The effect of the mortgage rate on the house prices in the Netherlands

(1)

The effect of the mortgage rate on the

house prices in the Netherlands

16-02-2018

Eric Koelmans, 10557768 Supervisor: Drs. P.V.Trietsch BSc Business Economics Finance and Organization Amount of EC’s: 12

(2)

2 | P a g e

Abstract

This thesis investigates the effect of the mortgage rate on the house prices in the

Netherlands. This effect is compared to the effect of the mortgage rate on the house prices in the four largest municipalities of the Netherlands. The determinants that are needed in order to investigate the house prices, are quarterly used from 1995 till 2017. Therefore, a time series regression is used for each area. Each time series analysis is using data from the past, to make a prediction for the future. The results show in all five areas, that the

mortgage rate have a negative influence on the house prices. However, this negative effect is only significant in the municipalities and not in the Netherlands as whole.

Statement of originality

This document is written by student Eric Koelmans, 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

(3)

3 | P a g e

1. Introduction

1.1 Describing the problem

The housing market in the Netherlands is an important factor of the Dutch economy as in June 2017, the mortgage debt was about 760 billion euro (CBS, 2018). In the past years, both the house prices and the mortgage rate have changed significantly in the Netherlands. According to the Dutch Central Bureau of Statistics, the average house prices have increased from €93.750 in 1995 to €267.868 in November 2017. The average house prices have

increased even more in the four largest municipalities; Amsterdam, Rotterdam, Utrecht and ‘s Gravenhage where the average house prices in Amsterdam have increased from €270.507 in 2012 to €358.976 in 2016. According to Ger Jaarsma, the increase of the Dutch house prices are detrimental for first-time buyers (NVM, 2017). This is because the mortgage lending regulation tightened after the economic crisis in 2013 and some have the

attendance of a student debt (Senne Janssen, 2017). Therefore, more savings are needed to purchase a house. Besides that, the supply-side of the housing market is decreasing,

especially in the large municipalities. The tight supply leads to above average house prices.

1.2 Factors that drive house prices

Changes of macro-economic factors play an important role in affecting the house prices (Beltratti and Morana, 2010). Gross Domestic Product (GDP), consumer price index (CPI), interest rate and consumer trust are factors that determine the house prices (Goodhart and Hofmann, 2008). Moreover, the diminishing mortgage interest rate is one of the main factors driving house prices according to Aoki et al. (2004). The mortgage costs determine both the higher house prices and the lower financing costs. Due to the lower mortgage rate in the Netherlands, the interest expenditure decreased. According to Senne Janssen, it is therefore an attractive opportunity for investors to profit from the current housing market (2017). All in all, the lower mortgage rate did improve the affordability of the houses in the Netherlands which in turn increased the demand. Higher demand eventually leads to an increase in house prices in the long-run (Green and Hendershott, 1996).

(5)

5 | P a g e

1.3 Research question

The question that will be answered in this thesis is: what is the effect of the mortgage rate on the house prices in the Netherlands in the period 1995 to 2017? This thesis also

investigates if the effect of the mortgage rate for the Netherlands as a whole differs from the effect in the four largest municipalities.

1.4 Data and methodology

The data used in this thesis is provided by the Dutch Central Bureau of Statistics (CBS) and the ‘Nederlandse Vereniging voor Makelaars’ (NVM). The mortgage rate is used as the main explanatory variable of the house prices. The macro-economic factors function as control variables, which can be divided into four groups: macro-economic factors, demographic factors, supply and financial factors. The control variables that will be fixed are GDP, unemployment rate, consumer trust, consumer price index and loan to value. The control variables that vary for each area are house supply and the population density. Furthermore, the effect of the mortgage rate on the house prices will be investigated quarterly during the period 1995 till 2017. This time frame is necessary in order to investigate the long-term effect between the mortgage rate and the house prices.

1.5 Regression

The research question will be answered with the use of a time series analysis. A total of five regressions will be provided in order to compare the results between the four municipalities and the Netherlands. The hypothesis is that the mortgage rate influences the house prices negatively. Moreover, the mortgage rate is expected to have less effect on the house prices in the Netherlands compared to the four municipalities. This is due to the fact, that the house price increased less in the Netherlands as a whole compared to the municipalities while the mortgage rate is the same for each area.

1.6 Structure

In chapter two of this thesis, a review of the existing literature of the effect of the mortgage rate on house prices is given. Besides that, this review will establish the house price

determinants. Chapter three describes the data and the methodology that is used. Chapter four contains the results and the analysis. Lastly, the conclusion is given in chapter five.

(6)

6 | P a g e

2. Literature

This section describes the determinants of house prices. In the past years, the structure of the housing market and its determinants has been researched. It is investigated how house prices are affected and by which extent of the determinant. This literature review is ordered by determinants. It starts with a visual conceptual model after which each determinant is described.

2.1 Visual conceptual model

In the model below, all the determinants according to the written literature are listed. The dependent variable is the average house price. The expected positive relation is marked with a ‘+’ and a negative relation with the ‘-‘ sign.

Figure 1: Visual conceptual model

Mortgage rate

Main explanatory variable

Supply

Unemployment

_rate

Consumer trust

Loan-to-value

Population density

Average

house price

Gross domestic

product

(7)

7 | P a g e

2.1.1 Mortgage rate

The mortgage rate is the percentage of interest that have to be paid on the amount that is lent by the mortgage institution (McQuinn and O’reilly, 2007). It is determined by conjunctural and structural factors. GDP and residential property prices are examples of the demand-side, while funding methods is a supply factor that influences the mortgage rate (Sørensen and Lichtenberger, 2007). According to Sørensen and Lichtenberger, country-specific features like the loan-to-value ratio, also affects the mortgage rate positively. More credit risk is inducing a higher mortgage rate.

McQuinn and O’reilly researched the relation between house prices and the role of income and interest rate. Two of the main determinants that drove up the house prices are income levels and interest rate. According to Mcquinn and O’reilly it is common in empirical models that both income levels and interest rates are insignificant or give a wrong

coefficient. Therefore, they came up with a model in which demand is characterized by the amount an individual can borrow. This in turn, is determined by both the income level as the interest rate (McQuinn and O’reilly, 2007). The dynamic ordinary least squares methodology is used to determine the estimators of a single equation: 𝑌𝑡 = 𝛽0 + 𝛽1 𝑋1,𝑡 + 𝛽2 𝑋2,𝑡+… + ε𝑡.

To check the model for autocorrelation, they use the Breusch-Godfrey test. A negative and significant effect, on long term between the house prices and the amount that can be borrowed, is found.

Another paper that shows the mortgage rate has a significant effect on the house prices, is from Goodhart and Hofmann (2008). They analysed 17 industrialised countries and compared them quarterly during the period of 1970 till 2006. With the use of a fixed-effects panel VAR, they found that GDP, CPI and the interest rate have a significant effect on house prices. Furthermore, they concluded that this effect was even stronger in times of booming house prices.

Additionally, is concluded that the appreciation expectations are an important determinant as well of the house prices (Harris, 1989). In the 1970’s, both the house prices and interest rates increased on average in the United States. He found that an increase of the nominal interest rate, can enlarge the expectations of a further increase. Therefore, an increase of the nominal interest rate can lead to an increase of home ownership. The real interest rate is negatively correlated with the change of house prices. (Harris, 1989)

(8)

8 | P a g e

2.1.2 GDP

The gross domestic product (GDP) is the summary statistic of the economic activity over a period of time (Konchitchky and Patatoukas, 2014). According to Konchitcky and Patatoukas, it can be measured by the income approach that equals all generated income by products and services. Égert and Mihaljek researched the determinants of house prices in Central and Eastern Europe and investigated whether conventional fundamental

determinants influence the house prices in eight transition economies and nineteen OECD countries. They showed the GDP per capita is an alternative for income per capita as there is a strong correlation (Égert and Mihaljek, 2007). Therefore, it is expected that the GDP has a positive influence on the house prices as an increase in economic activity encourages a higher demand in the housing market. A panel DOLS estimation is used to research the long-term relationship. Leads and lags are added to the regression to account for endogeneity of the regressors. The error correction portrays the relationship between house prices and the independent variables. Égert and Mihaljek proved that the income elasticity is significantly higher in the transition countries compared to the nineteen OECD countries and that the GDP is positively correlated with the house prices (2007).

In the article ‘Determinants of real house price dynamics’ by Capozza (2002), the real house prices are estimated with the use of a panel data set of 62 metro areas. The

determinants of the long-run equilibrium equation are estimated by using both OLS and a panel data estimator that controls for fixed effects. His empirical estimates show that an increase of real income with one percent, leads to almost a half percent increase of house prices. (Capozza, 2002)

2.1.3 Supply

In the regression of Capozza, supply consists of land that is available for

development. It explores two kinds of hypothesis, transactions-based explanations and supply based theories. Transaction based explanations means that transaction frequencies are used in the housing market. This is due to the heterogeneity in the house price

dynamics. An economic shock can lead to higher transaction frequencies in one area compared to other areas. On the contrary, a positive economic shock that increases the demand of housing stock can affect the house prices differently. This is due to the fact that it

(9)

9 | P a g e differs for each area to rapidly increase supply at low costs. Therefore, supply factors do influence price dynamics negatively in time series analysis (Capozza, 2002). He finds a negative and significant effect on the house prices. (Capozza, 2002)

2.1.4 Unemployment rate

The unemployment rate can be defined as the share of the labour force that is unemployed (Meen, 2010). Economic changes like population, employment and income are used as determinants in order to measure the house price changes. Clapp and Giaccotto investigated the influence of economic variables on local house price dynamics. A long-run ordinary least squares equation is used to define the house price changes. An increase of the unemployment rate in a specific region means that less people are certain of having an income. This could decrease the demand in the housing market, and affects the house prices (Clapp and Giaccotto, 1993). Therefore, house prices respond negatively to unemployment. The negative coefficient is even more sensitive in high priced areas. (Clapp and Giaccotto, 1993)

Additionally, Capozza used the unemployment rate as a determinant for the house price changes In the OECD countries. In all countries, the coefficients were negative and significant different from zero. (Capozza, 2002)

2.1.5 Consumer trust

In times of a bubble, overconfident households purchase houses that they normally could not afford (Case and Shiller, 2003). Resulting from the expectation of a price increase in the future. Confidence is an important factor in determining house price changes (Case and Shiller, 2003). Case and Shiller investigated whether there was a housing bubble in the United States and other countries with the use of a questionnaire. This survey investigated several aspects of home buying behaviour. A fall in house price is detrimental for

homeowners. This decrease could lead to a deterioration of the consumer trust and the purchase of houses is postponed. Now with the expectation of having lower prices in the future and to avoid capital loss. (Case and Shiller, 2003)

Furthermore, there is a strong effect of the Dutch index of consumer confidence on the house prices (Rouwendal and Longhi, 2008). Rouwendal and Longhi tried to find an

(10)

10 | P a g e explanation for the short-run developments in the housing market. They assumed that the irrational forces are presumably responsible for the house price changes during a bubble, however not the spatial pattern. Therefore, they took psychological variables, like optimism or pessimism, in consideration in order to explain the house price changes in the short-run. With the use of a OLS regression, they found that the economic fundamentals had a

moderate influence on the house price changes in 1999 and 2000. Even common housing market indicators were not able to explain the house price increase in the short-run. (Rouwendal and Longhi, 2008)

2.1.6 CPI

The consumer price index is used to change the nominal house prices into real house prices (Goodhart and Hofmann, 2008). According to Goodhart and Hofmann, inflation has to be taken into account, with the use of CPI, to get an unbiased estimation of the house price changes. It is expected that an increase of the CPI induces lower house prices as it become more expensive over the years. In an empirical analysis, it is found that CPI had a significant effect on house prices (Goodhart and Hofmann, 2008). An increase of the CPI will mainly influence supply. It will cause the real GDP to fall and this in turn leads to a decrease of house prices, money and credit in real terms. The only variable that will increase with a higher CPI, is the nominal interest rate. These results show a multidirectional relationship between the variables. (Goodhart and Hofmann, 2008)

Furthermore, all nominal house prices are also deflated by the CPI by Égert en Mihaljek (2007). In their regressions, the real house prices are used, instead of the nominal house prices. The increase in house prices, now include both price changes and inflation.

2.1.7 Loan to value

The loan to value is the share of the mortgage in percentage of the house price. A tightening of the LTV is reducing the access to finance (Igan and Kang, 2011). Therefore, it is expected that a higher LTV leads to more demand and higher house prices. Igan and Kang investigated the effect of macro prudential measures on house prices and real estate booms. It shows the impact of loan-to-value on house price dynamics, mortgage credit availability and the

(11)

11 | P a g e variable was used. The independent variable is a matrix of control variables with macro-economic factors, monetary policy and expectations. They concluded that, reducing the LTV leads to a decline of the house price appreciation. The appreciation however, does slow down later compared to housing market activity. Adjustments to the LTV rules take 3 months to have a significant effect on the housing market activity and 6 months on the house prices. Furthermore, tightening the LTV rules will lead to lower expected house price changes and thus curb the effect of postponing house purchases. (Igan and Kang, 2011)

2.1.8 Population Density

Population density is the number of inhabitants living per squared kilometre and is usually increasing in major cities which is called the urbanization effect (De Bruyne and Van Hove, 2013). De Bruyne and Van Hove investigated the change in house prices with spatial variation. The article focuses on house prices in different geographical locations and municipalities. The results of this study show that geographical factors affect the house prices. Moreover, municipalities with a high population growth have a larger significant effect on house prices than municipalities with a low population growth. It can be concluded that the location of a house is important as well in determining the house price. The closer to a provincial capital, the higher the house prices. (De Bruyne and Van Hove, 2013)

(12)

12 | P a g e

2.3 Summary determinants

Variable

Literature

article

Description

Hypothesis

Methodology

(R²)

Mortgage rate - Mcquinn & O’reilly - Goodhart and Hofmann - Significant effect on house prices in long term - Effect even stronger in booming house prices H0: Mortgage rate coefficient equals 0 H1: Mortgage rate coefficient is negative Dynamic ordinary least squares (DOLS) (0,95) Gross domestic product - Égert and Mihaljek

- House prices are highly determined by GDP

- GDP is alternative for income per capita

H0: GDP coefficient equals 0 H1: GDP coefficient is positive A panel DOLS estimation (0,85)

Supply - Capozza - Supply factors

influence price dynamics in time series analysis H0: Supply coefficient equals 0 H1: Supply coefficient is negative Ordinary least squares regression (0,65) Unemployment rate - Clapp and Giaccotto - House prices respond negatively to unemployment - More sensitive in high priced areas H0: Unemployment rate coefficient equals 0 H1: Unemployment rate coefficient is negative Ordinary least squares regression (0.76)

Consumer trust - Case and Shiller - Rouwendal and longhi - Confidence is an important factor in determining house price changes H0: Consumer trust coefficient equals 0 H1: Consumer trust coefficient is positive - Questionnaire - Ordinary least squares regression (0,93)

Loan-to-value - Igan and Kang - A tightening of the LTV rules will significantly lead to lower price appreciation H0: LTV coefficient equals 0 H1: LTV coefficient is postive Ordinary least squares regression Population density - De Bruyne and Van Hove - Higher population density increases the house prices

H0: Population density coefficient equals 0 H1: population density coefficient is positive Ordinary least squares regression (0,73)

(13)

13 | P a g e

3. Data and Methodology

In this section, all the data is provided. In table 2, a summary is given of the descriptive statistics of the variables.

Mean Std. Dev. Min Max Inhabitants

Real house prices

(The Netherlands)

€156.615 28326 €89.792 €197.067 17.194.371

(Amsterdam)

€189.768 40320 €93.957 €276.999 853.312

(‘S Gravenhage)

€139.275 31631 €68.141 €182.929 526.439

(Rotterdam)

€120.576 23568 €65.321 €153.889 639.587

(Utrecht) €159.739 32509 €81.864 €206.428 344.384 Mortgage rate 5.20% 1.07 2.85% 7.77% - Nominal GDP €136.729 30102 €79.042 €183.744 - Real GDP €106.992 12030 €79.042 €123.710 - Unemployment rate 5.27% 1.24 3.3% 8.1% - Consumer trust -0.91 20.92 -39.67 35.33 - Loan-to-value 103.45 13.8 89.42 147.87 - House supply (The Netherlands) 76237 42304 18820 148732 - House supply (Amsterdam) 3289 2080 506 7130 - House supply (’S Gravenhage) 2274 1202 666 4651 - House supply (Rotterdam) 2326 1424 429 4680 - House supply (Utrecht) 1376 976 274 3716 - Population Density (Netherlands) 482 15.43 455 507 - Population density (Amsterdam) 4591 215.98 4330 5111 - Population density (‘s Gravenhage) 6111 337.03 5511 6522 - Population density (Rotterdam) 2894 44.42 2831 2986 - Population density (Utrecht) 3323 396 2664 3840 -

Table 2: Summary statistics of variables Sources: CBS, DNB survey and NVM

(14)

14 | P a g e

3.1 Real House prices

The nominal house prices are provided by the Dutch Central Bureau of Statistics. The price changes of the municipalities and the Netherlands itself have been compared quarterly from 1995 till 2016. Therefore, each variable has 91 observations. The nominal house prices of each area are determined by taking the average of the selling prices. The descriptive statistics of the nominal house prices are shown in table 7 in appendix 1.

However, in this thesis the real house prices are used as dependent variable. The house prices are adjusted for inflation by the use of the consumer price index (CPI). This index shows the movement of prices of the same goods and is also provided by the CBS. The CPI of the first quarter of 1995 is used as benchmark. Now the real effect of the mortgage rate on house price is investigated, without having changes in the price itself. The descriptive statistics are shown in table 2. By comparing the descriptive statistics of both the nominal and the real house prices, it can be concluded that the real house prices have a lower mean. This means that the prices itself have increased during the period of 1995 till 2017, which is called inflation.

As can be seen in figure 2, the real average house price change of the Netherlands is during the period of 1995 till 2016 in between the four municipalities. The average nominal house price is in Amsterdam the highest and especially after 2014.

3.2 Mortgage rate

The mortgage rate is the main explanatory variable in this thesis. According to ‘De Nederlandsche Bank’, 52% of Dutch households chose for five and ten-year fixed rate mortgages in 2016 (DNB, 2016). In this thesis, the 10-year mortgage rate is used due to the fact that the rates are more fluctuating and have a broader range.

The interest rate is provided by both the Dutch CBS and ‘De Nederlandsche Bank’ (DNB). DNB is providing all the different mortgages rates, with different maturities, in the Netherlands quarterly from 2003 till now. Therefore, the residual term is provided by the CBS. The Dutch CBS is proving three different mortgage rates: low, average and high. As in this thesis the ten percent mortgage is used, the high mortgage rates coincide with the interest rates from the DNB. The mortgage interest rate is in the third quarter of 2017 almost three times as small as in the first quarter of 1995.

(15)

15 | P a g e Figure 2 shows the diminishing mortgage rate whilst at the same time the house prices are increasing in all areas. This figure suggests the correlation will be negative between those two variables. Therefore, the alternative hypothesis in this thesis will be that the coefficient of the mortgage rate is negative as expected in the hypothesis in section 2.1.

3.3 Gross domestic product

The gross domestic product is measured by adding the total amount of produced goods and services, minus the amount of goods that is used during the process. It is the gross added value of a country over a certain period of time. The Dutch CBS is providing the nominal GDP each quarter from 1995 till 2017. However, the real GDP is used as a determinant of the house prices. To adjust the nominal GDP for inflation, the consumer price index is used. The real GDP is calculated by the nominal value divided by the decimal form of the price index.

In table 2 and figure 4 of appendix 2, the descriptive statistics of both the nominal and real GDP are shown. In the first quarter of 1995 the nominal GDP amounts €79.042. This amount equals the GDP per quarter. The total GDP per year is divided into four quarters. At the end of 2017, this amount grew more than 150%. The real GDP on the contrary, started also at €79042, however was €123.710 at the end of 2017. This shows that, excluding inflation leads to both a lower maximum and mean.

Figure 2 – House price changes & mortgage rate Source: CBS 0.00% 1.00% 2.00% 3.00% 4.00% 5.00% 6.00% 7.00% 8.00% 9.00% € - € 50,000.00 € 100,000.00 € 150,000.00 € 200,000.00 € 250,000.00 € 300,000.00 19 951 19 962 19 973 19 984 20 001 20 012 20 023 20 034 20 051 20 062 20 073 20 084 20 101 20 112 20 123 20 134 20 151 20 162 20 173

House prices & mortgage rate

Real house price Netherlands Real house price Amsterdam Real house price 's Gravenhage Real house price Rotterdam Real house price Utrecht Mortgage rate

(16)

16 | P a g e

3.4 House supply

The house supply is provided by the ‘Nederlandse Vereniging van Makelaars en Taxateurs’ (NVM). It is measured by the housing stock that is available for sale for each area. The NVM took the median of each quarter to determine the house supply.

Table 2 shows that Amsterdam is on average the municipality with the highest number of houses that are available for sale, whilst Utrecht has on average the lowest average. In both figure 5 and 6 of appendix 2, an increase of the house supply is noticeable from 1995 till the end of 2013. It can be concluded that the economic crises did have an impact on the housing market. From the beginning of 2009 till the end of 2013, the house supply increased significantly. After 2013, the house supply diminished in all areas, especially in Amsterdam.

3.5 Unemployment rate

The unemployment rate is measured as the share of the labour force that is unemployed and is quarterly provided by the CBS. The employment rate is only available for the

Netherlands as a whole. Therefore, the same rate is used for all five areas. Besides that, the unemployment rate was only available yearly from 1995 till 2003. During this period, the yearly rate is used for each quarter of that specific year. From 2003 till 2017 the

observations are determined quarterly.

According to table 2, the average unemployment rate during the period of 1995 till 2017 is 5.27%. Figure 7 of appendix 2 shows that the majority of the time, the employment rate is decreasing whilst the house prices are increasing and vice a versa. This shows that there is a negative correlation as expected in section 2.1.

3.6 Consumer trust

The CBS is providing the consumer trust each month. It is in the Netherlands based on the economic and financial situation of each individual. Each month, 1000 random households are asked to answer five questions. They can be answered by: improved, not changed, do not know or worsened. The questions are the following:

(17)

17 | P a g e 1. The financial situation of the past 12 months

2. The financial situation of the coming 12 months

3. Whether it is a good time to purchase expensive goods 4. The economic situation of the past 12 months

5. The economic situation of the coming 12 months

The consumer trust is measured by the difference between the percentage of the consumer that answered ‘improved’ and the percentage that answered ‘worsened’. The average of the last 22 years is -0.91. In the first quarter of 2000, the consumer trust reached the highest amount of points, which equals 35.33. During the economic crisis, the beginning of 2013, the consumer trust reached rock-bottom: -39.67. As shown in figure 8 of appendix 2, the

consumer trust is positive again from 2015 till now.

3.7 Loan to value

The data that is needed to measure the loan to value ratio is provided by the DNB

Household Survey. It studies economic aspects of financial behaviour. The panel survey was launched in 1993 in which 2000 households participated. The loan-to-value was measured by the amount of the mortgage divided by the purchase price of the house times 100. The DNB Household survey is held yearly. For this reason, the LTV of each year is used quarterly in order to get the same observations as the other variables.

The mean of the loan-to-value was 103.45 during the period of 1995 till 2017. In 2018, the ratio will be limited to 100% to lower risk and stabilize the financial sector (DNB, 2015).

3.8 Population density

The population density is measured by the number of inhabitants per squared kilometre and is provided by the CBS. Highly densely populated areas have a higher demand for the house market. Besides this, it is more challenging to increase the supply as it is already crowded.

The surprising fact about the data of the population density is that ’s Gravenhage is more densely populated than Amsterdam. It has a mean of 6111 while Amsterdam only has 4591 inhabitants per squared kilometre. The Netherlands only has a population density mean of 482.

(18)

18 | P a g e

3.9 Correlations and multicollinearity

In a regression model with multiple variables, there is a possibility that independent variables are correlating. Multicollinearity occurs when two variables are highly correlated and leads to biased estimated coefficients (Mansfield and Helms, 2012). In appendix 3, the correlation matrixes of each area are provided. A measure to detect multicollinearity is the Variance Inflation Factor (VIF). The rule of 10 is used to give a sign of severe multicollinearity (O’Brien, 2007). With a VIF value of 10 or higher, variables are omitted from the regression to reduce the collinearity. In table 2, the Variance Inflation Factors results are shown. Only in the regression of the Netherlands as whole, severe multicollinearity is detected. Therefore, variables with the highest Variance Inflation Factor will be omitted from the regression until all variables turn below a VIF value of 10. Therefore, population density will be omitted first from the regression as shown in table 10 of appendix 3. Now the highest VIF value is real GDP with 9.96 and the problem of multicollinearity is solved according to O’Brien.

Variables The Netherlands (VIF) Amsterdam (VIF) ‘S Gravenhage (VIF) Rotterdam (VIF) Utrecht (VIF) Mortgage rate 6.74 3.87 4.04 4.38 4.11 Real GDP 11.78 6.41 7.49 9.77 6.56 House supply 11.84 8.38 4.59 6.14 4.01 Unemployment rate 3.14 3.08 2.57 4.49 2.40 Consumer trust 1.88 3.92 3.05 1.97 3.50 Loan-to-value 1.12 1.26 1.29 1.17 1.25 Population density 21.67 4.44 2.76 3.62 3.06 Mean 6.42 4.48 3.68 4.51 3.56

Table 3: VIF results Source: CBS

(19)

19 | P a g e

3.10 Methodology

A single time series equation is used to investigate the relation between house prices and the mortgage rate. The basic time series model is:

𝑌𝑡 = a + b𝑋𝑡 + 𝑒𝑡

In this equation, Y is the dependent variable and b𝑋𝑡 the explanatory variable. The intercept is shown by a. The estimators are predicted using ordinary Least squares. The advantage of using a time series analysis is that it shows how variable behaved in the past. Besides that, it can make a prediction for 𝑌𝑡 = a + b𝑋𝑡−1 + 𝑒𝑡.

The control variables are important determinants to predict the house prices. Some of them are fixed and are equal in all five regressions. Others are variable and differ for each area. The regression for each area is:

𝐻𝑜𝑢𝑠𝑒 𝑝𝑟𝑖𝑐𝑒𝑠𝑡 = 𝛽0 + β1 𝑀𝑟𝑡+ β2 𝐺𝐷𝑃𝑡 + β3 𝐻𝑆𝑡 + β4 𝑈𝑅𝑡 + β5 𝐶𝑇𝑡 + β6 𝐿𝑇𝑉𝑡 + β7𝑃𝐷𝑡 + 𝑒𝑡

Variable

Description

𝐻𝑜𝑢𝑠𝑒 𝑝𝑟𝑖𝑐𝑒𝑠𝑡 The house price prediction at a certain time for each

area.

𝑀𝑟𝑡 The 10-year mortgage rate at time t.

𝐺𝐷𝑃𝑡 The real Gross Domestic Product at time t

𝐻𝑆𝑡 The houses that are available for sale in each area at

time t.

𝑈𝑅𝑡 The unemployment rate at time t.

𝐶𝑇𝑡 The consumer trust at time t.

𝐿𝑇𝑉𝑡 The Loan-to-value at time t.

𝑃𝐷𝑡 The population density in each area at time t.

These variables are according to the literature review in section two, important

determinants of the house prices. There will be five regressions in total. One each for the Netherlands, Amsterdam, ‘s Gravenhage, Rotterdam and Utrecht.

(20)

20 | P a g e

4. Results and Analysis

In this section, the results are shown. The estimations are listed in the table below. The table contains all five areas marked from (1) till (5). For each regression, data from 1995 till 2017 is quarterly used. Dependent variable: Real house prices (1) The Netherlands (2) Amsterdam (3) ‘S Gravenhage (4) Rotterdam (5) Utrecht Mortgage rate -871 (1803) -4703** (2190) -5231* (1945) -4540* (1674) -3887** (1476) Real GDP 2.08* (0.27) 1.68* (0.21) 1.71* (0.24) 1.23* (0.21) 2.06* (0.16) House supply -0.14** (0.06) -3.64* (1.34) 2.52 (1.91) 1.75 (1.15) -0.38 (1.60) Unemployment rate -6140* (1369) -18685* (1433) -2753** (1300) -5386* (1319) -5979* (1066) Consumer trust -298* (65) -214** (105) -47 (89) -99** (46) -148** (64) Loan-to-value 258* (63) 297* (103) -19 (63) 180* (50) 109*** (59) Population density (Omitted) 89* (10) -21* (5.20) 22 (30) -5.13 (3.32) Constant -45362 (38243) -293032* (62129) 121794.1** (56157) -43631.85 92889 -3250 (28973) Observations 91 91 91 91 91 R-squared 0.9060 0.9381 0.9326 0.9274 0.9477 Prob > F 0.0000 0.0000 0.0000 0.0000 0.0000

Table 5: Regression results

*, **, *** denote significance at 1%, 5% and 10% levels respectively. Heteroskedastic robust standard errors are in the parenthesis

(21)

21 | P a g e

4.1 The mortgage rate

The mortgage rate is, in the Netherlands as whole, negative but insignificant. This means that there is insufficient evidence to reject the null hypothesis and thus the coefficient equals 0. The four largest municipalities on the other hand, do have a significant estimation and are negative as suspected. Amsterdam and Utrecht are significant at a 5% level and Rotterdam and ‘s Gravenhage are significant at a 1% level. An increase of the mortgage rate with 1%, will affect the house prices in ‘s Gravenhage the most. The house prices will

decrease with €5231.

4.2 GDP

The real GDP has a positive coefficient and is significant at 1% for each area. This

corresponds to the literature as a higher GDP means that the economy is improving and welfare increases. As a consequence, it stimulates the demand of the housing market and house prices increase. An increase of the GDP with 1, will increase the house prices in the Netherlands with €2,08. Rotterdam faces the lowest increase of the house prices with €1,23.

4.3 House supply

The price dynamics of the housing market are determined by both demand and supply factors. The house supply variable however, is only significant in the Netherlands and Amsterdam. Compared to the other three municipalities, both the Netherlands and

Amsterdam had the largest decrease of supply from 2013 until nowadays that could explain their significance. An increase of the supply with 1, decreases the house prices in the

Netherlands with only €0,14 and in Amsterdam with €3,64.

4.4 Unemployment rate

The unemployment rate for each area is negative and significant. In the Netherlands, the coefficient of the unemployment rate equals -6140 and is significant at a 1% level as well as Amsterdam, Rotterdam and Utrecht. In ‘S Gravenhage it is significant at 5%. With an

increase of the unemployment rate with 1%, the house prices will decrease with €18.685 in Amsterdam. The high coefficient was expected according Clapp and Giaccotto. In ‘s

(22)

22 | P a g e

4.5 Consumer trust

In all areas, the consumer trust has a negative influence on the house prices. That is the opposite of what is expected with the written literature. An explanation could be due to the fact that the consumer trust is based on the past, the future and not just the current

situation. In this model, the estimation results are significant for the Netherlands, Amsterdam, Rotterdam and Utrecht. For ‘S Gravenhage there is insufficient evidence to reject the null hypothesis. An increase in consumer trust with 1, reduces the house prices the most in the Netherlands with €298.

4.6 Loan-to-value

The loan-to-value variable is significant for each area except for ‘s Gravenhage. In the other areas, it has a positive coefficient which means that a higher LTV ratio leads to higher

average high prices. In the Netherlands, Amsterdam and Rotterdam it is significant at 1% and Utrecht at 10%. Igan and Kang expected that a tightening of the LTV will significantly lead to lower price appreciation. The estimated results do correspond with the expectations. If the loan-to-value increases with 1% than the average house price increases in the Netherlands with €258. In Amsterdam, the house prices will increase with €297 and Rotterdam with €180.

4.7 Population density

The population density variable is omitted in the regression of the Netherlands as it was highly correlated with other independent variables. In Amsterdam and ‘S Gravenhage it is significant at a 1% level. For the remaining areas, there is insufficient evidence and so the null hypothesis cannot be rejected. If the population density increases with 1 than the average house prices increases with €89 in Amsterdam and decreases with €21 in ‘s

Gravenhage. According to the data that is used, ‘s Gravenhage has the most inhabitants per square kilometer during the entire period compared to Amsterdam. The coefficients

however, are the opposite of each other while the literature expected the coefficient to be positive.

(23)

23 | P a g e

4.8 Breusch-Godfrey Test

The Breusch-Godfrey test is used to check the regressions for autocorrelation.

Autocorrelation means that the errors are correlated with each other. This can lead to inefficient regression coefficients because the OLS standard errors are affected. The

coefficients are still unbiased but not ‘Best Linear Unbiased Estimated’ (BLUE). It is similar to the Durbin Watson test. However, the Breusch-Godfrey test is more powerful.

The hypothesis for the Breusch-Godfrey Test is:

H0: There is no serial correlation

H1: There is sufficient evidence for autocorrelation

Since the data that is used in this thesis is quarterly, the lag equals four. The null hypothesis will be rejected if p-value < 0.05. The results are shown in table 19 of appendix 4. For all five areas, the p-value equals 0.000. This means that the null hypothesis will be rejected in all regressions and serial correlation is present. Therefore, the robust standard errors are used to correct for the serial correlated errors.

4.9 R-squared

The R-squared is the statistical measure that shows the variance of the real house prices that is explained by the dependent variables. It is the goodness of fit between the observed values and the model. The estimation results show the R-squared for each regression. The Netherlands has the lowest R-squared: 0.9060. However, compared to the existing literature it is still above average. Utrecht has the highest R-squared: 0.9477. All five R-squared results show that the data fits the model.

(24)

24 | P a g e

4.10 Hypothesis results

Hypothesis

Results

Mortgage rate coefficient is negative All the coefficients of the mortgage rate in the regressions are negative. H0 is rejected Mortgage rate have smaller effect in Netherlands

compared to four largest municipalities

Only in the Netherlands, the coefficient of the mortgage rate is insignificant. There is sufficient evidence to reject the null hypothesis

GDP coefficient is positive In all five regressions, the GDP coefficient is positive and significant. H0 is rejected

Unemployment rate coefficient is negative Each regression has a negative unemployment rate coefficient. H0 is rejected

Consumer trust coefficient is positive Not enough evidence to reject null hypothesis. All regressions have negative coefficients.

LTV coefficient is positive The LTV coefficient is negative and insignificant in ‘s Gravenhage. This means the null hypothesis can be rejected for all the areas except ‘s Gravenhage. Population density coefficient is positive H0 is rejected in Amsterdam and ‘s Gravenhage. In ‘s

Gravenhage however, the coefficient is negative. Table 6: Hypothesis results

(25)

25 | P a g e

5. Conclusion

This thesis has been investigating the effect of the mortgage rate on the house prices in the Netherlands. This effect is compared with the effect on house prices in Amsterdam, ‘s Gravenhage, Rotterdam and Utrecht to add value to the existing literature. The mortgage rate has diminished significantly whilst at the same time the average house prices have increased. As a consequence, the demand-side of the housing market increases due to the low financing costs. At the same time, there is a tight supply-side. This have resulted in a shortage of low-cost housing for first-time buyers.

The existing literature was used to answer the research question by describing the determinants of the house prices. Moreover, it determined the conforming regression to predict the effect of the mortgage rate on the house prices. The hypothesis was: the mortgage rate influences the house prices negatively and has a bigger effect on booming house prices. To answer the hypothesis, a time series analysis was used for each area to estimate the house prices. Each variable was observed quarterly from 1995 till 2017.

According to the estimation results, it can be concluded that in each area the

mortgage rate influences the house prices negatively. However, the effect is only significant for the municipalities. In ‘s Gravenhage and Rotterdam it is significant at a 1% level and in Amsterdam and Utrecht at 5%. Besides that, the coefficient of the mortgage rate is significantly higher for the municipalities compared to the Netherlands as a whole. That partly explains the higher average house price increase of the four cities. A reason for this could be due to private investors. The diminishing mortgage rate improves the affordability and therefore the house prices increase which makes it attractive and profitable. Investors have low monthly expenses due to the low mortgage rate and high revenues because of high demand and tight supply, especially in large cities. This all together, leads to a shortage of low-cost housing for first-time buyers.

The limitations of this thesis are that income per capita, for each area separately, is not available from 1995 till 2017. Therefore, the national GDP is functioning as the

alternative. Also, the unemployment rate is only available nationally and cannot be specified for each area specifically. Further research could include the impact of private investors. Investors gain form the mortgage rate and have an upward pressure on the house prices.

(26)

26 | P a g e

Appendix 1: Nominal house prices

Mean (€) Std. Dev. Min (€) Max (€) The Netherlands 200456 50699 89792 267464 Amsterdam 243977 71592 94401 417652 ‘s Gravenhage 179976 56334 68141 274037 Rotterdam 155138 43597 65321 232031 Utrecht 205584 58967 82130 308078

Table 7: The nominal house prices Source: CBS

(27)

27 | P a g e

Appendix 2: Data line charts

€ - € 100,000.00 € 200,000.00 € 300,000.00 € 400,000.00 € 500,000.00 19 951 19 962 19 973 19 984 20 001 20 012 20 023 20 034 20 05 1 20 062 20 073 20 084 20 101 20 112 20 123 20 134 20 151 20 162 20 173

Nominal house price changes

Netherlands Amsterdam s Gravenhage Rotterdam Utrecht

Figure 3 - Nominal house price changes Source: CBS € - € 50,000.00 € 100,000.00 € 150,000.00 € 200,000.00 19 951 19 96 2 19 973 19 984 20 001 20 012 20 023 20 034 20 051 20 062 20 073 20 084 20 101 20 112 20 123 20 134 20 151 20 162 20 173

GDP

Real GDP Nominal GDP

Figure 4 – Nominal and real GDP Source: CBS

Figure 5 – House supply Netherlands Source: NVM 0 20000 40000 60000 80000 100000 120000 140000 160000 19 951 19 962 19 973 19 984 20 001 20 012 20 023 20 034 20 051 20 062 20 073 20 084 20 101 20 112 20 12 3 20 134 20 151 20 162 20 173

House supply The Netherlands

(28)

28 | P a g e Figure 7 – Unemployment rate and house price change

Source: CBS 0.00% 2.00% 4.00% 6.00% 8.00% 10.00% € - € 50,000.00 € 100,000.00 € 150,000.00 € 200,000.00 € 250,000.00 € 300,000.00 19 951 19 963 19 981 19 993 20 011 20 023 20 041 20 053 20 071 20 083 20 101 20 113 20 131 20 143 20 161 20 173

Unemployment rate and house price change

House prices Netherlands Unemployment rate

Figure 8 – Consumer trust Source: CBS -60.00 -40.00 -20.00 0.00 20.00 40.00 19 951 19 961 19 971 19 981 19 991 20 001 20 011 20 021 20 031 20 041 20 051 20 061 20 071 20 081 20 091 20 101 20 111 20 121 20 131 20 141 20 151 20 161 20 17 1

Consumer trust

Consumer trust Figure 6 – House supply of municipalities

Source: NVM 0 1000 2000 3000 4000 5000 6000 7000 8000 19 952 19 962 19 972 19 982 19 992 20 002 20 012 20 022 20 032 20 042 20 052 20 062 20 072 20 082 20 092 20 102 20 112 20 122 20 132 20 142 20 152 20 162 20 172

House supply of municipalities

Amsterdam s Gravenhage Rotterdam Utrecht

(29)

29 | P a g e

Appendix 3: Correlations and VIF results

3.1 The Netherlands correlation matrix and VIF results:

Real house price Mortgage rate Real GDP House supply Unemployment rate Consumer trust Loan- to-value Population density Real house price 1.000 Mortgage rate -0.495 1.000 Real GDP 0.8669 -0.7187 1.000 House supply 0.5348 -0.5756 0.7714 1.000 Unemployment rate -0.433 -0.2500 -0.132 0.3187 1.000 Consumer trust -0.379 0.1033 -0.333 -0.5863 -0.1337 1.000 Loan-to-value 0.2878 0.0618 0.1028 0.0083 -0.1443 -0.1638 1.000 Population density 0.6515 -0.8445 0.5864 0.8704 0.2179 -0.3646 -0.015 1.000 Variable VIF Population density 21.67 House supply 11.84 Real GDP 11.78 Mortgage rate 6.74 Unemployment rate 3.14 Consumer trust 1.88 Loan-to-value 1.12 Mean VIF 8.31 Variable VIF Real GDP 9.96 House supply 7.71 Mortgage rate 3.59 Unemployment rate 3.13 Consumer trust 1.87 Loan to value 1.11 Mean VIF 8.31

Table 8: Correlation matrix of The Netherlands

Table 9: VIF results of the Netherlands Table 10: VIF results of the Netherlands after omitting population density

(30)

30 | P a g e 3.2 Amsterdam correlation matrix and VIF results:

Real house price Mortgage rate Real GDP House supply Unemployment rate Consumer trust Loan- to-value Population density Real house price 1.000 Mortgage rate -0.621 1.000 Real GDP 0.8592 -0.7187 1.000 House supply 0.3605 -0.4868 0.6561 1.000 Unemployment rate -0.441 -0.2500 -0.132 0.3332 1.000 Consumer trust -0.139 0.1033 -0.333 -0.7702 -0.1337 1.000 Loan-to-value 0.1007 0.0618 0.1028 0.1111 -0.1443 -0.1638 1.000 Population density 0.4603 -0.7700 0.5649 0.5338 0.4864 -0.1123 -0.263 1.000 Variable VIF House supply 8.38 Real GDP 6.41 Population density 4.44 Consumer trust 3.92 Mortgage rate 3.87 Unemployment rate 3.08 Loan-to-value 1.26 Mean VIF 4.48

Table 11: Correlation matrix of Amsterdam

(31)

31 | P a g e 3.3 ‘s Gravenhage correlation matrix and VIF results:

Real house price Mortgage rate Real GDP House supply Unemployment rate Consumer trust Loan- to-value Population density Real house price 1.000 Mortgage rate -0.692 1.000 Real GDP 0.9470 -0.7187 1.000 House supply 0.5864 -0.3589 0.6154 1.000 Unemployment rate -0.166 -0.2500 -0.132 0.3143 1.000 Consumer trust -0.389 0.1033 -0.333 -0.6561 -0.1337 1.000 Loan-to-value 0.1553 0.0618 0.1028 0.0325 -0.1443 -0.1638 1.000 Population density -0.610 0.2003 -0.493 -0.3731 0.1959 0.6275 -0.426 1.000 Variable VIF Real GDP 7.49 House supply 4.59 Mortgage rate 4.04 Consumer trust 3.04 Population density 2.76 Unemployment rate 2.57 Loan-to-value 1.29 Mean VIF 3.68

Table 13: Correlation matrix of ‘s Gravenhage

(32)

32 | P a g e 3.4 Rotterdam correlation matrix and VIF results:

Real house price Mortgage rate Real GDP House supply Unemployment rate Consumer trust Loan- to-value Population density Real house price 1.000 Mortgage rate -0.655 1.000 Real GDP 0.9387 -0.7187 1.000 House supply 0.6702 -0.4517 0.7382 1.000 Unemployment rate -0.261 -0.2500 -0.132 0.2512 1.000 Consumer trust -0.376 0.1033 -0.333 -0.6211 -0.1337 1.000 Loan-to-value 0.2110 0.0618 0.1028 0.0722 -0.1443 -0.1638 1.000 Population density 0.1117 -0.4822 0.2157 0.4408 0.7545 0.3126 -0.226 1.000 Variable VIF Real GDP 9.77 House supply 6.14 Unemployment rate 4.49 Mortgage rate 4.38 Population density 3.62 Consumer trust 1.97 Loan-to-value 1.17 Mean VIF 4.51

Table 15: Correlation matrix of Rotterdam

(33)

33 | P a g e 3.5 Utrecht correlation matrix and VIF results:

Real house price Mortgage rate Real GDP House supply Unemployment rate Consumer trust Loan- to-value Population density Real house price 1.000

Mortgage rate -0.631 1.000 Real GDP 0.9433 -0.7187 1.000 House supply 0.4567 -0.2655 0.5492 1.000 Unemployment rate -0.317 -0.2500 -0.132 0.2654 1.000 Consumer trust -0.369 0.1033 -0.333 -0.6625 -0.1337 1.000 Loan-to-value 0.1907 0.0618 0.1028 -0.0298 -0.1443 -0.1638 1.000 Population density -0.558 0.1022 -0.427 -0.2100 0.3712 0.5810 -0.400 1.000 Variable VIF Real GDP 6.56 Mortgage rate 4.11 House supply 4.01 Consumer trust 3.50 Population density 3.06 Unemployment rate 2.40 Loan-to-value 1.25 Mean VIF 3.56

Table 17: Correlation matrix of Utrecht

(34)

34 | P a g e

Appendix 4: Breusch-Godfrey Test results

Breusch-Godfrey test for

autocorrelation

Chi2 Df Prob > chi2

The Netherlands 51.823 4 0.0000

Amsterdam 35.296 4 0.0000

‘s Gravenhage 40.734 4 0.0000

Rotterdam 36.002 4 0.0000

Utrecht 30.437 4 0.0000

(35)

35 | P a g e

References

1. Aoki, K., Proudman, J., & Vlieghe, G. (2004). House prices, consumption, and monetary policy: a financial accelerator approach. Journal of financial

intermediation, 13(4), 414-435.

2. Beltratti, A., & Morana, C. (2010). International house prices and macroeconomic fluctuations. Journal of Banking & Finance, 34(3), 533-545.

3. Capozza, D. R., Hendershott, P. H., Mack, C., & Mayer, C. J. (2002). Determinants of

real house price dynamics (No. w9262). National Bureau of Economic Research.

4. Case, K. E., & Shiller, R. J. (2003). Is there a bubble in the housing market?. Brookings

papers on economic activity, 2003(2), 299-342.

5. Charles Goodhart, Boris Hofmann; House prices, money, credit, and

the macroeconomy, Oxford Review of Economic Policy, Volume 24, Issue 1, 1 March 2008, Pages 180–205,

6. Clapp, J. M., & Giaccotto, C. (1994). The influence of economic variables on local house price dynamics. Journal of Urban Economics, 36(2), 161-183.

7. Davis, E. P., & Zhu, H. (2011). Bank lending and commercial property cycles: some cross-country evidence. Journal of International Money and Finance, 30(1), 1-21.

8. De Bruyne, K., & Van Hove, J. (2013). Explaining the spatial variation in housing prices: an economic geography approach. Applied Economics, 45(13), 1673-1689.

9. Égert, B., & Mihaljek, D. (2007). Determinants of house prices in central and eastern Europe. Comparative economic studies, 49(3), 367-388.

10. Fitzpatrick, T., & McQuinn, K. (2007). House prices and mortgage credit: Empirical evidence for Ireland. the manchester school, 75(1), 82-103.

11. Glaeser, E. L., & Shapiro, J. M. (2003). The benefits of the home mortgage interest deduction. Tax policy and the economy, 17, 37-82.

(36)

36 | P a g e

12. Green, R., & Hendershott, P. H. (1996). Age, housing demand, and real house prices.

Regional Science and Urban Economics, 26(5), 465-480.

13. Harris, J. C. (1989). The effect of real rates of interest on housing prices. The Journal

of Real Estate Finance and Economics, 2(1), 47-60

14. Harter-Dreiman, M. (2004). Drawing inferences about housing supply elasticity from house price responses to income shocks. Journal of Urban Economics, 55(2), 316-337.

15. Igan, D., & Kang, H. (2011). Do loan-to-value and debt-to-income limits work? Evidence from Korea.

16. Verbruggen, J. (2015). Effects of further reductions in the LTV limit. De Nederlandsche

Bank, Vol. 13 -2.

17. Konchitchki, Y., & Patatoukas, P. N. (2014). Accounting earnings and gross domestic product. Journal of Accounting and Economics, 57(1), 76-88.

18. Langere rentevaste periode bij nieuwe woninghypotheek populairder. DNB. (2016). Retrieved from https://www.dnb.nl/nieuws/nieuwsoverzicht-en-archief/statistisch-nieuws-2016/dnb341004.jsp#

19. Mansfield, E. R., & Helms, B. P. (1982). Detecting multicollinearity. The American

Statistician, 36(3a), 158-160.

20. McQuinn, K., & O'Reilly, G. (2008). Assessing the role of income and interest rates in determining house prices. Economic modelling, 25(3), 377-390.

21. Meen, G. (1999). Regional house prices and the ripple effect: a new interpretation.

Housing studies, 14(6), 733-753.

22. Nederlandse Vereniging van Makelaars en Taxateurs. (2017). NVM woningmarktcijfers 2de kwartaal 2017. Retrieved from

https://www.nvm.nl/actueel/persberichten/2017/20170712cijfers2017q2

23. O’brien, R. M. (2007). A caution regarding rules of thumb for variance inflation factors. Quality & Quantity, 41(5), 673-690.

(37)

37 | P a g e 24. Reichert, A. K. (1990). The impact of interest rates, income, and employment upon

regional housing prices. The Journal of Real Estate Finance and Economics, 3(4), 373-391.

25. Rigobon, R., & Sack, B. (2004). The impact of monetary policy on asset prices. Journal

of Monetary Economics, 51(8), 1553-1575.

26. Rouwendal, J. (2007). Mortgage interest deductibility and homeownership in the Netherlands. Journal of Housing and the Built Environment, 22(4), 369-382.

27. Rouwendal, J., & Longhi, S. (2008). The effect of consumers' expectations in a booming housing market: space-time patterns in the Netherlands, 1999– 2000. Housing Studies, 23(2), 291-317.

28. Senne Janssen (2017). Prijsvorming op de amsterdamse woningmarkt. ING

Economisch Bureau, Retrieved from

https://www.ing.nl/media/ING%20Prijsvorming%20op%20Amsterdamse%20woning markt_tcm162-117993.pdf

29. Sørensen, C. K., & Lichtenberger, J. D. (2007). Mortgage interest rate dispersion in the

euro area. European Central Bank.

30. Zietz, J., Zietz, E. N., & Sirmans, G. S. (2008). Determinants of house prices: a quantile regression approach. The Journal of Real Estate Finance and Economics, 37(4), 317-333.

The effect of the mortgage rate on the house prices in the Netherlands