotheses are proposed:

3. Methodology

The type of study is non-experimental and cross-sectional, descriptive and correlational. The data used corresponds to the Mexican National Survey of Financial Inclusion 2021 (^{INEGI, 2022a}). In this survey, the population under study is made up of people in the age range of 18 to 70, who live in Mexico, in the national geographic coverage of urban-rural cut, divided into six geographical areas: Northwest, Northeast, West Bajío, Mexico City (CDMX), South and Central South and East. The total sample of 15,291 people was obtained between June 28 and August 13, 2021 through personal interviews at the respondents’ home.

The data for this research correspond to 7,251 people who declared receive income from work, as suggested by Hasler et al. (²⁰¹⁸) and Cziriak (²⁰²²). They are between the ages of 18 and 65 and were not pensioners at the time of the survey. For the analysis, variables whose data were reported by the national survey and whose relationship with financial fragility was sufficiently supported in the scientific literature were considered. To measure the main variable of this research, financial fragility, question 4.10 of the survey is used: "If you stopped receiving an income, how long would you be able to cover your expenses with your savings? The response options are: a) less than a week/no savings, b) at least a week, but less than a month, c) at least a month, but less than three months, d) at least three months, but less than six months, e) six months or more. The responses "no answer" and "don't know" were not considered for the analysis.

For the statistical strategy, a dichotomous variable is designed, as in Cziriak (²⁰²²) and Lusardi et al., (²⁰¹¹), which takes the value of 1 for responses a), b), c), or d) and a value of 0 for answer e). The value of 1 indicates that the person is “financially fragile” and 0 indicates that the person is “not financially fragile”. From this, the dichotomous Probit model is built.

The concept of financial literacy used in this research is the one proposed by Lusardi (²⁰¹⁹), who defines it as the abilities a person has to apply three fundamental concepts in the decision-making process of savings and investment. In its operational form, three questions are used to measure financial literacy i) the calculation of interest and understanding the capitalization of interest; ii) inflation and its effect on purchasing power and iii) the benefit of risk diversification (^{Lusardi, 2019}; ^{Cziriak 2022}). The first two measure the individual's ability to perform a calculation and the third assesses whether the respondent knows the relationship between diversification and risk.

To measure the financial literacy of the respondent, three questions from the survey are used (question 4.7.3, savings diversification; question 13.3, compound interest and question 13.4, inflation knowledge). For each question, a dichotomous variable is designed. The value of 1 is assigned if the respondent answers correctly, and 0 otherwise. From this, an indicator of financial literacy determined as the sum of correct answers is designed, whose range goes from 0 to 3, (^{Lusardi and Mitchell, 2011}; ^{Cziriak 2022}). To determine the impact of financial literacy on financial fragility, the dichotomous Probit regression model is used.

Financial inclusion is operationalized according to the standard first-level measure, indicated by having an account in a formal financial institution (^{Grohmann and Menkhoff, 2020}), with indicators of having a savings account and a sophisticated account (^{Lusardi et al., 2010}). For the financial education proxy, question 4.5 is used: Have you taken a course on how to save, how to make a budget or how to use credit responsibly? Regarding informal savings, the possession of any of the forms of informal savings indicated in question 5.1 of the survey is considered. The proxy variable of financial shocks is measured from an overspending in relation to the income received each month, indicated in question 4.3. For the proxy of assets possession, owning a home, car or land are considered (^{Friedline and West, 2016}). Sociodemographic characteristics of the respondent are included: gender, age, educational level, marital status, work income, region of residence, town size, job position. Table 2 presents the operationalization of the variables, their codification and their references.

4. Results

In the total sample, the majority are men (58.20%), 73.74% are under 47 years old, 48.15% have secondary school as the highest level of education, 34.34% are married, 26.55% obtain an income of less than 233.37 US dollar per month (corresponding to the first quartile of income), 66.90% reside in an urban area, 67.33% are employ and 24.37% are self-employed. Regarding the financial information of the respondents, 21% have a savings account in a bank or a financial institution, 62.1% save informally, less than 3% have a sophisticated account, 9.54% declare that they have received financial education, 55.22% declare that they are asset owners and 48.64% declare having had an overspending experience.

Table 3 presents the numerical characteristics of the sample, divided by groups regarding the condition of financial fragility and by demographic and economic characteristics. Of the total sample, 88.33% is financially fragile. This percentage is made up of those who, in the event of losing their main source of income, cannot cover their expenses with their savings for six months. Those who could cover their expenses with their savings from six months or more, comprise only 11.67%. More women than men (29.46% and 26.35%, respectively) declared that they have no savings or that if they stopped receiving income they would not be able to cover their expenses for a week. In a greater proportion, people in the 48-67 age range, and those living separated, divorced or widowed are considered financially fragile. Respondents with the highest educational level (20.06% with master and 28.42% with doctorate) state that they could cover their expenses with their savings for six months or more, while 42.80% and 34.03%, corresponding to people with primary and secondary studies, could cover their expenses less than a week.

Of those who do not have assets, 30.95% could cover their expenses with their savings for less than a week, compared to 24.98% of those who do have assets. The same situation for the 31.13% who reside in a rural location, higher than the 25.93 that live in a city, but the percentage of those who are not financially fragile is higher among those who live in a city (12.33% vs 10.33% living in rural area). Regarding the geographic region and job position, the Northwest has the largest number of financially very fragile people (34.39%) as well as the day laborer position (48.29%). Quartile 1 and quartile 2, are considered financially very fragile (40.31% and 30.63% respectively), compared to 14.04% of quartile 4.

Table 4 presents the numerical characteristics of the sample, divided by groups regarding the condition of financial fragility and financial literacy, financial education and financial condition. The 21.55% of the respondents answered the three financial literacy questions correctly, 69.80% understand the concept of diversification, 38.11% can calculate compound interest and 76.52% understand the effect of inflation on purchasing power. The 47.17% of the respondents who obtained zero correct answers, those who do not have a formal or informal savings account (31.48% and 47.01%, respectively), the 29.24% who lack financial education and the 37.74% who indicated having an overspending, can´t cover their expenses with their savings for more than a week.

On the opposite, respondents who can cover their expenses for six months or more are those who answered the three questions correctly (15.43 vs. 4.85% who can cover their expenses for this period and didn´t answer any correctly). There are also those who save; those who have not had an overspending (16.78 vs. 6.27%) and those who have received financial education (25.58 vs. 10.20%).

Table 5 presents the estimation of five Probit models, which include the sociodemographic variables indicated in table 3, the financial literacy, financial inclusion, financial education and savings holdings indicators. Model 1 includes the financial literacy indicator. From the results, it is identified that the coefficient of the financial literacy indicator has a negative and statistically significant effect (β=-0.1009, p<0.01) on the probability of being financially fragile, after incorporating the sociodemographic variables, which supports the Hypothesis 1 of this research. Model 2 includes financial inclusion indicators. Having a savings account has a negative and significant effect (β=-0.3865, p<0.01) on the probability of being financially fragile, which supports hypothesis 2 of this research, considering the definition of financial inclusion in its most basic form.

Model 5 includes the financial education indicator. Financial education has a negative and significant effect (β=-0.2991, p<0.01) on the probability of being financially fragile, which supports hypothesis 3 of this research. Model 3 includes the informal savings indicator. Informal savings have a negative and significant effect (β=-0.4202, p<0.01) on the probability of being financially fragile, which supports hypothesis 4 of this research. Model 4 includes the overspending and assets owning variable. Overspending has a positive and significant effect (β=0.4373, p<0.01) on the probability of being financially fragile; asset ownership has a negative and significant effect (β=-0.2029, p<0.01). In this model, the gender variable (male) has a positive and significant effect (β=0.0834, p<0.10).

For the marginal effect analysis, the results of model 6 are used, which includes all the study variables. Men are 1.02% more likely to be financially fragile, compared to women. The probability of being financially fragile decreases with a higher educational level (bachelor's, master's or doctorate), in 4.79% and 7.34% compared respectively, compared to those who have basic education. According to their marital status, those who are single are less likely to be financially fragile (2.30%), compared to those who are married. Those who reside in an urban location, compared to those who reside in a rural location, have a higher probability (2.0%) of being financially fragile. Those who reside in the Northeast and West region are less likely to be financially vulnerable compared to the Mexico City region. As for their job position, those who are day laborers are more likely (4.19%) to be financially fragile, compared to those who are employees, while those with an employer job position are less likely (4.74%). With a higher income, people are less likely to be financially fragile. Those who receive an income corresponding to quartile 4 are less likely (9.40%) to be financially fragile, compared to those in quartile 1.

Respondents who answered all questions correctly are less likely to be financially fragile. For each question with a correct answer, the probability of being financially fragile decreases by 1.12%. Having a savings account decreases the probability of being financially fragile by 5.34%, and having informal savings decreases the probability by 5.27%. Those who had an overspending in the previous year are more likely (6.38%) to be financially vulnerable than their counterparts. The possession of assets decreases the probability (2.21%) of being financially fragile, compared to those who do not have assets. Finally, those who have taken a course in financial education are less likely (5.06%) to be financially fragile, compared to those who indicated that they have not taken any courses.

Table 6 shows the summary of hypotheses, models and research results.

5. Discussion

The empirical results of this research show the importance of financial literacy, financial inclusion, financial education and having savings to face financial fragility. According to the definition of financial fragility used, 88.33% of the sample is financially fragile and only 11.67% is not considered financially fragile. The result regarding the fragility group exceeds the percentage found by other studies (^{Arellano et al., 2020}; ^{OECD, 2020a}; ^{Hasler et al., 2018}; ^{Cziriak, 2022}; ^{Friedline and West, 2016}). In our results, 21% of those surveyed have a savings account at a bank or financial institution, similar to the 25% reported by Demirgüç-Kunt et al. (²⁰²²); 62.1% save informally, higher than the 23% reported by OECD (^2020a). The 9.54% declare to be financially educated, which differs from some abroad populations like the 21% reported by Wagner (²⁰¹⁹) and the 6% found by Friedline and West (²⁰¹⁶) and 48.64% declare having had an experience of overspending, similar to the 46% found by Hasler et al. (²⁰¹⁸).

The Probit model results provide evidence that financial literacy has a negative and significant effect on the probability of being financially fragile. This result is consistent with Cziriak (²⁰²²), Lusardi et al. (²⁰²¹), Nikolaos & Christos (²⁰²⁰), Hasler et al. (²⁰¹⁸), Lusardi et al. (²⁰¹¹), and differs from Ramli et al. (²⁰²²) and Cardona-Montoya et al. (²⁰²²). The marginal effect of financial literacy on the probability of being financially fragile is 1.54%, while in Hasler et al. (²⁰¹⁸) the effect varies between 2% and 7% depending on the characteristics of the population. Likewise, it differs from the result obtained by Friedline and West (²⁰¹⁶), whose marginal effect is 176%.

The results of the model provide evidence that the holding of savings (formal and informal) is the second most robust predictor. Having formal savings decreases the probability (5.34%) of being financially fragile. The direction of the effect coincides with Roa and Villegas (²⁰²²) and Friedline and West (²⁰¹⁶), but in the latter, the effect is 80.4%, while in Ramli et al. (²⁰²²) the relationship between both variables is very low. In the case of informal savings, the marginal effect is 5.27%, whose direction of effect (negative) coincides with Roa and Villegas (²⁰²²).

The results of this research also show a negative and significant relationship between financial education and financial fragility (with a marginal effect of 5.06%). The direction of the effect coincides with Cardona-Montoya et al. (²⁰²²), while it differs, in the magnitude of the effect, with Lusardi et al. (²⁰¹¹) and Friedline and West (²⁰¹⁶), with 10.2% and 33.9%, respectively. The possession of assets has a negative and significant effect on the probability of being financially fragile, whose marginal effect is 2.21%. This result coincides in the direction of the effect with Friedline and West (²⁰¹⁶) and Cziriak (²⁰²²), but differs in the magnitude of the marginal effect of 56.8% and 13%, respectively.

From the Probit model, it is evident that those who indicated having had a negative financial shock on income during the last year, are more likely to be very financially fragile (with a marginal effect of 6.38%); whose direction of effect coincides with Ramli et al. (²⁰²²), Cziriak (²⁰²²) and Hasler et al. (²⁰¹⁸). In the latter, the magnitude of the effect varies between 4% and 10% depending on the characteristics of the population.

Regarding sociodemographic characteristics, in our model, there is no significant difference between men and women, regarding financial fragility. This result differs from other researches like Cziriak (²⁰²²); Lusardi et al. (²⁰¹¹) and Hasler et al., (²⁰¹⁸). Likewise, it is more likely that people are very financially fragile according to their sociodemographic characteristic: lower income (first quartile), low educational level (primary), according to their marital status (single) and according to their employment position (day laborers). In general terms, these results are consistent with Cziriak (²⁰²²), Hasler et al. (²⁰¹⁸), Roa and Villegas (²⁰²²), Friedline and West (²⁰¹⁶) and Lusardi et al. (²⁰¹¹).

6. Conclusion

This research adds to the growing literature that demonstrates how financial literacy, financial education, financial inclusion and savings holdings (formal and informal), are variables significantly related to the possibility that a person who stopped receiving income could be able to cover their expenses for a period of at least six months.

The results of this research are consistent with those found by previous studies and highlight the importance that it must have for the state to design strategies that contribute to promoting financial education and financial inclusion. To achieve this goal, the evaluation of both variables is a fundamental first step. The national survey collects information that allows us to broadly understand the level of financial inclusion of Mexicans. However, in terms of financial education, the survey only ask if the participants have taken any course on budgeting, savings or credit. Having more information about the financial topics that Mexicans have studied and about the teaching-learning strategy would allow their effectiveness to be measured and would support the design of a national financial training strategy.

From the statistical estimates, a high percentage of Mexicans (88.33%) is in an evident financial fragility condition. Around two thirds (62.16%) save informally, in contrast to 21.46% of people who have a savings account at a bank or a financial institution. Likewise, one out of ten people in the sample declared to be financially educated and one out of five people answered the three financial literacy questions correctly, which indicates low financial literacy in Mexico. Coinciding with international results, the educational level is a significant variable related to the habit of saving and in this sense fundamental to face a financial emergency. Almost half of the sample analyzed in this study has secondary school as their maximum educational level. Therefore, it is recommended to include in school curricula, no later than secondary school, training in personal finance and the development of mathematical skills that increase people's analytical capacity for greater financial decision-making.

Sociodemographic variables are also related to the probability of being very financially fragile. It is more likely that people, male, with lower income (first quartile), low educational level (primary), according to their marital status (single) and according to their employment position (day laborers) are very financially fragile. Considering these determinants in the design of the educational strategy can help focus effort on vulnerable populations, who need to resolve the short term and at the same time include uncertainty and the long term in their plans.