The impact of the credit crisis, 2007-2010, on Dutch pension funds

(1)

1

Bachelor thesis

The impact of the credit crisis, 2007-2010, on

Dutch pension funds

Date: 21‐06‐2018

Academic year: 2017‐2018

Supervisor: Spyridon Terovitis

Semester 2 periods 4 & 5

Name: Tim Noltes

Student number: 11027355

Abstract.

This research examines whether the credit crisis, 2007‐2010, has an effect on the investments of Dutch pension funds. In order to explore whether the credit crisis has an effect on the value change of investment and whether this effect is different between different types of pension funds, two different models are used. The sample of this research consists of three types of pension funds during the period between 2003 and 2014. This period is divided into three timeframes so that a distinction can be made between before, during and after credit crisis. The main finding of this research is that the credit crisis has a negative effect on the value change of investments. However, this effect is not different between different types of pension funds.

(2)

2

Statement of Originality.

This document is written by Tim Noltes who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document are original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of the completion of the work, not for the contents.

(3)

3

1. Introduction

The fundament of the Dutch retirement pension system consists of three pillars (Bikker and De Dreu, 2009). The first pillar is the Old Age Pension act (Dutch: Algemene ouderdomswet, AOW). This is a general basis pension for people older over 65 years in 2009. However, from 2010 the retirement age will gradually increase to 67 years in 2025 (De Grip, Fouarge & Montizaan, 2013). Because of this, the total pension contribution, paid through premiums by the workforce, will increase. The second pillar is the compulsory participation in the company. Employees save this pension in addition of the Old Age Pension Act. The third pillar is tax‐supporter scheme, which is regulated individually and is in addition of the other two pillars.

However, besides the fact that the retirement age is rising, mentioned by De Grip, Fouarge and Montizaan (2013), the percentage of Dutch people over the age of 65 relative to the workforce is expected to increase from 20 to 40 percent over the period 2009‐2030 (Kakes & Broeders, 2006). This is a problem for pension funds and for the society, because the retirement benefits of the pensioners will increase as well. This is one reason that pension funds play an important role in obtaining sustainable society (Amalric, 2006). Another reason is that pension funds have become dominant players in the financial market (Sandberg, 2013). Therefore, it is important that they make socially responsible investments. These investments take into account social, ethical and

environmental aspects. Pension funds also have an obligation towards present and future pensioners to pay retirement benefits.

Pension reforms are an important topic of policy debates for many years and became even more important as a result of the credit crisis (Bikker, Steenbeek & Torracchi, 2012). One reason for this is because retirement savings is a considerable asset for most of the people in developed countries, like the Netherlands. Financial stable pension funds are even more important in those countries.

Another reason is that given the objective of pension funds, to maximise their ability to meet their obligations, they are concerned about long‐term economic growth (Amalric, 2006). Economic growth will decline during a financial crisis. Hoque, Andriosopoulos, Andriosopoulos and Douady (2014)

(5)

5 argue that financial institutions take more risk and had worse returns during the credit crisis.

According to McKibbin and Stoeckel (2010), the bursting of the housing bubble, which is a cause of the credit crisis (Acharya, Philippon, Richardson & Roubini, 2009), has led to a decline in wealth and spending and an increase in defaults on loan held by financial institutions. They claim that the capital cost increased because of the rising risk. Also Jensen & Johannesen (2017) argue that the financial crisis caused a decline in corporate investments, employment and consumption. A reason for this is tightening of banks. Advanced economies, like the Netherlands, are highly levered and therefore depend on credit to sustain consumption. Furthermore, there is high quality of financial reporting and auditing during a financial crisis (Cimini, 2014). Therefore, the management earning, which is the ability of managers to manipulate the earnings, will decline. Another effect of a financial crisis is that the volatility of returns increased in financial markets during the credit crisis (Schwert, 2011). The volatility implies the risk of investments (Berk and DeMarzo, 2014, pp. 317‐319). The cost of capital will increase because of the increasing risk.

However, Duijm and Steins Bisschop (2017) suggest that short price movements do not have impact on pension funds because of their long‐term investments. They are like shock absorbers during financial distress. Furthermore, Bikker, Steenbeek and Torracchi (2012) argued that the Netherlands have a well‐developed pension system and therefore there is possible no effect of the financial distress on pension funds.

There are many different and conflicting signals from literature about pension funds during financial distress. Therefore, the research question that this paper aims to address is: Does the credit crisis, 2007‐2010, have an effect on the investments of Dutch pension funds? In order to get a

conclusion, there a two sub‐questions. The first sub‐question will be: Does the credit crisis affect the change in value of investments of Dutch pension funds? This question will be answered by means of a model with as dependent variable the value change of investments. The dependent variables are a crisis dummy variable, which is used for data during the crisis, and an after crisis dummy variable, which is used for data after the crisis. According to Alter and Schüler (2012), the credit crisis started

(6)

6 in June 2007 and ended in May 2010. In addition, there are an industry and a company dummy variable, because according to Bikker and De Dreu (2009), there are three types of pension funds. Namely, industry funds, company funds and professional funds. Bikker and De Dreu (2009) argue that industry funds are more efficient in managing of their investment operations and that professional funds have the highest cost of investments. Therefore, the type of pension funds will have an affect on the value change of investments. Finally, there is a control variable, number of pension funds, added to explain the dependent variable and so that no bias occurs. The hypothesis in this research is that there is no effect of the credit crisis. So if the coefficient of the Crisis dummy variable is different from zero, the hypothesis will be rejected, which implies that the claim that there is no effect of the credit crisis on the investments of Dutch pension funds will be rejected.

The second sub‐question is: Is the effect of the credit crisis different between the types of pension funds? In the second question, two interaction variables, “Crisis*Industry” and

“Crisis*Company”, will be added. These interaction variables indicate whether the effect of the credit crisis on the value change of investments depends on the type of pension funds. This is relevant for the research because there is a difference in efficiency of managing investments operations between types of pension funds, mentioned by Bikker and De Dreu (2009). There is a different effect of the credit crisis if the coefficients are different from zero.

The data used in this research is public data from CBS StatLine. Most of the data is available in the balance sheets and income statements of accumulated types of pension funds from the period 2003‐2014. This is a longitudinal research, which means that the data will be divided into three timeframes. These timeframes will be compared with each other. This data is only available annually and there is no public data about individual Dutch pension funds for this period. Therefore no statements are made in this paper about individual pension funds.

The results of this research show a negative effect of the credit crisis on the value change of investments of Dutch pension funds. This implies that the value change of investments is lower during the credit crisis. This research also shows that the value change of investments is higher after

(7)

7 the crisis compared with before and during the credit crisis. Furthermore, industry funds have higher value change of investments compared to other types of pension funds. However, company funds do not have an influence on the value change of investments. Finally, this research shows that there is no difference in the effect of the credit crisis, 2007‐2010, between types of pension funds.

To get an answer for the research question, the next section goes into details of previous literature about pension funds. After this section, the models will be described, and the way of testing will be explained. In the fourth section, the results will be showed and explained. In the last section, there will be the conclusion of this research. It contains also future research suggestions and shortcomings of this paper.

2. Literature

In this section the previous literature of pension funds and their performance will described. The literature mentioned in this section will contribute to the research because they have studied variables of the models used in this paper.

Pino and Yermo (2010) find out, by analyzing return on investments of all 34 OECD countries that on average the return on investments of pension funds of all OECD countries were negative with ‐ 21.4 percent. They argue that the return of investments in 2008 is lower among pension funds of countries with a ratio of equity to total assets over a third. The Netherlands has a ratio more over third. However, Pino and Yermo (2010) find out that on average the returns on investments of pension funds of the OECD countries were positive in 2009.

Sias (1996) explored that pension funds in general invest more in high‐risk investments, so with high volatility. There are two reasons for this. First, pension funds invest in investments with higher volatility because they want to outperform the market. Sias (1996) find only little evidence for this. Second, the volatility of securities is higher because of more institutional investors invest in them. Sias (1996) find more support for this reason. However, Faugère and Shawky (2003) find out that institutional investors prefer to invest in low‐risk investments during a decline of Nasdaq in the

(8)

8 period of March‐November 2000. So, pension funds prefer to invest in investments with low volatility during market stress. This evidence is consistent with statement of Duijm and Steins Bisschop (2017) that pension funds are like shock absorbers. Also, Pino and Yermo (2010) argue that a lot of pension funds decided to change their portfolio composition during the credit crisis to protect assets against financial distress. Pension funds decreased their equity risk and increased their investments in bonds and other asset classes.

The research of Bollar, Wittig and Kohler (2016) shows that the performance of pension fund in Swiss recover from the credit crisis, but that this recovery took longer than the asset performance suggest. They used asset performance of all Swiss pension funds and of one single Swiss pension fund with respect to a liability benchmark. In addition to the risk factors of assets, this measurement also includes the risk factors of the liabilities. Bollar, Wittig and Kohler (2016) find out that the return of assets increased after the credit crisis. However, during the crisis the liabilities increased, causing an increase in risk. Because of this, the recovery of the performance of Swiss pension funds after the crisis took shorter without considering the liability performance.

According to Bikker and De Dreu (2009), there are three main types of Dutch pension funds, namely industry funds, company funds and professional funds. By using 10,000 observations of all Dutch pension funds during the period 1992‐2004, they find out that the administrative and investment costs, operating costs, of industry are lower compared to company funds and

professional funds. A reason for this is that industry funds in general are more efficient in managing of their investments. The research also shows that the administrative and investment costs of company funds are higher compared to industry funds, but are lower than professional funds. The reason this is that company funds have specific benefits, such as tax gains, but these benefits come at high costs if the size of the pension fund is small. Bikker and De Dreu (2009) claim that professional funds have the highest cost of investments. They are more expensive because they have the smallest size. The research of Bikker and De Dreu (2009) shows that size of pension funds affects the

(9)

9 operating costs. They claim that these costs decline the return on investments and increase the cost of retirement security.

3. Data and Methodology

In this section the research method will be described. Furthermore, the models are shown and all the variables are explained. Two models will be used to answer the research questions. First, the model for the total investments of the pension funds is explained. Second, the model of change in value of investments will be described. The last model includes interaction variables to test whether the effect of the credit crisis differs between types of pension funds. Finally, the data will be

explained.

3.1 Methodology

To test what effect the credit crisis has on Dutch pension funds, the data is divided into three timeframes. Before the crisis is 2003‐2006, 2007‐2010 is during the crisis and 2011‐2014 is after the crisis. Because the data is determined at the end of each year, the whole year affected by the credit crisis is used to determine the effect of this crisis on the dependent variables.

Furthermore, it is important that the independent variables are not correlated with each other. Correlation implies the coefficient of the specific variable does not indicate the effect accurate. Therefore, the multicollinearity will be tested with the vif, variance inflation factor. If this value is greater than 10 there is significant multicollinearity. Multicollinearity implies that one independent variable can be linearly predicted from another independent variable. An interaction variable will be added to improve the estimates of regression coefficient. The correlation matrix will be used to do this.

Before the research question will be answered, the sub‐questions will be answered. The first sub‐ question is: Does the credit crisis affect the change in value of investments made by Dutch pension funds? The value change of investments model is used for this sub‐question. The second sub‐

(10)

10 question is: Is the impact of the credit crisis different between the types of pension funds? The value change of investments with crisis to pension fund type interaction variables model is used to answer this sub‐question. The models will be described in the following sections. The hypotheses will be tested with a significance level of 0.100. Hence, if the p‐value is lower than 0.100 there is a significant effect of the particular independent variable. To get the results, OLS, ordinary least squares,

regressions will be used. Both models are linear, so this type of regression shows representative results. An advantage of OLS regressions is that it is easy to implement and to interpret.

3.2 Empirical models: Value change of investments

According to Hoque, Andriosopoulos, Andriosopoulos and Douady (2014), financial institutions perform worse during financial distress. They argue that financial institutions take more risk and their returns decline. However, Duijm and Steins Bisschop (2017) argue that pension funds make long‐ term investments. Therefore, the credit crisis would have less or no effect on investments of pension funds. The following model is used to test whether the credit crisis affects the value change of investments of Dutch pension funds:

ℎ ( , ) = 0 + 1 ∗ + 2 ∗ +

3 ∗ + 4 ∗ + 5 ∗

+ ( , )

The dependent variable is value change of investment. It can be a negative amount, for example because of a fall in share prices in which pension funds have invested. The value change of

investment is measured over year t.

The “Crisis dummy” is used for the data during the credit crisis. According to Alter and Schüler (2012), the credit crisis starts in 2007 with financial stress. They argue that the credit crisis ends in 2010, when European governments set up a rescue funds. This shifted the risk of defaults from banks to governments. Thus, for the years 2007 until 2010 this dummy variable has a value of 1 and

(11)

11 otherwise a value of zero. The “After crisis dummy” is used for the years after the crisis. Hence, this dummy variable gets a value of 1 for the years 2011 until 2014. Despite of that the credit crisis started in June 2007 and ends in May 2010, the whole years are included in the “Crisis dummy”. This is because the data is only available at the end of each year and there is no information available about the distribution of the variables during the years. So, all years affected by the credit crisis are included in the dummy variable.

The “Industry funds dummy” and the “Company funds dummy” are included in this model because this indicates the type of pension fund. According to Bikker and De Dreu (2009), industry funds are more efficient in managing their investments and therefore have lower investment costs. Professional funds have the highest investment costs. Investment costs deduct return on

investments and with it the value of investments. The “Industry funds dummy” has a value of 1 for industry funds and zero otherwise. The “Company funds dummy” has a value of 1 for company funds and zero otherwise. If both dummy variables are zero, it is about professional and other funds.

The number of pension funds is a control variable. This variable has been added to explain the dependent variable, value change of investment, and so that no bias occurs. Number of pension funds is the total number of pension funds in the Netherlands. Epsilon is the error term of this model. The hypothesis of this question is:

 H0: β1=0, the credit crisis does not affect the value change of investments.  H1: β1≠0, the credit crisis does affect the value change of investments.

3.3 Empirical models: Value change of investments with crisis to pension funds type interaction variables

For this research it is also useful to test what effect the credit crisis has on a specific type of pension fund. According to Bikker and De Dreu (2009), the return on investments differs between the types of pension funds. Therefore, the second model of this research is:

(12)

12

ℎ ( , ) = 0 + 1 ∗ + 2 ∗ +

3 ∗ + 4 ∗ + 5 ∗ ∗ + 6 ∗

∗ + 7 ∗ + ( , )

The dependent variable is value change of investment of fund type i, with i is industry funds, company funds or professional and other funds and t is the particular year.

This model is similar to the previous model, but two interaction variables have been added. First, “Crisis*Industry” is included in this model. This interaction variable has a value of 1 for industry funds during the crisis and zero otherwise. Second, “Crisis*Company” has a value of 1 for company pension funds during the credit crisis and zero otherwise. Those interaction variables indicate whether the effect of the credit crisis on the value change of investments depends on the type of pension fund. The variable number of pension funds is similar to the previous model. Epsilon is the error term. The hypothesis of this model is:

 H0: β5= β6=0, there is no different effect of the credit crisis on the value change of different types of pension funds.

 H1: β5≠0 and/or β6≠0, there is a different effect of the credit crisis on the value change of different types of pension funds.

3.4 Data

The data used in this research is coming from CBS StatLine and is publicly available. The data for the value change of investments is coming from income statements per type of pension funds and therefore this is a quantitative research. The data is accumulated to use it for all Dutch pension funds. For all variables data is collected for the period 2003‐2014, so this is a longitudinal research. There is no public data of individual Dutch pension funds available for this period. Therefore, no statements at firm level will be made. Furthermore, all the data is determined at the end of each year.

(13)

13

Table 1: number of pension fund

This data is coming from CBS StatLine. The table shows the number of pension funds in the Netherlands. The most common type of pension fund is the company fund. Furthermore, the number of pension funds of all the three types declined over time. Huang and Mahieu (2012) argue that the biggest fund group outperforms the smallest fund group. Therefore, the pension funds are

consolidating.

4. Results.

In this section the results of the research will be shown and explained. First, the model of value change will be described. Finally, the model of value change of investments with crisis to pension fund types interaction variables will be shown and interpreted.

4.1 Results: Value change of investments

As mentioned in the previous section, data and methodology, it is important that the vif has no value above 10, because this would imply that there is multicollinearity. An interaction variable will be added to improve the estimates of the regression coefficients. However, the variance inflation factor is below 10 for all variables (see Appendix 1). Therefore, it can be assumed there is no

Year Industry funds Company funds

Professional

funds & others Total

2003 103 759 15 877 2004 102 722 17 841 2005 103 683 16 802 2006 103 650 16 769 2007 96 604 14 714 2008 95 547 14 656 2009 87 479 13 579 2010 82 417 13 512 2011 77 364 13 454 2012 74 327 13 414 2013 71 295 13 379 2014 68 268 12 348

(14)

14 multicollinearity and no interaction variable will be added. Hence, we can use an OLS regression of the value change of investment model to estimate the effects of the independent variables. The table below shows this regression.

Table 2: OLS regression of value change of investments

R‐squared = 0.3295 Independent variables Coefficient Standard error t P>|t|

Crisis dummy -1.85e+10 9.46e+09 -1.95 0.060

After crisis dummy 2.47e+10 1.08e+10 2.28 0.030

Industry dummy 2.18e+10 1.17e+10 1.86 0.072

Company dummy -9.37e+09 2.97e+10 -0.32 0.755

Number of pension funds 3.66e+07 5.58e+07 0.66 0.517

Constant -5.81e+09 1.02e+10 -0.57 0.573

The “Crisis dummy” has a p‐value of 0.060, which is lower than 0.100. Thus, the coefficient of the crisis period is significantly different from zero and H0 has been rejected. This means that the crisis affects the value change of investment. The coefficient is negative, so the value change of investments was lower during the crisis. This result is consistent with the research of Cimini (2014), Hoque, Andriosopoulos, Andriosopoulos and Douady (2014), McKibbin and Stoeckel (2010) and Schwert (2011). The value change of investments is € 18.5 billion lower during the credit crisis. The “After crisis dummy” coefficient is also significant different from zero. The p‐value of 0.030 is lower than 0.100. The coefficient is 24.7 billion, which implies that the value of change of investments is € 24.7 billion higher after the crisis compared to before the crisis. The value change of investments is € 43.2 billion higher after the crisis compared to during the crisis. Furthermore, the “Industry dummy” has a p‐value of 0.072, which is lower than 0.100. The value change of investments of Industry funds is € 218 million higher than other types of funds. This is consistent with the findings of Bikker and De

(15)

15 Dreu (2009). However, the “Company dummy” and the number of pension funds do not differ from zero. So, they do not have a significant effect on the dependent variable. Also, the intercept does not differ from zero, which means that if all the variables are zero, the value change of investments will also be zero. Furthermore, the R‐squared of this regression is 0.3295. This implies that the

independent variables of this model explain 32.95 percent of the variance of the value change of investments.

4.2 Results: Value change of investments with crisis to pension funds type interaction variables

The first regression shows a non‐significant coefficient of the “Company dummy”. This implies that the value change of investment of company funds is not different than other types pension funds. Therefore, it is not relevant to test whether the crisis has a different effect on the value change of investments of company funds relative to other pension fund types. This is why the “Crisis*Company” interaction variable will be excluded in the second empirical model. However, the coefficient of the “Industry dummy” is significant with a significance level of 0.100. Thus, it is relevant to explore whether the “Crisis*Industry” interaction variable has a significant effect on the value change of investments. So, the model will be:

ℎ ( , ) = 0 + 1 ∗ + 2 ∗ +

3 ∗ + 4 ∗ + 5 ∗ ∗ + 6 ∗

∗ + ( , )

The variance inflation factors for this model are below 10 (see Appendix 2). So, there is no reason to assume that there is multicollinearity in this model. Therefore, the OLS regression can be used to estimates the coefficients of the model. The regression below shows these estimates.

(16)

16

Table 3: OLS regression of value change of investment with crisis to pension funds type interaction variables

R‐squared = 0.3772 Independent variables Coefficient Standard error t P>|t|

Crisis dummy 8.69e+09 1.14e+10 -0.77 0.450

After crisis dummy 2.47e+10 1.06e+10 2.33 0.027

Industry dummy 3.15e+10 1.32e+10 2.39 0.024

Company dummy -9.51e+09 2.91e+10 -0.33 0.747

Number of pension funds 3.69e+07 5.47e+07 0.67 0.505

Crisis*Industry -2.93e+10 1.97e+10 -1.49 0.147

Constant -9.09e+09 1.02e+10 -0.89 0.382

The “Crisis*Industry” interaction variable has a p‐value of 0.147. This p‐value is above 0.100 and therefore the coefficient is not significantly different from zero. H0 has not been rejected and therefore there is no different effect of the credit crisis on different pension funds. This is not consistent with the expectation of this research. Furthermore, the included interaction variable has changed the results. The coefficient of the “Crisis dummy” is non‐significant instead of significant. Namely, the p‐value of the “Crisis dummy” is 0.450, which is greater than 0.100. The “After crisis dummy” and the “Industry dummy” are still significant, with a p‐value of 0.027 and 0.024. Therefore, both variables have an effect on the value change of investments. The coefficient of the “After crisis dummy” is 24.7 billion, which implies that the value change of investments is € 24.7 billion higher after the crisis than before and during the credit crisis. The coefficient of the “Industry dummy” is 31.5 billion, which implies that the value change of investments of Industry funds is € 31.5 billion higher than company and professional funds. In addition, the coefficient of the number of pension funds is not significantly different from zero. So, there is no influence of the number of pension funds to the change of investments of the pension funds accumulated per type of pension fund. The

(17)

17 intercept is also not significant, which means that if all independent variables are zero, the value change of investments will also be zero. Finally, the R‐squared is 0.3772. Thus, the variance of the value change of investments is 37.72 percent explained by the independent variables of this model.

5. Conclusion.

The purpose of this thesis was to explore whether the credit crisis, 2007‐2010, affect the

investments of Dutch pension funds. Two sub‐questions were used to get an answer to this research question. Based on literature, it was expected that the credit crisis does have an effect on the value change of investments. It was also expected that the credit crisis does have a different effect on different type of pension funds. To see if these expectations are true, a dataset of 36 observations was used. OLS regressions and a significance level of 0.100 are used to get a result.

The first regression shows a significant negative effect of the credit crisis on the value change of investments, which is the answer on the first sub‐question. There are several reasons for this effect. First, Schwert (2011) argues that the volatility increased during the credit crisis. Therefore, the cost of capital increased because of the rising risk during the credit crisis (McKibbin and Stoeckel, 2010). Second, the corporate investments decreased during the crisis (Jensen & Johannesen, 2017) Third, management earnings declined during the crisis (Cimini, 2014). The quality of financial reporting and auditing increased and thus it became more difficult for managers to manipulate the earnings.

The first regression also shows a significant positive effect after the crisis on the value change of investments. A reason for this can be the rising retirement age since 2010. The pension contribution increased, so more can be spent. In addition, the bank rescue schemes shifted risk from banks to governments (Alter & Schüler, 2012). As a result, the investments of financial institutions increased. Furthermore, the effect of industry funds on the value change of investments is significant. The reason behind this finding is that industry funds are more efficient in managing of their investments. Because of this the cost of capital is lower compared to other funds. In addition, the effect of

(18)

18 of capital nor the lowest (Bikker and De Dreu, 2009). Therefore, the value change of investments will be around the average of all pension funds. Finally, the number pension funds do not affect the value change of investments. This because the pension funds are consolidating (Huang and Mahieu, 2012). The consolidated pension funds can invest about as much as the pension funds separately.

The second regression shows a non‐significant effect of the “Crisis*Industry” interaction variable on the value change of investments. This implies that there is no different effect of the credit crisis on the value change of investments of different types of pension funds, which is the answer on the second sub‐question. Furthermore, the effect of the crisis changed from significant negative effect to a non‐significant effect. A reason for this is can be that the effect of the crisis is divided between the interaction variable and the “Crisis dummy”. In this regression, the value change of investments is also higher after the crisis compared to before and during. The effect of industry funds has also remained significant. Furthermore, the effect of company funds and number of pension funds has remained insignificant.

In short, there is an effect of the credit crisis, 2007‐2010, on the investments of Dutch pension funds. During the credit crisis, their value change of investments is lower compared to before and after the crisis. This effect is not different between different types of pension funds.

A limitation of this research is that there is no data about individual pension funds available for the period between 2003 and 2014. Therefore, no statements can be made about individual pension funds. It would be a suggestion for further research to examine whether there is an effect of the credit crisis on individual pension funds. Another limitation is that the pension system varies per country, so the results of this research cannot be generalized to other countries. Hence, it would be a good idea to explore whether the credit crisis affects the pension funds in other countries. Finally, because the data is determined at the end of each year, the whole year affected by the credit crisis is used to determine the effect of this crisis on the dependent variables. However, Alter and Schüler (2012) argue that the credit crisis started in June 2007 and ended in May 2010. It is not possible to accurately determine the values of the variables in June or May.

(19)

19 Reference list

Acharya, V., Philippon, T., Richardson, M., & Roubini, N. (2009). The financial crisis of 2007‐2009: Causes and remedies. Financial markets, institutions & instruments, 18(2), 89‐137. doi:10.1111/j.1468‐0416.2009.00147_2.x

Alter, A., & Schüler, Y. S. (2012). Credit spread interdependencies of European states and banks during the financial crisis. Journal of Banking & Finance, 36(12), 3444‐3468.

doi:10.1016/j.jbankfin.2012.08.002

Amalric, F. (2006). Pension funds, corporate responsibility and sustainability. Ecological Economics,

59(4), 440‐450. doi:10.1016/ecolecon.2015.11.009

Berk, J., & DeMarzo, P. (2014). Corporate Finance (3rd ed.) Harlow, England: Pearson

Bikker, J. A., & De Dreu, J. (2009). Operating costs of pension funds: the impact of scale, governance, and plan design. Journal of Pension Economics & Finance, 8(1), 63‐89.

doi:10.1017/S1474747207002995

Bikker, J. A., Steenbeek, O. W., & Torracchi, F. (2012). The Impact of Scale, Complexity, and Service Quality on the Administrative Costs of Pension Funds: A Cross‐Country Comparison. Journal

of Risk and Insurance, 79(2), 477‐514. doi:10.1111/j.1539‐6975.2011.01439

Bolla, L., Wittig, H., & Kohler, A. (2016). The liability market value as benchmark in pension fund performance measurement. Journal of Pension Economics & Finance, 15(1), 90‐111.

doi:10.1017/S147474721400033X

Cimini, R. (2015). How has the financial crisis affected earnings management? A European study.

Applied economics, 47(3), 302‐317. doi:10.1080/00036846.2014.96982

De Grip, A., Fouarge, D., & Montizaan, R. (2013). How sensitive are individual retirement expectations to raising the retirement age?. De Economist, 161(3), 225‐251. doi:10.1007/s10645‐013‐9205‐5

(20)

20 Duijm, P., & Steins Bisschop, S. (2017). Short‐termism of long‐term investors? The investment

behaviour of Dutch insurance companies and pension funds. Applied Economics, 1‐12. doi: 10.1080/00036846.2017.1420898

Faugère, C., & Shawky, H. A. (2003). Volatility and institutional investor holdings in a declining market: A study of Nasdaq during the year 2000. Journal of Applied Finance Fall-Winter, 32‐42.

Hoque, H., Andriosopoulos, D., Andriosopoulos, K., & Douady, R. (2015). Bank regulation, risk and return: Evidence from the credit and sovereign debt crises. Journal of banking & finance, 50, 455‐474. doi:10.1016/j.jbankfin.2014.06.003

Huang, X., & Mahieu, R. J. (2012). Performance persistence of Dutch pension funds. De Economist,

160(1), 17‐34. doi:10.1007/s10645‐011‐9176‐3

Jensen, T. L., & Johannesen, N. (2017). The Consumption Effects of the 2007–2008 Financial Crisis: Evidence from Households in Denmark. American Economic Review, 107(11), 3386‐3414. doi:10.1257/aer.20151497

Kakes, J., & Broeders, D. (2006). De houdbaarheid van het Nederlandse pensioenstelsel. Occasion

studies, 4(6). Retrieved from https://www.dnb.nl/binaries/occ%20stud6_tcm46‐146651.pdf

McKibbin, W. J., & Stoeckel, A. (2010). The global financial crisis: Causes and consequences. Asian

Economic Papers, 9(1), 54‐86. doi:10.1162/asep.2010.9.1.54

Pino, A., & Yermo, J. (2010). The Impact of the 2007‐2009 Crisis on Social Security and Private Pension funds: A threat to their financial soundness?. International Social Security Review, 63(2), 5‐30. doi:10.1111/j.1468‐246X.2010.01359.x

Sandberg, J. (2013). (Re‐) Interpreting Fiduciary Duty to Justify Socially Responsible Investment for Pension Funds?. Corporate Governance: An International Review, 21(5), 436‐446.

doi:10.1111/corg.12028

Schwert, G. W. (2011). Stock volatility during the recent financial crisis. European Financial

(21)

21 Sias, R. W. (1996). Volatility and the institutional investor. Financial Analysts Journal, 52(2), 13‐20.

(22)

22

Appendix 1: Table vif value change of investments.

Appendix 2: Table vif value change of investments with crisis to pension funds type

interaction variables.

Variable Vif

Company dummy Number of pension funds Industry dummy

After crisis dummy Crisis dummy 9.88 8.93 1.53 1.47 1.00 Variable Vif Company dummy Number of pension funds Industry dummy

Crisis*Industry Crisis dummy After crisis dummy

9.88 8.93 2.02 2.00 1.50 1.47

The impact of the credit crisis, 2007-2010, on Dutch pension funds

Bachelor thesis