Sage Journals: Discover world-class research

Abstract

Borrowing is an integral part of most households’ economy and is a commonplace to raise one's consumption potential. But it also brings the risk of payment problems. Borrowing, and any subsequent payment problems, has a social class dimension, as social class relates to consumption patterns, wealth and earnings. We apply a Bourdieuan class analysis to investigate differences in the onset of payment problems among the adult Norwegian population, and we investigate whether the social classes have different risk factors for payment problems. We use longitudinal register data and the results show that payment problems are more common among those in the lower classes and in the economic middle classes. Moreover, the economic middle classes are more exposed to payment problems in case of life events that incur income loss. Better knowledge about whom is at risk for payment problems, and under what circumstances, can improve welfare policies.

Keywords

Payment problems social class register data life events unsecured debt

Introduction

Debt, income, and wealth constitute the financial situation of households. Hence, in line with wealth and income, debt has a social class dimension (Day, 2020; Sparkes, 2019; Sullivan, 2012). On the one hand, borrowing can increase the flow of capital through investments in education, real estate, or other rent-bearing items and promote social class mobility (Sullivan, 2012). On the other hand, borrowing can finance consumption and allow the indebted to maintain or imitate a certain type of classed lifestyle. Thus, borrowing distorts and blurs class lines (Sparkes, 2019; Sullivan, 2012). The allure and pitfall are that borrowing is invisible—until default. Borrowing, especially unsecured consumer borrowing, is associated with the risk of payment problems, but thus far, the prevalence of payment problems is under-researched from the social class perspective (Sparkes, 2019).

In this study, we perform a Bourdieuan class analysis to investigate the onset of payment problems in the adult Norwegian population. The Bourdieuan class scheme is a structured social space in which people are positioned along two axes: vertical depending on the capital they possess and horizontal depending on their cultural or economic capital dominance (Bourdieu, 2010: 122–123). A key feature of Bourdieu's class analysis is that social positions exhibit homogeneity in dispositions, a certain way of life shared between people who belong to the same class. The utility of Bourdieuan class analysis is that it closely ties capital to lifestyle with an emphasis on the expression of subjective class positions through consumption, and this process of distinction is critical for drawing class boundaries.

Because persons from higher social classes generally have higher-paying jobs, higher employability, and better health compared to persons in lower social class positions, they are less likely to be exposed to risk factors for payment problems (Wiedemann, 2021). Yet, the middle classes have the highest debt-to-income ratios (Bazillier et al., 2021; Sullivan, 2012; Sullivan et al., 2020; Weller, 2012; Wiedemann, 2021), which can increase their vulnerability to the onset of payment problems owing to job loss, ill health, or other life events. Although payment problems are more common among low-income households (Davis & Cartwright, 2019), due to the lack of longitudinal data, one cannot know whether the current class position–typically measured through income and occupation at the time of survey–corresponds to subjective class identity because people may have experienced downwards social mobility before the onset of payment problems (Sullivan, 2012; Warren and Thorne, 2012). We use longitudinal data to escape this coincidence in cross-sectional data and use a person's highest achieved social class position as a measure of their social class belonging.

An even less examined topic is how payment problems relate to horizontal distinctions within the middle and upper classes, for instance, between those who belong to the cultural classes (e.g. lecturers, artists, bureaucrats), professional classes (e.g. doctors, dentists, architects), and economic classes (e.g. corporate directors, managers, financial brokers). Incumbents of the economic classes are more often exposed to volatile income, and the use of unsecured loans to finance status-seeking consumption is more common in the middle to upper-middle income groups (Sparkes, 2019: 1419). Because the economic class fractions have more expensive and consumption-oriented lifestyles than those of the cultural and professional class fractions (Flemmen et al., 2018; Jarness, 2015, 2017), the economic middle classes are perhaps more likely to resort to credit to maintain their lifestyles in face of income shortfall.

We use longitudinal individual-level data for 2008–2018 to explore the onset of payment problems. These data have been compiled from the administrative registers of public authorities, and they are exceptionally well suited for our purpose because they cover the entire population and provide complete overviews of individuals, including their finances, employment details, family details, and health status, and allows us to observe the timing of events. We use the registry of state-enforce wage garnishments as the key indicator of payment problems. Wage garnishments is a serious and objective indication of payment problems, and the registry covers all occurrences so non-response or social eligibility bias is not an issue. We use the Oslo Register Data Class (ORDC) scheme (Hansen, Flemmen and Andersen, 2009), which is an operationalization of Bourdieu adapted to the Norwegian register data and occupational structure. The ORDC scheme distinguishes 13 social classes by capital volume and capital composition.¹

The combination of registry data and a uniform debt-collection system presents us with a unique opportunity to study the antecedents of payment problems. We contribute new knowledge on how different social classes are exposed to payment problems and whether the analyzed risk factors have different consequences across the class structure. In addition, we present new knowledge on the interrelationships between debt, work, health, and payment problems from the life-course perspective. After a long period of economic stability, Norwegians have been facing high inflation and rising interest rates since 2022 (SSB, table 12880). Therefore, any classed patterns revealed in this research would likely be amplified in a worsening economic situation because evidence suggests that rising income inequality fortifies the classed consequences of debt and payment problems as the lower and middle classes become indebted to maintain consumption (Bazillier et al., 2021).

Debt and debt collection in the Norwegian welfare state

Norway has a universal and redistributional welfare state model. Norway provides a high replacement rate on sickness and unemployment benefits, and the national insurance and social security benefits ensure that all individuals receive an income to cover living expenses. Moreover, Norway has public financing for health and education. Medical consultations, including prescribed medications, are free of charge once the yearly expenses surpass about 3000 NOK (ca. 300 USD). Norwegian educational institutions do not charge tuition fees, and childcare is subsidized and means tested. The Norwegian State Educational Loan Fund provides financing at low interest rates to adult students. Accordingly, the correlations between social class and payment problems found in Norway are expected to be more pronounced in other economies where these services are privatized, and benefit replacement rates are less generous.

Norwegian households have the highest debt-to-income ratio among the Organisation for Economic Co-operation and Development (OECD) countries (OECD, 2024). In June 2022, the total secured and unsecured borrowings of Norwegian households amounted to 4.08 billion NOK (SSB, table 11599), and the income-to-borrowing ratio shows that debt surpassed income by 233%. Although about 96% of the lending volume in the private market was secured against real estate, approximately 32% of households had unsecured loans, mostly consumer loans and interest-bearing credit card debt (Poppe and Kempson, 2021). For most of the 2010s, the interest rates on secured loans ranged from 2% to 4% and those on unsecured loans ranged from 14% to 25%. However, rising housing values have allowed homeowners to use equity borrowing for financing consumption, such as cars and lifestyle goods (Borgeraas et al., 2016).

Payment problems in the population clearly vary with economic cycles. In the “debt crisis” years (1988–1993), 13% of Norwegian households reported recurrent payment problems. In the years before the 2008–2009 financial crisis, 4% of the households reported payment problems. This proportion increased during the financial crisis and stabilized at 6% after the crisis. The scope of payment problems is significantly higher among those who use unsecured credit products, such as consumer loans (17%) and credit cards (40%) compared to those who only use mortgages (2%). Statistics obtained from the Norwegian Financial Supervisory Authority show default on 20% of the lending volume, and this figure is growing (Poppe and Kempson, 2021).

Process of debt collection in Norway

If a person defaults on their debt obligations, the first reaction is a reminder from the creditor. If the person continues to default after 28 days, the claim is forwarded to a debt-collection agency. The initial claim accumulates penalty rates in this process. The debt-collection agency gives the debtor another 14 days to pay before the claim is taken to the judicial system. If the person still defaults on their debt obligations, private creditors can call upon the bailiff (namsmannen) to enforce the payment according to the Enforcement Act (LOV-1992-06-26-86) and Creditor's Security Act (LOV-1984-06-08-59). The National Collection Agency (NCA) handles any state claims, such as taxes, fines, or child support. Enforcement can include execution lien, wage garnishments, and forced sales to cover creditors’ claims. The bailiff is legally required to register wage garnishments in the National Register of Mortgaged Movable Property (Løsøreregisteret). On a yearly basis, the wages of around 200,000 persons are garnished. Their employers, or the Norwegian Welfare Administration (NAV) in the case of benefit receipts, deduct money before payment. The employee or benefit recipient retains about 920€ (more for single parent providers) to cover living expenses, and the rest is transferred to the creditors. To discontinue wage garnishments, the debtor must repay all their debt to their creditors or undergo a voluntary or forced debt-settlement procedure in line with the Debt Settlement Act (LOV-1992-07-17-99). However, debt settlement is reserved for those who are considered permanently unable to repay debts and where such settlement does not compromise the general sense of fairness. In this context, 33–50% of the applicants are rejected (Poppe, 2022). After 5 years of debt settlement, creditors relinquish all remaining debts and delete any payment remarks.²

Social class, borrowing, and payment problems

We review research that investigate how socio-economic status and social class relate to attitudes and behaviors that increase a person's leverage and lead to actual payment problems. We do not apply a strict definition of social class because previous research use different measures of income, education, and work to position individuals.

Studies have linked social class to credit uptake through two main motivations, namely replace income and finance consumption. Those in lower social class positions use credit and debt more often to make ends meet. Below-median income is associated with increased use of payday loans in the USA (Lim et al., 2014), and university-educated young people in the UK use loans and credit to pay for everyday essentials (Davis and Cartwright, 2019). Low-wage work, temporary employment, and (lack of) unemployment benefits do not cover living expenses. Unstable and shifting life situations represent a major risk factor for over-indebtedness and payment problems in Denmark (Hohnen et al., 2020) and Finland (Oksanen et al., 2016). Health issues, especially mental health issues, increase the risk of payment problems (Bakkeli and Drange, 2024; Dackehag et al., 2019; Sulivan et.al., 2020).

Research on the consumption pathway places emphasis on expenditure cascades, where “increased expenditure by some people leads others just below them on the income scale to spend more as well” (Frank et al., 2014: 57). At the societal level, rising income inequality is correlated with household credit expansion, and this effect becomes substantially stronger when top-level incomes increase at the expense of the middle class instead of the lower classes (Bazillier et al., 2021: 21; Carr and Jayadev, 2015; Frank et al., 2014). Therefore, the middle-class increases borrowing to “keep up with the Joneses,” which subsequently increases the risk of payment problems. Through qualitative studies, Sparkes (2019) and Sullivan (2012) highlighted how individuals use credit to ameliorate feelings of inferiority stemming from internalized class divisions and to climb the class ladder by using borrowed money. The downside is that they cannot sustain these habits.

High income, high education levels, and professional status have been found to be associated with more positive attitudes towards credit and with greater credit usage (Chien and Devaney, 2001; Rendall, Brooks and Hillenbrand 2021; Poppe et al., 2016). Next, positive attitudes are linked to high-risk financial decisions, such as taking more credit to repay existing debts. This is the net effect of other personal, emotional, and situational factors. According to Rendall et al. (2021), high-earning individuals are more confident about their ability to manage and repay debt, even in periods of financial strain. The use of credit exposes them to the risk of payment problems. Poppe et al. (2016) reported that consumer loans were three times more prevalent among the defaulters who applied for debt settlement in 2011 compared to those in 1999, and this increase was especially marked among those with higher incomes. Moreover, over-consumption was the cause of 33% of the debt-settlement applications, and records indicate that credit uptake to relieve income shortfall propelled these debt problems (Poppe et al., 2016). Wiedemann (2021) showed that high-income households in Denmark were more likely to borrow money to tide over income deficiencies because of the limited benefits available to high-earning segments and easy access to credit. Yet, most households were able to repay their debt upon returning to the labor market.

In sum, a review of the literature indicates that individuals from the lower classes are the most exposed to payment problems owing to a lack of income to cover living expenses. Yet, the middle class has been the center of attention in the international research literature because they rely on work income to manage high debt burdens. This reliance increases their financial vulnerability if they experience a drop in income, for instance, owing to divorce, ill health, or unemployment. Second, a culture of escalating consumption creates expenditure cascades that increase leverage among those with relatively low incomes. Yet, we lack information about payment problems and how life events relate to payment problems across classes.

Theoretical framework

The Bourdieuan class scheme is a structured social space comprising agents’ capital amount and capital composition, and changes in their capital over time (Bourdieu, 2010). Bourdieu's (1990) concept of habitus covers these structured and structuring dispositions. Habitus is the internalization of class positions with the associated lifestyles, ambitions, “reasonable” behavior, and choices. Therefore, trajectories in the social space tend to follow conventional patterns depending on what is viewed as “appropriate” and within reach, although the trajectories of individuals can diverge.

Borrowing and payment problems are tied to social class in (at least) three ways. The first is through access to economic capital in terms of wealth, income, and credit rating. Those in higher social classes, especially those in the economic fractions of the middle and upper classes, have more possessions to borrow against and higher incomes to service such borrowing; therefore, they can carry higher levels of borrowing without the risk of default. Moreover, evidence from the U.S. shows that stagnant wages and rising living costs result in increased credit usage among the lower-middle and working classes. Hence, differences in stocks and capital flows can increase the vulnerability of the lower classes to payment problems and reduce it among the middle and upper classes (Leicht, 2012). Because our data contain extensive information about work and household income, secured and unsecured debt, and wealth, we can explore the influence of monetary resources for the onset of payment problems.

The second route between social class and payment problems is through habitus. On the one hand, the middle and upper classes are less likely to face payment problems because they have more resources and are more knowledgeable about the financial field; together, these factors lower their chances of credit entrapment (Lusardi & Mitchell, 2014). On the other hand, they harbor different norms towards borrowing and credit. Those with limited means are more likely to maintain an ascetic “save before you buy” morality, whereas those belonging to the middle class are more likely to harbor a morality of credit and consumption (Bourdieu, 2010: 309; Swedberg, 2011). More extensive credit usage does not necessarily cause payment problems. However, research shows that persons with a positive attitude towards borrowing—those who think it is acceptable to use credit to finance consumption—are more likely to choose high-risk strategies during financial difficulty, which can increase their risk of payment problems (Rendall et al., 2021). People in middle-class positions are more optimistic about their economic prospects and ability to repay debt compared to those in lower social class positions. Based on the trend of rising consumer debt in debt-settlement applications submitted by people from higher-income groups, Poppe et al. state: “Obviously, the more resourceful the borrower is, and the more acceptable this type of borrowing is in one's socio-cultural surroundings, the more likely s/he is to take out consumer loans and credit – i.e. for as long as s/he is considered creditworthy by the lenders” (2016: 27). Therefore, we anticipate that the middle-class habituses can protect themselves against payment problems owing to their superior knowledge and understanding of the financial game while also being more vulnerable to payment problems because of their stronger risk-taking tendency. Financial risk-taking and credit usage are more likely among the economic class fractions (Bourdieu, 2010: 123).

The third route that links social class to payment problems is the quest for mobility and maintenance of class position. In Distinction, Bourdieu closely ties consumption to cultural and economic capital, and thus, class (Swedberg, 2011). Sparkes (2019: 1422) draws attention to the role of credit in Bourdieu's Distinction and emphasizes that feelings of deficit relative to higher classes are intensified in cross-class interactions. Credit is a means for lower classes to attain what the group just above them has, and this element has mostly been overlooked by researchers in cultural class analysis. A key feature of Bourdieuan class analysis is the division between objective and subjective class positions. The objective class position, typically measured by occupation, can change from one day to the next because people experience occupational advancement, degradation, or job loss. In the experience of the downwardly mobile, their objective class positions do not reflect their internalized, subjective class positions (Poppe, 2008). These people can continue to perceive themselves as part of the middle class and strive to maintain the consumption patterns of that class position, even though their wage incomes can no longer support that lifestyle (Hodson et al., 2014). Bazillier et al. (2021) argues from a purely economic standpoint that consumption preferences are part of a utility function for maintaining past consumption levels (“habit formation”) in the face of income shortfall and to reference one's own consumption to those just above in the income distribution (“keeping up with the Joneses”). This upwards and downwards social reflexivity justifies the use of credit by a person to not fall below past consumption levels and preferably attain the consumption patterns of the reference social groups (Bazillier et al., 2021: 132). A key insight from Sparkes (2019: 1429) is that credit usage allows one to buy an image of class mobility that affects their view of themselves (“I’m part of the elite with this device”) and how others view them. Credit usage serves two main mechanisms–it ameliorates feelings of inferiority stemming from internalized class divisions, and it helps climb the class ladder, albeit with borrowed money. The downside is that such people cannot sustain these habits. Sullivan. (2012) argues that debt and credit have emerged as important covariates of social class and social stratification. Borrowing allows one to manifest social class divisions or even to blur such distinctions because it detaches spending from income–at least for a while.

Analytical model

The review of the literature in combination with our theoretical lens leads to two analytical models. The first model is a mediation model (solid lines in Figure 1) wherein social class has a direct and an indirect path to payment problems through selective association with other known risk factors. Specifically, we expect the relationship between social class and payment problems to weaken after we adjust the analyses for total debt burden and unsecured loans, labor-market status and income instability, onset of health issues and divorce, as well as other demographic variables (such as age, gender, and region of residence). The second model is a moderation model (dashed lines in Figure 1). According to this model, we expect that social class conditions the relationship between risk factors and payment problems. The relative influence of life events on payment problems depends on an individual's resources, which are measured in terms of social class status in this study.

Figure 1.

Analytical models.

Data and methods

The data used in this study originate from administrative registers and are made available for research through Statistics Norway. Different public authorities collect and register information about the populations within their jurisdictions for administrative purposes. A personal identifier allows us to link the different registers and create individual records with information about employment, income, debt, and other aspects.

Dependent variable

We define payment problems at the first encounter of wage or benefits garnishments within the calendar year. This information is obtained from the National Register of Mortgaged Movable Property (løsøreregisteret). We use a dichotomous indicator of absence or presence of wage garnishments (0 = absence, 1 = presence), independent of the duration of this coercive measure. We conduct robustness checks with a threshold of a minimum of six months with wage garnishment during the year (consult the supplementary documentation).

Independent variables

We measure social class by using the ORDC scheme (Hansen et al., 2009). This Bourdieuan-inspired class scheme combines information about occupation, labor-market status, and income (including welfare transfers) to identify social class and distinguish between the elite, upper- and lower-middle classes, working classes, and those living mainly on welfare transfers. The occupations of the elite and middle classes are further divided into the cultural, professional, and economic fractions.³ The ORDC has 13 categories (see Supplementary Figure 1 for a graphical display of the class scheme) and the documentation report explains how occupations are classified (Hansen et al., 2009). We use income information obtained from the Tax and National Insurance register (FD-trygd) and occupational information obtained from STYRK-98, the Norwegian equivalent of ISCO-88 codes, to classify individuals. Because occupation and income can shift from one year to the next, we consider the highest held class position of an individual as the indicator of their social class belonging.

Our analytical models specify five main explanatory variable sets related to (i) borrowing, (ii) employment status, (iii) marital status and household income, (iv) health, and (v) wealth. All these variable sets use information collected on an annual basis and are, therefore, time-variant within persons.

Data on borrowing are obtained from the annual tax returns of individuals and the register of mortgaged movable property. This register contains data on loan volume and debt interest rates. First, we calculate a person's debt-to-income ratio based on their total, annual debt volume against income. Owing to outliers (persons with excessive debt-to-income ratios), we transform this measure to percentiles (1–100) and include it in regressions as a curvilinear function to the third power. Next, we use the debt interest rate to identify unsecured borrowing. Any debt with interest rate exceeding 8% is considered unsecured borrowing.⁴ We apply a logarithmic transformation to the volume of unsecured debt and include this transformed variable in the regressions as a curvilinear function to the second power. Third, we include a dummy to identify a 20% increase in unsecured borrowing from one year to the next to identify persons facing exaggerated financial risk. In addition, we include the lagged value of this variable in the regressions because any payment problems can take time to evolve. Finally, we include a continuous variable pertaining to the number of creditors because debtors who have many creditors may be facing more complex and precarious economic situations.

We identify employment status based on information about annual wage income from the Tax register and any receipt of unemployment benefits from the National Insurance register. First, we include the logarithm of annual wage in the regressions as a curvilinear function to the second power. Wage income includes all earnings during the year and any benefits received in place of wages such as parental benefits, unemployment benefits, and sickness benefits. Next, we measure unemployment in terms of the receipt of unemployment benefits and distinguish between short- and long-term unemployment based on the amount received.⁵ Short-, medium-, and long-term unemployment correspond to 0 < 10% of annual wage income, 10 < 50% of annual wage income, and >50% of annual wage income, respectively. Finally, we include a dummy variable to identify a 20% drop in work income from the previous year. This variable captures changes in a person's employment situation that can set off payment problems. Again, because payment problems take time to evolve, we include the lagged value of this variable in the regressions.

Information about marital status is obtained from the National Insurance register. We construct a dichotomous measure with score 0 for single or divorced persons and score 1 for married or cohabitating couples. We include the logarithm of annual household income⁶ in the regressions as a curvilinear function to the second power. Household income measures the flow of money in the family. We include a dummy to identify a 20% drop in household income from the previous year. As with the other variables, we lag this variable in the regressions.

We use two dichotomous measures for health issues and distinguish between mental⁷ and somatic health issues based on diagnosis groups. This information is obtained from the patient registry of hospitalization or from the data of polyclinical consultations with the specialized secondary health service. The two measures take the value 1 if a person has been under treatment during the year and 0 otherwise. Hence, these measures narrow in on conditions that are more acute, severe, long-lasting, or too medically complicated to be solved by the primary health service. We also include the lag of these variables in the regressions to account for the delay between a negative change in life situation and payment problems.

Finally, we adjust the regressions for taxable net wealth, including fixed and financial assets. This information is obtained from the Tax registry. We include the logarithm of gross wealth as a curvilinear function to the second power. We include a dummy variable to identify a 20% drop in wealth from the previous year, and this variable is lagged to measure any delayed consequences of wealth reduction.

All our measurements related to income, wealth, and debt are in Norwegian kroner (NOK). We take the natural logarithms of those numeric values to normalize distributions. As the log of zero is undefined, we add (x + 1) before transformation.

Control variables

The control variables are gender (0 = man, 1 = woman), first- or second-generation immigrant (0 = no, 1 = yes), education level (categories are as follows: secondary education, high/school or vocational education, lower-tertiary education, higher-tertiary education, and missing information), and student (0 = no, 1 = yes). These pieces of information, except gender and immigrant status, are time-variant.

Data structure and analysis

We analyze the data using discrete-time models. These models are appropriate for application to data that have a panel-data structure with interval-censored data in person-years (Rabe-Heskett and Skrondal, 2012). The discrete-time models account for censoring (no event during observation window) and accommodate the incorporation of time-varying variables. The dependent variable takes the value 1 if a person has any wage garnishments during the calendar year. We use cloglog models because payment problems occur rarely. Given a small number of events, the cloglog models yield better estimates than the logit and probit models.⁸ We model time dependence with dummy variables.

Our observation window covers 2008 to 2018. The sample includes all Norwegian residents of ages 18 to 70 years. Individuals are included from the year 2008, upon turning 18 years, or upon immigration. Individuals are excluded upon their first encounter with payment problems (no estimate of repeated events), turning 70, death, emigration or censoring. In the main analysis, we consider the entire population because everyone is at the risk of payment problems. This file consists of 3,070,710 individuals and 33,482,938 person-year observations. The analysis time is based on age because people are not at the risk of payment problems until they attain the age of majority.⁹ When time is measured in age, left-censoring is not an issue (Allison, 1984: 57). The model yields an age-specific incidence rate, as the data cumulatively computes the risk of failure in each age-group. We perform robustness analyses when individuals are included upon the first observation of debt (heightened risk). The results of these robustness analyses are like those of the population sample. Supplemental Tables S1 and S2 summarize the results obtained using the cloglog models and Cox proportional hazards, in addition to presenting the estimated probability of payment problems, not only onset.

We use different post-estimation diagnosis tools to investigate model fit, such as goodness-of-fit and receiver-operator curves (ROC).¹⁰ The ROC serves as a graphic display of model sensitivity vs. specificity. The area under the curve of the final model is 0.85, which is satisfactory.

Ethics

This project received ethics approval from the Norwegian Data Protection Service (SIKT), Norwegian Data Protection Authority, Regional Committee for Medical and Health Research Ethics (REK), and various data owners, namely the Norwegian Health Directorate, Norwegian Tax Administration, and Brønnøysund Register Centre.

Results

Descriptive results

Tables 1 and 2 present descriptive statistics of the time-variant and time-invariant characteristics, respectively.

Table 1.

Descriptive statistics categorical variables.

	Wage, garnishment		Total, N
Incidence of wage garnishments	456,416	1.36%	33,482,938
Social class
Cultural upper-class	4,998	0.59%	846 702
Professional upper-class	6,711	0.37%	1 792 546
Economic upper-class	7,495	0.72%	1 038 867
Cultural upper-middle	10,169	0.48%	2 110 901
Professional upper-middle	34,000	0.69%	4 924 942
Economic upper-middle	25,290	1.10%	2 292 677
Cultural lower-middle	9,635	0.96%	1 004 924
Professional lower-middle	38,697	1.10%	3 508 499
Economic lower-middle	43,161	2.00%	2 162 014
Skilled workers	103,141	1.84%	5 608 494
Unskilled and semi-skilled workers	99,182	2.19%	4 530 500
Farmers, fishermen, foresters	1,079	1.62%	66 446
Welfare, transference	72,858	2.03%	3 595 426
Sex
male	293,398,	1.76%	16 684 691
female	163,018	0.97%	16 798 247
Immigrant
No	362,786	1.26%	28 858 756
Yes	93,630	2.02%	4 624 182
Education level
Secondary education,	239,051	2.36%	10 138 744
High school/vocational training	141,797	1.29%	10 966 115
Lower university,	50,303	0.58%	8 623 075
Higher university,	11,147	0.35%	3 215 810
Missing	14,118	2.62%	539 194
Student
No	434,130	1.46%	29 684 494
Yes	22,286	0.59%	3 798 444
20 pct increase in unsecured debt
No	387,127	1.26%	30 673 110
Yes	69,289	2.47%	2 809 828
Unemployed
No	105,357	1.66%	6 357 224
Short	313,014	1.19%	26 252 141
Medium	21,059	3.86%	545 735
Long	16,986	5.18%	327 838
20 pct drop in work income
No	392,057	1.30%	30 264 383
Yes	64,359	2.00%	3 218 555
20 pct drop in household income
No	380,749	1.29%	29 613 938
Yes	75,667	1.96%	3 869 000
During the year: ever had mental health diagnosis
No	397,473	1.25%	31 802 060
Yes	58,943	3.51%	1 680 878
During the year: ever had somatic health diagnosis
No	266,308	1.36%	19 613 937
Yes	190,108	1.37%	13 869 001
20 pct drop in gross wealth
No	389,865	1.27%	30 697 624
Yes	66,551	2.39%	2 785 314

Table 2.

Descriptive statistics continuous variables.

	No wage garnishment		Wage garnishment		Total
	Mean	Std.dev	Mean	Std.dev	Mean	Std.dev
Debt-to-income ratio, percentiled	39.1	33.3	42.3	30.3	39.2	33.3
Logarithm of unsecured debt (NOK)	3.1	4.6	5.8	5.3	3.1	4.7
Number of debt cases/creditors	2.8	2.8	4.9	4.4	2.9	2.8
Logarithm of work income (NOK)	2.8	1.6	2.4	1.6	2.8	1.6
Logarithm of household income (NOK)	4.0	0.7	3.7	0.7	4.0	0.7
Logarithm of gross wealth (NOK)	10.8	5.6	6.7	5.9	10.7	5.6

Table 3 presents the distributions of total debt,¹¹ total secured and unsecured debts, and income-to-debt ratio¹² according to social class. The debts of the economic fractions are clearly higher than those of the cultural and professional fractions. This is consistent with the view that people belonging to the economic fractions engage in elevated financial activity. The debt-to-income ratios of skilled workers and the lower classes are lower than those of the middle and upper classes, and payment problems are relatively more frequent in those lower classes.

Table 3.

Debt and wage deductions according to social class.

Social class	Debt in NOK	Unsecured debt (NOK)	Secured debt (NOK)	Debt-to-income ratio	Wage deductions
Cultural upper-class	1,115,735	55,587	1,357,876	2.0	0.6%
Professional upper-class	1,491,077	71,522	1,753,424	2.1	0.4%
Economic upper-class	2,591,318	126,677	2,913,917	2.2	0.7%
Cultural upper-middle	830,187	55,561	1,036,425	1.8	0.5%
Professional upper-middle	1,019,740	67,660	1,246,782	1.9	0.7%
Economic upper-middle	1,463,998	96,706	1,665,607	2.4	1.1%
Cultural lower-middle	571,358	50,211	788,546	1.6	1.0%
Professional lower-middle	679,942	60,577	928,316	1.6	1.1%
Economic lower-middle	750,417	72,398	1,011,782	2.0	2.0%
Skilled workers	705,940	69,837	981,470	1.4	1.8%
Unskilled and semi-skilled workers	489,621	68,351	753,399	1.2	2.2%
Farmers, fishermen, foresters	572,361	78,878	828,447	1.1	1.6%
Welfare, transference	348,945	59,986	582,875	1.0	2.0%

Figure 2 depicts the baseline hazard survival function of payment problems in the adult population from a cloglog model with age as the only predictor.

Figure 2.

Baseline hazard ratios.

The baseline hazard displays a convex shape over age. The baseline hazard rises until 25 years of age where it reaches a hazard probability of 2,15%, and remains at a relatively high level until the early 40 s, before it gradually declines. The lowest baseline hazard probabilities are found among the oldest persons aged 70.

Regression results: mediation model

The regression results obtained using the mediation model are listed in Table A1. Figure 3 graphically presents the hazard rates of payment problems according to social class status when all the other variables in the model are held at the mean.¹³ The upper left panel shows the gross difference in payment problems according to social class position, adjusted only for gender, education level, and immigrant status. The upper right panel adjusts for borrowing levels. The lower left panel adjusts for work income, drop in work income, unemployment and health status. The lower right panel adjusts for marital status, household income, drop in household income and for wealth and drop in wealth.

Figure 3.

Average marginal effects of onset of payment problems associates with social class position. Upper left: net effect of social class. Upper right: with additional controls for debt. Lower left: with additional controls for work income, unemployment and health. Lower right: with additional controls for marital status, household income and wealth. Figures correspond to Table A1 models 1, 2, 4, and 6, respectively.

Any remaining differences in the hazard rates of payment problems of the social classes likely stem from unobserved variations in social networks, consumption, lifestyles, and financial risk-taking, which affect their vulnerabilities to the onset of payment problems. All models include dummy variables for age to account for the time trend.

The results depicted in Figure 3 highlight a clear social class gradient. Onset of payment problems are more common in the lower social classes than in the upper social classes. The unadjusted estimates indicate a rather high hazard rate of payment problems among the people on welfare benefits. Next, we observe the differences in onset of payment problems depending on capital composition in the lower-middle class, where the economic fraction is at the highest risk of payment problems. However, in the upper-middle and upper classes, the cultural fraction is at the highest risk compared to the economic and professional fractions.

Our comparisons of the adjusted and unadjusted estimates reveal that the higher hazard rates among the lower classes can partly be attributed to their greater exposure to risk factors such as debt, income volatility, unemployment, and health issues. After adjusting for these factors, the differences between social classes in terms of the hazard of payment problems decreases. This result is obtained because of a composition effect–the higher social classes are less exposed to life events that increase the hazard of payment problems. However, after adjusting for these factors, the hazard of payment problems increases and equalizes across classes. Therefore, the adjusted estimates change the results, as hypothesized.

Figure 4 depicts the average marginal change in the hazard rate of onset of payment problems associated with various risk factors.

Figure 4.

Average marginal change in the hazard for onset of payment problems associated with life events.

Long-term unemployment and mental health issues are by far the most likely to cause onset of payment problems. The hazard rate of payment problems is the highest in the initial year of disease, but we observe a substantial hazard rate of payment problems in the following year as well. Somatic illnesses increase the hazard rate of payment problems in the short and long terms but not nearly as much as mental illnesses. This is commensurate with the literature (Bakkeli and Drange, 2024). Illnesses that lead to hospitalization and specialized treatment can increase payment problems through three main avenues, namely loss of earnings,¹⁴ medical expenditure,¹⁵ and the illness itself (Sullivan et al., 2020, ch. 5). One possible reason for the large gap in the hazard rate of payment problems between those with somatic and mental illnesses is that persons with mental disorders lack the capacity to maintain their finances and would be less likely to receive and accept help from debt advisors (Bakkeli and Drange, 2024). Additionally, heavy spending can be a part of the diagnosis for mental disorders.

The hazard rate of payment problems associated with a drop in wealth and an increase in unsecured debt reveals an interesting pattern. An increase in unsecured debt reduces the hazard rate of payment problems in the year of uptake but increases it in the following year. One interpretation is that people use credit to tide over income loss (Wiedemann, 2021) and misjudge their ability to repay loans. This is commensurate with the literature (Poppe et al., 2016). A > 20 percent drop in wealth from one year to the next increases the hazard rate of payment problems, but the effect intensifies in the following year. Therefore, when people reduce savings, it is likely that their cash flows are unbalanced with less income than expenses. In the next section, we investigate whether these risk factors vary across social class backgrounds.

Regression results: moderation model

Table A2 lists the regression results obtained using the separate models for the economic, professional, and cultural class fractions and the working classes. Figure 5 depicts the average marginal change in the hazard rate of payment problems associated with different life events across the class fractions and working classes. Our analyses reveal two findings. First, each life event decreases or increases a person's hazard rate of payment problems in the same direction across all classes, although by different magnitudes. These patterns are similar across the upper, upper-middle, and lower-middle classes within each fraction, but they are the most pronounced in the lower-middle class (results not shown).

Figure 5.

Average marginal change in the hazard of onset of payment problem associated with life events. Marginal effects are estimated across the cultural, professional, and economic class fractions and the “working” classes.

The most striking result in Figure 5 is that life events have relatively stronger effects on payment problems in the economic fractions of the upper and middle classes compared to those in the cultural and professional fractions. Especially, mental health issues, long-term unemployment, and drop in wealth expose these economic fractions to risk. This finding can support the hypothesis that these fractions are more susceptible to income deficiencies, and perhaps, their private economy has lower margins compared to those of the professional and cultural classes. Moreover, the economic fractions are susceptible to the highest hazards of payment problems owing to a drop in wealth, which can indicate that they are more willing to take risks with investments. The risk profile of the working classes indicates a higher likelihood of the onset of payment problems in response to changes in labor-market activity, such as a drop in work income and medium- and long-term unemployment. Health issues, too, increase their hazard rates of payment problems. Finally, the uptake of unsecured debt (L1) is associated with a higher hazard rate of payment problems in the working classes.

Discussion

In this research, we investigated the extent to which payment problems are associated with social class positions to reveal differences in vulnerability in response to life events. We tested two main hypotheses, which we discuss in turn.

First, we investigated to what extent social class differences in payment problems were caused by different risk exposures. According to our results, payment problems exhibited a clear social gradient. The lower-ranked classes were at a higher hazard of onset of payment problems than the higher-ranked classes. This is consistent with the theory and previous research because lower-ranked classes have lower incomes and limited savings, and therefore, have less money to cushion against economic setbacks (Poppe et al., 2016). However, after adjusting for factors related to the level of borrowing, income, employment, health, and family status, these inter-class differences diminished but did not disappear. Hence, the higher social classes are at lower hazard of payment problems because generally, they have more secure employment, better health, and more income and wealth than the lower classes. Despite our attempt to make the classes more comparable by considering observable resources and debt obligations, several social class differences remained. An interesting pattern emerged (c.f. Figure 3, lower right panel) because those in the higher social classes were at a lower hazard of onset of payment problems compared to those in the lower social classes, except for the economic class fractions, for which the hazard rate of onset of payment problems was at par with those of the lower classes. We attributed this pattern to unobserved differences in class, and we elaborate on it in the following paragraphs.

The higher social classes have more external reserves to draw on to avoid payment problems, and family, friends, or even employers can provide access to capital (Bourdieu, 2010; Poppe, 2008). Access to limited economic resources and overspending to cover everyday essentials are frequent precursors are frequent precursors of payment problems among lower-income households (Poppe et al., 2016). Next, the upper and middle classes are more likely to understand the complexity of the financial field, plan their economy, and stay away from “bad deals” (Hayes, 2020; Swedberg, 2011). Yet, the higher social classes can be overconfident in their ability to manage financial risk (Hayes, 2020: 30), and according to the literature, the higher classes have more positive attitudes towards debt and credit and choose more risky strategies in strained situations (Rendall et al., 2021). Lower risk aversion can be one explanation for the higher hazards of payment problems among the economic fractions of the upper and middle classes, all else being equal. Furthermore, the “expenditure cascade” suggests that interpersonal comparisons with the consumption patterns of those just above oneself lead to increased debt uptake and spending. Among the welfare recipients and lowest social classes, this pressure might be lower because they cannot afford such spending, but among those in the middle class, where internal wage inequality is likely greater and people work and live in closer proximity to those with higher earnings, lower-ranked individuals can easily observe their higher-ranked colleagues’ consumption. Sullivan (2012) argued that debt and credit have become important covariates of social class and social stratification because borrowing allows one to blur class distinctions by detaching spending from income, at least for a while. Hence, even if we adjust our results for borrowing, work income, and household income and wealth, the class differences reflect different levels of access to means, differing economic situations, and differing valuations of consumption. Although we cannot observe these dynamics in our aggregated data, our findings in combination with those of previous studies and the theory support the plausibility of such mechanisms.

Second, we investigated whether life events had different consequences across social classes. We hypothesized that the economic fractions of the middle class would be at a greater risk of payment problems owing to life events such as divorce, unemployment, and income shortfall. This class fraction has more excess income with which to carry debt, expectations of rising future income, and greater willingness to take on high levels of debt, perhaps as a strategy to increase wealth. Moreover, these fractions harbor a consumption-oriented lifestyle. However, carrying debt on income flow increases their susceptibility to payment problems. In short, life events have similar effects on all class fractions, but these effects are more intense for the economic class fractions, as anticipated. First, short-term unemployment reduces the hazard of payment problems, but moderate- and long-term unemployment increase this risk. This can imply that people are better able to adjust to temporary income loss, whereas longer periods of income loss are more challenging. Interestingly, short-term unemployment reduces payment problems by the greatest extent in the cultural fraction. One interpretation of this result is that those in the cultural sector are more used to intermittent periods without work, for instance, freelance musicians, actors, and artists, and they can accommodate such periods. Long-term unemployment increases the risk of payment problems across all classes but to the greatest extent among the economic class fractions. The maximum unemployment benefit available to a person is ca. 461 000 NOK or 30 000 € in 2024¹⁶. Undoubtedly, long-term unemployment becomes more economically burdensome for those who used to earn a high wage. Second, an increase in unsecured borrowing reduces payment problems in the short term but increases them in the long term, more so among the economic class fractions. This finding aligns with the observation made by Sullivan (2012) that persons do not usually use credit for luxury consumption but to tide over an income interruption. Indeed, mid-level managers have been advised to use credit to maintain their lifestyle to increase their chances of landing a similar job and forestalling the public signs of downward social mobility (Sullivan, 2012: 54). Sparkes (2019) clearly demonstrated how persons use credit to maintain their previous lifestyle and sense of self after experiencing class degradation or to alleviate experiences of relative deprivation compared to their friends. An example of the first is parents who do not wish their children to see that they have fewer financial resources after divorce and desire to maintain the same relative class position as that of their previous, and more well-off, spouse. An example of the second is persons who use credit cards to finance eating out with friends instead of suggesting a cheaper restaurant because it would reveal a lower-class position and concomitant feelings of defeat/subordination. However, the use of credit implies that the dinner becomes more expensive for those who, in principle, cannot afford it compared to those who can.

Mental health issues substantially increase the risk of payment problems, and the magnitude of this increase is on a par with that resulting from long-term unemployment. A striking result is that mental health disorders increase the hazard rate of payment problems first and foremost among the working classes and the economic class fractions. The effects of mental health disorders on payment problems are considerably weaker for the professional and cultural fractions of the middle and upper classes. We can only speculate on the reason behind this finding, but a possible explanation is that those in the working classes have fewer social and economic resources to assist them. Persons experiencing mental health disorders are less likely to receive and accept help (Bakkeli and Drange, 2024). Once the debt-collection system comes into play, it can be difficult to halt the process. Those from higher social origins might have more knowledge, networks, and contacts to assist them with this situation. However, that does not explain why the economic class fraction is almost as much at risk as the working class. On average, the economic class fractions have substantially higher debts and higher unsecured debts, signalling a more complex private financial situation that can spiral into payment problems. To speculate further, perhaps financial risk-taking, mental health, and payment problems are more closely associated in this class fraction, and access to credit can be a problem if the disorder itself leads to overspending. Mental health issues increase payment problems among persons with sound economic conditions before the onset of disease (Dackehag et al. 2019).

Conclusion

We applied a social class lens to payment problems and firmly concluded that social class is associated with payment problems, even beyond differential exposure to risk factors. The social class gradient in payment problems implies that the higher social classes have more resources at their disposal to fend off payment problems. The middle classes experience payment problems, particularly in the aftermath of job loss, income loss, or divorce. Therefore, the combination of high debt obligations and low or no wealth increases their vulnerability to payment problems when they experience discontinuous income flows. An important nuance is that the economic middle-class fractions are more exposed than their professional and cultural counterparts. The lower classes are less affected by income or employment disruption than the middle and upper classes. This implies that the lower classes are more exposed primarily because of their continuous low-income levels.

This research follows individuals to the first incidence of payment problems, but their histories continue. Future studies should investigate the repeated payment problems and, eventually, debt settlements and how these issues affect marriages, housing, and labor-market careers, in addition to variations across social class, age, and gender. We performed event-history analyses and used lagged variables to investigate the effects of life events on the onset of payment problems. Although these analyses provided some evidence of the temporality of payment problems, the underlying method was unable to flesh out the evolution of spiralling trajectories towards payment problems. In future studies, methods such as sequence analysis should be used to investigate the “career patterns” of payment problems and over-indebtedness, in addition to combining data sources to merge the strengths of register data for considering structural factors comprising survey data on habits, knowledge, strategies, and personality traits.

Supplemental Material

sj-docx-1-asj-10.1177_00016993241309762 - Supplemental material for Payment problems through a social class lens

Supplemental material, sj-docx-1-asj-10.1177_00016993241309762 for Payment problems through a social class lens by Ida Drange and Nan Zou Bakkeli in Acta Sociologica

Footnotes

Data availability statement

The registry data are protected according to national law, and only available for research upon application to data authorities and ethic committees. The scripts to produce these results are available upon request.

Ethical statement

This research has been ethically approved by the Norwegian Data Protection Services (SIKT) and the Regional Ethics Committee (REK) for health research.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research is part of the WellDebt-project, financed by the Norwegian Research Council, grant no. 302884.

ORCID iD

Ida Drange

Nan Zou Bakkeli

Supplemental material

Supplemental material for this article is available online.

Notes

Author biographies

Ida Drange is an assistant professor at the Centre for the Study of Professions at Oslo Metropolitan University, where she obtained her PhD in 2013. Her research covers topics within the fields of labor market and organizational sociology, social inclusion, and diversity.

Nan Zou Bakkeli obtained a PhD in sociology from the University of Oslo, where her research focused on social inequality and medical sociology. She is currently a senior researcher at Consumption Research Norway, Oslo Metropolitan University. Her research explores social inequality, health studies, and comparative welfare systems using social epidemiology and statistical methods.

References

Allison

(1984) Event history analysis. Thousand Oaks, CA: Sage Publications, Inc.,

Bakkeli

Drange

(2024) Relationship between payment problems and health: A nation-wide register study in Norway. Sociology of Health & Illness 46(7): 1364–1399.

Bazillier

Héricourt

Ligonnière

(2021) Structure of income inequality and household leverage: Cross-country causal evidence. European Economic Review 132: 103629.

Borgeraas

Poppe

Lavik

(2016) Consuming the home. Walking the thin line between welfare and catastrophe. Italian Sociological Review 6(1): 87.

Bourdieu

(1990) The logic of practice (R. Nice, Trans.). Stanford, CA: Stanford University Press.

Bourdieu

(2010) Distinction: A social critique of the judgement of taste. Abington: Routledge.

Carr

Jayadev

(2015) Relative income and indebtedness: Evidence from panel data. Review of Income and Wealth 61(4): 759–772.

Dackehag

Ellegård

L-M

Gerdtham

U-G

Nilsson

(2019) Debt and mental health: New insights about the relationship and the importance of the measure of mental health. European Journal of Public Health). 29(3): 488–493.

Davis

Cartwright

(2019) Deferred lives. In Featherstone

(Ed.) The Sociology of Debt). Bristol University Press; JSTOR. 1st ed, 91–118.

10.

Day

(2020) The short happy life of the affluent working class: Consumption, debt and embourgeoisement in the age of credit. Capital & Class 44(3): 305–324.

11.

Flemmen

Jarness

Rosenlund

(2018) Social space and cultural class divisions: The forms of capital and contemporary lifestyle differentiation. The British Journal of Sociology 69(1): 124–153.

12.

Frank

Levine

Dijk

(2014) Expenditure cascades. Review of Behavioral Economics 1: 55–73.

13.

Hansen

Flemmen

Andersen

(2009) The Oslo register data class scheme (ORDC). Final Report from the Classification Project. Memorandum 1. Oslo: University of Oslo, p. 22.

14.

Hayes

(2020) The behavioral economics of pierre bourdieu. Sociological Theory 38(1): 16–35.

15.

Hodson

Dwyer

Neilson

(2014) Credit card blues: The middle class and the hidden costs of easy credit. The Sociological Quarterly 55(2): 315–340.

16.

Hohnen

Gram

Jakobsen

(2020) Debt as the new credit or credit as the new debt? A cultural analysis of credit consumption among danish young adults. Journal of Youth Studies 23(3): 356–370.

17.

Jarness

(2015) Modes of consumption: From ‘what’ to ‘how’ in cultural stratification research. Cultural Sociology and New Forms of Distinction 53: 65–79.

18.

Jarness

(2017) Cultural vs economic capital: Symbolic boundaries within the middle class. Sociology 51(2): 357–373.

19.

Leicht

(2012) Chapter 11: Borrowing to the brink: Consumer debt in America. In Porter

(Ed.) Broke: How Debt Bankrupts the Middle Class). Redwood City: Stanford University Press. pp. 195–217.

20.

Lim

Bickham

Broussard

(2014) The role of middle-class status in payday loan borrowing: A multivariate approach. Social Work 59(4): 329–337.

21.

Lusardi

Mitchell

(2014) The economic importance of financial literacy: Theory and evidence. Journal of Economic Literature 52(1): 5–44.

22.

Oksanen

Aaltonen

Rantala

(2016) Debt problems and life transitions: A register-based panel study of Finnish young people. Journal of Youth Studies 19(9): 1184–1203.

23.

Poppe

(2008) Into the Debt Quagmire: How defaulters cope with severe debt problems. Doctoral thesis. University of Oslo.

24.

Poppe

Randi

Elling

(2016) The dangers of borrowing in the age of financialization. Acta Sociologica). 59(1): 19–33.

25.

Poppe

Kempson

(2021). Borrowing, Poor Health, Labour Market Marginalisation and Payment Problems. Conceptual Paper for the WellDebt Project Oslo: OsloMet - Oslo Metropolitan University, p. 26.

26.

Poppe

Lavik

Borgeraas

(2016) The dangers of borrowing in the age of financialization. Acta Sociologica 59(1): 19–33.

27.

Rabe-Heskett

Anders

(2012) Multilevel and longitudinal modeling using stata. Volume II: Categorical responses, counts, and survival. Stata Press. 3rd ed.

28.

Rendall

Brooks

Hillenbrand

(2021) The impacts of emotions and personality on borrowers’ abilities to manage their debts. International Review of Financial Analysis 74: 101703.

29.

Sparkes

(2019) Borrowed identities: Class(ification), inequality and the role of credit-debt in class making and struggle. The Sociological Review 67(6): 1417–1434.

30.

Statistisk Sentralbyrå (2012) Statistikkbanken kildetabell 12880.

31.

Sullivan

(2012) Debt and the Simulation of Social Class. New York: Oxford University Press.

32.

Sullivan

Warren

Westbrook

(2020) The Fragile Middle Class. New Haven and London: Yale University Press; JSTOR. https://doi.org/10.2307/j.ctvwcjdxd

33.

Swedberg

(2011) The economic sociologies of pierre bourdieu. Cultural Sociology 5(1): 67–82.

34.

Warren

Thorne

(2012) A vulnerable middle class. In Porter

(Ed.). Broke: How Debt Bankrupts the Middle Class. Stanford, CA: Stanford University Press. 1st ed., pp. 25–39.

35.

Weller

(2012) Unburdening America’s middle class. Challenge 55(1): 23–52.

36.

Wiedemann

(2021) Indebted Societies: Credit and Welfare in Rich Democracies. Cambridge University Press; Cambridge Core.

37.

OECD (2024), Household debt (indicator). (Accessed on 03 January 2024)

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.19 MB