Ekonomiaz N.º 100, 2º semestre, 2021 Financial vulnerability of households: evidence for the Basque Country and the rest of Spain One of the consequences of the COVID-19 pandemic has been the loss of the main source of income for many households due to the lockdown of the workforce and the suspension of part of the economic activity. The aim of this paper is to analyze the importance of fi- nancial literacy and personal experiences with the financial system on financial vulnerabil- ity. We show evidence for households in the Basque Country and in comparison with the rest of Spain. Our research shows that controlling for multiple socio-demographic factors, households with lower financial literacy and who had some disagreement with the finan- cial system are more likely to be very vulnerable. These effects are larger in the Basque Country than in the rest of Spain. Una de las consecuencias de la pandemia de la COVID-19 ha supuesto la pérdida de la princi- pal fuente de ingresos de muchos trabajadores generando graves consecuencias financieras en sus hogares. El objetivo de este trabajo consiste en analizar la importancia de la educación fi- nanciera y las experiencias personales con el sistema financiero sobre la vulnerabilidad financi- era en los hogares del País Vasco y si existen diferencias con respecto al resto de España. Nues- tra investigación muestra que, controlando por múltiples factores sociodemográficos, los hogares con menores conocimientos financieros y que han tenido algún desacuerdo con el siste- ma financiero tienen más probabilidad de ser muy vulnerables, y estos efectos son mayores en el País Vasco frente al resto de España. COVID-19ren pandemiaren ondorioetako bat langile askoren diru-sarreren iturri nagusia galtzea izan da, eta ondorio finantziero larriak eragin ditu haien etxeetan. Lan honen helburua honako hau da: finantza-hezkuntzak eta finantza-sistemarekin izandako esperientzia pertsonalek Euskadiko etxekoen unitateen finantza-kalteberatasunari buruz duten garrantzia aztertzea, eta ea alderik ote dagoen Espainiako gainerako lurraldeekin alderatuta. Gure ikerketak erakusten duenez, faktore soziodemografiko askoren ondorioz, finantza-ezagutza txikiagoa duten eta finantza-sistemarekin desadostasunen bat izan duten etxekoen unitateek oso kalteberak izateko probabilitate handiagoa dute, eta ondorio horiek handiagoak dira Euskadin Espainiako gainerako lekuetan baino. * Spanish versión available at https://euskadi.eus/ekonomiaz 95 Ekonomiaz N.º 100, 2º semestre, 2021 Alfonso Arellano BBVA Research and Complutense University of Madrid Noelia Cámara BBVA Research Table of contents 1. Introduction 2. The Financial Capability Survey and financial literacy measurement 3. Does financial literacy and personal experiences affect financial vulnerability? 4. Conclusions and policy recommendations Bibliographic references Appendix Keywords: financial vulnerability, financial literacy, experience effect, households. Palabras claves: vulnerabilidad financiera, educación financiera, experiencias personales, hogares. JEL codes: G51, G53 Entry Date: 2021/06/11 Acceptance Date: 2021/10/11 The opinions expressed and arguments employed are those of the authors and should not be reported as representing the official views of BBVA. The authors also thank the editor and two anonymous review- ers for their constructive comments and insights. 1. INTRODUCTION Starting in the first quarter of 2020, the pandemic of COVID-19 and the subse- quent crisis generated severe economic consequences in Spain. The country experi- enced one of the largest GDP contractions among major advanced economies as well as an important increase in an already high unemployment rate.1 Economic literature points out that this crisis is creating a similar employment destruction pattern to pre- vious ones affecting more to those «classic» vulnerable groups in the Spanish labor market: younger, female, less educated, and temporary workers (Dolado et al., 2020). As a result, financial conditions of many Spanish households are seriously compro- 1 Unemployment rate increased from 13.8 percent in the last quarter of 2019 to 16.1 percent in the last quarter of 2020 and GDP fell to 9.1 percent in that period (National Statistics Institute, INE). ALFONSO ARELLANO, NOELIA CÁMARA 96 Ekonomiaz N.º 100, 2º semestre, 2021 mised. In response to the pandemic consequences, the Spanish government and the European Authorities established policies that sought to mitigate the economic wreck- age. The goal of these interventions was to provide a safety shield to the most affected groups and limit deeper effects on inequality (Aspachs et al., 2020). Policies, such as the short-time work schemes, public loan guarantees for companies, forbearance peri- od on loans for mortgage holders and suspended evictions of renters, provided a buff- er to temporarily soften the effect of losses on household’s finances, limiting the eco- nomic consequences of the pandemic. Despite these measures, many Spanish households lost substantial income and wealth and this situation may worsen in the future once the policies are withdrawn. A relevant question is whether those more financially literate households made better savings and investment decisions in the past, so they can more easily with- stand economic shocks to cope with times of crisis, or whether financial vulnerabili- ty is only a matter of income. We study two channels that can influence financial decision making and its results in terms of financial vulnerability. First, decisions can be made through learning experience that can be approximated by financial lit- eracy. But financial decisions can also be made from personal experiences with the financial system. We analyse such relationships for households in the Basque Coun- try and the rest of Spain, using the microdata from the Survey of Financial Compe- tences (ECF onwards) elaborated by the Bank of Spain. One way to measure finan- cial vulnerability is through the time that households are able to subsist in the event of losing their main source of income (generally their salary) without this loss repre- senting a major shock to the household, such as a change of home, the rescue of a pension fund, or the sale of homes. The financial vulnerability distribution of households prior to the COVID-19 crisis (Table A1 in Appendix A) in the Basque Country was as follows: 20.4 percent of households reported being highly financially vulnerable (holds out for less than 3 months), 31.4 percent of households were clas- sified as financially vulnerable (3-12 months) and 48.2 percent of households can be considered financially secure (above 12 months). This distribution worsens off in the case of the rest of Spain, and the percentages are 33.8, 32.4 and 33.8 percent, re- spectively. Our research builds on different strands of the literature. We contribute to the literature that analyses the role of financial literacy on decision making and finan- cial outcomes (Van Rooij et al., 2011; Lusardi and Mitchell, 2014; Angrisani et al., 2020; Clark et al., 2020; among others). These works find that knowledge of basic fi- nancial concepts, such as the interest rate, compound interest rate, inflation, as well as mastering simple aspects of calculation, has a positive effect on decision making. Although it is difficult to establish a causal link between financial capabilities and fi- nancial behavior, both the analyses with instrumental variables and the experimen- tal approaches suggest that financial literacy plays a role in influencing financial de- cision making, and causality goes from knowledge to behavior. FINANCIAL VULNERABILITY OF HOUSEHOLDS: EVIDENCE FOR THE BASQUE COUNTRY AND THE REST OF SPAIN 97 Ekonomiaz N.º 100, 2º semestre, 2021 Our contribution focuses on the effect of financial literacy on household finan- cial vulnerability. We wonder whether respondents who were more financially liter- ate were better financially prepared and then better able to absorb financial setbacks associated with the pandemic. Additionally, we explore the effect of experiences on financial vulnerability. Do past personal experiences with the financial system shape current individ- uals’ saving and investment outcomes? Standard models in economics assume that individuals are endowed with stable saving propensity, unaltered by econom- ic experiences. Thus, these models based predictions of economic and financial decision-making on variables such as wealth, income, employment, education, and other demographics characteristics. In contrast, the psychology literature ar- gues that personal experiences, especially recent ones, exert a strong influence on personal decisions (Nisbett and Ross, 1980; Weber et al., 1993; Hertwig et al., 2004). A strand of the literature in economics suggests that the cultural and politi- cal environment in which individuals grow up affects their preference and belief formation, such as the level of trust in financial institutions, stock market partici- pation, and preferences over social policies (Guiso et al., 2004 and 2008; Osili and Paulson, 2008). Growing evidence in macrofinance suggests long-lasting effects of personally experienced outcomes on beliefs (Laudenbach et al., 2019; Malmendier et al., 2020). This literature points out that emotional tagging plays a crucial role in assigning weights in the belief formation process. Personal experience of economic boom and bust periods (i.e. big traumas) or other day-to-day financial experiences (i.e. small traumas) shape financial behavior and might alter risk attitudes and financial in- vestment decisions for decades to come. Thus, traditional financial models are miss- ing a key element, a way to capture that the personal experiences alter indivi-dual decision-making. Since we have cross-sectional data and cannot measure the impact of major eco- nomic and financial events, we try to capture the effect of day-to-day experiences through reported information on whether our reference person in the household had any disagreement with financial institutions. These considerations of the role of financial literacy and personal experiences in household vulnerability are key to understanding how households build resilience to deal with shocks of diverse nature and design policies based on mechanisms for improving household welfare. The rest of the paper is organized as follows. Section 2 describes the survey and explains the construction of financial literacy indicators and how to measure indi- vidual experiences with the financial system. Section 3 explores the effect of the aforementioned variables on the household’s financial vulnerability. Finally, Section 4 summarizes the findings and presents policy recommendations. ALFONSO ARELLANO, NOELIA CÁMARA 98 Ekonomiaz N.º 100, 2º semestre, 2021 2. THE FINANCIAL CAPABILITY SURVEY AND FINANCIAL LITERACY MEASUREMENT 2.1. The Financial Capability Survey We use data from the Survey of Financial Competencies (ECF), carried out by the Bank of Spain.2 This survey belongs to a joint initiative of the Bank of Spain and the National Securities Market Commission within the framework of the Financial Education Plan and is included in the National Statistical Plan. The ECF is part of an international project coordinated by the OECD’s Interna- tional Financial Education Network, which allows the comparison of the finan- cial skills of the Spanish population with those of a wide range of countries. It was carried out between the fourth quarter of 2016 and the second quarter of 2017 with the aim of measuring the financial competencies of the adult popula- tion in Spain. The ECF covers information about demographic and economic characteristics and focuses on the financial knowledge and ownership of finan- cial products among the population between 18 and 79 years old. By means of face to face personal interviews, information was collected both on the financial knowledge of the person chosen at random to respond to the interview, and on the knowledge available in his or her home.3 It contains 8,554 valid interviews and is representative at both country and region level (Bover et al., 2019). In or- der to achieve the objective of a representative sample, the ECF includes a varia- ble with weights for each interview obtained according to the National Statistics Institute (INE) guidelines. We use these weights to adjust the descriptive statis- tics and the results from econometric modelling. Data from ECF allow a rigorous and comprehensive description of the finan- cial knowledge in Spain, as well as the relationship between the financial system and its users. In this paper, we use variables that refer to both households and individuals. For most of the variables that refer to individuals, the ECF allows to identify two target individuals: the interviewed person and the person who knows the proper- ties, financial products and debts of the household as a whole, as well as the rea- sons for the decisions that have been made regarding expenses and investments (i.e. informed person). For this purpose, the ECF includes two blocks of identical questions if the interviewed person does not coincide with such an informed per- son. We work with the characteristics of the informed individuals. Although the ef- fect of distinguishing between the respondent and the informed individual is rela- tively small (the percentage of informed people who do not correspond to the respondent is 6.1 percent of the total sample), we make this decision under the as- 2 For further details, documents and presentations based on the ECF, including access to microdata, user guide and questionnaire, are available on the Bank of Spain (2016). 3 Face to face personal interviews avoid biases introduced by other methods commonly used. FINANCIAL VULNERABILITY OF HOUSEHOLDS: EVIDENCE FOR THE BASQUE COUNTRY AND THE REST OF SPAIN 99 Ekonomiaz N.º 100, 2º semestre, 2021 sumption that a person who is not familiar with the aforementioned issues may not have an impact on financial decision making and therefore, her influence on the household’s financial status would be limited. Hereinafter, for the sake of simplici- ty, we will refer to the informed individual as the individual. The household characteristics focus on its composition and size, ownership of the main dwelling and other similar illiquid assets, the level of annual gross income and planning for expenses. The characteristics of individuals focus on socio-demo- graphic factors, such as gender, age, level of education, labor market situation, as well as financial knowledge and financial experiences.4 Comparing the descriptive statistics of the most important socio-demographic characteristics between the Basque Country and the rest of Spain (Table A1), it is possible to appreciate some groups that also appear when dealing with social ine- quality (e.g., Dolado et al., 2020; Aspachs et al., 2020). Gender issues are usually concerning. As the level of financial vulnerability in the household decreases, the proportion of women decreases in the two territories. For the most financially vulnerable households, individuals are more concen- trated among younger adults, both in the Basque Country and in the rest of Spain. However, for the most financially resilient households, the proportion of older in- dividuals is higher, and even more in the Basque Country than in the rest of Spain. This result is partly conditioned by the differences in the age-gender-pyramid of the two territories considered (INE, 2021). We incorporate education level among socio-demographic covariates to con- trol for the potential relationship between financial literacy and education (Behr- man et al., 2012). The level of education presents a negative relationship with the degree of financial vulnerability of the household, both in the Basque Country and the rest of Spain. The composition of the highly vulnerable households stands out, which is concentrated to a greater extent in primary education for the rest of Spain, while the share of secondary education increases for the same type of households in the Basque Country. The employment situation of the individual may be also connected to the level of financial vulnerability of the household. Employment-related status (employee and self-employed) are overrepresented in households with lower financial vulner- ability (to a lesser extent in the latter group). On the opposite side, we find unem- ployment and individuals not in the labor force except for retirement. These results are common to both of the territories considered. The case of retirement is special, because it shows a very similar behavior to that of employment-related status in the Basque Country, as opposed to the rest of Spain, where the result is less clear. This finding is compatible with the amount of pensions by region, which is the 4 The financial experience is the unique variable affecting only the interviewed person, instead of the informed person. This experience can be interpreted as a proxy of the variable of interest. ALFONSO ARELLANO, NOELIA CÁMARA 100 Ekonomiaz N.º 100, 2º semestre, 2021 highest in the Basque Country compared to the Spanish regions (1,231.5 euros/ month in 2019, MTES, 2021). Regarding household structure, single-person households and those without a partner and with children below 18 in the household have a higher proportion among the most financially vulnerable households. The same result occurs for those households without a partner who live with other adults. The numbers are higher in the Basque Country than in the rest of Spain. The home ownership status and other real estate properties shows an enriched vision of the household’s financial situation. These variables provide information from the past long-run financial behavior of the household that complement the short-run nature of the financial vulnerability concept, based on liquid assets. Non-financial assets, such as the ownership of the main dwelling (regardless of how it was acquired), other houses, land and garages, are inversely related to the level of financial vulnerability of the household, both in the Basque Country and in the rest of Spain. The level of gross annual household income is inversely related to the most fi- nancially vulnerable status. This conclusion can be seen in both territories. We include a dummy variable indicating whether the household plans for its expenses. Those households with expenditure planning in the Basque Country are overrepresented in the most vulnerable group and underrepresented in the greater financial resilience group. In contrast, the proportion of the households that plan for expenses in the rest of Spain is higher than those which do not, regardless of the vulnerability level, and decreases gradually as financial resilience increases. 2.2. Measuring financial literacy and experiences Financial literacy refers to the knowledge of some financial concepts that help to make informed financial decisions. To measure the financial literacy in a household, we take as a reference person the one who is familiar with the household finances. We consider two levels of financial literacy. Basic level is related to savings and ad- vanced level is related to investment. Basic financial literacy considers the consumer in his role as a saver and consists of questions based on interest rate, a basic concept of risk and inflation related to financial transactions and day-to-day decision-mak- ing. These questions are in essence the basic financial literacy questions proposed by Lusardi and Mitchell (2008, 2011a, 2011b, 2014), excluding the risk diversification question. In addition to this basic financial knowledge, it is interesting to know if there is an optimal degree of financial literacy that allows the consumer to move forward and also acquire the role of investor. To this end, we consider the role played by more complex financial questions referring to the profitability of investment funds FINANCIAL VULNERABILITY OF HOUSEHOLDS: EVIDENCE FOR THE BASQUE COUNTRY AND THE REST OF SPAIN 101 Ekonomiaz N.º 100, 2º semestre, 2021 and risk diversification, among others. We include the risk diversification question in the advanced literacy, investment-related block since, as stated by Lusardi and Mitchell (2014), risk diversification is the concept people have the most difficulty grasping. An additional reason to separate questions related to basic concepts (i.e. savings) from the more difficult ones is that the latter are often noisy proxies of the true level of financial knowledge, as shown by Van Rooij et al. (2011). In our empir- ical work, we address random measurement error in covariates, by building indices of financial literacy rather than taking these questions at face value. This intends to avoid biases towards zero in the coefficient estimates. Table 1 contains the questions included to build these two indicators. Table 1. DESCRIPTION OF QUESTIONS RELATED TO FINANCIAL LITERACY INDICATORS   Question Answers B a si c fi n a n ci a l lit e ra cy Question e0600: Now imagine that the five siblings had to wait a year to obtain their share of the €1,000, and that inflation that year was 1%. Within one year they will be capable of buying More than they could buy today with their share of money The same amount Less than what they could buy today Don’t know No reply Question e0700: If you lend €25 to a friend one night and the friend returns €25 to you the following day, what interest have you charged on the loan? A number Question e0800: Let’s suppose you deposit €100 in a savings account with fixed interest of 2% per annum. In this account there are no commissions or taxes. If you make no other deposit in this account and do not withdraw any money, how much money will there be in the account at the end of the first year, once interest has been paid? A number Question e0900: Once again, if you make no deposit or withdrawal, once the interest has been paid to you, how much money will there be in the account after five years? Over €110 Exactly €110 Less than €110 It is impossible to say with the information given Question e1001: Do you consider the following statement to be true or false? An investment with a high return is also likely to be high risk. True False Don’t know No reply Question e1002: Do you consider the following statement to be true or false? High inflation means that the cost of living is rising rapidly. True False Don’t know No reply …/… ALFONSO ARELLANO, NOELIA CÁMARA 102 Ekonomiaz N.º 100, 2º semestre, 2021   Question Answers A d va n ce d f in a n ci a l lit e ra cy Question e1003: Do you consider the following statement to be true or false? Generally, it is possible to reduce the risk of investing on the stock exchange by purchasing a wide variety of shares. True False Don’t know No reply Question e1101: This card show the market value of three investment funds in which €10,000 were invested six years ago. Assuming that the commissions and expenses are the same for all the funds, which fund obtained the best return after six years? Fund 1 Fund 2 Fund 3 Don’t know No reply Question e1102: Which would have been the fund with the bast return if the investment had had to be withdrawn at the end of three years? Fund 1 Fund 2 Fund 3 Don’t know No reply Question e1200: A 15-year mortgage normally requires higher monthly payments than a 30-year mortgage, but the total interest paid over the duration of the loan will be lower. True False Don’t know No reply Source: Own elaboration on ECF. In order to build the two financial literacy indices, we perform a factor analy- sis for the basic and advanced financial literacy questions respectively. Since our variables are dichotomous, we use a polychoric correlation matrix rather than us- ing standard methods based on Pearson’s correlation matrix. The resulting factor structure reveals that for basic financial literacy, the first factor contains 85 per- cent of the total information in our dataset. Similarly, most of the information in the group of variables defining advanced financial literacy is gathered in the first factor that accounts for 92 percent of the total variation. Thus, we build the two indicators with their first factor, respectively. This result suggests that, for both basic and advanced literacy, the selected questions represent a single latent con- cept (i.e. financial literacy), which is desirable when working with factor analysis to construct indices. Table 2 presents the percentage of the quartile distribution of our financial literacy indicators and some sociodemographic characteristics. We observe that women score lower than men in the two financial literacy levels. This result holds for the Basque Country and the rest of Spain. The most educated are the ones with higher scores in financial literacy. However, around 70 percent of people with primary education concentrate below the median for the two financial literacy indicators. We observe a similar distribution for the two ter- ritories studied. Regarding age, both the youngest and oldest in the Basque …/… FINANCIAL VULNERABILITY OF HOUSEHOLDS: EVIDENCE FOR THE BASQUE COUNTRY AND THE REST OF SPAIN 103 Ekonomiaz N.º 100, 2º semestre, 2021 Table 2. BASIC AND ADVANCED LITERACY ACROSS DEMOGRAPHICS   Basque Country Rest of Spain     1st quartile 2nd quartile 3rd quartile 4th quartile 1st quartile 2nd quartile 3rd quartile 4th quartile Basic financial literacy                 Gender Man 12,98 20,03 34,96 32,04 20,79 22,64 33,22 23,35 Woman 27,51 23,77 34,57 14,15 34,12 28,15 27,46 10,27 Age < 25 years old 34,56 14,81 26,72 23,92 35,11 32,21 22,86 9,82 25-34 33,40 22,88 24,89 18,83 30,04 30,03 27,31 12,62 35-44 12,64 25,09 39,88 22,40 23,73 25,75 32,33 18,19 45-54 14,00 19,52 37,32 29,16 20,40 24,91 33,51 21,19 55-64 17,76 24,53 36,23 21,48 26,04 23,71 32,31 17,94 > 64 years old 26,49 20,14 33,58 19,79 39,83 21,59 26,07 12,51 Education level Primary or below 38,16 26,72 26,87 8,25 42,74 27,21 23,03 7,03 Secondary 14,56 19,96 40,57 24,91 20,21 27,79 34,35 17,65 University or above 7,28 18,90 36,84 36,99 9,48 19,25 38,27 32,99 Advanced financial literacy               Gender Man 14,02 29,08 26,06 30,84 20,91 27,68 22,49 28,92 Woman 29,66 31,03 20,29 19,02 33,06 32,18 18,28 16,48 Age < 25 years old 13,00 15,64 41,48 29,88 17,92 32,90 30,99 18,19 24-35 24,06 36,07 28,93 10,94 17,97 31,43 28,92 21,68 35-44 19,09 27,39 26,27 27,25 19,84 30,66 21,43 28,07 45-54 19,03 25,69 26,52 28,76 22,39 31,34 19,72 26,56 55-64 17,45 36,02 15,96 30,56 32,93 29,02 16,60 21,45 > 64 years old 34,80 32,05 13,89 19,25 48,65 26,20 12,29 12,86 Education level Primary or below 38,42 33,16 17,54 10,88 42,97 31,12 14,98 10,93 Secondary 17,26 27,48 29,22 26,04 18,27 32,07 24,41 25,26 University or above 9,22 29,77 21,86 39,16 9,61 25,20 24,82 40,37 Note: Figures represent weighted percentages of households. Source: Own elaboration on ECF. ALFONSO ARELLANO, NOELIA CÁMARA 104 Ekonomiaz N.º 100, 2º semestre, 2021 Country are more financially literate than in the rest of Spain. The two age groups are considered traditionally financially vulnerable. People between 35 and 54 years old score at the highest level for basic and advanced financial literacy. In addition to financial literacy, we introduce information about the personal experiences and attitudes that can shape financial vulnerability through decision making. In line with the literature cited in Section 1, personal experiences in the fi- nancial field might have an impact on the financial outcomes of the household, as shown in D’Acunto et al. (2020) with shopping in inflation expectations or stock market crisis in financial participation (Malmendier and Nagel, 2011). Due to data limitations, we cannot incorporate information about financial crises, but we try to control for the effect of «small traumas», such as having an important disagreement with a financial institution about a financial product (Table 3). Table 3. EXPERIENCE EFFECT ACROSS DEMOGRAPHICS     Basque Country Rest of Spain Disagreement with a financial institution No Yes No Yes Gender Man 86,83 13,17 78,34 21,66 Woman 88,58 11,42 83,32 16,68 Age < 25 years old 100,00 0,00 95,58 4,42 25-34 85,21 14,79 81,05 18,95 35-44 81,80 18,20 72,97 27,03 45-54 86,37 13,63 78,57 21,43 55-64 85,36 14,64 81,50 18,50 > 64 years old 96,16 3,84 88,58 11,42 Education level Primary or below 91,40 8,60 84,91 15,09 Secondary 87,29 12,71 78,75 21,25 University or above 83,98 16,02 76,33 23,67 Note: Figures represent weighted percentages of households. Source: Own elaboration on ECF. Our hypothesis is that people who had a bad experience may have a different rela- tionship with the financial system in terms of trust and their financial decisions might FINANCIAL VULNERABILITY OF HOUSEHOLDS: EVIDENCE FOR THE BASQUE COUNTRY AND THE REST OF SPAIN 105 Ekonomiaz N.º 100, 2º semestre, 2021 be affected, resulting in suboptimal financial participation or undersaving. It would lead to higher levels of financial vulnerability. We can find a signal of past experiences with the financial system in the question of whether in the last five years people had a disagreement with a financial institution about a financial product for more than thir- ty euros, not including any dispute that could have been resolved with a phone call.5 According to the ECF, on average nearly 20 percent of adults stated having had a dis- pute with the financial system in Spain. Table 3 shows that the proportion of people who had a disagreement with the financial system about a product is lower in the Basque Country than in the rest of Spain. This proportion is similar for men and women in the Basque Country and considerably higher for men than women in the rest of Spain. The proportions in terms of age follow a similar pattern in the two terri- tories. Finally, the proportion of people who had a dispute increases with the educa- tion level, both in the Basque Country and in the rest of Spain. 3. DOES FINANCIAL LITERACY AND PERSONAL EXPERIENCES AFFECT FINANCIAL VULNERABILITY? In this section, we test whether the financial literacy level of the households and their previous personal experiences with the financial system have an impact on household vulnerability, as it is suggested by the descriptive results in Section 2. In order to test this hypothesis, Table 4 presents the results of a multivariate regression analysis that shed some light on the importance of financial literacy and personal experiences on household vulnerability, after controlling for other factors. In our model, the dependent variable is the financial vulnerability of the house- hold i (Vi), which is determined by our covariates of interest, financial literacy (BL and AL) and personal experiences (E), and controlled by socio-demographic and household characteristics (X). Due to the discrete nature of our dependent variable, we estimate the following ordered probit model for financial vulnerability: Vi = dj if and only if δj-1 < βBLBLi + βALALi +βEEi + X’βX + εi ≤ δj (1) where dj is an unknown real number, and we observe instead δj, j = 0,...,4; δ0 is taken as -∞ and δ4 is taken as ∞. Finally, εi is an unobserved normal error term, with mean 0 and variance-covariance matrix Σ. Moreover, in Equation (1) we control for geographical effects that may have an impact on household vulnerability. We estimate two sets of models, the first one specific for households in the Basque Country (columns (1), (2) and (3) in Table 4) 5 In this question, there is 0.47 percent of households where the information comes from the inter- viewed person who is not the informed one. ALFONSO ARELLANO, NOELIA CÁMARA 106 Ekonomiaz N.º 100, 2º semestre, 2021 and the second one (columns (4), (5) and (6) in Table 4) for households living in the rest of Spain, which incorporates regional fixed effects.6 The dependent variable, which represents the degree of financial vulnerability, takes on three values and reflects whether the household is highly financially vul- nerable (Vi = 1, households that can handle below 3 months without the main source of income), financially vulnerable (Vi = 2, between 3 and 12 months) or fi- nancially safe (Vi = 3, above 12 months). Thus, larger values of this variable are in- terpreted as a better financial resilience. Our dependent variable represents the same in all the models. Table 4 shows the estimates using three specifications. Table 4. MULTIVARIATE ANALYSIS OF FINANCIAL VULNERABILITY: ORDERED PROBIT RESULTS (1) (2) (3) (4) (5) (6) Dependent variable: Financial vulnerability Gender (Reference: Man) Woman 0,0117 0,0405 0,0574 -0.0554* -0,0391 -0,0443 Age groups (Reference: <25 years old)       25-34 -0,0366 0,0053 0,1062 0.1752** 0.1817** 0.2007*** 35-44 -0,4313 -0,4062 -0,3184 0.1517** 0.1626** 0.1883** 45-54 -0,3938 -0,3738 -0,3271 0,0764 0,0903 0,1081 55-64 0,1181 0,1276 0,1712 0.1792** 0.2070*** 0.2223*** > 64 years old 0,1231 0,1534 0,1685 0.4108*** 0.4592*** 0.4671*** Education level (Reference: Primary Education)     Secondary Education -0,0224 -0,0544 -0,0744 0.1372*** 0.1180*** 0.1236*** University Education 0.3749** 0.3343* 0.3188* 0.3833*** 0.3538*** 0.3603*** Employment status (Reference: Unemployed)       Employee -0,0002 -0,0208 -0,1085 0.2094*** 0.2128*** 0.2102*** Self-employed 0,1031 0,0958 0,0613 0.1253** 0.1261** 0.1328** Retired 0,2017 0,2050 0,1398 0,0725 0,0734 0,0692 Other inactive situation -0,2130 -0,2266 -0,2815 0.1207** 0.1297** 0.1250** Household composition (Reference: Single person)   With partner -0,0995 -0,0796 -0,0533 -0.1321** -0.1314** -0.1311** 6 We take advantage of the regional level representativeness of the ECF in order to analyse separately the Basque Country. …/… FINANCIAL VULNERABILITY OF HOUSEHOLDS: EVIDENCE FOR THE BASQUE COUNTRY AND THE REST OF SPAIN 107 Ekonomiaz N.º 100, 2º semestre, 2021 (1) (2) (3) (4) (5) (6) No partner, with other adults -0,3840 -0,3687 -0,3662 -0.1422** -0.1395** -0.1495** No partner, with children < 18 -0,4312 -0,4567 -0,4160 -0.2695** -0.2726** -0.2644** Home ownership status (Reference: Ownership through purchase mainly) Ownership through inheritance or gift (mainly) 0,3160 0,3344 0,3000 -0.1297** -0.1270** -0.1301** Rental or lease purchase agreement -0.7457*** -0.7325*** -0.7290*** -0.4869*** -0.4859*** -0.4924*** Free use (includes usufruct and similar) -0,4556 -0,5110 -0,5766 -0,0749 -0,0719 -0,0782 Other real estate (Reference: No)       Yes 0,2235 0,2237 0.2504* 0.3009*** 0.2944*** 0.3023*** Annual household income (Reference: < 9,000€) 9,001-14,500 0.7935** 0.7699** 0.8399** 0.2693*** 0.2583*** 0.2599*** 14,501-26,000 1.0585*** 1.0339*** 1.0772*** 0.5689*** 0.5466*** 0.5470*** 26,001-44,500 1.6071*** 1.5752*** 1.6248*** 0.8555*** 0.8272*** 0.8280*** 44,501-67,500 2.1429*** 2.1197*** 2.1826*** 1.1424*** 1.1075*** 1.1081*** > 67,500€ 1.8410*** 1.7987*** 1.9318*** 1.1398*** 1.1073*** 1.1028*** Do not know / No answer 1.3539*** 1.3445*** 1.4336*** 0.6372*** 0.6235*** 0.6224*** Planning expenditures (Reference: No)       Yes -0,0140 -0,0042 -0,0026 -0.1049*** -0.1031*** -0.1006*** Financial literacy             Basic / Saving 0.6745*** 0.5977** 0.6568** 0.2407*** 0.1569** 0.1707*** Advanced / Investment - 0,2822 0,3497 - 0.2317*** 0.2341*** Disagreement with a financial institution (Reference: No)   Yes - - -0.4766** - - -0.1522*** Regional fixed effects No No No Yes Yes Yes Observations 430 430 430 7661 7661 7661 R-squared 0,2268 0,2284 0,2355 0,128 0,1293 0,1304 Note: The estimates are controlled by robust standard errors and we have included weights for individuals in the estimation process. * Significant at 10% level, ** Significant at 5% level, *** Significant at 1% level. Source: Own elaboration on ECF. …/… ALFONSO ARELLANO, NOELIA CÁMARA 108 Ekonomiaz N.º 100, 2º semestre, 2021 The first specification includes the basic financial literacy indicator (columns (1) and (4)), the second specification also includes the advanced financial literacy indicator (columns (2) and (5)), and in the third specification, we add the personal experiences (columns (3) and (6)). In addition to these estimates in Table 4, we report the marginal effects from the third specification of the ordered probit analysis for our variables of in- terest (financial literacy and personal experiences) as well as household income.7 These figures provide further details to understand the magnitude effects of these covariates. Table 4 shows that basic financial literacy is strongly and positively associated with financial resilience, and the results are statistically significant at conventional levels for both the Basque Country and the rest of Spain. Therefore, those house- holds with knowledge about saving-related financial questions are less likely to be financially vulnerable. When it comes to knowledge related to investment, we also find a positive and significant effect on financial resilience, but only for the rest of Spain. In the Basque Country, the estimate for this variable retains the expected sign but it loses statistical significance at the conventional levels. Moreover, the inclusion of investment-related knowledge in the regression increases the estimated value of saving-related knowledge in the two territories, which may point to the idea that saving and investment knowledge reinforce each other. The results depicted by the marginal effects in Figures 1 and 2 illustrate the im- pact of seven equidistant values of these two financial literacy indicators on the proba- bility of belonging to each of the three options of financial vulnerability (i.e. highly vulnerable, vulnerable and safe). For those households in the Basque Country (Figure 1, upside), the probability of being highly financially vulnerable is two times larger when the levels of savings-related concepts are the lowest compared to the highest lev- el of savings knowledge (0.33 versus 0.16). Thus, this estimated probability decreases at most 17 percentage points as the basic financial literacy improves in the household. However, the probability of belonging to the group of financially safe households is 0.55 for the highest level of basic financial literacy and 0.32 for the lowest value. This estimated probability increases at most 23 percentage points with improvements of this literacy level. For those financially vulnerable households, the negative relation- ship of the estimated probability across levels of basic financial literacy is consistent with the previous result and occurs in a narrower range. However, it shows a non-lin- ear process of declining that accelerates for higher levels of basic financial literacy, un- like the linear trend of high financial vulnerability. The marginal effects of the basic fi- nancial literacy for the households in the rest of Spain share the same qualitative results but their impacts are lower compared to those in the Basque Country (Figure 1, bottom). For highly vulnerable households, the difference in the estimated proba- bilities between the lowest and the highest level of basic literacy is above 6 percentage points (0.38 versus 0.32), while the equivalent probability increases at most 6 percent- age points (from 0.29 to 0.36) for those financially safe. 7 Marginal effects for the rest of the covariates are available upon request. FINANCIAL VULNERABILITY OF HOUSEHOLDS: EVIDENCE FOR THE BASQUE COUNTRY AND THE REST OF SPAIN 109 Ekonomiaz N.º 100, 2º semestre, 2021 Figure 1. MARGINAL EFFECTS OF BASIC FINANCIAL LITERACY ON FINANCIAL VULNERABILITY Basque Country Rest of Spain Note: The vertical axis reflects the effect on the probability of belonging to each financial vulnerability level. Vertical lines at the edge of each point represent the 95% confidence intervals. Source: Own elaboration on ECF. The pattern of marginal effects and the associated confidence intervals for the in- vestment-related knowledge shed some light into the loss of significance of this variable in the Basque Country (Figure 2, left). For each vulnerability type, the associated mar- ginal effect, for the highest level of investment-related knowledge, is contained in the confidence interval of the lowest level. Thus, we cannot distinguish across categories ALFONSO ARELLANO, NOELIA CÁMARA 110 Ekonomiaz N.º 100, 2º semestre, 2021 with enough discrimination as to make conclusions in terms of investment knowledge. Once we control for the other variables in the models, households in the Basque Coun- try may have polarized levels of advanced financial literacy and, in this case, only basic financial literacy can be used as a good predictor for financial vulnerability. In contrast, the results for the households in the rest of Spain (Figure 2, bottom) show narrower Figure 2. MARGINAL EFFECTS OF ADVANCED FINANCIAL LITERACY ON FINANCIAL VULNERABILITY Basque Country Rest of Spain Note: The vertical axis reflects the effect on the probability of belonging to each financial vulnerability level. Vertical lines at the edge of each point represent the 95% confidence intervals. Source: Own elaboration on ECF. FINANCIAL VULNERABILITY OF HOUSEHOLDS: EVIDENCE FOR THE BASQUE COUNTRY AND THE REST OF SPAIN 111 Ekonomiaz N.º 100, 2º semestre, 2021 confidence intervals, confirming that the predicted probability of belonging to either high vulnerability or safety is statistically different as we compare differentiated options of investment knowledge. In contrast to the Basque Country, the households in the rest of Spain require a higher discrimination in terms of financial literacy to capture the ef- fect of financial literacy on financial vulnerability. Figure 3. MARGINAL EFFECTS OF DISAGREEMENT WITH A FINANCIAL INSTITUTION ON FINANCIAL VULNERABILITY Basque Country Rest of Spain Note: The reference group is constituted by those without disagreement with a financial institution. The vertical axis reflects the effect on the probability of belonging to each financial vulnerability level. Vertical lines at the edge of each point represent the 95% confidence intervals. Source: Own elaboration on ECF. The variable associated with personal experiences in the field of finance is in- troduced in columns (3) for the Basque Country and (6) for the rest of Spain. Our estimates suggest that the concept of «experience effects» is strong and negatively ALFONSO ARELLANO, NOELIA CÁMARA 112 Ekonomiaz N.º 100, 2º semestre, 2021 related to financial resilience. Households who had a disagreement with the finan- cial system about a financial product are less likely to scale through the highest categories of our dependent variable (i.e. they are more likely to be financially vul- nerable). This finding helps explain the impact of lack of trust in the financial in- stitutions on financial vulnerability via decision making from experience. Margin- al effects, displayed in Figure 3, show that this impact is stronger in the Basque Country than in the rest of Spain. Taking into account that the reference group are households without any disagreement, the estimated probability of being highly vulnerable increases above 10 percentage points in the Basque Country and below 5 percentage points in the rest of Spain for those households that had a dis- agreement. If we focus on financially safe households, the estimated probability is reduced by about 14 percentage points in the Basque Country and by less than 5 percentage points in the rest of Spain. Among the rest of the covariates, another interesting factor conditioning financial vulnerability is household income, which has been widely studied in the inequality lit- erature. We find this effect consistent with each vulnerability level and significant at the conventional levels in all cases. In particular, households with higher income are more (less) likely to be resilient (highly vulnerable) and this probability is increasing (declining) with the income level. Our results confirm that, conditioning on the finan- cial literacy level and personal experiences, household income has an impact on finan- cial vulnerability, both in the Basque Country and in the rest of Spain. This result sug- gests a complementary role of income in this issue, also after conditioning for additional income-related variables, such as labor status and household composition, so we discard the hypothesis that income is the only factor of financial vulnerability. On the other hand, the effect of education level allows us to distinguish differ- ential impacts for secondary and tertiary education in the rest of Spain (primary ed- ucation is the reference). However, we can only discriminate the impact of attained tertiary education versus any lower education level in the Basque Country. The esti- mated coefficients of education on financial vulnerability are positive in all cases but their figures are lower than that from personal experiences (in absolute value). This result is in line with the findings in psychology literature that argues that personal experiences, especially recent ones, exert a greater influence on personal decisions than information in books or via education (Hertwig et al., 2004). In addition to equation (1), where only linear relationships are considered, we include the interac- tion between financial literacy and education level, because it is possible that the ef- fects of financial literacy depend on the level of education and vice versa. However, the interactions are not significant at conventional levels.8 Finally, we find interesting results on financial vulnerability when controlling for home ownership status. As expected, those households with a rental home are more 8 Results are available upon request. FINANCIAL VULNERABILITY OF HOUSEHOLDS: EVIDENCE FOR THE BASQUE COUNTRY AND THE REST OF SPAIN 113 Ekonomiaz N.º 100, 2º semestre, 2021 financially vulnerable than those owning a home, either through purchase, inherit- ance, gift or free use (i.e. usufruct) regardless of the territory considered. Columns (4) to (6) also show that in the rest of Spain, households with an inherited or gifted home are more likely to be financially vulnerable than those who bought it. This result sug- gests that this group of households, in the rest of Spain, are not showing any capacity of affording a relevant dwelling investment. Such an event reflects a negative influence on financial resilience compared to those who purchased their home, reinforcing the idea that the way in which the home was acquired is informative to predict our varia- ble of interest. Moreover, the ownership of other real estate properties also presents a similar behavior on financial vulnerability as the home-ownership does. The estimates are higher and more statistically significant for the households in the rest of Spain than those in the Basque Country. Overall, the estimates of the financial literacy indicators and personal experi- ence with the financial system are statistically significant after controlling for socio- demographic characteristics and present asymmetric effects on vulnerability. We compare two extreme situations in the Basque Country, a household with the lowest value of basic and advanced literacy indicators and disagreement with a financial in- stitution versus another that did not experience any disagreement with a financial institution and shows the highest levels of basic and advanced literacy. For the sec- ond household type, the estimated probability of belonging to the highest vulnera- bility is reduced by 36 percentage points and the increase of the estimated probabili- ty of being financially safe is above 49 percentage points. The impacts for households in the rest of Spain are below 18 and above 19 percentage points, respec- tively. It implies that both financial literacy and personal experiences with the finan- cial system may prove useful tools to mitigate observed differences in vulnerability and also better off safety among households. We are aware that there could be endogeneity in our models whether vulnera- bility and financial literacy are both the result of some other unobserved factors af- fecting these variables (risk aversion, self-esteem, innate ability, intelligence, and motivation). Also, spillovers of a learning-by-doing process may exist, from using financial products to save or invest towards improving knowledge, which might cause reverse causality. In this case, the magnitude of our estimates can be consid- ered a lower bound of the effects of financial literacy on financial vulnerability (Lusardi and Mitchell, 2014). 4. CONCLUSIONS AND POLICY RECOMMENDATIONS COVID-19 crisis has brought about the importance of financial resilience of households to cope with the economic consequences of the pandemic. The ability to deal with shocks is highly correlated to the financial buffer of households in the form of liquid assets accumulated and such ability has been put to the test. ALFONSO ARELLANO, NOELIA CÁMARA 114 Ekonomiaz N.º 100, 2º semestre, 2021 We study two channels that can influence financial decision making and its out- comes in terms of vulnerability. On one hand, financial decisions can be made through learning experience that is approximated by financial literacy (decisions from description). Two levels of financial literacy are considered, one associated with savings (i.e. basic financial literacy) and the second one, more advanced, relat- ed to investment knowledge. On the other hand, financial decisions can also be made from personal experiences with the financial institutions. Our research shows that financial knowledge and personal experiences with the financial system are im- portant in explaining the financial vulnerability of households. This finding holds even conditional on other important factors that also impact on the variable of in- terest and are fully recognized in the literature, such as household income. The two variables of interest, financial literacy and personal experiences, present asymmetric effects on vulnerability. In the rest of Spain, households with better financial knowl- edge (both saving and investment-related) are less likely to be vulnerable and those experiencing disagreements with the financial system about a financial product in- crease their likelihood of being vulnerable. For households in the Basque Country, we only find statistically significant effects for saving-related knowledge and per- sonal experience with the financial system. Such effects show the same signal and are higher (in absolute value) than for those in the rest of Spain. The absence of sig- nificant impact of the investment-related knowledge may be justified by a more po- larized vision of financial literacy level in the Basque Country. Our findings raise concern about the ability of households to develop their fi- nancial resilience in order to be better able to handle such shocks. This belief has been tested during the COVID-19 crisis. It is essential for households to become more financially literate in order to become less vulnerable and build trustable rela- tionships with the financial system since decisions from experience may be more important than decisions from description (i.e. knowledge built from books or edu- cation) when leading to choice behavior. Several questions remain unexplained such as the role of some covariates in re- gional subsamples (i.e. advanced financial literacy in the Basque Country), the di- rection of causality between financial literacy and vulnerability, or the effect of other important financial experiences, such as past economic crises, on financial out- comes. Solving these questions are interesting issues for further research. In addi- tion, it will be interesting to analyze the long-term ramifications of the current COVID-19 induced crisis through the lens of experiences and its effects on financial behavior. Basic concepts from neuroscience that explain how experiences «re-wire» us, and their implications for individual risk-taking and financial decisions, are es- sential to be incorporated in the analyses. Having lived through the COVID-19 cri- sis has already generated a change in our thinking patterns and decision making ca- pacity, so even if we were to return to pre-crisis circumstances, our decision making process may be different. FINANCIAL VULNERABILITY OF HOUSEHOLDS: EVIDENCE FOR THE BASQUE COUNTRY AND THE REST OF SPAIN 115 Ekonomiaz N.º 100, 2º semestre, 2021 Our results are useful in informing policymakers to support efforts to improve household resilience by increasing financial literacy and creating sound legal frame- works to protect interactions between consumers and the financial sector. Financial education cannot eliminate socioeconomic inequalities itself, but it can equip peo- ple with the knowledge to create resilience, and better deal with shocks and plan for the future. Considering the Public Sector as a promoter of social welfare, apart from developing measures mitigating income inequality levels (especially in situations such as the COVID-19 crisis), we recommend fostering its role as a comprehensive and proactive provider of financial education along the life cycle of citizens (e.g. through schools, universities, employee training, and other social programs), going beyond a limited and reactive role associated to financial consumer’s protection. Collaboration with the private financial sector through public-private partnerships in financial education constitute a classic tool to fight against financial vulnerability of households. In this sense, the EBA (2020) provides an annual report of financial education initiatives in the EU. Our insights show other important action points for the financial system. In a context of reduced interest rates in Europe and new competitors from Big Tech companies, the focus of the traditional private financial institutions needs to be reo- rientated from a product-based strategy to financial health providers. A bad person- al experience with a financial institution generates high costs not only in the short- run credibility for both parties, but it generates other implications compromising social welfare through an increase of financial vulnerability of households and re- ducing financial stability of the system. The areas involved are very diverse, such as technology, psychology and economy. ICT and Big Data techniques have the poten- tial to treat new data sources to identify the microeconomic fundamentals to pre- vent and solve disagreements quickly, improving service quality. The Spanish Fi- nancial Sandbox recently created at the end of 2020 increases the opportunities to test financial innovations without jeopardizing the client relationship. ALFONSO ARELLANO, NOELIA CÁMARA 116 Ekonomiaz N.º 100, 2º semestre, 2021 Angrisani, M.; Burke, J.; Lusardi, A.; Motto- la, G.r. (2020): The Stability and Predictive Power of Financial Literacy: Evidence from Longitudinal Data. NBER Working Paper 28125, NBER, Cambridge, MA. Aspachs, O.; Durante, R.; Graziano, A.; Mes- tres, J.; montalvo, j.g.; reynal-querol, m. (2020): Real-Time Inequality and the Welfare State in Motion: Evidence from COVID-19 in Spain. CEPR Discussion Paper DP 15118, CEPR, London. Bank of Spain (2016): The Survey of Financial Competences (ECF) (2016). https://www.bde. es/bde/en/areas/estadis/estadisticas-por/en- cuestas-hogar/relacionados/encuesta-de- comp/ Behrman, J.R.; Mitchell, O.S.; Soo, C.K.; Bra- vo, D. (2012): How Financial Literacy Affects Household Wealth Accumulation. American Economic Review, 102(3): 300–304. Bover, O.; Hospido, L.; Villanueva, E. (2019): The Survey of Financial Competences (ECF): Description and Methods of the 2016 Wave. Occasional paper No. 1909, Bank of Spain. Clark, R.L.; Lusardi, A.; Mitchell, O.S. (2020): Financial Fragility during the COV- ID-19 Pandemic (No. w28207). National Bu- reau of Economic Research. D’Acunto, F.; Malmendier, U.; Ospina, J.; Weber, M. (2020): Exposure to Grocery Pric- es and Inflation Expectations. Journal of Polit- ical Economy, 129(5): 1615-1639. Dolado, J.J.; Felgueroso, F.; Jimeno, J.F. (2020): The Spanish Labour Market at the Crossroads: COVID-19 Meets the Meg- atrends. CEPR Discussion Paper DP 15452, CEPR, London. Eba (2020): EBA Financial Education Report 2019 / 2020. EBA/Rep/2020/12. European Banking Authority, Paris. https://www.eba. europa.eu/sites/default/documents/files/doc- ument_library/News and Press/Press Room/ Press Releases/2020/EBA identifies trends and lessons learned in financial education and lit- eracy initiatives in its second Financial Edu- cation Report/EBA Financial Education Re- port 2019-2020.pdf Guiso, L.; Sapienza, P.; Zingales, L. (2004): The Role of Social Capital in Financial Devel- opment, American Economic Review, 94(3): 526–556. — (2008): Trusting the Stock Market. Journal of Finance, 63(6): 2557-2600. Hertwig, R.; Barron, G.; Weber, E.U.; Erev, I. (2004): Decisions from Experience and the Effect of Rare Events in Risky Choice, Psycho- logical Science, 15: 534-539. Ine (2021): Continuous Register Statistics. https:// www.ine.es/dynt3/inebase/en/index.htm?type = p c a x i s & p a t h = / t 2 0 / e 2 4 5 / p 0 8 / & f i l e = pcaxis&dh=0&capsel=1 Laudenbach, CH.; Malmendier, U.; Niessen- Ruenzi, A. (2019): Emotional Tagging and Belief Formation: The Long-Lasting Effects of Experiencing Communism. AEA Papers and Proceedings, 109: 567-71. Lusardi, A.; Mitchell, O.S. (2008): Planning and Financial Literacy: How Do Women Fare?. American Economic Review, 98(2): 413- 17. — (2011a): Financial Literacy and Planning: Im- plications for Retirement Well-Being. In Lusardi, A., and Mitchell, O. S. (eds.) Finan- cial Literacy: Implications for Retirement Secu- rity and the Financial Marketplace, 17-39. Ox- ford University Press, Oxford and New York. — (2011b): Financial Literacy around the World: An Overview. Journal of Pension Economics and Finance, 10(4): 497-508. — (2014): The Economic Importance of Finan- cial Literacy: Theory and Evidence. Journal of Economic Literature, 52(1): 5-44. Malmendier, U.; Pouzo, D.; Vanasco, V. (2020): Investor Experiences and Financial BIBLIOGRAPHIC REFERENCES FINANCIAL VULNERABILITY OF HOUSEHOLDS: EVIDENCE FOR THE BASQUE COUNTRY AND THE REST OF SPAIN 117 Ekonomiaz N.º 100, 2º semestre, 2021 Market Dynamics. Journal of Financial Eco- nomics, 136(3): 597-622. Malmendier, U.; Nagel, S. (2011): Depression Babies: Do Macroeconomic Experiences Af- fect Risk Taking? The Quarterly Journal of Economics, 126(1): 373-416. Mtes (2021): Anuario de Estadísticas del Ministe- rio de Trabajo y Economía Social 2019. NIPO: 117-20-049-3. https://www.mites.gob.es/es/ estadisticas/anuarios/2019/index.htm Nisbett, R.E.; Ross, L.D. (1980): Human Infer- ence: Strategies and Shortcomings of Social Judgment, Englewood Cliffs, N.J.: Prentice Hall. Osili, U.O.; Paulson, A.L. (2008): Institutions and Financial Development: Evidence from International Migrants in the United States, Review of Economics and Statistics, 90(3): 498- 517. Van Rooij, M.; Lusardi, A.; Alessie, R. (2011): Financial Literacy and Stock Market Partici- pation, Journal of Financial Economics, 101(2): 449-427. Weber, E.U.; Böckenholt, U.; Hilton, D.J.; Wallace, B. (1993): Determinants of Diag- nostic Hypothesis Generation: Effects of In- formation, Base Rates, and Experience, Jour- nal of Experimental Psychology: Learning, Memory, and Cognition, 19: 1151-1164. ALFONSO ARELLANO, NOELIA CÁMARA 118 Ekonomiaz N.º 100, 2º semestre, 2021 APPENDIX Table A1. FINANCIAL VULNERABILITY ACROSS DEMOGRAPHICS Basque Country Rest of Spain     < 3 months 3-12 months > 12 months < 3 months 3-12 months > 12 months Financial Vulnerability 20,40 31,35 48,25 33,81 32,35 33,84 Gender Man 40,18 45,47 52,83 44,28 48,43 51,49 Woman 59,82 54,53 47,17 55,72 51,57 48,51 Age < 25 years old 4,95 8,25 3,62 7,34 6,63 3,99 25-34 21,31 12,99 9,77 15,40 16,70 11,96 35-44 27,27 20,08 19,47 21,15 24,45 21,89 45-54 20,57 25,39 20,99 22,51 23,32 23,02 55-64 10,45 21,32 22,74 17,53 14,90 19,24 > 64 years old 15,47 11,96 23,41 16,08 13,99 19,92 Education level Primary or below 42,97 43,20 23,60 58,83 40,81 33,22 Secondary 45,91 32,03 36,55 30,73 33,53 31,36 University or above 11,12 24,76 39,86 10,43 25,65 35,42 Labour market situation Self-employed 10,74 12,18 13,89 9,18 12,38 12,22 Employee 40,69 49,51 48,33 34,89 47,23 48,46 Unemployed 17,35 9,39 5,74 22,11 12,82 7,65 Retired 7,57 13,82 22,69 15,78 13,18 19,55 Other inactive situation 23,64 15,10 9,35 18,03 14,39 12,12 Household structure Single person 12,42 7,10 8,89 10,51 7,86 8,97 With partner 51,79 69,67 72,71 60,83 65,97 68,80 No partner, with other adults 30,41 21,04 16,86 26,19 24,03 21,27 No partner, with children < 18 5,38 2,20 1,54 2,47 2,13 0,95 …/… FINANCIAL VULNERABILITY OF HOUSEHOLDS: EVIDENCE FOR THE BASQUE COUNTRY AND THE REST OF SPAIN 119 Ekonomiaz N.º 100, 2º semestre, 2021 Basque Country Rest of Spain     < 3 months 3-12 months > 12 months < 3 months 3-12 months > 12 months Home ownership status Ownership through purchase (mainly) 55,84 80,80 84,94 55,40 68,76 79,31 Ownership through inheritance or gift (mainly) 2,37 8,30 9,40 11,05 8,52 9,18 Rental or lease purchase agreement 36,84 10,89 4,41 28,35 19,07 7,88 Free use (includes usufruct and similar) 4,95 0,00 1,25 5,19 3,66 3,63 Other real estate No 80,87 71,92 48,56 76,95 63,44 48,50 Yes 19,13 28,08 51,44 23,05 36,56 51,50 Annual household income < 9,000€ 25,27 4,57 0,57 26,77 9,02 5,44 9,001-14,500 23,32 13,99 6,38 29,67 19,76 11,59 14,501-26,000 30,51 35,67 18,07 22,89 29,81 23,38 26,001-44,500 11,23 27,70 32,23 10,11 21,38 27,94 44,501-67,500 2,03 5,62 23,45 1,65 7,80 14,24 > 67,500€ 1,34 3,00 10,74 0,82 3,55 8,63 Do not know / No answer 6,30 9,44 8,56 8,10 8,67 8,78 Household plan for expenses No 44,58 46,96 58,27 30,99 34,48 41,59 Yes 55,42 53,04 41,73 69,01 65,52 58,41 Note: Figures represent weighted percentages of households. Source: Own elaboration on ECF. …/… Financial vulnerability of households: evidence for the Basque Country and the rest of Spain