Negative interest rates and individuals’ risk-taking behaviour

(1)

Abstract

Several central banks have implemented negative interest rate policies. Partly caused by these policies, current interest rates on savings are historically low. At several banks, wealthy depositors even have to pay their banks for the possibility to store their savings. This raises the question how individual savers would react on negative interest rates on their savings. This experimental study analyses how negative interest rates impact risk-taking behaviour of individual savers. Negative interest rates on savings cause people to invest more in risky alternatives to saving, while a comparable declining interest rate in positive territory does not lead to an increase in individuals’ risk-taking behaviour. The experimental results might be explained by loss probability aversion and the diminishing sensitivity principle of prospect theory. In addition, the impact of negative interest rates on individuals’ risk-taking behaviour depends on personal characteristics. The level of risk-taking of women increases when the interest rate on savings becomes negative, while negative interest rate do not change the risk-taking behaviour of men. Furthermore, people who score low on financial literacy increase their asset investments more in the case of a negative interest rate than people who score high on financial literacy.

Master’s thesis in Economics Radboud University Nijmegen

Name: Tijmen de Grutter Student nummer: 4487435

Specialisation: Economics, Behaviour and Policy Supervisor: dr. Janneke Toussaint

Second reader: dr. Dirk-Jan Janssen

Date: 9 July 2020

Negative interest rates

and individuals’

(2)

1

List of abbreviations

AFM Autoriteit Financiële Markten (Authority Financial Markets)

EC European Commission

ECB European Central Bank

EU European Union

DNB De Nederlandsche Bank (Dutch Central Bank) GDP Gross Domestic Product

NIR Negative interest rate

NIRP Negative interest rate policy OLS Ordinary Least Squares

(5)

4

1. Introduction

The European Central Bank (hereafter: ECB) has employed negative reserves deposit rates as a monetary policy to stimulate economic growth since June 2014. The ECB currently charges banks for the excess reserves banks park at the ECB, instead of paying banks interest on the parked reserves. Over the last years, several other central banks, such as the Danish Central Bank, the Bank of Japan and the Swedish Riskbank, have implemented negative interest rate policies as well (Arteta, Kose, Stocker, & Taskin, 2018).

Partly as a consequence of the negative interest rate policies of central banks, the interest rates on saving accounts are at historic lows (Baars, Cordes, & Mohrschladt, 2019). Most depositors currently receive an interest rate close to or at zero on their savings. At some banks, depositors whose savings exceed a certain amount have to pay their bank for the possibility to store their money at the bank. For example, the three largest Dutch banks charge negative interest rates from wealthy savers (“Na ING en ABN Amro”, 2020, February 11).

At the end of April 2020, the ECB announced that the ECB interest rates will remain at their present level, or will even become lower, until the inflation outlook in the Euro area will converge towards the target level of 2% (ECB, 2020, April 30). The coronavirus pandemic hampers economic growth and leads to a steep drop in the inflation rate, caused by a decrease in demand and low oil prices. The European Commission (EC) forecasts an inflation rate of 1.1% in 2021, which is far below the target level of 2% (European Commission, 2020, May 6). Therefore, the reserve deposit rate will probably remain negative the coming period. This might lower bank’s profitability in such a way, that commercial banks will be put in the position in which they have to consider whether to pass on the negative interest rate to all customers, including small depositors. This would mean that also common and small savers would have to pay their bank for the ability to store savings at the bank (ING Economic and Financial Analysis, 2015).

The topic of negative interest rates has been studied regularly in the academic field over the last years. However, most of the literature on negative interest rates deals with the impact of such policies on banks (See, among others, Basten & Mariathasan, 2018; Bohn, Plantefè, Poppensieker, & Schneider, 2020; Demiralp, Eisenschmidt, & Vlassopoulos, 2017; Heider, Saidi, & Schepens, 2018; Jobst & Lin, 2016; Tan 2019). The impact of negative interest rates on the behaviour of individual savers has not received much attention in the scientific literature yet. Few experimental studies, with conflicting results, consist on the topic of the impact of negative interest rates on risk-taking behaviour of individuals (Baars et al., 2019; Bracha 2016; Ma & Zijlstra, 2018). In this thesis, risk-taking behaviour is defined as investing in risky alternatives to saving.

The topic of negative interest rates on savings has become the subject of a political discussion recently. For example, the ‘Dutch Consumer Association’1 advocates a ban on negative interest rates charged from small savers (Donat, 2019, September 13). The Dutch Finance Minister argued in a letter to the Parliament that negative interest rates on (small) savings are undesirable. However, according to the Dutch Minister of Finance, a prohibition of negative interest rates is

(6)

5

not necessary yet (Hoekstra, 2019, December 5). For the discussion on how the government should act upon the possibility of banks charging negative interest rates from savers, it is important to know how savers would behave when they are confronted with a negative interest rate. Until now, it is highly uncertain how the behaviour of individuals would change under negative interest rates (Bech, & Malkhozov, 2016).

Since the subject has been underexposed in the literature so far, and because of the societal relevance of this topic, the following main research question is studied in this thesis:

What are the consequences of negative interest rates on individual savers’ risk-taking behaviour?

This study adds new dimensions to the literature on the topic of negative interest rates and individuals’ risk-taking behaviour, as it is the first study to analyse whether gender effects and/or the level of financial literacy influence the reaction of individual savers on negative interest rates.

The impact of a negative interest rate on risk-taking behaviour is studied by an online experiment, in which students of Radboud University Nijmegen participated. In this experiment, participants were asked to allocate an amount of money between saving on a saving account and investing in a riskier asset. This allocation task was repeated with varying interest rates and different possible asset returns. Participants were confronted with two positive and two negative interest rates. The results of this experiment show that individuals are more willing to invest in risky alternatives to savings, if the interest rate on savings is negative. This means that negative interest rates on savings lead to increased risk-taking behaviour of individual savers. This positive relationship between negative interest rates and risk-seeking behaviour mainly concerns women and individuals who score low on financial literacy.

The structure of this master’s thesis is as follows. In the next chapter, the background of the negative interest rate policies of central banks and the impact of these policies on banks are discussed (Chapter 2). Subsequently, the literature on the impact of negative interest rates on the behaviour of individuals is reviewed. In this chapter, both the empirical findings and the underlying theories are discussed (Chapter 3). Chapter 4 deals with the research questions and hypotheses. Thereafter, the methodology is described. In particular, the experimental design and the variables used in the analysis are discussed (Chapter 5). Subsequently, the results of the analysis are given. This chapter starts with a description of the results of the general analysis. Subsequently, the importance of the personal characteristics (financial literacy, gender, risk aversion, and loss aversion) are discussed (Chapter 6). The following chapter discusses the results in the light of the research questions. Furthermore, the discussion chapter examines how the experimental results relate to the reviewed literature. In addition, the limitations of this study and possible directions for further research are given (Chapter 7). The thesis ends with a conclusion (Chapter 8).

(7)

6

2. Background

2.1 Developments in the interest rate

Over the last decades, there has been a decreasing pattern of the interest rate (see Figure 1). Tokic (2017) states that, after a globally rising pattern between the 1960’s and 1980’s, interest rates have been declining since the 1980’s. This means that for decades, interest rates have been becoming lower and lower towards the nowadays interest level near and below zero per cent. According to Tokic (2017), demographic circumstances and increased globalization might explain the decreasing interest rates. The most common explanation for the downward pattern of interest rates is a lack of investments together with excessive savings (Arteta, Kose, Stocker, & Taskin, 2016; Tokic, 2017).

Figure 1 (From Tokic, 2017: Figure 2)2

2.2 Motivation of central banks’ negative interest rate policies

Aims of negative interest rate policies

Several central banks, including the ECB, implemented a negative interest rate policy (NIRP) by bringing the deposit rate into negative territory (Jackson, 2015). With the negative deposit rate, banks have to pay the central bank for the excess reserves they park at the ECB.3 The main argument of the ECB supporting their NIRP is that these policies tackle deflationary pressures and stimulate economic growth (Arteta et al., 2018; Jackson, 2015). These goals of the negative interest rate policies correspond with the main objective of the monetary policy of the ECB to maintain price stability, with an target inflation level of approximately two per cent.4

Mechanism of negative interest rate policies

The mechanism behind the stimulating impact of negative interest rates policies on inflation and economic growth is largely the same as the stimulating impact of a decrease in the interest rate in positive territory. The negative interest rates set by the central bank cause a decline of the interest rate on savings. With lower, or even negative, interest rates on savings, saving is discouraged. A negative interest rate creates an incentive for individuals to save less and consume more. Furthermore, the negative interest rates of central banks lead to cuts in the

2_{Figure 1 shows the development of the Effective Federal Funds Rate in the United States (US). This rate is set} by the Federal Reserve and is the benchmark for all short-term interest rates in the US (Tokic, 2017).

3_{https://www.ecb.europa.eu/explainers/tell-me-more/html/why-negative-interest-rate.en.html}_. 4_{https://www.ecb.europa.eu/mopo/html/index.en.html}_.

(8)

7

interest rates that financial intermediaries, such as banks, require from lenders and borrowers, which stimulates lending and borrowing. The shift from saving to consumption and investments caused by the NIRP creates inflationary pressures (Arteta et al., 2018; Kimball, 2019). Because of the beforementioned mechanisms, the ECB expects that its negative interest rate policies will stimulate economic growth in the Euro area.5

Low inflation rate

The inflation rate in the European Union (EU), and more specifically the Euro area, was just above zero, at the moment that the ECB introduced its negative interest rate policy in June 2014 (see Figure 2). The inflation rate was far below the target level of two percent and was still decreasing. Because it was forecasted that inflation within the euro area would remain far below the target inflation level, the ECB decided to decrease the three main interest rates on which it can act.6 As the deposit facility rate was already at zero percent, the decrease in this interest rate caused the deposit rate to reach negative territory.7

Figure 28_{: Inflation rate in the EU (blue line) and Euro area (orange line)}

2.3 Why would a bank charge negative interest rates from deposit holders?

Banks’ profitability

Negative interest rate policies might cause a decline in the profitability of banks. While negative interest rate policies can be beneficial for the economy by stimulating economic growth, these policies also have disadvantages for financial institutions. The net-interest margins of banks are eroded by these kind of policies. Lower interest rates set by central banks normally also lead to banks declining the interest rates that savers receive on their savings. However, banks are reluctant to pass on negative interest rates to customers, because banks are afraid of the consequences this might have for the behaviour of individual savers (Heider et al., 2018). The net-interest margins of most banks within the scope of central banks employing negative interest rate policies have been declining since 2015 (Bohn et al., 2020).

5_{https://www.ecb.europa.eu/explainers/tell-me-more/html/why-negative-interest-rate.en.html}_.

6_{These are the main refinancing operations facility, the marginal lending facility for overnight lending to banks,} and the deposit rate. See https://www.ecb.europa.eu/explainers/tell-me-more/html/why-negative-interest-rate.en.html.

7_{https://www.ecb.europa.eu/explainers/tell-me-more/html/why-negative-interest-rate.en.html}_. 8_{This figure is retrieved from}

(9)

8

It is questionable whether banks are able to mitigate the impact of negative interest rate policies on their profitability in the long run. So far, most banks have been able to mitigate the negative impact of negative interest rate policies on their profitability, for example by increasing lending volumes, profiting from higher asset prices and saving costs by improving operational efficiency. However, further mitigation measures are limited. Therefore, in the long run, a continuation of the negative interest rate policies of central banks could pose a threat for the profitability of the banking sector (Heider et al., 2018; Turk, 2016). On the other hand, the chance of negative interest rates charged to small savers might have become lower since the implementation of the two-tier system of the ECB on 30 October 2019. This tiering system leads to an exemption of the negative deposit rate for a part of the excess liquidity holdings of banks. Because commercial banks have to pay less interest on their excess liquidity holdings, the two-tier system eases the pressure of the negative interest rate policies on banks’ profitability (ECB, 2019, September 12).

Risk-taking by banks

The lower profitability of banks caused by the negative interest rates policies of central banks has led to more risk-taking by banks. To maintain their profit levels, banks are becoming more involved with risky practices. For example, research has found that high-deposit banks finance riskier firms in a negative interest rate environment (Heider et al., 2018). Furthermore, banks have shifted assets towards longer maturities and liabilities towards shorter maturities, leading to an increase in interest rate risk. Negative interest rates have also caused banks to hold riskier assets. Risk-taking mainly increased at the banks which are affected the most by the negative interest rate policies (Basten & Mariathasan, 2018).

The increase in risk-taking by banks caused by the negative interest rate policies can increase systemic risk in the banking sector. Systemic risk has been increasing in the European banking sector since the ECB employed its negative interest rates policies (Kurowski & Rogowicz, 2017). The impact of negative interest rate policies on the systemic risk of a bank depends on the business model of the bank. Negative interest rate policies are expected to mainly increase the systemic risk of small and traditional banks, which rely mostly on deposit funding and have less diversification of income streams (Nucera, Lucas, Schaumburg, & Schwaab, 2017).

Indirect effect on individual savers

Individual savers might indirectly be confronted with the impact of negative interest rates policies on banks. This effect is indirect, because individuals are faced with the negative interest rate policies of central banks via the bank where they hold their saving account. Because the NIRP lowers the profitability of banks, commercial banks might (partly) pass on the costs of the negative interest rate policies to savers, by lowering the interest rate on savings.9

Several banks already charge negative interest rates from customers with large deposits. Eventually, the negative interest rate policies of central banks can bring the profitability of banks under such pressure, that banks become forced to charge negative interest rates from all depositors (ING Economic and Financial Analysis, 2015). As of 1 July 2020, all savers at the Dutch Triodos Bank have to pay €2,- per month for holding their saving account at the bank. With the interest rate on this saving account already being zero per cent, these costs related the

(10)

9

holding the saving account lead to a negative return on savings. Therefore, the obligation to pay €2,- monthly can be seen as a disguised negative interest rate (De Waard, 2020, April 17). It is not inconceivable that, in the near future, negative interest rates will also apply to savers with average and small savings at other banks. Multiple Dutch banks have changed the terms and conditions of their saving accounts, which makes it possible to charge negative interest rates in the future (Spierenburg, 2019, December 2019).

2.4 Prohibition of negative interest rates

The possibility and desirability of a prohibition of negative interest rates charged to small and average savers is debated in several countries. Proponents of such a prohibition warn for the hazard of a bank run. Banks would come into trouble, if many savers would simultaneously withdraw their money from their saving accounts when interest rates on savings become negative. This would cause instability of the financial sector, and especially the banking sector (ING Economic and Financial Analysis, 2015). Furthermore, the public generally considers negative interest rates to be unjustified and unacceptable. It is seen as a basic principle of the banking system to get rewarded for saving behaviour by receiving interest on savings. A negative interest rate on savings leads to the directly opposite of this principle (Van Dooren, Toolens, & Manschot, 2019). Despite of the ongoing debate on the desirability of a prohibition of negative interest rates on (small) savings, no national government has implemented a regulatory prohibition of negative interest rates yet (Hoekstra, 2019, December 5).

The German Minister of Finance has stated that it should be studied whether small savers can be protected by law against negative interest rates. This is a reaction of the Finance Minister on a proposal of the Prime Minister of the German Free State of Bavaria to introduce a prohibition on negative interest rates. Several economists are sceptical about the proposed prohibition, mainly because it could lead to instability of the banking system (“Scholz will prüfen”, 2019). In the Netherlands, some political parties have raised questions about the possibility of a prohibition of negative interest rates charged to depositors. The Dutch Finance Minister concluded that a prohibition currently is not necessary, partly because banks will be reluctant to charge negative interest rates on deposits below €100.000 (Hoeksta, 2019, December 2005). Furthermore, such a prohibition on negative interest rates also has clear disadvantages. Firstly, a prohibition of negative interest rates would be a significant restriction of the free market. Secondly, it could have negative consequences for the Dutch economy, as it can weaken the economic stimulus of the monetary policies of the ECB. Thirdly, a prohibition of negative interest rates could have a negative impact on banks’ profitability, which would lead to instability of the banking sector (Hoekstra, 2019, December 2005).

(11)

10

3. Literature review

3.1 Theoretical expectations

3.1.1 Decreasing interest rates and household savings

It would be expected that household savings have been decreasing over the last decade, because the interest rate has been declining over this period. Figure 3 shows that the household saving rate has indeed decreased between 2009 and 2018. However, since the first quarter of 2018, the household saving rate has been increasing in the Euro area (Eurostat, 2019, July 3).

An interest rate becoming negative is seen as a discouragement of saving, because people have to pay their bank for the ability to store their savings, instead of receiving interest on their savings (Kimball, 2019). The unexpected result that the household saving rate increased recently, despite of the historically low interest rates, raises the question whether the theoretical expectation that negative interest rates would lead to lower household savings would hold true in the real world (Efendić et al., 2019).

Figure 3: This figure is retrieved from Eurostat (2019, July 3). 3.1.2 Zero lower bound and the demand for cash

Zero lower bound theory

It is argued in the literature that banks are reluctant to pass through negative interest rates to households, because they fear a bank run. Theoretically, it is expected that savers would withdraw money from their saving account when savers have to pay for storing money at the bank, while it is possible to hold cash to avoid the negative interest rate (Heider et al., 2018). The zero lower bound is especially of importance with respect to household deposits, because these deposits are relatively small and therefore the costs of holding cash would be relatively low (Eisenschmidt & Smets, 2018). Another reason why banks might be reluctant to charge negative interest rates on household savings, is that no bank wants to be the first to introduce negative interest rates on savings, because this could lead to many customers switching to a bank that does not charge a NIR yet. On the other hand, if one bank introduces a negative interest rate on all savings, others are likely to follow, because negative interest rates will mainly damage the image of the first bank that charges negative interest rates.

(12)

11

Demand for cash

Negative interest rates on savings are theoretically expected to lead to an increase in the demand for cash (Humphrey, 2016). Because the monetary policies of the central banks lead to lower interest rates on savings, the difference between holding cash and having bank deposits have become negligible for most savers (Danish Payments Council, 2016; Madaschi & Pablos Nuevo, 2017). However, the negative monetary policies have not led to a significant increase in the demand for cash in Denmark and Sweden. Both the nominal amount of cash and the value of cash in circulation as a percentage of GDP has been declining over the last decade in both Sweden and Denmark. This downward trend of the demand for cash has also continued after the introduction of the negative interest rate policies of the Danish and Swedish central bank (Danish Payments Council, 2016; Madaschi & Pablos Nuevo, 2017; Sveriges Riskbank, 2016). That the demand for cash did not increase, despite of the negative monetary policy rates, might be caused by the relatively low use of cash in these countries. Therefore, it is possible that in other countries the impact of negative interest rate policies on the demand for cash would be stronger. Negative interest rate policies might be an explanation for the increasing cash in circulation in the UK and the Eurozone (Danish Payments Council, 2016; Madaschi & Pablos Nuevo, 2017). Furthermore, it is important to keep in mind that the interest rate on small deposits has not become negative in these countries yet. Therefore, it remains unclear what the effect of a negative interest rate on household deposits will be on the demand for cash.

3.1.3 Prospect theory

Based on prospect theory it can be argued that a negative interest rate on savings would lead to an increase in risk-seeking behaviour of individuals. According to prospect theory, individuals are risk-averse in the gain domain and risk-seeking in the loss domain (Kahneman & Tversky, 1979). Following the diminishing sensitivity principle of prospect theory, negative interest rates are expected to cause an increase in risk-taking behaviour. According to the diminishing sensitivity principle, people are the most sensitive to changes close to the reference point. With respect to interest rates and risk-taking behaviour, the reference point depends on the expected return on savings and investments, which is based on past experiences (Baars et al., 2019; Ma & Zijlstra, 2018). How further away a gain or loss is from the reference point, the more the sensitivity of individuals diminishes. Baars et al. (2019) assume that an annual return on savings and investment of zero is the reference point of savers. Therefore, an interest rate that switches from positive to negative is a loss that takes place in the steepest part of the value function. Individuals are highly sensitive to such a change. Therefore, the utility of saving on a saving account is devaluated heavily by savers when the interest rate becomes negative. This devaluation would probably be higher than the devaluation of the alternative, which is investing in a riskier asset (Baars et al., 2019).

An example, based on the experimental design of Baars et al. (2019), is used to explain this mechanism. An individual can, for example, choose between saving on a saving account and investing in a risky asset. Imagine that the interest rate on the saving account drops from +1% to -1%. The alternative risky asset has two possible returns with 50% probability each. These possible returns change from 24% or -11% to 22% or -13%. The change in the return of the saving account is closer to the reference point of zero per cent than the changes in the possible returns of the risky asset. Therefore, it is expected that people would opt more frequently for the risky asset if they are confronted with a negative interest rate on their saving account (Baars et al., 2019).

(13)

12

3.1.4 Other theories on negative interest rates and risk-taking behaviour

Loss probability aversion

Based on loss probability aversion theory, an increasing effect of negative interest rates on risk-taking behaviour is expected. Loss probability aversion theory assumes that a switch from the gain to the loss domain causes a disproportionally value devaluation. With negative interest rates, the probability of being faced with a loss on the saving account is 100%. If an individual invests in a risky asset, the probability of having a loss is less than 100%, for example 50%. To avoid a certain loss, more people would opt for the risky asset in a negative interest rate environment (Baars et al., 2019).

Saliency

Risk-taking behaviour of individuals might increase when the interest rate on savings drops, because the risk premium of a risky asset is more salient when interest rates are low. The salience of the difference between a 1% certain return and a 6% average return might be higher than the salience of the difference between a 5% risk-free return and an average return of 10% on the risky asset, despite of the constant risk premium (Ma & Zijlstra, 2018). This reasoning can also be used to argue that risk-taking behaviour would increase when interest rate on savings become negative. A difference between a return on a saving account of -1% and an expected return of +4% on a risky asset, might be more salient to individual savers than the difference between an interest rate on savings of +1% and a 6% expected return on a riskier investment.

Psychological consequences of negative interest rates

According to a report of the Volksbank,10 a negative interest rate would lead to an irrational increase in risk-taking behaviour of individual savers. A NIR leads to a threat of the psychological need for justice, autonomy and security. Additionally, implementing negative interest rates on savings would decrease the trust in the bank, which results in a decrease of saving intentions (Van Dooren et al., 2019).

3.2 Empirical findings

3.2.1 Surveys and experiments

Surveys

Both surveys and experiments are used to study the impact of negative interest rates on saving behaviour of individuals. The possibility of interest rates on savings becoming negative and the possible consequences of this, is a topic that has received a lot of media attention. To find out individual savers’ intentions when interest rates become negative, several (financial) institutions and TV-programmes have conducted surveys. In these surveys, respondents are asked whether they would continue saving at their bank if their bank starts charging negative interest rates. If people state to have the intention to withdraw money from their saving account under such circumstances, it is questioned how they would use their money instead.

While surveys can give useful insights in the impact of negative interest rates on individuals’ saving behaviour, it has the disadvantage that the results say more about intentions than about real choices. A limitation of the survey method is the possible intention-behaviour gap, which

(14)

13

means that it is uncertain whether participants in reality will act the same way as they declared they will in the survey (ING Economic and Financial Analysis, 2015).

Experiments

The literature on the topic of negative interest rates and individuals’ risk-taking behaviour also consists of a few experimental studies. An advantage of using experiments compared to surveys, is that in experiments actual choices of participants are studied. Researchers can get closer to the actual behaviour of subjects with experiments than with surveys, because participants of experiments receive a payment based on their performances. While the internal validity of experiments is mostly high, experiments frequently have a lack of external validity. Experiments simplify the reality, which leads to a decrease of the generalizability of the results towards the real world scenario (Friedman & Cassar, 2004).

3.2.2 Impact of negative interest rates on the level of savings

The majority of savers intend to withdraw at least some of their savings from their saving account, when their bank implements a NIR on their savings (ING Economic and Financial Analysis, 2015; Meijer, 2020; Pooters & Zijlstra, 2017; Radar, September 23, 2019). While most people have not made adjustments to their saving behaviour as a consequence of the declining interest rate yet, most people state that they will act when interest rates become negative. Only between ten and twenty per cent of the subjects stated to have the intention to continue saving the same way as they do nowadays. The most given reason why these people continue saving at the bank is the safety of storing money at a bank (ING Economic and Financial Analysis, 2015; Pooters & Zijlstra, 2017; Radar, September 23, 2019). A small portion of the people surveyed even stated to save more when interest rates on savings become negative, because they would have to save more to meet their saving goals (ING Economic and Financial Analysis, 2015). Furthermore, another approximately ten percent of the people intend to switch from saving on a saving account to saving on a long-term deposit (Meijer, 2020; Pooters & Zijlstra, 2017; Radar, September 23, 2019). This means that these people would have the intention to continue saving, but in a different way than they do nowadays. However, more than half of the people would search for an alternative to saving if the interest rate on saving becomes negative (ING Economic and Financial Analysis, 2015; Meijer, 2020; Pooters & Zijlstra, 2017; Radar, September 23, 2019).

In contrast to the findings of the surveys discussed above, Efendić et al. (2019) state that savers might be more tolerant towards negative interest rates than would be expected. In this study, participants are confronted with a hypothetical situation in which a new policy is implemented, which entails that savers are required to pay their bank for keeping savings at the bank. Subjects are asked whether they would be willing to pay the negative interest rate in several situations, or whether they would withdraw their savings from the saving account. The willingness to pay is questioned for six cases: two different amounts of money ($1000 and $35.000) and three different interest rates (-0.3%, -0.6% and -1%). It is stressed in the instructions that all banks charge negative interest rates, which makes it impossible to avoid the negative interest rate by switching to another bank.11 It is important to note that in the real world, it could be that some

11_{The instructions of the study of Efendić et al. (2019) are retrieved from} https://osf.io/7ftak/?view_only=bf579396c3b741f8aaaf5b47025aa43b and https://osf.io/7ftak/?view_only=806070cf8f5e46d89fe76d6ac30496fa.

(15)

14

banks implement negative interest rates before other banks, which could create incentives for savers to switch from bank.

The results of the experiments conducted by Efendić et al. (2019) show that depositors are relatively tolerant towards negative interest rates. As expected, participants were the most tolerant towards an interest rate of 0.3%, and the least accepting towards an interest rate of -1.0%. More than half of the people would continue saving at the bank when faced with an interest rate of -0.3%. Participants were willing to pay the required payment in about 40% of all the choices made (Efendić et al., 2019). This means that a large proportion of the savers has the intention to withdraw savings from their saving account if a negative interest rate is charged from depositors. Furthermore, it follows from this study that individuals with a lower amount of savings are more likely to accept negative interest rates. It also follows from this study that older people accept negative interest rates less (Efendić et al., 2019). This result corresponds with the survey study of ING (ING Economic and Financial Analysis, 2015). In addition, men seem to be less tolerant towards negative interest rates than women. The overall conclusion of the paper is that individuals seem to be more tolerant towards negative interest rates than theoretically assumed. However, still about half of the subjects states to save less if negative interest rates on savings are introduced (Efendić et al., 2019).

To sum up, it follows from the literature that negative interest rate on savings would cause the intended decrease in household saving levels, as most people state that negative interest rates would lead to a decrease in their saving intentions. However, the several studies on this topic have different results with respect to the extent to which savers would withdraw savings from their saving account (Efendić et al., 2019; ING Economic and Financial Analysis, 2015; Pooters & Zijlstra, 2017; Radar, September 23, 2019). Furthermore, it is important to keep in mind that the results might be caused by the intention-behaviour gap, as most of the studies finding that people would withdraw savings when confronted with negative interest rates are survey studies. It remains questionable whether indeed the majority of the savers would withdraw (some of) their savings, when a negative interest rate on savings is introduced.

3.2.3 Alternatives to saving in a NIR environment

Small effect of negative interest rates on consumption

Only between five and twelve per cent of the people who would withdraw money from their saving account in a negative interest rate environment has the intention to increase their level of consumption (ING Economic and Financial Analysis, 2015; Pooters & Zijlstra, 2017; Radar, September 23, 2019). One study even found that less than one per cent of the respondents has the intention to increase their spending when interest rates become negative (Efendić et al., 2019). As negative interest rates would probably not stimulate spending that much, the intended increase in inflation might not be achieved by the negative interest rate policies. Furthermore, the increase in consumption that a negative interest rate could cause, would probably be offset by savers who would increase their savings when interest rates turn negative. However, it is important to note that richer and older respondents stated relatively more frequently to have the intention to increase their consumption if interest rates become negative. Because rich and older people on average have many savings, the negative interest rate policies might be a considerable stimulus of consumer spending after all (ING Economic and Financial Analysis, 2015).

(16)

15

More popular destinations of withdrawn savings

Paying off debts seems to be a more attractive alternative to saving than increasing the level of consumption when negative interest rates are implemented. Around twenty per cent of the respondents has the intention to use their savings to pay off debts, if they have to pay the bank for the possibility to store savings. It is mainly stated that savings will be used to pay off mortgage debt (Meijer, 2020; Pooters & Zijlstra, 2017; Radar, September 23, 2019). One of the most popular alternatives to savings in a negative interest rate environment is to hoard cash at home or in safe. Between 20% and 37% of the respondents stated to have the intention to hoard cash if negative interest rates on savings are implemented (ING Economic and Financial Analysis, 2015; Meijer, 2020; Pooters & Zijlstra, 2017; Radar, September 23, 2019).

3.2.4 Risk-taking behaviour

Declining interest rates

Several experimental studies have studied whether the demand for risky assets increases when the risk-free interest rate (for example the interest rate on deposits) decreases, but stays positive. In these experimental studies, participants are asked to allocate their endowment between a risk-free asset with a certain return and a risky asset with an uncertain outcome. The average expected return on the risky asset is higher than the average return on the risk-free asset. This reflects the risk premium of the stock market. Participants in the experiment are asked to allocate their endowment with several interest rates and expected returns (Ganzach & Wohl, 2018; Lian, Ma, & Wang, 2018; Ma & Zijlstra, 2018).

Most of the studies have found that a drop in the interest rate causes an increase in the share of money that is allocated to the risky assets (Ganzach & Wohl, 2018; Lian et al., 2018; Ma & Zijlstra, 2018). In a low interest rate environment, savers will look for yield. Therefore, they invest more in riskier assets to get higher returns on their savings (Lian et al., 2018). When the risk-free rate is lower, the risk premium is considered to be more meaningful to investors, compared to the situation in which the risk-free rate is higher. Therefore, people prefer the riskier assets more often in a low interest rate setting (Ganzach & Wohl, 2018).

A declining interest rate did not lead to more risk-taking in the experiment studies of Baars et al. (2019) and Bracha (2019). Baars et al. (2019) did not find a significant difference in the share of money invested in the risky asset between the situation in which the risk-free return is 3% and the situation in which the risk-free return is 1%. The findings of Baars et al. (2019) and Bracha (2016) contradict with the results of the above discussed other literature (Ganzach & Wohl, 2018; Lian et al., 2018).

Negative interest rates

Many savers have stated to intend to use at least some of their savings to invest in riskier investments, if the interest rate on savings becomes negative (Efendić et al., 2019; ING Economic and Financial Analysis, 2015; Meijer, 2020; Pooters & Zijlstra, 2017; Radar, September 23, 2019). One of the surveys on this topic has found that, while in 2016 most people stated that they would hoard cash, in 2017 most people had the intention to invest in alternative assets when negative interest rates are charged. In 2016 and 2017, respectively 33% and 27% of the respondents stated to hoard cash and respectively 20% and 28% of the respondents stated to invest in riskier alternatives if interest rates become negative (Pooters & Zijlstra, 2017). As

(17)

16

many respondents have the intention to invest in alternative, riskier investments, negative interest rates would probably lead to an increase in risk-taking behaviour of individuals (Efendić et al., 2019; ING Economic and Financial Analysis, 2015; Meijer, 2020; Pooters & Zijlstra, 2017; Radar, September 23, 2019; Van Dooren et al., 2019). Whether the negative interest rate is framed as a negative interest rate or as storing costs does not significantly influence individuals’ saving and risk-taking intentions (Van Dooren et al., 2019).

The literature also consists of a few experimental studies that have investigated risk-taking behaviour of individuals in a negative interest rate environment (Baars et al., 2019; Bracha 2016; Ma & Zijlstra, 2018). 316 business students of a German university participated in the experimental study of Baars et al. (2019). Subjects were asked to allocate 60.000 monetary units between an asset with an annual risk-free return and a risky asset that has a 50% probability of a positive return and a 50% probability of a negative return. The expected return on the risky asset is higher than the expected return on the risk-free asset, which indicates the risk premium. Each participant played six independent rounds, which means that the experiment has a within-subject design The experiment consists of three different combinations of annual returns. Firstly, the annual return of the risk-free asset is 3% and the annual return of the risky asset is 26% or -9%, both with 50 per cent probability. In the second scenario, 1% annual interest is earned on the riskfree asset and the annual returns on the risky assets are 24% and -11%. Lastly, the interest rate on the risk-free interest rate is negative (-1%) and the risky asset annual returns are 22% and -13%. Each participant makes two decisions per combination of annual returns. Once with a one year time horizon and once with a ten year time horizon. In all the rounds, the loss probability, annual risk premium and annual return volatility of the risky assets are the same (Baars et al., 2019).

In the experiment of Bracha et al. (2016), subjects had to allocate funds between two portfolios, Portfolio X and Portfolio Y. Portfolio X has a sure return, while Portfolio Y is a lottery with two or three possible returns. Participants were randomly assigned to the negative or the positive frame group, which means that this experiment has a between-subject design. In the negative frame, a negative interest rate is charged on the funds allocated to Portfolio X. In the positive frame, Portfolio X has a sure positive return. The share of the money allocated to the risky asset is compared between the positive frame and the negative frame. In contrast to the beforementioned experiments on the impact of interest rates on risk-taking behaviour, no risk premium is included in this study. Therefore, the average return on the risky portfolio is equal to the average return on the risk-free portfolio (Bracha, 2016).

Some experimental studies have found that people opt more often for a riskier alternative of saving at a saving account, when the interest rate on savings becomes negative. In Germany as well as in the US and the Netherlands, risk-taking is higher in a negative interest rate environment than in a positive interest rate environment. This means that negative interest rates seem to lead to an increase in risk-taking behaviour (Baars et al., 2019; Ma & Zijlstra, 2018). For example, Baars et al. (2019) have found that around half of the endowment is invested in the risky asset when the interest rate is positive, while with a negative interest rate of -1%, over 60% of the money is invested in the risky asset. The order of the different rounds of the experiment did not influence the results (Baars et al., 2019). These results are in line with the diminishing sensitivity principle as discussed in §3.1.3.

(18)

17

In contrast to the beforementioned literature, Bracha (2016) did not find a significant difference in risk-taking between the positive and negative interest rate treatment group. On average, 48.5 per cent of the funds are allocated to the risky investment in the positive frame and 51.8 per cent of the funds are allocated to the risky portfolio in the negative frame. Because these differences are not significant, no increase of risk-seeking in a negative interest rate environment is found in this experiment. The main finding of the paper of Bracha (2016) is that subjects mostly act in a risk-neutral manner. This indicates that risk-seeking behaviour would not increase when negative interest rates on savings are implemented (Bracha, 2016).

However, a limitation of this experimental study is that the investment setting does not include a risk premium. The lack of a risk premium might explain why the results of this experiment differ from the other literature (Baars et al., 2019; Ma & Zijlstra, 2018). The studies with a risk premium included are closer connected to the real world scenario. In the real world the average expected returns on risky alternatives mostly are higher than the interest rate on savings. On average, the return on risky assets, such as investments in stocks, exceeds the average return on savings. When a risk premium is included in the experimental design, the higher expected return on riskier assets is taken into account.

3.2.5 Conclusion on empirical studies

Many savers have the intention to withdraw money from their saving account when interest rates turn negative. This is an indication that negative interest rate policies would have the intended impact of decreased savings. However, only a small proportion of the savers would increase spending as a consequence of negative interest rates. More popular alternatives of saving in the case of negative interest rates are hoarding cash at home or a safe and paying off debts. Both the surveys and most of the experimental studies have found that many savers would invest in riskier alternatives to savings when interest rates on savings become negative (Baars et al., 2019; ING Economic and Financial Analysis, 2015; Ma & Zijlstra, 2018; Pooters & Zijlstra, 2017; Radar 2019, September 23). This indicates that a negative interest rate on deposits would cause an increase in individuals’ risk-taking behaviour.

However, the results of the experimental studies on the impact of negative interest rates on individuals’ risk-taking behaviour are inconsistent. Most articles are in line with the diminishing sensitivity principle of prospect theory, as the majority of the literature that deals with this topic has found that risk-taking behaviour increases when the interest rate becomes negative (Baars et al., 2019; Ma & Zijlstra, 2018). The result of Baars et al. (2019) that, while a drop in the interest rate in positive territory does not increase risk-taking, risk-taking does increase when the interest rate becomes negative, could indicate that individuals psychologically value a drop in the interest rate in positive territory differently than a drop in the interest rate which results in a negative interest rate. In contrast to the other literature, Bracha (2016) has not found increased risk-seeking in a negative interest rate environment. However, this might be caused by the lack of a risk premium in the study of Bracha (2016). Concluding, the impact of decreasing interest rates, and especially negative interest rates, on risk-taking behaviour of individuals is still unclear, since the existing studies have had different results.

(19)

18

4. Research problem

4.1 Research questions

Little is known yet about the impact of negative interest rates on the risk-taking behaviour of individual savers. Firstly, the literature is inconsistent with respect to the answer on the question whether risk-taking behaviour of individuals would change if interest rates become negative.12 Secondly, the impact of negative interest rates on risk-taking behaviour has not been experimentally studied with a sample that includes people without experience with investments yet. Thirdly, it is still unclear whether the impact of negative interest rates on risk-taking behaviour depends on individual characteristics, such as gender and financial literacy.

The following research question is studied in this master’s thesis:

What are the consequences of negative interest rates on individual savers’ risk-taking behaviour?

To be able to answer the research question, several questions are investigated. These sub-questions are shown below.

1. Does a relationship between the interest rate on savings and risk-taking behaviour of individual savers exist?

2. Do negative interest rates lead to an increase in the risk-taking behaviour of individual savers?

3. Does the level of the negative interest rate influence risk-taking behaviour of individual savers?

4. Does the risk-seeking behaviour of individuals in a negative interest rate environment depend on the level of financial literacy?

5. Do men react differently on negative interest rates on savings than women, with respect to risk-taking behaviour?

6. Does the impact of negative interest rates on risk-taking behaviour depend on the level of risk aversion?

7. Does the impact of negative interest rates on risk-taking behaviour depend on the level of loss aversion?

4.2 Hypotheses

The following hypotheses are formulated, based on the existing literature.

Hypothesis 1: A decline in the interest rate leads to more risk-taking behaviour.

According to the majority of the literature, a decline in the return on a safe investment, such as a saving account, causes people to search for yield. Individuals are expected to invest more in riskier alternatives, if the interest rate on savings is lower (Ganzach & Wohl, 2018; Lian et al., 2018; Ma & Zijlstra, 2018). Therefore, the hypothesis is that there is a negative relationship between the level of the interest rate on savings and individuals’ investments in risky assets.

(20)

19

Hypothesis 2: Risk-taking behaviour of individual savers increases if the interest rate on savings becomes negative.

This hypothesis deals specifically with the impact of an interest rate switching from positive to negative on risk-taking behaviour of savers. The literature is divided with respect to this topic. However, most studied did find strong support for a relationship between negative interest rates and individuals’ risk-taking behaviour (Baars et al., 2019; Ma & Zijsltra, 2018).13_{Theoretically,}

it can be argued based on the diminishing sensitivity principle, which is a part of Prospect theory, that risk-taking will increase if the interest rate becomes negative.14 Thus, it is expected that negative interest rates lead to more risk-taking behaviour. As breaching the zero bound is assumed to lead to a psychological shock and cause a reconsideration effect, it is expected that an interest rate becoming negative will have a stronger impact on the level of risk-taking of individual savers than a decline in the interest rate in positive territory.

Hypothesis 3: The more negative a negative interest rate, the more individuals make risky investments.

People seem to be less tolerant towards lower negative interest rates than towards negative interest rates just below zero. More people state to have the intention to withdraw savings from their saving account in the scenario of an interest rate on savings of -1%, compared to when people are confronted with an interest rate of -0.3% (Efendić et al., 2019). This indicates that the influence of negative interest rates on saving behaviour depends on the size of the NIR. However, according to the diminishing sensitivity principle, people are mainly sensitive to changes that happen close to the reference point (Kahneman & Tversky, 1979). If it is assumed that a return of zero is seen as the reference point, it would be expected that a change in the interest rate from +1% to -1% leads to more behavioural changes than a change in the interest rate from -1% to -3%. To sum up, the hypothesis is that risk-taking behaviour of individuals is influenced by the level of the negative interest rate. However, I expect that a decline in an already negative interest rate increases the risk-taking behaviour of individual savers less than a decline in the interest rates that leads to a switch from a positive interest rate towards a NIR.

Hypothesis 4: The impact of negative interest rates on risk-taking behaviour depends on the level of financial literacy.

People with lower levels of financial literacy are expected to make less rational financial decisions than people with higher levels of financial literacy. It might be that people with lower financial literacy levels are more sensitive to the psychological shock that is caused by an interest rate becoming negative. Therefore, even if the risk premium of a risky asset remains the same, people with lower levels of financial literacy are expected to adjust their saving behaviour in a NIR environment. It is therefore expected that people with lower levels of financial literacy increase their risk-taking behaviour in a negative interest rate environment more than people with higher levels of financial literacy.

However, it can also be argued that people with lower financial literacy levels, compared to people with higher levels of financial literacy, are less likely to invest in stock markets to search for yield, if the interest rate on savings becomes negative. People with less knowledge about

13_{For a comprehensive discussion of the existing literature on the relationship between (negative) interest} rates and risk-taking behaviour of individuals, see the previous chapter.

(21)

20

the workings of the financial markets are less likely to invest in the stock market (Van Rooij, Lusardi, & Alessie, 2011). Therefore, it could be that people with low levels of financial literacy will be reluctant to invest in riskier alternatives of saving on a saving account, even if the interest rate on savings becomes negative. To conclude, the effect of financial literacy on the relationship between negative interest rates and risk-taking behaviour is uncertain, as arguments can be given both for a positive and for a negative relationship.

Hypothesis 5: The impact of negative interest rates on risk-taking behaviour differs between men and women.

Several studies have found strong evidence for gender differences with respect to risk-taking in the financial markets. On average, women are more risk-averse than men (Byrnes, Miller, & Schafer, 1999; Charness & Gneezy, 2012; Dwyer, Gilkeson & List, 2002). Because men are more accepting towards risks, it is expected that men will switch more towards riskier alternatives to saving when the interest rate on savings become negative. This is in line with the literature that has found that men are less tolerant towards negative interest rates than women (Efendić et al., 2019). Therefore, the expectation is that negative interest rates increase the risk-taking behaviour of men more than the risk-risk-taking behaviour of women.

Hypothesis 6: The impact of negative interest rate on risk-taking behaviour differs between people with low and high risk avoidance.

Risk averse people will behave less risky with respect to financial decisions than risk-seeking people. It can be argued that people who score low on risk aversion are willing to invest in riskier alternatives of saving when interest rates on savings become negative, while risk averse people will remain reluctant to make risky investments. This would mean that risk-taking behaviour of individuals with weak risk avoidance increases relatively much in a NIR environment. However, Baars et al. (2019) did not find significant differences with respect to negative interest rates and risk-taking behaviour of individuals with weak and strong risk avoidance.

Hypothesis 7: Risk-taking behaviour of relatively loss averse people will increase more when interest rates become negative.

People with a high level of loss aversion will be more reluctant to invest in risky investments than people with low loss avoidance. Therefore, it is expected that strong loss averse individuals invest less in risky investment with possible losses compared to weak loss avoidance individuals. However, if negative interest rates on savings are implemented, a loss will also be obtained on savings. To prevent this certain loss, investing in riskier alternatives of saving will become more attractive in the perspective of individuals with strong loss avoidance. Baars et al. (2019) have shown that negative interest rates only impact the risk-taking behaviour of people who score high on loss avoidance.

(22)

21

5. Methodology

5.1 Experimental design

Part 1 of the experiment

In the first part of the experiment, participants are asked to allocate an amount of money between a saving account and a risky asset. This is similar to earlier experimental studies on the impact of negative interest rates on risk-taking behaviour (Baars et al., 2019; Bracha, 2016; Ma & Zijlstra, 2018).15 The risky asset has a 50% probability of a positive return and a 50% probability of a negative return. The interest rate on the saving account and the possible asset returns vary among the different rounds of the experiment. The risk premium of the risky assets is stable at 5%. The risk premium of the asset can be measured with the following formula: average expected return on the risky asset – interest rate on saving account. A risk premium of 5% is chosen, because this risk premium is comparable to the real world risk premium on the stock market. In both Europe and the US, the annual risk premium of stocks currently fluctuates around 5 per cent (Baars et al., 2019; Dieperink, 2020; Stegeman, 2019). Furthermore, the volatility of all the risky assets in the experiment will be 11%. Volatility is the degree to which the asset value fluctuates. The further the possible outcomes of the risky asset are apart, the higher the volatility of the asset (Taylor, 2011). The chosen level of volatility is based on the volatility of the stock price index in the US of 2017, which is the most recent information I could find on the volatility of the stock market.16

At the beginning of the experiment, the participants are told that they have an hypothetical amount of €10.000 on their saving account. The initial savings are chosen to be a bit below the median level of household savings in the Netherlands, because the impact of negative interest rates on individuals with average and small savings has the specific attention of this study.17

The experiment consists of four rounds (see Table 1). In every round of the experiment, subjects choose how much of their endowment they save on an interest bearing saving account and how much they invest in a risky asset. The interest rate on the saving account is +3% in the first round. The two possible returns of the investment in the asset are +19% and -3%. In the following rounds, the endowment depends on the outcomes of the previous round. Furthermore, the interest rate on the saving account decreases with 2 per cent per round. The possible returns on the risky assets change in such way, that the risk premium (+5%) and volatility (11%) remain constant. To avoid that people will change their behaviour in the last round, purely because it is the last round, for example by taking more risk to give the outcome a last boost, the participants are not informed about the precise number of rounds of the experiment. The participants are told that part 1 of the experiment consists of five rounds at the most.

15_{The experimental design of this study largely corresponds with the experiment of Baars et al. (2019). §5.4} discusses on which parts this experiment differs from the experimental study of Baars et al. (2019). I will mainly discuss the reasons why I made some other decisions with respect to the experimental design.

16_See_{https://fred.stlouisfed.org/series/DDSM01USA066NWDB}_{for the annual stock volatility in the US from} 1998 until 2017.

17_{In 2019, the median savings in the Netherlands was €13.000 and the average household savings was €40.000} (Meinema, 2019, October 25).

(23)

22

Table 1: Overview of part 1 of the experiment Interest rate on saving

account

Possible returns on risky asset (both with 50% probability)

Round 1 +3% +19% / -3%

Round 2 +1% +17% / -5%

Round 3 -1% +15% / -7%

Round 4 -3% +13% / -9%

A conceptual overview of the experimental design is shown below.

Round 1:

Saving account Risky asset +19% return X +3% return X 100% probability -3% return Round 2:

Saving account Risky asset

+17% return X +1% return X

100% probability

-5% return

Round 3:

+15% return X -1% return X

100% probability

(24)

23

Round 4:

+13% return X -3% return X

100% probability

-9% return

Dynamic experimental design

In contrast to the existing experimental studies, this experiment has a dynamic design. The dynamic design of the experiment is based on an experimental study on the economic consequences of a Tobin tax (Hanke, Huber, Kirchler, & Sutter, 2010). This experiment is dynamic, because the endowment of every round depends on the outcome of the previous round. Therefore, the several rounds of the experiments are interconnected. The outcomes of the previous round are set as the default choice for the subsequent round. This means that subjects need to make an active choice to switch from saving to investing and vice versa. If participants do not make an active choice, the amount of money on the saving account and the amount of money invested in the asset will completely depend on the outcome of the previous round. To make adjustments to this allocations, participants of the experiment have to make an active choice. This is comparable to the real world, because in the real world the endowments of people are already, for example, on a saving account. To invest the money in the stock market, people have to make an active choice to switch from saving to investing. The dynamic design of the experiment increases the external validity of the study, because the choice architecture is closer to the real-world situation.

Within-subject design

The experiment has a within-subject design, which means that every participant of the experiment is confronted with all the treatments (Charness, Gneezy, & Kuhn, 2012). All subjects are confronted with all the combinations of interest rates and asset returns. The participants are confronted with two positive interest rates (+3% and +1%) as well as with two negative interest rates (-1% and -3%). With a within-subject design, causal relationships can be studied by analysing the way individual behaviour changes when the experimental circumstances change (Charness et al., 2012). In this experiment, this means that the relationship between (negative) interest rates and risk-taking behaviour can be studied by examining how risk-taking changes if the interest rate on the saving account declines.

A disadvantage of a within-subject design compared to a between-subject design could be that time and order effects might influence the participants’ behaviour. If an participant is exposed to all treatments, the participant’s experience in the first treatment might influence his or her behaviour in the following treatments. However, in a quit similar experiment, Baars et al. (2019) have found that the order of the varying interest rate scenarios did not influence the level of risk-taking of participants. Therefore, time and order effects do not seem to be of importance

(25)

24

in these kind of experiments. Furthermore, the participants in this research will be confronted with a declining interest rate pattern, because this fits with the real world, as the interest rate has been declining for decades (Tokic, 2017).

The respondent’s level of risk aversion is measured in the second part of the experiment. The level of risk aversion is the extent to which someone avoids risks. A risk averse person is averse towards the dispersion in a distribution (Wang, Rieger, & Hens, 2017). In the literature, several experiments have been used to study how people deal with risks (Crosetto & Fillipin, 2013; Eckel & Grossman, 2008; Holt & Laury, 2002).

The Eckel & Grossman task is used to measure risk aversion in this experiment. This incentivized task is chosen, because of its simplicity. Subjects will have no difficulties in understanding what is expected from them, which is especially important in an online experiment. In this task, participants choose one of the six alternative gamble options (see Table 2). Every gamble option has a 50% probability of a high payoff and a 50% probability of a low payoff. The expected return and standard deviation rises from gamble option 1 until gamble option 5. Payoff option 6 has the same expected return as option 5, but has a higher standard deviation. The higher the standard deviation of the gamble option, the riskier this choice is. A risk-averse person would choose an option with a low expected return and a low standard deviation. A risk-neutral person would choose option 5, because option 5 and 6 have the highest expected return and option 5 is less risky than option 6. Risk-seeking subjects would opt for option 6 (Dave, Eckel, Johnson, & Rojas, 2010). The chosen gamble options provide an index of the level of risk aversion of the participants. Choosing gamble option 1 corresponds with the highest level of risk aversion and choosing gamble option 6 corresponds with the lowest level of risk aversion (Eckel & Füllbrunn, 2015).

Table 2: Overview of part 2 of the experiment: risk aversion task

Payoff low Payoff high Expected return Standard deviation

Gamble option 1 €2,00 €2,20 €2,10 0.1 Gamble option 2 €1,60 €2,80 €2,20 0.6 Gamble option 3 €1,20 €3,40 €2,30 1.1 Gamble option 4 €0,80 €4,00 €2,40 1.6 Gamble option 5 €0,40 €4,60 €2,50 2.1 Gamble option 6 €0,00 €5,00 €2,50 2.5

Loss aversion means that losses have more impact on people than equal-sized gains. This is caused by the fact that the value function is steeper in the negative domain compared to the positive domain (Gächter, Johnson, & Herrmann, 2010; Tversky & Kahneman, 1991).

A lottery task, based on the task in Gächter et al. (2010), is used to measure the level of loss aversion of a participant (see Table 3). In this lottery task, people choose whether they want to participate in five heads and tails game. If the coin turns up heads, the result is a loss. If the coin turns up tails, the result is a win. For every lottery, the participant decides to accept or reject the lottery (Gächter et al., 2010). The only aspect that changes between the lotteries, is the amount which is lost if the coin turns up heads. If the coin turns up tails, the earnings are €2,00 constantly. If people behave rational, they would be willing to accept lottery 1 until lottery 3,

Negative interest rates and individuals’ risk-taking behaviour

Negative interest rates

and individuals’

Table of contents

List of abbreviations

1. Introduction

2. Background

3. Literature review

4. Research problem

5. Methodology