The Influence of Social Media on Consumer Credit Behavior #BuyNowPayL8r

(1)

#BuyNowPayL8r

The Influence of Social Media on Consumer Credit Behavior

Lättman, Henrik; Rosdahl, Anton

Abstract:

Amplified by the introduction of the smartphone, which is now integrated in nearly every aspect of daily life, consumer credit-financed purchases are more accessible than ever and can now be spontaneously made in a matter of minutes. The aim of this thesis is to investigate the relationship between various financial decisions and time spent on social media (SMT). We theorize unprecedented exposure to the consumption and carefully managed narrative of others displayed on social media systematically make social comparisons unfavourable, implicitly causing young adults to overspend using consumer credit. Based on our hypothesized narrative, we stepwise explore the predictive power of SMT on (1) attitudes towards financing non-durable goods by instalment, (2) instalment lending, and (3) payment troubles. Through a web survey we collect individual measures of SMT, relative preferences, time preferences and various other variables we conjecture affect people’s economic attitudes and decisions in this context, such as feelings of inferiority as a prerequisite for envy. Our results show that SMT has a significant, positive effect on attitudes towards financing consumption of non-durable goods but does not further predict actual instalment lending or payment troubles. Also, individual time preferences are significant predictors of both attitudes towards- and realized instalment lending, whereas inferiority is positively significant for both instalment lending and payment troubles. Though, we refrain from drawing any strong inferential conclusions from these findings as clarified in the paper.

Keywords: social media, consumer credit, indebtedness, behavioral economics, time preference, social

comparison, relative- income and consumption, normative perceptions, envy.

Bachelor’s thesis in Economics, 15 credits Spring Semester 2020

Supervisor: Oben Bayrak Department of Economics

(2)

2

3 6.2 Variables of interest ... 50 6.2.1 Dependent variables ... 50 6.2.2 Independent variables ... 52 6.2.3 Control variables ... 53 7. Results ... 55 7.1 Regression modelling ... 58 8. Conclusions ... 63 References ... 64 Appendix I ... 69 Appendix II ... 75 Appendix III ... 80 Appendix IV ... 86 List of tables Table 1. 5-point Likert scale survey questions ... 32

Table 2. Social media shopping ... 34

Table 3. Payment methods ... 34

Table 4. Sources of debt ... 35

Table 5. Total amount of current debt ... 35

Table 6. Payment troubles I ... 36

Table 7. Payment troubles II ... 37

Table 8. Structure of the incentivized choice experiment ... 38

Table 9. Results from the incentivized choice experiment ... 39

Table 10. Results from hypothetical choice experiment ... 43

Table 11. Variable list ... 54

Table 12. Spearman Rank Correlation Test ... 55

Table 13. Regression output I ... 59

Table 14. Regression output II ... 61

List of figures Figure 1. Consumption loans ... 9

Figure 2. Consumption loan lending in Sweden, 2008 – 2017. ... 9

Figure 3. Debt collection process ... 12

Figure 4. Hypothetical scheme ... 26

Figure 5. Survey structure ... 28

Figure 6. Income distribution ... 29

Figure 7. Social media usage by measurement form ... 30

Figure 8. Social media usage by gender ... 31

(4)

4

1. Introduction

During the past decades, the way we finance our consumption has changed dramatically. A decade ago, given that the vendor even offered it, an online purchase would require a consumer to sit down by their computer to place the order, pay the full amount upfront, then wait a week or more for it to arrive. Today, the same purchase can be made on the phone while standing in line at a coffeeshop, requires no advance payment and can be delivered to your door within a few days.

While the number of indebted people in Sweden has decreased due to a booming economy and low interest rates, the total debt amount per individual is rising, along with the number of people receiving their first payment injunction from the Swedish Enforcement Agency (Kronofogden, 2019a). In conjunction with these patterns, regular consumables are more frequently being paid for by instalment instead of immediately (Finansinspektionen, 2018a; Finansinspektionen 2019) and there are increasing numbers of debts accruing from small, credit-financed purchases as well as large unsecured loans ending up in the registers of the enforcement- and various debt collection agencies (Kronofogden, 2019b).

In this thesis, we investigate whether traditional theories of social comparison can be applied in a social media context and whether feelings of envy towards others may trigger

overspending among young adults. Given that we spend about 2.5 hours per day on social networking online (Statista, n.d.), it is important to provide knowledge on how consumption behaviors are shaped in these digital environments.

Through a web survey we collect individual data from more than 400 respondents, with the large majority being students at the Gothenburg School of Business Economics and Law. Initially, we perform a correlation test to explore the underlying emotions elicited from spending time on social media. We then perform an exploratory analysis to see whether time spent on social media platforms positively affects;

1. Attitudes towards financing consumption of non-durable goods by instalment 2. Instalment lending

(5)

5

The predictive power of SMT on our three dependent variables is stepwise tested in chronological order to our hypothesized narrative using eight ordinal probit- and logit regression models. Also, we further examine whether present-biased time preferences, characterized by systematic impatience and impulsiveness, might help explain the increasing amounts of debt originating from flexible payment-schemes in e-commerce.

(6)

6

2. Background

In this section, we outline the market for consumer credit, different products, their

characteristics, usefulness and attached risks in terms of debt accumulation, as well as their role in the modern economy. Lastly, we walk through previous trends in rapidly rising digital culture that inspired us to conduct this thesis.

2.1 Indebtedness in Sweden

As stated in the introduction to this paper, young people between ages 18-25 and 26-34 have shown increasing difficulties regarding paying their debts on time, and an increasing number are ending up in the system of the Enforcement Authority1_{. While many of these debts are}

quite small, they are at the same time an indication that many people live with small financial margins and are at risk if the economy takes a downward turn (Kronofogden, 2019a;

Kronofogden 2019c). This is exactly what we have seen worldwide during the spring of 2020 as a result of the Covid-19 pandemic.

When we talk about indebtedness and problematic debt in this paper, we are taking a wider view than to simply look at default rates. Larsson, Svensson & Carlsson (2016) emphasize debt can be a large problem for the individual long before, or even without, a default. Given that it takes time to accrue severely problematic debt and increasingly common instalment purchases are often a part of these debts, it is important to look at young people in this context. The process to becoming so indebted that one seeks help is often protracted. In Sweden, every municipality is required by law to offer free budget and debt counselling. The average client using these services is a middle-aged man, while young people are a rare occurrence (Larsson et al. 2016). This does, however, not necessarily mean that the debt of people is a non-issue. From interviews with debt counsellors, Larsson et al. (2016) also learned that it takes time to accrue debt so severe that one asks for help in handling it. The estimated time is ten to fifteen years.

(7)

7

According to Finansinspektionen (2019), consumption loans account for 5 percent2_of

Swedish households’ total debt. While these sums are not considered a significant threat to financial stability, for certain individuals and households the interest they carry may be. Households already spend a substantial part of their income on interests and amortization. If their loans are large3_{, interest and amortization make up for a quarter of households’}

disposable income on average (Finansinspektionen, 2019). At the end of 2019, the annual growth rate of consumption loans was 8,3 percent, in part attributed to a booming economy (Finansinspektionen, 2020). Regardless, this growth is significantly higher than the growth rate of wages during the same time (Medlingsinstitutet, 2020; Medlingsinstitutet, n.d.) and should be seen in the context of households’ overall debt ratio, which has been growing steadily for more than two decades (SCB, 2019). People who accrue an unsustainable level of debt have usually tried to solve their situation themselves over 10-15 years, oftentimes

leading them to accrue additional debt as a means of making payments on the initial debt (Larsson, et al. 2016).

From the aforementioned observations follows that the debt of an individual often becomes a problem long before a potential default. Overindebtedness is related to several health

problems, such as an increased risk for both physical and mental illness, not limited to the person bearing the debt. Both the people with problematic amounts of debt and their children are, among other things, significantly more prone to commit suicide (Socialstyrelsen, 2013) making overindebtedness an intra-generational public health problem instead of just an economic one.

2_{The definition of consumption loans varies between different contexts. Finansinspektionen include loans that}

are collateralized by e.g. vehicles, agricultural properties and financial instruments in their reports frequently referred to throughout this section. When including such loans, consumption loans account for 18% of households total debt as a share of disposable income (Finansinspektionen, 2019). However, they are not relevant to the purpose of this study and we solely report non-collateralized debt figures by the definition of consumption loans adopted from SCB (2018) unless stated otherwise.

3_{Unsecured loans are referred to as large when the debt amount exceeds 100,000 SEK. Clarified by}

(8)

8

2.2 The market for credits

Credit enables consumption smoothing, allowing households to maintain their long-term level of consumption when their current income does not allow it, thus increasing flexibility and freedom in terms of financial planning (Roszbach, 2004). Although the market for credits has become highly diversified during the past few decades due to technological and financial ingenuity (SOU 2013:78), there are mainly two types of loans: collateral and non-collateral loans. They are similar in the sense that they both offer credit in exchange for interest and often other administrative fees, but for a collateral loan to be granted the borrower has to pledge a property, vehicle or some other asset to the lender as a security to help ensure that the borrower keeps up with their financial obligation (Finansinspektionen, 2019). Collateral and non-collateral loans can either be amortized straight (linearly) or as an annuity (non-linearly) and can either be revolving or issued with a fixed maturity date.

In this thesis we focus solely on consumption loans, defined by SCB (2018) as loans issued by non-governmental agents to non-firm households which do not require any collaterals, that have grown into a diversified and profitable market for lenders with various loan offerings directed towards different groups of consumers (Sveriges konsumenter, 2018).

Market characteristics for consumption loans are arguably problematic by nature, offering high-income individuals that are more creditworthy better loan terms, e.g. lower interest rates, than for low-income individuals that are not as credit worthy, all else equal.

Furthermore, consumer credit business models increase profits when additional fees are triggered due to delayed payments, motivating keeping customers with limited repayment capacity just about financially afloat while creating incentives for businesses to promote expensive, however flexible, payment schemes. In an article posted in SvD (Bratt, 2017, 11 May), Finansinspektionen comments on their report Consumer protection in financial

markets, raising concern for increasing overindebtedness fueled by unsound lending,

insufficient credit search, and the heavy marketing of flexible payment schemes from consumer credit institutes, their collaborating stores and businesses.

(9)

9

Notes: Invoices are excluded but can either be revolving or fixed.

2.3 Consumption loans

New lending of non-collateral loans, referred to in this thesis as consumption loans, has been steadily growing since the financial crisis of 2008, as shown in figure 2. According to SCB (2018) consumption loans account for approximately 5% of Swedish households’ total debt. Compared to loans collateralized by housing, cars, financial instruments or any other asset, consumption loans are typically issued with shorter maturity terms and higher interest rates as a premium to compensate the lender for the increased risk attached to relying solely on the borrower’s repayment capacity. However, a consumption loan is not necessarily interest bearing, e.g. an invoice is per definition a consumption loan.

Consumption loans Fixed maturity loans Unsecured

loans payment-planFixed

Revolving loans

Credt card

debt payment planRevolving

Figure 1. Consumption loans

(10)

10

2.3.1 Unsecured loans

In order to avoid confusion, we define unsecured loans as instalment loans issued with a fixed maturity date along with a pre-specified repayment- and amortization plan, payed as a lump sum to the debtor’s bank account. According to Finansinspektionen (2019), new lending of unsecured loans tends to grow with the economy and have been doing so at a rate of between three and ten percent annually since 2010. They are generally quite large and can be used for several purposes, including pure consumption and housing purposes such as to finance down payments, renovation or furnishing. However, unsecured loans are particularly interesting in the sense that they are the most occurring solution used to pay off other loans

(Finansinspektionen, 2019). Paying off other loans by taking new loans may be an indication of risky credit behaviour, accumulating more debt in order to continue living beyond current means, but a large proportion of the turnover in unsecured loans is due to better terms offered by a different credit institute.

2.3.2 Revolving loans

If there is no predetermined maturity of the loan it is referred to as revolving or open-ended. After a certain number of days, typically around 30, a minimum payment is required and the debtor can then choose to pay any amount equal to or exceeding that limit, offering high flexibility in terms of repayment. After a full or partial amortization, that same amount will then again be available for borrowing. These are the consumer credits of main concern in this thesis due to their characteristics, increasing popularity and particularly common in terms of spiralling debt accumulation (Sveriges konsumenter, 2018; SOU 2013:78).

Credit cards are typical revolving credits. After the credit application has been granted, the debtor can continuously make purchases for any amount within a specified credit limit without having to ask nor inform the lender. The same arrangement is offered by various stores, on- and offline, in collaboration with consumer credit institutes.

(11)

11

below 2,000 SEK is just about 30 percent of the loan size, whereas debts in the range of 2,000 – 5,000 SEK cost about 7% on average. However, these figures are averages and include such revolving credits that have triggered costs, but they could have been paid in full shortly thereafter, thus should be interpreted with caution.

Effective interest rates vary drastically depending on the purchase amount, amortization scheme and length, timing of repayments and the individual’s payment reliability. Credit institute Klarna offers a wide variety of payment schemes and currently handle the majority of e-commerce payments in Sweden, according to their Sweden Manager (Realtid, 2019). As an illustrative example, a 1,000, 2,500 and a 10,000 SEK purchase on a twelve-month flexible payment-plan offered by Klarna yield effective interest rates of 121, 41, and 9.4 percent respectively (Klarna, n.d.).

Credit cards on the other hand often come with perks, such as travel insurance and cashback bonuses, and their effective interest rates vary widely depending on the drawdown and granted credit amount, which itself depends on the consumer’s needs and repayment

reliability. Hence, comparisons are problematic, but as a rough illustration, one of the largest credit institutes in Sweden, Handelsbanken, offers an effective interest rate of 29.7% for a 50% drawdown on a twelve month granted credit amount of 5,000 SEK, and 14.6% for the same terms but for a granted credit amount of 20,000 SEK on their credit card called “Allkortet” (Handelsbanken, 2017).

(12)

12

2.3.3 Prerequisites for good credit practice

Credit companies’ revenues and costs are uncertain at the time of sale. Access to qualitative information that enable effective credit scoring and pricing is thereby vitally important when determining whether an application for credit should be approved or not, and what risk-adjusted loan terms to offer. In order to ensure sound credit practice4_{, institutes generally}

outsource part of their credit search to external parties with better access to information on applicants’ declared income, property holdings, addresses, misappropriated credit use, other credit commitments and total debt balances etc. (UC, n.d.). In a recent report on consumer protection, Finansinspektionen (2020) raises concerns about repeatedly irresponsible lending to credit-unworthy individuals and households due to insufficient credit search. In their most recent sample from 2018, almost 15% of all consumer loans5_{generated debt collection}

demands, although these figures are likely underestimated due to turnover from factoring (Finansinspektionen, 2019).

Regulatory restraints and lack of co-operation between institutes and credit search companies cause absence of up-to-date information on current income and debts among individuals to obstruct effective credit scoring. Finansinspektionen (2020) have pushed for the government to consider introducing a central debt register to come to terms with these issues. For

individuals who have trouble keeping up with their financial obligations, the process from due payment to suffer the more severe consequences from ending up in the registers of KFM is presented in a simplified6_{manner below.}

4_{Lenders are obliged to base their judgement on qualitative credit search and only grant credit if they assess that}

the consumer has sufficient conditions to fulfill their financial obligations (Finansinspektionen, 2020).

5_{These figures include consumption loans collateralized by other assets than housing.}

6_{This is the standard procedure for collecting debts due. However, the lender has no obligation to send a}

payment reminder before sending a collection claim demand. For detailed overlook, see Datainspektionen (n.d).

Payment Injunction Payment Reminder

Collection Claim Demand Payment Default

(13)

13

Private individuals who have not been able to pay neither a payment reminder nor collection claim demand will likely receive a payment injunction from KFM. Credit search companies then issue a mark that remains for three years in their registers as a warning sign for lenders, strictly limiting the marked individual’s possibility of being granted any further credit during this period (Datainspektionen, n.d). Not only is it necessary to maintain a clean payment complaint record to be able to enter agreements for rental housing, cell phone-subscriptions or other credit, being marked in the registers of credit companies is associated with a

considerable sense of shame. In a study specifically conducted to investigate the connection between digital consumption and overindebtedness among young adults, Larsson et al. (2016) interview debt counsellors in Swedish municipalities who experience “it is more acceptable to confess to alcohol or drug abuse than to confess a payment default that has resulted in a mark in the Enforcement Authority’s register” (p, 57-58), arguing people will do anything to avoid it. When possible, borrowing from relatives and other acquaintances is a common solution to cover short-term liquidity needs, but desperate individuals tend to reach out for more

consumer credit wherever their application may be granted. From this point of view, credit searches are often counterproductive, motivating behaviours that continuously seem to cause individuals to accumulate spiralling debt rather than facilitating sustainable lending (Larsson, et al. 2016).

Due to recent technological development within application- and extensive data processing systems, lenders today have far more information on their customers, allowing them to

develop differentiated loan offerings, scoring- and pricing methods more accurately (Caufield, 2012). According to Caufield (2012), most companies today use risk-based pricing methods, predicting the probability of default for each consumer based on their financial, personal, and demographic characteristics. In some default-risk models, details such as what hour of the day the application is made and even the name of the applicant’s email address provider is

(14)

14

2.4 Digital trends

The revolutionary technological development of the Internet has had major impact on our daily lives, changing the way we work, consume, socialize and share information.

KFM analyst Johan Krantz predicts the trend of rapid growth in payment injunctions sourcing from flexible payment schemes offered by Klarna, Qliro, and other popular providers of consumer credit in e-commerce checkouts in Sweden will likely persist (Bränström, 2019, 16 December). Debt collection agency Lowell (2019) share that view, highlighting the

problematic influence of social media where consumer behaviors of young adults are being shaped.

Inspired by Taylor and Strutton (2016), we define social media as digital platforms used for interactive purposes that allow users to construct a profile within a bounded system where they share content with each other. In Sweden, 90% of young internet users in the age of 15-25 use social media on a daily basis with the most popular platforms being Facebook, Instagram and Snapchat (Internetstiftelsen, 2019). This fundamental shift has attracted

businesses to increase their presence on digital platforms, spending more money and effort on social media marketing (Stephen, 2016). For instance, influencers7_{promote sponsored}

products and embody brand identities by showing off desirable lifestyles in partnerships with brands. Also, e-marketers can use the digital trace we leave behind when using social media to specifically tailor advertisement that show up in our feeds based on user activity.

The field report performed by Lowell (2019) called Social media is pushing for

overconsumption8_{strongly indicates that those who purchase things they stumble upon in}

their social media feeds have higher tendencies to spend above their means using consumer credit. In fact, almost 60% of respondents in their sample who financed a social media-inspired purchase by lending admit they regularly consume more than they can afford. Larsson et. al (2016) argue young adults using social media tend to click ads, check in at stores and do more research on products using their cell phones more than others, making them particularly susceptible to online marketing.

7_{Someone who influence their large number of followers on social media in exchange for personal rewards from}

brand partnerships (Audrezet, de Kerviler & Moulard, 2018; Abidin, 2015).

(15)

15

Whether driven by changing attitudes towards credit, FOMO9_{, short-term impatience or some}

other influence, young adults who draw their consumption inspiration from social media show tendencies of overspending. Given the considerable amount of time we spend using social media and its ever-changing nature, up to date economic research on these topics are important to provide policy makers and the public with knowledge to maintain solid consumer protection.

9_{A trendy acronym for “Fear of Missing Out”, which is closely related to social comparison and the well-known}

(16)

16

3. Theoretical Framework and Previous Research

In this section, we introduce the reader to relevant theoretical background in the behavioral economic field in terms of explaining credit financed consumption. Then we present

somewhat overlapping academic psychological literature of particular relevance in explaining subjective well-being in a digital social media context.

3.1 Intertemporal choice

Trade-offs between current and future benefits/costs are central in both personal and public decision making. Discounted-utility models in various forms handle such time-considerations, providing decision-makers with theoretical ground to make well-informed intertemporal choices. Parameters incorporated into these models, along with their many implicit

assumptions, must be chosen with great caution because they can have significant effect on outcome, especially over long time-horizons (Shane, Ewenstein & O'Donoghue, 2002).

The intuition behind discounting monetary values (money) is quite tangible. Receiving 100 SEK today is, under normal circumstances10_{, worth more than receiving 100 SEK in one year.}

This is simply because one could, in principle, deposit the money into a savings account, wait for one year, then withdraw the 100 SEK plus interest earnings gained during that time. Choosing to instead collect the 100 SEK in one year would imply missing out on these interest earnings, which in economic terms is referred to as an opportunity cost. The savings interest rate in this illustrative example, where we ignore inflation and consider bank savings risk-free, would be the discount rate used to calculate how much 100 SEK received in one year is worth, in terms of today. The net present value would then be less than 100 SEK, where the amount size in decreased value is determined by the time delay and interest rate. The reader is encouraged to imagine replacing money in the above example with, say, the utility gained from purchasing the latest iPhone, or their grandchildren’s utility gained from environmental quality, to see how one’s perception of how the net present value of that good might change and to better understand the implications that occur when dealing with longer time-horizons and more complex issues involving uncertainty, ethical considerations and

10_{The effective interest rate of many bank savings accounts in Sweden today are negative, which may arguably}

(17)

17

individuals’ subjective utility. Dietz, Hepburn & Stern (2008) provide meaningful insight on such issues and perhaps most accurately describe utility in a more traditional economic manner as welfarism, although in the context of this thesis subjective well-being may seem more appropriate. In the remaining part of this section, we primarily focus on determinants of utility and how they affect economic behavior.

In this simple multiple-period utility discount function,

𝑈_! = 𝑢_!+ 𝑢!"# (1 + 𝜌)+ 𝑢_!"$ (1 + 𝜌)$… 𝑢_!"% (1 + 𝜌)% (1)

the value of total utility in terms of today 𝑈_! equals utility gained today plus utility gained in all future periods 𝑡 + ⋯ 𝑛, exponentially discounted by their corresponding discount factors (1 + 𝜌). The discount rate 𝜌 reflects how much the individual values utility gained today over utility gained in the next future period. Similar to how cash flows are discounted to maximize expected return on investment, the rational individual behaves as to maximize the sum of all future utilities (Samuelson, 1937). All possible motives underlying an individual’s time preferences are compressed into a single parameter 𝜌, making the model

comprehensible, yet limited and filled with questionable assumptions. For instance, the

exponential model presented above assumes utility gained from all possible utility bundles are discounted by a constant rate, regardless of size, in all time periods. All else equal, the

exponential utility function assumes discount factors are equal for all 𝑡, implying it should not matter whether the delay itself occurs today, six months or 100 years from now.

(18)

18

3.1.2 Time preferences

Ericson and Laibson (2018) argue that “all animals, including humans, tend to pursue instant gratification, even when such immediate rewards are obtained by foregoing a substantially larger amount of delayed gratification” (p, 5). In other words, individuals tend to value the present over the future, even though the absolute value of future utility might be substantially larger than the absolute value of present utility. Such tendencies, causing utility gained from purchasing something today to systematically outweigh the discounted displeasure of future payments, are referred to as present bias. However, Ericson and Laibson (2018) prefer to use the term positive time preferences to avoid insinuating the behavior is necessarily irrational nor a mistake. In discrete hyperbolic discounting form, time preferences can be

mathematically expressed

𝑈!= 𝑢!+ 𝛽[𝛿𝑢!"#+ 𝛿$𝑢!"$+ 𝛿&𝑢!"&… 𝛿%𝑢!"%] (2)

where the total utility 𝑈! equals the sum of flow utilities 𝑢 in periods 𝑡. The bias parameter 𝛽 indicates to which extent one systematically over- or undervalues the future, and parameter 𝛿 is the individual discount factor, equivalent to discount factor 𝜌 in equation (1). Generally, the bias parameter and individual discount rate is assumed to be 0 < 𝛽 < 1 and 0 < 𝛿 ≤ 1 respectively, hence future utility flows are less valued than present (Shane et al. 2002).

(19)

19

argue significantly correlates with the degree of credit card borrowing in a given sample. On the other hand, possible explanatory factors for such inconsistencies include risk-aversion, because the future involves uncertainty, or the need for short-term liquidity to be able to keep up with one’s financial obligations. Such implications are discussed further in section 5.4.

Per definition, the shape of the hyperbolic discount function simultaneously motivates short-term impatience and long-run patience. However, as time goes on and the delay approaches, the hyperbolic discount function predicts the individual will yet again prefer the instant reward when faced with the exact same prerequisites. In other words, present biased individuals are dynamically inconsistent because all future selves value utility bundles differently from present self, causing future selves to systematically fail to follow up on the intentions of present self. For instance, Laibson (1997) empirically shows positive time preferences can help explain why a considerable number of households allocate long-term savings at a relatively low growth rate while simultaneously accumulating sizable credit-card debts at higher cost.

Analyzing patterns of consumption for non-durable goods among UK households, Stephens (2006) empirically provides large-scale evidence for such short-term impatience, showing households who get their income paid monthly have higher tendencies to overspend shortly after paycheque receipt, compared to weekly paid households, even though the timing of cash flows are perfectly consistent and previously known. Kuchler and Pagels (2018) reach similar conclusions, using financial data from American individuals bank accounts and credit cards, showing that systematic patterns of eagerness and lack of self-control cause individuals to significantly fail to repay their credit card debt according to plan, however provides further insight highlighting that individuals who are self-aware of their future short-term impatience show better ability to repay as planned.

(20)

20

3.2 Relative Income and Consumption

Traditional economic theory assumes only absolute quantities of income and consumption matter for individual well-being. However, there is extensive empirical evidence and a growing consensus among researchers supporting relative position perceived through interpersonal comparisons have significant impact on subjective well-being, thus economic behavior (Johansson-Stenman, Carlsson & Daruvala, 2002; Frank, 2005a). Also, marginal utility from absolute income and consumption diminishes as countries become richer (Layard, Mayraz, & Nickell, 2008; Clark, Frijters & Shields, 2008), implying absolute parameters are utility-dominant only up to some certain point of wealth. As standard of living increases world-wide, interdependent utility functions might further influence important policy issues in the future, such as for optimal taxation (Frank, 1985a), environmental quality and the

provision of other public goods (Brekke & Johansson-Stenman, 2008; Cornforth, 2009), whereas the authors of this thesis aim focus towards behaviors that may lead to

overconsumption and unsustainable use of financial credit.

In an experimental survey performed by Solnick and Hemenway (1998), Harvard students and faculty staff were asked to choose between two hypothetical stages of the world, where only the allocation of some good differed between the two stages. In the “positional” state, respondents had more of the good in question compared to others in society. In the “absolute” state however, both respondents and others in society had more of that same good than in the “positional” state, but respondents had less than others. In this sample, results show positional concerns regarding income are extremely important, more so for students than for faculty staff. Frank (1985b) even argues income is more positional than leisure, hence if the standard economic assumption that utility gains from income rather comes indirectly from

consumption, then per definition relative consumption must matter for individual utility (Alpizar, Carlsson & Johansson-Stenman, 2005).

(21)

21

(Mason, 2000), or even compete in the marriage market (Frank, 2005b). Following this theoretical reasoning, students at Harvard in the above example may therefore, at least partly, make a rational choice in sacrificing substantial proportions of real purchasing power,

attractiveness and education in exchange for increased relative standing. On the other hand, Akay (2019) shows certain personal traits, such as neuroticism, are strongly associated with positionality, which implies relative concerns may also source from personality traits generally perceived less desirable.

Positional value of particular goods differs between time-periods, cultures and individuals, although consumption has always played a vital role in the embodiment of personal identity (Taylor & Strutton, 2016). In general terms, “goods that effectively signal a desirable characteristic of the owner are more positional than other goods” (Johansson-Stenman & Martinsson, 2006, p, 131). Naturally, visible goods are generally more positional than non-visible goods, simply because they reveal information directly by observation. More specifically, Johansson-Stenman and Martinsson (2006) exemplify cars, jewellery and housing as typical positional goods, however emphasize less conspicuous goods may have better signalling value for other ideals, such as concerns for values beyond materialistic ones. Positional goods historically11_{consumed by privileged minorities in primarily Western}

societies include top-hats, front-row theatre tickets, luxurious leisure activities in general, and various rarities such as art, gems and tulips, motivated by desire to signal social status and economic power rather than practical needs (Mason, 2000; Schneider, 2007).

For whatever reason there is additional utility to be gained from consuming conspicuously, other individuals with high relative concerns are subjectively worse-off when exposed to those who are seemingly better off, all else equal. When exposed to such consumption by others, consciously or not, one might compensate for the utility lost by increased spending to keep up with those who are seemingly better off (Krause et al. 2019). Bowles and Park (2005) empirically show the average amount of work hours within high standard of living OECD countries increases by the degree of income inequality, and further argue such reallocation of time reveal large-scale emulation tendencies among the populations to keep up with the socio-economically superior.

11_{For extensive overview on historical economic theory regarding conspicuous consumption, see Schneider}

(22)

22

In general, individuals who make upward social upward comparisons are more concerned with relative consumption than those who make downward social comparisons (Andersson, 2008), which motivates tendencies of upward emulation rather than downward distancing, causing subjective costs to spiral down the income distribution (Bowles & Park, 2005). In other words, low-income groups are more vulnerable of such inequalities, however Andersson (2008) provides insight emphasizing there is no indication of relative concerns to decrease with age nor depend on gender.

3.3 Envy

Concerns for relative consumption and income reveal individual preferences are not perfectly pure and completely disentangled from any influence of others, as traditionally assumed (Alpizar, et al. 2005). Pollak (1976) has provided important contributions in incorporating personal interdependency to economic modelling, allowing for more accurate approximates of economic behaviour in contexts where interpersonal social comparisons appear relevant for individual utility.

In its simplest functional form, the linear two-individual model of interdependent utility from consumption is mathematically expressed

𝑈_' = 𝑢_'(𝑐_') + 𝜀 𝑢₍(𝑐₍) (3)

where individual A’s total utility 𝑈' equals the utility gained from her own consumption 𝑢_'(𝑐'), plus the utility lost from the fraction 𝜀 of individual B’s utility from consumption 𝑢₍(𝑐₍). The size of 𝜀 may depend on the individual’s relative concerns, the good consumed by individual B and its corresponding degree of positionality, or some underlying

psychological influence mitigating feelings of envy. Extended versions of the model

(23)

23

Elster (1991) accurately describes envy as a negative emotion-based state of mind derived from unfavourable social comparisons towards “another person's success, happiness,

intelligence, health, good looks, sunny disposition, character, knowledge, wealth, before-tax income, after-tax income, material possessions, power, title, job, or status – or even his

freedom from enviousness” (p, 50), all of which can be widely defined as goods. Although, an important notation is that we define envy as the emotion elicited from the feeling of

inferiority which negatively affects individual utility, whether triggered by neuroticism or some other underlying psychological influence.

Smith (2004) lists four conditions for envy to arise which can be briefly summarized; a) apart from the desired good, the envious share personal similarities with their comparative

reference, b) the desired good is personally relevant, c) the desired good is imaginable but beyond reach, and d) the disadvantage is seen as unfair. In line with Smith’s (2004) similarity condition, Pollak (1976) assumes envy is stronger in his economic modelling when physical or emotional distance is shorter between the individuals and their comparative reference. Also, Elster (1991) distinguishes between conditions c) and d) by illustrating the difference between “it could have been me” and “it should have been me” (p. 52-53).

3.4 Social media

Bowles and Park (2005) argue the influence of a comparative reference group may be completely decoupled from its size, enabling conspicuous consumption of seemingly privileged minorities to affect interdependent utility functions of potentially much larger majorities. In a social media context, the job-description of influencers is to impact their large number of followers by showcasing certain products and services for potential customers (Krause et al. 2019; Chae, 2017) through modern versions of traditional word of mouth recommendations or product placement (Stephen, 2016). This form of marketing, where regular interaction with followers is common and the influencers share posts consisting of photos, videos and texts from their everyday lives, is generally perceived more authentic compared to traditional media (Abidin, 2015). As argued by Chae (2017), influencers emotionally appear closer to ordinary people than traditional celebrities.

(24)

24

mediate. Thus, it is within both parties interest the influencer can present a widely desirable lifestyle and that the marketed product is part of that narrative (Audrezet et al. 2018). Frequently showing off their more or less financially sponsored exotic vacations, gourmet meals and exclusively branded wearables (Taylor & Strutton, 2016), influencers often portrait the life young people can only dream of living (Chae, 2017). In a survey-based study

performed on German students, Krause et al. (2019) show exposure to Instagram positively predicts overspending, arguing feelings of envy mediated from observing the conspicuous consumption of others on Instagram trigger a self-enhancing behavioural response to close the perceived gap between oneself and envied others, often through increased consumption of such conspicuous goods.

However, it is important to note there are influencers directed towards different target groups and that they do not necessarily showcase abundant lifestyles, promote expensive nor

conspicuous products, but a considerable part of all content shared on social media evolves around consumption of goods (Krause et al. 2019). Frequent self-comparison through social media has empirically proven to provoke feelings of envy (Krause et al. 2019; Tandoc, Ferrucci & Duffy, 2015; Krasnova, Wenninger, Widjaja & Buxmann, 2013), sometimes even depression (Tandoc et al. 2015), and passive users seem more susceptible to such

unfavourable comparisons than active users (Krasnova et al. 2013). However, given the opportunity to carefully shape the portrait of our lives, the average social media user’s profile also generally present a more idealized self, highlighting and sometimes even staging positive experiences, possessions, attributes and accomplishments while, perhaps more importantly, leaving out the boring bits (Krause et al. 2019; Taylor & Strutton, 2016). In a survey performed by Chou and Edge (2012), active Facebook users were more likely to agree that “others have better lives than themselves” and less likely to agree that “life is fair”, arguing perceptions become skewed because social media users must use the limited heuristics that are available to form impressions of others in their network.

(25)

25

4. General Purpose & Hypotheses

In recent years, a large body of economic research has directed attention towards how businesses adapt to the large-scale relocation of social activity to online social media platforms, providing insight on how consumer preferences and behaviors are effectively shaped in digital environments (e.g. Audrezet et al., 2018; Krause et al., 2019; Stephen, 2016; Taylor & Strutton, 2016). On an individual level, psychological researchers frequently

contribute with scientific depth with regards to mental side-effects elicited from frequent social media usage (Chou & Edge, 2012; Krasnova et al., 2013; Tandoc et al., 2015), but to our knowledge, little economic research empirically examine whether the unprecedented exposure to the consumption, achievements and carefully managed narrative of others displayed on social media systematically trigger feelings that eventually may cause

consumers to overspend to such an extent that they accumulate unsustainable debt amounts.

In fact, Swedish debt collection agencies have raised public concern for the increasing debts among young adults emerging from expensive payment schemes commonly offered in e-commerce check-outs, specifically highlighting the problematic influence of social media and particularly influencer marketing (Lowell, 2019). Based on previous economic research and somewhat overlapping psychology literature, the purpose of this thesis is to investigate

whether such exposure systematically lead young adults to finance consumption unsustainable by their current means by using consumer credit. Our theory is that the inherent nature of social media and the endless social comparison that comes with it is the main source of emotions that trigger consumption desire in this context, and that flexible payment schemes lower the thresholds to act on that impulse. Even though we assume floods of underlying psychological influences combined, rather than one single emotion, trigger behavioural responses that may translate into credit-financed overconsumption, we narrowed it down to one, yet multi-layered, emotional source: envy.

(26)

26

insight on whether social media exposure might affect the interdependent utility functions we conjecture motivate credit-financed consumption, based on the presented theoretical

framework, previous literature and our preconceptions, we initially explore the following sub-hypotheses:

h: There is a positive relationship between the amount of time spent using social media and 1) consumption desire evoked from social media,

2) positional concerns, 3) envy towards influencers, 4) superficiality, and

5) feeling inferior to others.

The daily average amount of time spent on social media is the main explanatory variable used to widely capture these underlying influences provoked by unfavourable social comparisons, whereas positive time preferences provide the direct link to the impatience that increases propensity to use consumer credit. In addition, we believe attitudes towards consumer credit as a reasonable way of financing consumption of non-durable goods are more relaxed among frequent social media users, partly due to the heavy marketing of such flexible payment schemes in digital environments and the way they are often presented as a solution to short-term liquidity needs. In line with the sole purpose of this thesis, we stepwise formalize our three main hypotheses:

H1: Spending more time on social media increases the likelihood of being willing to purchase

non-durable goods on instalment.

H2: Spending more time on social media increases the likelihood of actually purchasing

goods in instalment.

H3: Spending more time on social media increases the likelihood of having payment troubles.

Social Media

Social Comparison

Credit-Financed

Consumption Payment Troubles

Envy Skewed Normative

Perceptions

(27)

27

5. Data

This section is structured as follows: first we outline the process of collecting our data. Then we present and briefly comment on relevant descriptive statistics to provide an overview our sample characteristics. We then walk through the experiments used to measure individual economic preferences. Lastly, we critically discuss quality of our data.

The data was collected using an online survey created in Google Forms. Initially, we sent a trial-version of the questionnaire to our Facebook-friends (N=81) which allowed us to test questions before sending out the real questionnaire to students. Apart from a few small technicalities, the key takeaway point was that more than one hypothetical question in repeated choice form was too heavy. We discuss such implications, weaknesses in our data collection process and provide suggestions for future researchers at the end of this section.

5.1 Collection process

A total of 447 respondents completed the online questionnaire which was distributed on May 6th_{via email to about 3,000 addresses collected from the internal system of Gothenburg}

University, resulting in a participation rate of about 15%. The opportunity to win a lottery price of 1,000 SEK was highlighted in the email in order to attract respondents. Once directed to the questionnaire through a clickable link, participants were informed that we hoped this would motivate them to answer carefully.

Moreover, participants were informed about the study purpose and the estimated time to complete the survey. They were also assured their answers would remain anonymous, whether choosing to participate in the lottery or not, and that we have no interest in knowing the answers of any specific person. Due to the sensitive nature of some questions this message was regularly repeated throughout the survey. In order to collect the email-addresses for those who wanted to participate in the lottery without breaking anonymity, respondents were then redirected to an external page to fill out their contact details. Participants in the Facebook pilot study were also included in the lottery and the winner was randomly drawn May 28th

(28)

28

The questionnaire was divided into five parts, roughly structured the following way:

The complete survey is found in Appendix IV, along with a copy of the email invitation and extensive descriptive statistics in addition to the sample characteristics of particular relevance we present throughout this chapter.

5.2 Descriptive statistics

Three participants prefer not to be defined by gender. One respondent has reported a larger disposable income than 40,000 SEK per month and only one respondent is more than 50 years old, hence these are all dropped for comparative reasons. As we will show, an additional 11 respondents are also dropped, leaving a total sample of 431 observations.

Our typical respondent is a female in the age of 21-25, who lives in one of the three largest cities in Sweden with her partner or spouse, is three years through university studies in business and economics, whose only source of income is study grants and loans. She has no debts apart from student loans, uses Instagram most frequently and spends more than two hours per day using social media

Out of 431 respondents; - 90% are students,

- 74% are 25 years of age or younger, - 60% are females, and

- 83% lives in a large city (Göteborg, Stockholm or Malmö).

The top three most frequent programs studied among university students in our sample are business and economics (49%), political science (19%) and law (17%). University studies are free to Swedish citizens, and most students receive financial support from the government

Estimated time to complete survey: 10 minutes

Demographics Internet Habits _{Circumstances}Financial Personality Multiple Choice _Experiment

(29)

29

during post-secondary studies in the shape of a mixed aid between study grants and loans, issued with generous interest rates in an already low-interest economy. In general, a full-time student receives roughly 3,000 SEK in grants plus 7,000 SEK in loans per month, although there are various supplementary allowances and income restrictions that could affect the amount size.

Regardless, respondents were instructed to choose the income interval of 10,001 – 11,000 SEK if study grants and loans is their only source of income. 35% of students have reported other incomes in addition to study grants and loans, but there are no differences worth

mentioning between demographical categories with regards to the income distribution shown in Figure 5.

In the second section of the survey, respondents were asked several questions regarding their internet habits, initially instructed to report their daily average time spent on social media. iPhone-users were encouraged to use the Screen time-tracking feature and report the measured daily average time spent on Social networking as an objective proxy for social media time, whereas users of Android or other smartphone operating systems were asked to use their equivalent feature. What specific social networks an Android-phone measures varies, but the

Social networking category in iOS compiles a daily average amount of time spent on all the

main social media platforms together based on user activity during the past week. Included

(30)

30

platforms are Facebook, Instagram, Snapchat, LinkedIn etc., but also applications mainly used for texting, such as WhatsApp and Messages. These apps are less relevant for the sole

purpose of this thesis and most people probably would not think of them as social media. Hence, time measures from 55% of the respondents who stated their screen-time feature reports may be somewhat overstated, compared to the 45% who stated their self-estimated measures.

As shown in figure 6, self-estimated measures seem slightly lower than those calculated by screen-time features. However, the distribution between respondents’ measurement reports are satisfactorily similar, thus we conclude reported measures overall or reliable

approximations of respondents’ real daily average time spent using social media.

In comparison, the daily worldwide average time spent on social networking per individual internet user was 2.5 hours in 2019, which equals a 60% increase since 2012 (Statista, n.d.). Furthermore, judging from Figure 7 below, females in our sample seem to spend more time on social media than males.

(31)

31

Figure 8 shows Instagram is the most frequently used social media platform among our respondents, followed by YouTube. Only one person chose Pinterest, and several respondents who answered “Other platform” specified they use the online forum Reddit most frequently. However, there are distinctive differences between genders regarding which platform is mostly used. 53% of females use Instagram most frequently, whereas 46% of males mostly use YouTube.

Figure 8. Social media usage by gender

(32)

32

The nature of content varies between social media platforms. When asked what sort of

content respondents actively follow and/or frequently engage with on social media, only 18% males stated they follow influencers whose content is characterized by either lifestyle, travels and luxury or by fashion, interior and design, whereas 62% females follow at least one of such profiles.

In the questions presented in table 1, respondents were instructed to rate their level of agreement to a given statement on a 5-point Likert scale (1 = Strongly disagree, 3 = Neither agree nor disagree, 5 = Strongly agree). Questions were given in different sections of the survey and are presented in chronological order.

Table 1. 5-point Likert scale survey questions

N = 431 Likert Scale Answers (%) Descriptive

# Statements 1 2 3 4 5 Mean P50 𝜎

1

I spend a lot of time surfing around in online stores

29.93 29.70 21.11 14.62 4.64 2.34 2 1.18

2

I spend a lot of time walking around in physical stores

46.64 34.11 14.15 4.41 0.70 1.78 2 0.90

3 I often comment content posted by others on social media

39.44 27.38 17.87 9.51 5.58 2.15 2 1.21

4 I often post pictures or other content on social media

27.61 29.47 17.40 11.83 13.69 2.55 2 1.36

5 People generally present a beautified version of their lives on social media

1.16 0.46 7.66 33.87 56.84 4.45 5 0.75

6 Social media makes me desire to consume

19.26 24.13 23.20 25.52 7.89 2.79 3 1.24

7

I envy the influencers’ life shown on social media

39.91 27.38 21.81 8.35 2.55 2.06 2 1.09

8

My life is inferior to influencers’ life shown on social media

41.53 25.52 20.42 10.21 3.32 2.06 2 1.11

9

I wish to live like influencers on social media

(33)

33

10 I always buy more than planned

22.04 35.96 20.19 13.69 8.12 2.50 2 1.21

11

I always manage to save money at the end of the month

8.35 15.31 14.62 19.03 42.69 3.72 4 1.37

12

I have someone close to me that would help me if I have payment troubles

1.62 2.09 6.26 20.42 69.61 4.54 5 0.83

13 I have a large circle of friends

6.96 16.24 32.48 27.15 17.17 3.31 3 1.14 14 I care a lot for my looks 1.16 5.10 27.38 47.80 18.56 3.77 4 0.85 15 I like luxury 7.42 16.01 32.25 29.70 14.62 3.28 3 1.12 16

The clothes and goods I have highly reflects my identity

13.69 29.47 32.02 18.10 6.73 2.75 3 1.11

17

Others in my

acquaintance are better off than I am

13.92 24.36 37.12 17.87 6.73 2.79 3 1.10

18 I have had less help than

others through life 32.95 21.58 26.45 12.53 6.50 2.38 2 1.24

Notes: 𝜎 = standard deviation, P50 = median, 10 questions excluded12_.

Questions 3 and 4 are adopted from Taylor and Strutton (2016) in order to measure the extent to which respondents interact with others online. Even though most respondents seem to spend considerable time using social media on a daily basis, they score fairly low on these questions. We also expected participants to spend more time surfing around in online stores and about 95% of respondents disagree they spend a lot of time in physical stores, implying they draw their inspiration on what to consume from elsewhere. About a third of all

respondents admit consumption desires are drawn from their social media usage.

Questions 6, 7 and 8 regarding envy towards influencers are directly adopted from Chae (2017) and respondents in our sample generally score fairly low points measured by the Likert scale. However, a remarkable share agrees people generally tend to present beautified

versions of their lives on social media. In terms of relative standings compared to others, reported answers are widespread. Even though 90% of respondents are students, a

considerable share claim they manage to save money each month and almost 70% strongly agree they have someone close to them that would help out in case of experiencing financial

12_{Several questions were completely excluded because of time-restraints, problematic distributions or they fail}

(34)

34

difficulties. The reader should note the survey was given in Swedish, meaning these questions have all been translated into English for presentational purposes.

Table 2 below shows many respondents draw inspiration on what to consume from social media and tend to purchase goods they come across on these platforms.

Table 2. Social media shopping

Have you bought something you saw on social media during the past year?

N = 431 Freq. Percent Cum. (%)

Marketed good/service Yes, multiple times 48 11.14 11.14 Yes, once in a while 178 41.30 52.44

Never 205 47.56 100

Non-marketed good/service I have come across

Yes, multiple times 46 10.65 10.65

Yes, once in a while 160 37.04 47.69

Never 205 47.45 100

Have used voucher from

sponsored posts Yes, multiple times Yes, once in a while 159 60 13.92 36.89 13.92 50.81

Never 212 49.19 100

In the last question regarding online shopping behavior, respondents were asked to state what payment methods they have used during the past two years. The attentive reader will notice we have sometimes used different timeframes. These are chosen carefully so that respondents should be able to remember their previous consumption behaviours and two-year timeframes are specifically used in questions regarding instalments due to the fact that 24-month

maturities are commonly offered in e-commerce checkouts.

Table 3. Payment methods

Have you made a purchase using any of the following payment methods during the past two years?

N = 431 Freq. Percent Cum. (%)

Invoice Multiple times 202 46.87 46.87 Once in a while 133 30.86 77.73 Never 96 22.27 100 Credit card Multiple times 87 20.19 20.19 Once in a while 28 6.50 26.69 Never 316 73.32 100

Instalment Multiple times Once in a while 15 55 12.76 3.48 16.24 3.48

Never 361 83.76 100

Quick loan

Multiple times - - -

Once in a while 2 0.46 0.46

(35)

35

Moreover, questions in survey section three regarding respondents’ financial circumstances are presented in chronological order in tables below, initially mapping out sources of debt.

Table 4. Sources of debt

What loans/debts do you currently have? (N=431)

Loans Freq. Percent

Instalment loan 18 4.18

Credit card debt 11 2.55

Unsecured loan 6 1.39

Other loans/debts 29 6.73

No loans 72 16.71

Student loans 357 80.77

Notes: Respondents can have several loans. The 19.23% respondents who have loans collateralized by cars or housing areexcluded.

A surprisingly small share of respondents in our sample seem to currently have any

consumption debt. In the next question, respondents stated their total amount of current debt, excluding student loans. Their answers are presented in table 5 where we have once again excluded respondents with loans collateralized by housing or cars.

Table 5. Total amount of current debt

Apart from student loans, what are your current debts/loans? (n=378)

Debt intervals (SEK) Freq. Percent Cum. (%)

No debts 329 87.04 87.04 1 – 500 4 1.06 88.1 1,001 – 1500 1 0.26 88.36 2,001 – 3,000 1 0.26 88.62 3,001 – 4,000 2 0.53 89.15 4,001 – 5,000 2 0.53 89.68 5,001 – 6,000 3 0.79 90.78 6,001 – 7,000 1 0.26 90.74 9,001 – 10,000 1 0.26 91.01 10,001 – 15,000 2 0.53 91.53 20,001 – 30,000 2 0.53 92.06 30,001 – 50,000 4 1.06 93.12 50,001 – 100,000 5 1.32 94.44 100,000 – 250,000 17 4.50 98.94 250,000 – 500,000 4 1.06 100

(36)

36

Total debt amounts are widespread, and an important notation is that about a quarter of respondents’ total debt amounts exceeding 100,000 SEK consists of “Other loans/debts”. These are issues we further address in section 5.4.

In the last part of the survey section regarding financial circumstances, respondents were asked several questions about how well they are able to keep up with their financial

obligations. Keeping in mind 90% of respondents are students with relatively low income, the degree of payment difficulties regarding specific expenses vary widely. Even though

recurring issues paying rent and cellphone subscriptions are somewhat common, respondents generally seem to manage paying their expenses on time.

Table 6. Payment troubles I

Have you had trouble paying any of the following expenses during the past two years? (N=431)

Expenses Answers Freq. Percent Cum. (%)

Rent

Never 390 90.49 100

Housing-cooperative fee

Never 425 98.60 100

Interest/amortization

Never 426 98.84 100

Electricity, water, gas

Never 425 98.61 100

Cellphone subscription Yes, multiple times Yes, once in a while 12 8 1.86 2.78 1.86 4.64

Never 411 95.36 100

Broadband, TV Yes, multiple times Yes, once in a while 2 6 0.46 1.39 0.46 1.85

Never 423 98.14 100

Credit card debt Yes, multiple times Yes, once in a while 13 2 0.46 3.02 0.46 3.48

Never 416 96.52 100

Other invoice Yes, multiple times Yes, once in a while 23 5 1.16 5.34 1.16 6.50

Never 403 93.50 100

Leasing fee

(37)

37

Alimony Yes, multiple times Yes, once in a while 2 - 0.46 - 0.46 -

Never 429 99.54 100

Backtax

Yes, multiple times - - -

Never 414 96.06 100

In order to collect specific data reflecting respondents realized payment defaults rather than self-estimated perceptions of payment troubles, we presented respondents with the following questions while repeatedly reminding respondents their answers would remain anonymous.

Table 7. Payment troubles II

1. Have you received a payment reminder during the past year? (N=431)

Freq. Percent

Yes 140 32.48

No 275 63.81

Don’t remember 16 3.71

1.1 What was the reason you received a payment reminder? 13_(n=140)

I could not afford to pay on time 4 2.86

I was away 8 5.71

Mistake / Misunderstanding 128 91.43

2. Have you received a debt collection demand during the past two years? (N=431)

Yes 39 9.05

No 386 89,56

Don’t remember 6 1.39

2.1 What was the reason you received a debt collection demand? (n=39)

I could not afford to pay on time 5 12.82

I was away 5 12.82

Mistake / Misunderstanding 29 74.36

3. Have you received a payment injunction from the Enforcement Authority (Kronofogden) during the past two years? (N=431)

Yes 6 1.39

No 4 98.61

Don’t remember - -

3.1 What was the reason you received a payment injunction from the Enforcement Authority (Kronofogden)? (n=6)

I could not afford to pay on time 3 50

I was away - -

Mistake / Misunderstanding 3 50

13_{Many respondents clarified in comments the reason for why they received payment reminders. Reoccurring}

(38)

38

Notably, most respondents who have received either a payment reminder, debt collection demand or a payment injunction state it was because of some sort of mistake or

misunderstanding. A majority of respondents strongly agreed to the fact about having someone close who would help out in case of having payment troubles, which might partly explain why surprisingly few in our sample have received payment reminders, collection claim notices and payment injunctions because of the fact that they could not afford to pay on time. The possibly large hidden number of indebted individuals not shown in official statistics reported by the Swedish Enforcement Authority and various debt collection agencies is what partly motivated the conduction of this thesis.

5.3 Measuring economic preferences

In the last survey section, respondents were initially given two hypothetical choice questions with regards to relative concerns. Then they performed an experiment of incentivized

multiple-choice questions with regards to time preferences before submitting.

5.3.1 Time preferences

The incentivized multiple-choice experiment used to measure respondents’ time preferences is directly adopted from Meier and Sprenger (2010). Respondents were given three multiple price lists to make a series of choices between an earlier, smaller payment 𝑋 in 𝑡 months, or a larger, later payment 𝑌 in 𝑡 + 𝑑 months. The experiment structure is presented in table 3, and the instructions were given as follows:

Your choices in this experiment determine the size and date of the prize payment if you win the lottery. The choices you make could mean a difference in payment of more than 700 SEK, so choose carefully! Your task is simple: For each possible number below, state whether you would like the earlier, smaller payment or the later, larger payment. The winner of the lottery will be randomly assigned a number between 1 to 22 and will be paid their chosen payment. Your answers in previous questions remain anonymous. Good luck!

Table 8. Structure of the incentivized choice experiment

Price list Choices (𝑋) Time Delay

1 7 𝑡# = 0 𝑑# = 1

2 8 𝑡_$ = 0 𝑑_$ = 6

3 7 𝑡& = 6 𝑑& = 1

The Influence of Social Media on Consumer Credit Behavior #BuyNowPayL8r

#BuyNowPayL8r