Factors that influences the house prices before and
during the crisis, case study of Amsterdam
Maxime Kerstholt
10795642
Economics and Business; specialization Economics
Supervisor: Chih-Chung Ting
University of Amsterdam
26
thJune 2018
Abstract The real estate market of Amsterdam is seen as a unique market, with a high demand and a scarcity of properties. This research compares the significance of several factors that predict house prices before and during the crisis in the capital city of the Netherlands. This is done by multiple regression analyses. Furthermore, a comparison with the year 2017 is made as well, to see if there are similarities with the year before the crisis. In both years, the house prices in Amsterdam were rising.
Page | 2
Statement of Originality
This document is written by Maxime Kerstholt 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 completion of the work, not for the contents.
Page | 3
Table of content
1. Introduction
page 42. Literature review
page 82.1 The ‘different’ house market of Amsterdam page 8 2.2 The impact of the global financial crisis on house prices page 8 2.3 The influence of different factors on house prices page 9
2.3.1 Mortgage interest rate page 10
2.3.2 Time on the market page 12
2.3.3 Location page 13
2.3.4 Size of population page 14
2.3.5 Seasonality page 14
3. Research Question and Hypothesis
page 164. Methodology
page 174.1 Data page 18
5. Results
page 206.
Discussion
page 266.1 The comparison with 2017 page 27
6.2 Further research page 28
7. Conclusion
page 30Page | 4
1. Introduction
Amsterdam and her ‘home system’ has gone through two major transformations in the past two decades. The first is a large-scale privatization of the housing stock starting in the nineties of the last century. These changes created the context for a subsequent period of restructuring. The bubble of rapidly rising house prices, came to an explosion in 2008. The second transformation started with the global financial crises in 2008 and limited access to and promotion within the Amsterdam housing market (Boterman et al., 2013).
A lot of research is done to see what factors influences rising house prices in different periods. Reichert (1990) focused mainly on the national economic factors, like the (mortgage) interest rates. He stated that during periods of housing price appreciation (for instance before the crisis period in 2007), the real housing costs commonly decline, which stimulates the demand for houses even more. Besides economic factors, Rosen (1979) found seasonality effects house prices as well. The founding was mainly that house prices will be at their lowest during the fourth quarter (winter). He concludes also that seasonality of the demand is the predominant variable and that constructors react to this seasonal demand in a way that’s logic – when demand is weak and building costs are high during the cold months (winter and fall), they will reducing the construction. Resulting in less properties built and less properties sold in this season (Rosen, 1979). Therefore, it is interesting to see if these factors influenced house prices before the crisis, when the house prices were rising, and during the crisis, when the house prices were declining.
To get a proper illustration of the real estate market in the Netherlands and Amsterdam, the following information is given. In general, prices in Amsterdam fall more steeply in times of economic crises than in the rest of the Netherlands, but also rise faster in times of economic boom. In 2009, when the Dutch economy shrank by as much as 3.5 per cent, the drop in prices in Amsterdam was above average and in the first quarter of 2013, the prices fell to the low point, with a decrease of 18 per cent (CBS). Different factors influenced the price of houses in that period, for instance the mortgage interest rate decreased during the crisis to stimulate the house market. Between the years of 2015 and 2017, the housing prices in Amsterdam rose by 46 per cent1. This is a significant rise, compared to the rest of the Netherlands, where there was only a 5 per cent increase (ESB).
1
Retrieved 4th of June 2018 from: https://www.folia.nl/wetenschap/115721/amsterdamse-huizenprijzen-stijgen-harder-dan-economisch-model-kan-verklaren
Page | 5
Also, house prices in Amsterdam are more volatile than average (Ambrose et al., 2013), which means that affordability also fluctuates more strongly. In the past twenty years, there have been two earlier periods in Amsterdam in which affordability deteriorated sharply, namely 1999-2001 and 2007-2009. Both periods were characterized by sharply rising house prices, followed by periods with price decreases (CBS). Two crises (the dot-com and the credit crises) with associated loss of confidence were the 'triggers' for the turnaround in Amsterdam.
In 2017, affordability declined again, making it plausible that the strong house price increases in Amsterdam will weaken in the coming years (ESB). When these sharp increases continue, the risk of a sudden price correction increases more and more. A new period of economic headwind, a strong rise in interest rates or another trigger could cause such a correction.
The goal of this research is to examine the factors that led to the various house prices in Amsterdam before (2007) and during the crises (2012) and see if there are significant correlations between these years. This is stated in the following research question of this thesis: Are the factors to predict house prices consistent for before the crisis (2007) and during the crisis (2012), focused on the case study of Amsterdam?
Besides that, I will study these years and the year 2017, to see if there is a comparison between the analysed factors. The year of 2017 was chosen because the house prices in that year sharply increased, similar to the year of 2007. However, the house prices in 2017 increased more than before the crises period (Figure 1). The reason behind this, is because of the crisis period that occurred from 2008 till circa 2013. There were too little properties built in this period, for some years no properties at all (CBS). The result is the scarcity of the properties now, what causes the rising house prices. Another explanation is the popularity of the city Amsterdam. The tourism is for instance sharply increased; in 2015 the number of tourists was grown by 56% compared to 2000.
Page | 6
Figure 1 House prices, averages transaction price in Amsterdam x €1000
Source: CBS
So far, little research has been done for the specific house market of Amsterdam, with the particular influencing factors that are analysed here and in the years already mentioned (2007 and 2012). Kakes et al (2004) and Heijden et al (1996) looked for the most part of their research for the house prices on the Dutch house market in general, Eichholtz et al (2012) and Ambrose et al (2013) investigated especially the aspect of long-term (>300 years) relations between house prices and rents in Amsterdam. The results of Kakes et al (2004) shows that stock prices fluctuations have an distributional impact on the Dutch house market and that different locations have a causal impact on house prices. Furthermore, Anglin et al (2003) focussed on the Time on the market of the property. This research stated that the time on the market influences the house price in the following way: to get the highest price for the property, the time on the market should be as short as possible.
Besides, a different research of Eichholtz et al (2014) focused mostly on whether a fundamental factor or a trend explains house prices, based only on the Herengracht in Amsterdam. Furthermore, Houben et al (2017) took the same time periods as this research (1990-2017), and concentrated on different factors like the free rental sector, population and
Page | 7
house characteristics. They found that house price increases can partially explained by house characteristics factors, but that population grow has a big impact.
So, this research will be the first to look at the specific house prices in Amsterdam before and during the crisis years, in addition to a comparison of these years with 2017. Five variables that dictate the price of a property in Amsterdam will be studied, which also is done in previous research for several factors. However, this is not researched for the year before and during the crisis and with this particular set of factors. These five factors are the following ones: Mortgage interest rate, Time on the market, Location, Size of Population and Season.
The data for most factors (House prices, TOM, Location and Season) is obtained from leden.nvm.nl, a secured environment for Dutch real estate agents. The data for the mortgage interest rate is found in the history of Obvion Hypotheken and the size of the population on the website of the municipality of Amsterdam. This will be further explained in paragraph 4, the methodology.
In this paper the setup will be as follows. The next paragraph presents a literature overview, followed by the research question and hypotheses. After this section, the methodology part including the research set-up and regression is next, with the results of the regression and the discussion following up. Finally, there will be a conclusion.
Page | 8
2. Literature Review
This chapter will start with an explanation of how the house market in Amsterdam differs from the rest of the Netherlands. Second, the impact of the global financial crisis on the house prices will be given. The last part is used to discuss the different influential factors, which had an impact on the house prices of Amsterdam, for the year before the crisis (2007) and during the crisis (2012).
2.1 The ‘different’ house market of Amsterdam
The Dutch house market is by definition very volatile (Ambrose et al., 2013), meaning that the prices of houses fluctuate rapidly between high and low prices in a short time span. At this moment the price correcting effect of new supply is almost nil. The price development is mainly driven by income and interest; because interest is extremely low, buyers can borrow more and also have low housing costs.
The residential property market behavior of the capital city Amsterdam works sometimes different compared to the rest of the country. In Amsterdam the average house prices are usually higher than in other cities, such as the difference in the price development pattern. For instance, when you look at the recent Global Financial Crisis, the house prices of Amsterdam declined more sharply, but also returned back to normal more rapidly than in other major Dutch cities like Utrecht, The Hague and Rotterdam. For example, between 2003 and 2006 the average property prices in Amsterdam decreased with approximately 11%, whereas in the same period there was an upward trend in the rest of the country (Teye, A.L. et al., 2017). Mentioning that in this period there were some global financial accidents, like the Enron and the Turkey crisis, it may seem the Amsterdam house market is more sensitive to current global economic circumstances than other areas of the country.
2.2 The impact of the global financial crisis on house prices
The collapse of the American housing market is often seen as one the most important reasons for the financial global crisis: the American credit market was carelessly dealing with the creditworthiness check of applicants, causing that unreliable borrowers in the lower segment could obtained risky mortgages. These mortgages were repacked and sold, to be repacked and resold again which is called the process of securitization. The consequence
Page | 9
was the high number of unfounded securities. Many banks had such unreliable credits in their portfolio, causing the collapse of housing prices in the US. The global nature of these banks also ensured that offices in European countries were closed, where massive redundancies occurred. Due to this redundancies, the unemployment rate increased, which influenced the consumer confidence in a negative way. In the end, there was a fall of the entire financial system and the trust in the system was gone (Machinea, 2009).
If we look at the Netherlands, the impact of this fall took many changes regard to credit opportunities and requirements. This has to do with the large household debt of the Netherlands, the highest of Europe. People couldn’t easily get a mortgage anymore, what lowered the demand for houses and what therefore decreased house prices as well. The effect of the fall of house prices was that people owned a house that was ‘under water’ (the value of the house is below its level of mortgage debt). This had not only a negative impact on consumption and saving behaviour, but also on the dynamics of the housing market: the house owners couldn’t move quickly if they wanted to, because of the residual debt that yielded what made it therefore increasingly difficult to borrow (Mrkaix et al., 2015). The impact of this fact was that house prices decreased even more. Concluded, the global financial crisis had thus an indirect impact on the house prices in the Netherlands, and so, on the house prices in Amsterdam as well.
2.3 The influence of different factors on house prices
There are various factors that influence house prices in Amsterdam, which can be divided into different sections. There are for example economic factors; household income, employment and the development of the economy that have an impact on the property prices. Also the mortgage interest rate is an important variable, because an increase of 1 per cent interest means 20.000 euro less mortgage2. Besides the economic factors, there are demographic factors as well, like the size of population, migration, aging etc. Furthermore the number of people who are leaving the city and moving/staying changed. Dignum (2011) explained that people try to reach the top of the housing ‘escalator’ and therefore stay in Amsterdam. The problem that grows in this case is that there are more people coming or staying than leaving. Moreover, too little new housing is being built. This is part of the government decisions factor. Besides these factors, there will be some influences for house prices without control over, the external factors. For instance, this is consumer confidence and
Page | 10
seasonality. As said in the introduction part, Rosen (1979) found seasonality effects house prices as well and is therefore interesting to look at.
In previous research of Anglin et al (2003) and Yavas et al (1995) is stated that the time on the market of a house, is also a factor that affects the price of the house. As a broker you want to strive to get the highest price, therefore it’s commonly accepted that the TOM should be as short as possible.
Concluded, there are many factors that have an impact on the price of a house. In the paper of Boelhouwer et al (1996) are the factors subsumed under five headings: ‘’the influence of government and other key institutions, demographic developments, developments in the supply of owner-occupied dwellings, the development of a number of economic variables and the dynamics of market forces’’. Each of these headings can be broken down into many different factors, but this paper has picked out five important factors, which are in previous studies also mentioned (see introduction) as influencing factors on house prices. However, there is no research done before with these five factors together. These five are the following: Mortgage interest rate, Time on the market, Location, Size of population and Season.
2.3.1 Mortgage interest rate
The mortgage interest rate is the interest that is charged on a loan, what is used to purchase a property. This rate is also linked with cycles of the national interest rate and can affect the homebuyer’s market drastically: as said before, an increase of 1 per cent interest means 20.000 euro less mortgage.
One of the main reasons of the increasing house prices before the years of the crisis was the so-called ‘hypotheekrenteaftrek’, mortgage interest deduction. This deduction is the tax advantage a homeowner gets on his taxable income, the mortgage interest you pay from the main residence. In addition, for every mortgage, there will be a mortgage debt. The relationship between mortgage debt and economic growth is formed by the surplus value on residential property. This is the difference between the market value of a house and the amount of the mortgage with which it is taxed (Notten, 2011).
The tax advantage had a great deal of influence on the consumer's choice of mortgage, as evidenced by the figures from 2007 in which more than half of all mortgages were repayment-free mortgages. Mortgage interest deduction made higher mortgage more
Page | 11
attractive, as a result of which it was seen as the cause of the high household debt, but also lead to increasing house prices (Mrkaix et al, 2015).
In the wake of the 2008 crisis, rising house prices meant that the value of total Dutch home ownership also increased, which implied that households had more assets. In addition, it was so easy to get an extra mortgage on the surplus value between 1990 and 2007 of the house, that during this period the mortgage was preferably one and a half times as large as necessary for the financing of purchase and investment in housing. On top of that, there was an extremely low average mortgage rate for a long time. Due to this low interest rate of 3.7 per cent in mid-2005 (CBS), it was not only attractive to take out new mortgages, but also against existing mortgages to close the low interest rate. As a result, a higher amount could be borrowed for the same monthly expenses. This is an important explanation why the house prices drove up in the period before the crisis (Notten, 2011).
Then the crisis period broke through. One of the factors of decreasing house prices in Amsterdam during the crisis was the increasing interest rate. A little explanation of how this turned out will first be given.
The easing of loan standards and the low entry rates that banks used (due to the reduction that the government had set up) led to a new target group of consumers who initially had insufficient income to take out a mortgage, but could now finance a house with a high mortgage debt without a problem (Bijlsma et al, 2014). The measures resulted into a sharp increase of demand for houses, as a result of which the house price began to increase too. Due to this increase of house prices, the demand for houses increased even more and so on. Of this reason, houses were seen as investment objects. The assumption was that a house would always transcend the investment, ‘because the price nevertheless rose annually, surely’. The turnaround came in 2008. The rising demand and prices of property eventually stagnated (R. Kaashoek, 2012). The cause was the increasing interest rate, partly due the government and on the other hand by the banks. They had started with low interest contracts, but after expiration, renewed with higher interest rates. In addition, house prices in Amsterdam were at this point so high, that buying a house became almost unaffordable. The result was a decline in demand for houses and higher supply of houses (CBS), because homeowners could no longer pay their ‘new’ mortgage interest contracts. This development led to a reduction in house prices and due to increased forced sales, many homeowners with a residual debt lagged behind (R. Kaashoek, 2012). Banks often had to written off a loss, what started the first financial problems in mortgage banks.
Page | 12
Since this global financial crisis, banks and lenders had more restricted criteria for mortgages (DNB, 2011). The most important guideline for the refined donation criteria is the new Mortgage Code of Conduct (Gedragscode hypothecaire financieringen), which started on the first of August 2011. The interest-only mortgage has been tied to limits from that day, and so it became less attractive to buy a house, what decreased demand and prices of the houses in Amsterdam in the year of 2012.
2.3.2 Time on the Market
In this subparagraph the factor Time on the market (TOM) will be discussed. The definition of the TOM is the actual time in days the house is on the real estate market before the negotiation between the seller and buyer ends. This means from the moment the property is on the internet as starting date, till the keys are in the new owner’s hands at the notary (Anglin et al., 2003).
As mentioned in previous subparagraph, banks and lenders had some new guidelines. Because of this restricted criteria in the guidelines, the houses for sale in Amsterdam in the year 2012 became even harder to sell and had for that reason also an increased time on the market. Anglin et al (2003) and Yavas et al (1995) stated that as a broker you want to strive to get the highest price, therefore it’s commonly accepted that the TOM should be as short as possible. In previous research of Rawal et al (2004) is mentioned: when the TOM is too high, the potential house buyers will get a suspicious mind. If there is a good located house with a fair price still not sold after a while, and so still on the market, this kept people thinking. This will gives an impact on the selling price and causality gives an effect on the time of the market (Rawal et al., 2004).
This was different in the year before the crisis. Because of the mortgage interest deduction and the low interest rate, the time on the market of the houses in Amsterdam was also affected for the year of 2007. The time on the market was shorter than years before and property was sold massively at registration, where buyers had to bid against each other. The time on the market in the last quarter of 2007 decreased from 35 to 29 days (CBS), even if the prices were rising (for instance, in 2004 this number of days was 85). This is supported by previous research of Cubbin (1974), where is concluded that higher house prices gives a lower time on the market.
Page | 13
2.3.3 Location
In previous research of Kiel & Zabel (2008) is stated that the value of houses depend on the location, because a house is unable to move. They list three important aspects of the location, namely the street, town and statistical area where is it located. These aspects can include distance to different facilities, parks, highways, duration to work etc. This research classified the location of a house in two categories: the absolute and relative location. The zip code is part of the absolute location and the street and town are part of the relative location. They looked to facilities and the accessibility to work or school in the neighbourhood. If these services are nearby, this will influence the house price positively. Also the safety index number is part of the location, as a more safety location (a higher safety index number) will increase the price of the house at that particular location. The municipality of Amsterdam monitors the location by its safety index number (Veiligheidsindex Amsterdam).
The results of other researches are pointing in the same direction: in general have an open (green) space, water and different landscape types a positive effect on the house price (Lutzenhiser et al., 2001; Daly et al., 2003; Bervaes & Vreke 2004; Van Dam et al., 2008). The effects vary, however, per characteristic and distance to such aspects. They also stated that a negative location, with low school quality or high crime rates influences the house prices negatively. This negative location is linked to a low safety index number.
In research of Visser en Van Dam (2006) the city of Amsterdam is highlighted. They said that the house price level in Amsterdam is higher anyway, but for some houses in specific locations and in certain residential environments, so much money is being paid that it lost the connection with psychical characteristics such as living space or the type of the house.
Of course the location of the houses is not changed before and during the crisis, but the value of the houses is increased/decreased for some locations. For instance the popularity of neighbourhoods or renovation and new construction projects of the municipality what will positively influences the neighbourhood and the safety index number of that location and therefore the house prices of that typical location in Amsterdam. The renovation and new construction projects decreased during crisis years compared to before crisis years (Amsterdam in Cijfers, 2007-2012). Given this fact, there was during the crisis a lower positively influence on location, and thus on the price of the property, then before the crisis years.
Page | 14
2.3.4 Size of population
In previous research of Ambrose et al (2013), which examines a long-run relation (355 years) between house prices and rents, is the size of the population also taken into account. They found that in the sub-period of 1781 to 1814 in Amsterdam, there was the strongest decline in house prices and rents, what was the only period in the recorded history in which there was consistently declining population. In other researches of Parker (2000) and Jud et al (2002), they concluded that house price appreciation is strongly affected by the grow or crimp of the population.
Given this researches, there will now be looked at the population of Amsterdam. The population in Amsterdam grew in 10 years’ time (2007-2017) with 100,000 inhabitants. In 2007 there were 742,884 people living in the capital city and in 2012 this number increased to 790,110 (CBS). Besides the statistics mentioned above, in the year of 2007 6,346 new homes were built in Amsterdam, much more than in the years before 2007. More than 31 per cent of these new homes are owner-occupied homes, 56 per cent of these homes were built by social landlords and the rest (13 per cent) were built by private landlords, these are mainly rental properties in the market segment. In a period of 10 years in the past, the housing stock has grown with 17,000 homes (from 367,000 in 1998 to 384,000 homes in 2008). During this period the number of owner-occupied homes more than doubled from 47,000 to 99,000. Since the end of last century active policies have been pursued to expand the owner-occupied housing stock in Amsterdam. The housing stock has been changing slowly since then (Amsterdam in Cijfers, 2008).
There were only 3000 houses built between the years of 2009 and 2013 in Amsterdam (Amsterdam in Cijfers, 2013). Cause was part due the many construction companies that had reduced their planning and construction capacity during the crisis. Concluded, there were in 2007 double as much houses built compared to the years between 2009 and 2013, but the size of the population between 2007 and 2012 grew with 47,226. Due to this higher number of inhabitants and the scarcity of houses in Amsterdam, this will give an negative effect (higher prices) on the house price.
2.3.5 Seasonality
In previous research of Tenreyro et al (2009), he argued that in spring and summer (warm season) there are more people involved on the house market compared to winter and fall (cold season). Sellers do not need to discount their prices in spring and summer, what they have to do in winter and fall for selling a house. Tenreyro uses a theory to explain this circle;
Page | 15
in summer and spring sellers can set a higher price for the house, because there a more potential buyers. Given this fact, the number of sellers will increase too, because they see their opportunities increasing for selling a house. For this reason, there will be a better match between the seller and buyer, because of the increased supply of both groups. In winter and fall there are less potential buyers, as a result the theory of Tenreyro will not hold in this period.
Besides this and as mentioned in the introduction part before, Rosen (1979) found seasonality effects house prices as well. The founding was mainly that house prices will be at their lowest during the fourth quarter (winter).
Another reason why there could be more houses sold in warm season, is because of families who prefer to settle in new place before a new school/work year starts.
Page | 16
3. Research Question and Hypotheses
In this paragraph the research question of this paper is stated. Besides that, a set of hypotheses is given to answer the research question.
As mentioned in the previous chapter, various factors have an influence on the house prices in Amsterdam. Boelhouwer et al (1996) subsumed different factors under five headings, but Houben et al (2017) focused primarily on the standard macro-economic factors. In regard of finding out what factors have a significant impact on the house prices in Amsterdam in different years, the following research question is stated in this paper:
Are the factors to predict house prices consistent for before the crisis (2007) and during the crisis (2012), focused on the case study of Amsterdam?
To answer this question, several hypotheses will be tested. In this paper the focus is on five different factors, what will give the following set of hypotheses:
Hypothesis 1: The same factors can be used to predict house prices before and during the crisis.
Hypothesis 2: The mortgage interest rate will predict significance for the house prices in Amsterdam in 2007, but not for the year 2012.
Hypothesis 3: The TOM will significantly predict the house prices of Amsterdam in 2007, but not for the year 2012.
Hypothesis 4: A positive location will significantly predict the house prices for both years compared to a neutral or negative location.
Hypothesis 5: The size of the population will significantly predict the house prices for 2012 (because of the scarcity of houses and the grew of inhabitants), but not for 2007.
Hypothesis 6: The warm season will significantly predict the house prices for both years compared to cold season.
Page | 17
4. Methodology
In this paragraph the research design is given. It will explain the type of research, the formula is given and the used variables are explained. In the subparagraph the data collecting is described.
In this paper, the following regression is performed to examine if the factors Mortgage interest rate, Time on the market, Location, Size of population and Season are significant to predict the house prices in Amsterdam before and during the crisis. This is done by doing a multiple regression analyses, which is generally purposed as the relationship between several independent variables (predictors) and a dependent variable (criterion). Conclusions can be drawn about individual predictors, with the focus on coefficients, their estimated standard errors and t-test probabilities. These statistics can be used to test the hypotheses to evaluate their relative importance or about the significant individual effect of such a predictor on the criterion (C.H. Mason, 1991). The test can be used with continue and ordinal data, if there is nominal data, this has to be established in a dummy variable.
The price of the houses is regressed on a constant, the independent variables Mortgage interest rate, Time on the Market and Size of Population and dummy variables Location and Season.
Pt = β0 + β1MortgageRate + β2 TOM + β3Location + β4Size + β5Season + εt
The Pt is the dependent variable price per square meter of the house on the date t. This
dependent variable is regressed on MortgageRate, the independent variable mortgage interest rate, what is the average daily mortgage interest in Amsterdam3; TOM is the independent variable Time on the Market which is given in full days; Location is a dummy variable who has the value one when it has a positive location and zero when the location is neutral or negative; Size is the independent variable Size of Population, which is the average daily grow/crimp of the population in Amsterdam4; Season is a dummy variable who has the value one when it is warm season (spring and summer) and zero when it is cold season (fall and winter) . εt is the error term at date t.
3
Retrieved 10th of June 2018 from: https://www.obvion.nl/Hypotheekrentehistorie.htm
Page | 18
4.1 Data
The prices per square meter, the TOM, the Location and Season are obtained from
www.leden.nvm.nl, which is a secured environment where Dutch real estate agents work
with. It provides much more additional data than the public site of NVM. Because of inside contacts I can make use of this professional program, with a given username and password.
For every year there are randomly picked out 83 properties of the database, what started on the first of January and with 4 or 5 days between. A total of 166 properties has been researched. Every single case is checked for the right property. That means, every houseboat or parking spot is removed from the used data set and only houses or apartments were left. Also the houses which were withdrawn, for sale and withdrawn again are left out.
Besides these left outs are also the ones with a lot of registrations removed. This is a marketing strategy of some real estate agents, what they use for every single house price change of the property. If there is such a change in the price, these agents will make a different registration then is used before, to make the TOM as short as possible. This will give a distorted picture of the TOM and therefore left out.
Further, for some properties in 2012 the Time on the Market was not known and was given a 0. Which will cause a different and skewed relationship with the price in the regression. The decision has been made to give these two properties the average number of days of the other 81 properties left.
For the location the ZIP code of Amsterdam (1000 AA – 1099 ZZ) is used to distinguish between a negative, neutral or positive location. This distinction is made using a monitor of the municipality, what is looking at the safety rate of the different neighbourhoods of Amsterdam. This is called the safety index and also used by big rent commissions (Veiligheidsindex Amsterdam, 2007-2012). Every neighbourhood received a safety rate number, which was in general 70. Based on the safety index, there are three groups made:
Group 1: Positive with zip codes 1071 AA – 1099 ZZ Group 2: Neutral with zip codes 1020 AA – 1070 ZZ Group 3: Negative with zip codes 1000 AA – 1019 ZZ
For the variable Season, the astronomical season is used. Winter is from December 21 till March 20, Spring from March 21 till June 20, Summer from June 21 till September 20 and
Page | 19
Fall from September 21 till December 20. Warm season is thus from March 21 till September 20 and cold season from September 21 till March 20.
The selected time frames are 01-01-2007/31-12-2007 and 01-01-2012/31-12-2012.
Table 1: Summary statistics before the crisis, 2007
Variable Obs Mean Std. Dev. Min Max
PriceM2 82 3633.751 1068.629 1594.737 8420 MortgageRate 82 4.758171 .1855198 4.41 5.04 TOM 82 59.97561 72.75105 1 421 Location 82 .4146341 .4956906 0 1 Size 82 11.47268 22.05948 -20.16 61.23 Season 82 .5121951 .5029273 0 1
Table 2: Summary statistics during the crisis, 2012
Variable Obs Mean Std. Dev. Min Max
PriceM2 82 3428.265 1091.876 1687.5 6755.725 MortgageRate 82 4.038902 .1091042 3.84 4.26 TOM 82 140.5122 157.9345 1 670 Location 82 .3658537 .4846329 0 1 Size 82 25.17766 31.86102 -19.47 81.42 Season 82 .5121951 .5029273 0 1
The mean for price per square meter in 2007 is 3633.751 (std. dev. 1068.629) and in 2012 3428.265 (std. dev. 1091.876), a difference of 205.486. Also, the maximum price per square meter is higher in 2007 compared to 2012, respectively 8420 and 6755.725. A difference of 1664.275. The mortgage interest rate mean is for 2007 4.758171 and for 2012 4.038902, 0.719269 higher in 2007.
For the Time on the market in 2007 there is a mean and maximum of days of respectively 59.97561 and 421. In 2012 these numbers are 140.5122 and 670. There is an average daily population grow of 11.47268 in 2007, with a minimum crimp of -20.16 and a maximum growth of 61.23 people per day. For 2012 the mean of the daily population grow is 25.17766, with a minimum crimp of -19.47 and a maximum growth of 81.42 people per day.
Page | 20
5. Results
In this paragraph the results of the correlations and regression will be discussed. Also the significance of variables for both years are given. After this, the results of the stepwise regression are given in the last tables.
The correlation results of the price per square meter and the five different variables before the crisis are summarized in Table 3.
Table 3: Correlations before the crisis, 2007
Price M2 MortgageRate TOM Location Size Season
Price M2 Pearson Correlation
Sig. (2-tailed) N 1 - 82 .049 .660 82 -.195 .080 82 .039 .025 -.158 .729 .824 .156 82 82 82
MortgageRate Pearson Correlation
Sig. (2-tailed) N .049 .660 82 1 - 82 .189 .089 82 -.216 -.147 .359** .051 .187 .001 82 82 82
TOM Pearson Correlation
Sig. (2-tailed) N -.195 .080 82 .189 .089 82 1 - 82 -.129 .053 .183 .248 .639 .099 82 82 82
Location Pearson Correlation
Sig. (2-tailed) N .039 .729 82 -.216 .051 82 -.129 .248 82 1 -.064 -.120 - .569 .285 82 82 82
Size Pearson Correlation
Sig. (2-tailed) N .025 .824 82 -.147 .187 82 .053 .639 82 -.064 1 -.171 .569 - .125 82 82 82
Season Pearson Correlation
Sig. (2-tailed) N -.158 .156 82 .359** .001 82 .183 .099 82 -.120 -.171 1 .285 .125 - 82 82 82 ** Correlation is significant at the 0.01 level (2-tailed)
A significance level of 0.05 is used in this regression, but the correlations with a significance level of 0.01 are highlighted with ** and are extremely significant. The correlations between the Price M2 and the variables MortgageRate, TOM, Location, Size and Season are respectively 0.049, -0.195, 0.039, 0.025 and -0.158. The strongest Pearson correlation of the given numbers is between the Price M2 and the TOM, which is a negative correlation. This means a rise in the price per square meter, will give a lower time on the market in days of a property. But none of these are significant. This implies that increases or decreases in the Price M2 does not significantly relate to increases or decreases in the five given variables.
Page | 21
The strength of the relationship between Season and MortgageRate is the strongest of Table 3 (.359) and is significant with a 0.01 level. The correlation between these two variables is positive, what says that the mortgage interest rate is higher for warm season, so for summer and spring.
The correlation results of the prices per square meter and the five different variables during the crisis are summarized in Table 4.
Table 4: Correlations during the crisis, 2012
Price M2 MortgageRate TOM Location Size Season
Price M2 Pearson Correlation
Sig. (2-tailed) N 1 - 82 -.150 .180 82 -.109 .329 82 .135 .104 -.136 .228 .354 .224 82 82 82
MortgageRate Pearson Correlation
Sig. (2-tailed) N -.150 .180 82 1 - 82 .011 .920 82 -.030 -.166 -.026 .791 .135 .819 82 82 82
TOM Pearson Correlation
Sig. (2-tailed) N -.109 .329 82 .011 .920 82 1 - 82 -.202 -.062 .109 .069 .579 .328 82 82 82
Location Pearson Correlation
Sig. (2-tailed) N .135 .228 82 -.030 .791 82 -.202 .069 82 1 -.062 -.120 - .581 .284 82 82 82
Size Pearson Correlation
Sig. (2-tailed) N .104 .354 82 -.166 .135 82 -.062 .579 82 -.062 1 .030 .581 - .792 82 82 82
Season Pearson Correlation
Sig. (2-tailed) N -.136 .224 82 -.026 .819 82 .109 .328 82 -.120 .030 1 .284 .792 - 82 82 82
There is also a significance level of 5% used in this regression, but in this case there isn’t a correlation with significance level of 0.01. The correlations between the Price M2
and the variables MortgageRate, TOM, Location, Size and Season are respectively -0.150, -0.109, 0.135, 0.104 and -0.136. The strongest Pearson correlation of the given numbers is now not between the Price M2 and the TOM as in 2007, but between the Price M2 and MortgageRate. This is a negative correlation, what means a rise in the price per square meter, gives a lower mortgage interest rate. There a none significant correlations. The correlation between Price M2 and the TOM is less negative, compared to 2007. This means that the time on the market
Page | 22
of the property (in full days) was larger in 2012 compared to 2007, when the Price M2 increased.
In tables 5 and 6 the regression results of before the crisis are given. The Price M2 is the dependent variable regressed on the five independent variables and a constant (intercept). The TOM in 2007 is the only value (0.099) what lays between the p-values 0.05 and 0.10, and is therefore marginally significant. The other variables are not significant. MortgageRate has the highest positive t-statistic (1.29) and TOM the lowest (-1.67).
Table 5: Regression results before the crisis, 2007
PriceM2 Coef. Std. Err. T P>|t| [95% Conf. Interval]
MortgageRate 909.2198 703.5754 1.29 0.200 -492.072 2310.512 TOM -2.80543 1.677945 -1.67 0.099* -6.147348 .5364884 Location 64.09943 246.0945 0.26 0.795 -426.0403 554.2392 Size 1.497964 5.505295 0.27 0.786 -9.466782 12.46271 Season -363.2848 256.1884 -1.42 0.160 -873.5283 146.9587 _cons -381.9056 3330.43 -0.11 0.909 -7015.031 6251.22
R-squared = 0.0739, F(5, 76) = 1.21, Number of obs = 82, *correlation is significant at the 0.10 level (2-tailed)
Table 6: Regression analysis; beta with significance before the crisis, 2007
Variable Beta Significance
1 MortgageRate 909.2 0.200
2 TOM -2.81 0.099
3 Location 64.1 0.795
4 Beta 1.50 0.160
Page | 23
In tables 7 and 8 the regression results of during the crisis are given. The Price M2 is the dependent variable regressed on the five independent variables and a constant (intercept). There are no significant variables. If we don’t look at the constant, Location has the highest positive t-statistic (0.95) and MortgageRate the lowest t-statistic (-1.19).
Table 7: Regression results during the crisis, 2012
PriceM2 Coef. Std. Err. T P>|t| [95% Conf. Interval]
MortgageRate -1342.793 1125.587 -1.19 0.237 -3584.593 899.0072 TOM -.4633702 .7874315 -0.59 0.558 -2.031676 1.104935 Location 243.1511 257.0535 0.95 0.347 -268.8152 755.1175 Size 2.995274 3.871297 0.77 0.442 -4.715082 10.70563 Season -263.6493 243.353 -1.08 0.282 -748.3289 221.0303 _cons 8887.452 4577.641 1.94 0.056 -229.7107 18004.16
R-squared = 0.0674, F(5, 76) = 1.10, Number of obs = 82
Table 8: Regression analysis; beta with significance during the crisis, 2012
Variable Beta Significance
1 MortgageRate -1342.8 0.237
2 TOM -0.463 0.558
3 Location 243.2 0.347
4 Beta 2.99 0.442
5 Season -263.65 0.282
In tables 9 and 10 the stepwise regression for both years is shown. In column 1 the dependent variable Price M2 is regressed only on one variable, the MortgageRate. In the second column, this regression is done for MortgageRate and TOM and so on. In both regressions the R-squared is higher as there a more regressions made. But in table 9 (2007) the R-squared rises with 0.0216 from (1) to (2), but only with 0.001 and 0.0028 in respectively (2) to (3) and (3) to (4). While adding Season in (4) to (5), the R-squared rises
Page | 24
with 0.0245. Thus, the R-squared is more effected (explained) by adding the TOM and Season, then by adding Location and Size. In table 10 (2012) the R-squared rises with 0.0115 from (1) to (2), and with 0.0123, 0.0068 and 0.0144 in respectively (2) to (3), (3) to (4) and (4) to (5). This is not the same effect as in 2007. For instance, adding Location gives in 2007 a R-squared increase of 0.001, but in 2012 a R-squared increase of 0.0123.
Before the crisis the MortgageRate coefficients are all positive, however during the crisis these are all negative. The explanation behind is, if the mortgage increase with for instance 1%, the results of the price per square meter of the house in 2007 shows an increase/decrease with the coefficient of that particular regression. For example regression (1), if the mortgage interest rate increases with 1%, the results of the price per square meter shows an increase with €284,06. In 2012 regression (1), if the mortgage interest rate increases with 1%, the results of the price per square meter shows a decrease with €1496,90. This is the case when MortgageRate is the only independent variable.
In the case of regression (2) in table 9, if the TOM increases with 1 day, the results of the price per square meter shows an decrease of €3,11. In table 10 this is the following: if the TOM increases with 1 day, the results of the price per square meter shows an decrease of €0.74, which is negligible small.
For regression (3) in table 9, Location is a dummy variable who has the value one when it has a positive location and zero when the location is neutral or negative. So, when there is a positive location, the results of the price per square meter shows an increase of €70,13. In table 10, when there is a positive location, the results of the price per square meter shows an increase of €255,03.
For regression (4) in table 9 and table 10, if the size of the population grows with one person, the results of the price per square meter shows an increase of respectively €2,62 and €2,89.
The last regression interpretation is (5), Season is a dummy variable who has the value one when it is warm season (spring and summer) and zero when it is cold season (fall and winter). In table 9, for regression (5), when it is warm season, the results of the price per square meter shows a decrease of €363,28. In table 10, when it is warm season, the results of the price per square meter shows a decrease of €263,65.
Page | 25
Table 9: Stepwise regression results before the crisis, 2007
PriceM2 (1) (2) (3) (4) (5) MortgageRate 284.056 (643.225) 514.439 (644.75) 551.643 (661.438) 608.164 (675.17) 909.22 (703.575) TOM -3.1077 (1.6442) -3.064 (1.661) -3.1235 (1.674) -2.805 (1.678) Location 70.133 (245.14) 81.017 (247.413) 64.099 (246.09) Size 2.621 (5.4837) 1.498 (5.505) Season -363.28 (256.19) _cons 2282.162 (3062.87) R-squared = 0.024 1372.35 (3052.96) R-squared = 0.0456 1163.627 (3156.33) R-squared = 0.0466 863.67 (3233.53) R-squared = 0.0494 -381.91 (3330.43) R-squared = 0.0739
Note: the numbers are the coefficients of the variables, the numbers between brackets are the std. errors.
Table 10: Stepwise regression results during the crisis, 2012
PriceM2 (1) (2) (3) (4) (5) MortgageRate -1496.90 (9474.10) -1484.753 (1106.77) -1453.73 (1107.2) -1311.77 (1126.49) -1342.79 (1125.59) TOM -.7425 (.7646) -.5845 (.7806) -.5396 (.7852) -.4634 (.7874) Location 255.031 (254.49) 270.70 (256.08) 243.15 (257.05) Size 2.893 (3.875) 2.995 (3.871) Season -263.65 (243.35) _cons 9474.10 (4469.86) R-squared = 0.0224 9529.36 (4471.81) R-squared = 0.0339 9288.55 (4478.14) R-squared = 0.0462 8630.32 (4576.64) R-squared = 0.0530 8887.45 (4577.64) R-squared = 0.0674
Page | 26
6. Discussion
In this paragraph the research purpose is given, together with the expected and unexpected findings of the regression results. In the first subparagraph the comparison with the year 2017 is given. In the second subparagraph ideas for future research are suggested.
The research purpose was to enlarge the knowledge about the factors that have an influence on the house prices in Amsterdam, before and during the crisis. These time periods were used to get a proper representation of the influence of a ‘normal’ year, compared to a ‘crisis’ year. Also, to can compare the increased house prices in 2007 with the year 2017, what will discussed in the next subparagraph. There is chosen for the real estate market of Amsterdam, because this is an incomparable market with the rest of the Netherlands and because the market is known for the tightness of it.
The research was focused on the dependent variable price per square meter and the independent variables; mortgage interest rate, time on the market, location, size of population and season. The mortgage interest rate was measured by the daily rate, the time on the market in full days. The location is divided in three different groups, positive/neutral/negative. The daily size of the population is taken by the monthly size of the population divided through the numbers of days of that particularly month. Last, for season there are two groups made, warm season (summer and spring) and cold season (winter and fall). To measure this, two databases were selected from the real estate website leden.nvm.nl. Each year contained of 83 randomly properties, which were checked for every single case.
To test the hypothesis, the regression on those databases were used. First, the summary statistics with among other things the mean and standard deviates were found. Second, the correlations with the all variables is measured and finally there is a stepwise regression done for both years. Only one hypothesis is true and was an expected finding; the TOM is marginally significant to predict house prices in the year 2007 and not in the year 2012, if there is a 0.10 significance level. In table 3 and 5 is shown that there is a negative correlation and t-statistic. This means, the higher the price per square meter, the lower the TOM. This is supported by Cubbin (1974). The demand in 2007 is so high (and supply low) that the houses will be sold quickly, even when prices are rising. Because of scarcity and popularity of properties in Amsterdam, this can be the explanation for the negative correlation and t-statistic of the price per square meter and the TOM.
Page | 27
An unexpected finding of this research is that the numbers for regression (5) in the stepwise regression results (table 9 and 10) are negative for the variable Season. This means, if it is warm season, the results for the price per square meter shows a decrease of €363.28 in 2007 and €263.65 in 2012.
Another unexpected finding is that the TOM in the regression results of 2007 (table 5) marginally significant is with the lowest P>|t|, but has in the regression results of 2012 the highest P>|t|. Also, when you look at the stepwise regression results of 2007 and 2012, you see a huge difference between the mortgage interest rate in (1) 2007 and the mortgage interest rate in (1) 2012. This means, if there is a mortgage interest rate increase of 1%, the results for the price per square meter shows an increase of €284,06 in 2007 and a decrease of €1496,90 in 2012.
Besides this, there was a more significance expected for the mortgage interest rate in 2007, the size of the population in 2012, a positive location in both years and for a warm season in both years.
6.1 The comparison with 2017
Last year, in 2017, there were several newspapers stated all the same thing. The house prices in Amsterdam rose enormously. ‘Housing prices Amsterdam rose 21 per cent in one year’, (Parool, July 2017), ‘Houses Keizersgracht have the same price tag as all houses at Terschelling’, (Volkskrant, April 2017), ‘Square meter price houses in Amsterdam are rising enormously’, (NRC, January 2017). In this subparagraph the factors that are researched in this paper for 2007 are linked to the pattern that occurs in 2017 on the real estate market. The reason for this, is that the house prices in Amsterdam are increasing again, like the case in 2007. The factor season is not discussed, because this is an external, not changing factor over time.
Hekwolter of Hekhuis et al. (2017) found that the recently house pricing rise is for the time being not credit-driven, compared to 2012 where this was the case. But the strong price increase indicates an overheated housing market. This fits with the traditional image that the Amsterdam housing market is more cyclical than that in the rest of the Netherlands. Van der Harst and de Vries (2017) found that a further reduction of the mortgage interest deduction and a reduction of the loan-to-value ceiling remain desirable in order to strengthen the underlying financing structure, but the influence of this in these last years is limited because the Amsterdam houses are more purchased by people’s own money. The mortgage interest rate in 2017 was decreased to an average of 2%, this was around 5% in 2007. Because of this
Page | 28
low interest rate, more people will be inclined to buy a house. Besides, the GDP rose by 3.3 per cent, consumption of households by 2.2 per cent and unemployment decreased from 6 per cent in 2016 to 4.9 per cent in 2017 (Staat van de Woningmarkt, 2017).
In the introduction part of this thesis, is figure 1 shown where the average transaction price rise is higher for 2017 compared to the rise in 2007. Due to this fact and the marginally significance of the TOM (with price per square meter) in 2007, for the year 2017 the TOM could be even lower in comparison with 2007.
Besides the TOM, the population is changed between 2007 and 2017. Immigration has led to strong population growth in 2016 and 2017 (ING Economisch Bureau). About half of this intake comes from the US, Europe, India or China. These people often have a good financial situation and therefore often opt for a purchase or free sector rental home. In the crisis years 2008-2013 foreign influx was much more limited. At that time, the metropolitan population grew alongside births, more because of domestic relocations. Many young people moved to Amsterdam, but only few locals left the capital. In 2017 there was for the first time since 2007 a more outflow of families than inflow, so there was a shift between the percentages of different inhabitants (ING Economisch Bureau). The size of the population is thus changed, what could give more pressure on house prices.
The number of new built properties in 2017 is 4916, this number was in 2007 3287. The most properties were built in the East (2017). The East of Amsterdam have ZIP codes between 1071-1099, what is the positive group. For this reason the positive group will relatively grow, compared to the neutral and negative group. This could influence the price per square meter, if the factor was significant in this research. However, this is not the case.
The reason why the house prices in 2017 are much higher compared to 2007, is due to the crisis period. There were too little properties built, for some years no properties at all. The result is the scarcity of the properties now, what causes the rising house prices. Besides, the mortgages interest rate is almost twice as low compared to 2007.
6.2 Further research
In this research there are a few limitations, considering the low significance of the output. For future research a bigger database can be established and more years could be add, so the results should be more informative and probably more significant. Besides, more variables can be added. Like the inflation, people’s confidence level or GDP. Also, a comparison of other big cities in different countries could be measured for future research too. Because Amsterdam has a unique and tight real estate market in the Netherlands, it can be
Page | 29
informative to have a look at the real estate market of London or Paris for instance. To see how these real estate market differs compared to Amsterdam in the years before and during the crisis, and to see the pattern that occurs after these years.
Page | 30
7. Conclusion
In this research, an multiple regression is used to see if there are significant factors that predict house prices in Amsterdam for before (2007) and during the crisis (2012). With an addition of a comparison of these years with the year of 2017. There will be looked to the hypotheses that were given in paragraph 3.
The first hypothesis, where is stated that the same factors can be used to predict the house prices for the two time series, is partly true. On the one hand, from previous research is concluded that every factor has an impact on the house prices, so every factor can partly predict the house prices during the different time series. On the other hand, in this research there is no significance found for the same variable in the two years. Hypothesis 2 is rejected, there is no significance found for the mortgage interest rate for both years. This is also the case for hypothesis 4, 5 and 6, where the variables location, size of population and season are used. There is no significance found in both years, what reject those hypothesis. Hypothesis 3 is the only one who is true if there is a 0.10 significant level (2-tailed); the TOM is marginally significant to predict house prices in the year of 2007, but not for the year 2012. To give answer to the research question, the factors to predict house prices for before (2007) and during the crisis (2012) in Amsterdam are not consistent.
I note that there need more research done with a bigger database of houses, to really confirm this hypotheses results. Using more variables and longer time series could give another result.
The comparison with 2017 is given as follows: the mortgage interest rate was lower in 2017 than 2007. Suggest is a lower TOM in 2017 compared to 2007, in connection with the higher transaction price in 2017. The size of the population is increased in 2017 and the positive location group is grown too. However, because of negligible significance for these factors in 2007, it is difficult to give a proper conclusion for the comparison between 2007 and 2017.
Page | 31
Reference list
Ambrose, B. W., Eichholtz, P., & Lindenthal, T. (2013). House prices and fundamentals: 355 years of evidence. Journal of Money, Credit and Banking, 45(2‐3), 477-491.
Amsterdam, Gemeente. "Amsterdam in cijfers 2007’’. Dienst Onderzoek en Statistiek (2008).
Amsterdam, Gemeente. "Stadsdelen in cijfers 2012." Amsterdam: Bureau Onderzoek en Statistiek (2013).
Anglin, P.M., Rutherford, R., & Springer, T. M. (2003). The trade-off between the selling price of residential properties and time-on-the-market: The impact of price setting. The Journal of Real Estate Finance and Economics, 26(1), 95-111.
Bervaes, J.C.A.M. & J. Vreke (2004). De invloed van groen en water op de transactieprijzen van woningen, Alterra-rapport, 959.
Bijlsma, M., Boone, J., & Zwart, G. (2014). Competition leverage: how the demand side effects optimal risk adjustment. The RAND Journal of Economics, 45(4), 792-815.
Boterman, W., Hochstenbach, C., Ronald, R., & Sleurink, M. (2013). Duurzame toegankelijkheid van de Amsterdamse woningmarkt voor starters. Amsterdam: University of
Amsterdam.
Cubbin, J. (1974). Price, quality and selling time in the housing market. Applied
Economics, 6, 171-187.
Daly, J., G. Stuart, D. Jenkins & F. Plimmer (2003). Consumer behaviour in the valuation of residential property: a comparative study in the UK, Ireland and Australia, Property Management 20, 295-314.
Eichholtz, P., Huisman, R., & Zwinkels, R. C. (2015). Fundamentals or trends? A long-term perspective on house prices. Applied Economics, 47(10), 1050-1059.
Eichholtz, P., Straetmans, S., & Theebe, M. (2012). The Amsterdam rent index: The housing market and the economy, 1550–1850. Journal of Housing Economics, 21(4), 269-282.
Glaeser, E. L., & Parker, J. A. (2000). Comments and discussion. Brookings Papers
Page | 32
Heijden, H. V. D., & Boelhouwer, P. (1996). The private rental sector in Western Europe: developments since the Second World War and prospects for the future. Housing
Studies, 11(1), 13-33.
Houben, Dröes en R. van Lamoen (2017) De prijsstijgingen op de Amsterdamse woningmarkt onder de loep. Economische Statische berichten
Huang, J., & Palmquist, R.B. (2001). Environmental Conditions, Reservation Prices, and Time on the Market for Housing. Journal of Real Estate Finance and Economics, 22, 203-219.
Jud, G. D., & Winkler, D. T. (2002). The dynamics of metropolitan housing prices. The journal of real estate research, 23(1/2), 29-46.
Kakes, J., & Van Den End, J. W. (2004). Do stock prices affect house prices? Evidence for the Netherlands. Applied Economics Letters, 11(12), 741-744.
Kiel, K.A., & Zabel, J.E. (2008). Location, location, location: The 3L Approach to house price determination. Journal of Housing Economics, 17, 175-190
Lutzenhiser, M., & Netusil, N. R. (2001). The effect of open spaces on a home's sale price. Contemporary Economic Policy, 19(3), 291-298.
Machinea, J. L. (2009). The international financial crisis: its nature and the economic policy challenges. cepal Review.
Mason, Charlotte H., and William D. Perreault Jr (1991). "Collinearity, power, and interpretation of multiple regression analysis." Journal of marketing research, 268-280.
Mason, C. H., & Perreault Jr, W. D. (1991). Collinearity, power, and interpretation of multiple regression analysis. Journal of marketing research, 268-280.
M. Bijlsma, J. Hers en R. Mocking (2013). ‘De Nederlandse woningmarkt. Hypotheekrente, huizenprijzen en consumptie,’ CPB Notitie, 15.
M. Hekwolter of Hekhuis, R. Nijskens & W. Heeringa,(2017). "The housing market in major Dutch cities," DNB Occasional Studies 1501, Netherlands Central Bank, Research Department.
M. Peihani (2015). ‘The Basel Committee on Banking Supervision. A post crisis assessment of governance and accountability’, Canadian Foreign Policy Journal 21:2 146-163, aldaar 146.
Page | 33
Mrkaix, M., M. Hassine en S. Saksonovos (2015). ‘Kingdom of the Netherlands’. IMF Country Report 14/328 1-66.
Ngai, L. R., & Tenreyro, S. (2014). Hot and cold seasons in the housing market. American Economic Review, 104(12), 3991-4026.
Notten, F., (2011). ‘Hypotheekschuld in Nederland’, Rapport CBS: de Nederlandse Economie.
Philips, P.C.B., S. Shi and J. Yu (2015). Testing for multiple bubbles: historical periodes of exuberance and collapse in the S&P 500. International Economic Review, 56(4), 1043-1078
P. De Vries & F. Van der Harst (2017) Grote regionale verschillen bij herstel woningmarkt. Kadaster nieuws.
P. Visser en F. Van Dam (2006). De prijs van de plek, woonomgeving en woningprijs. Ruimtelijk Planbureau, Den Haag, 30-93.
Rawal, S., & Rodgers, G.J. (2004). Growth and coagulation in a herding model. Physica A: Statistical Mechanics and its Applications, 344, 1-10.
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.
R. Kaashoek (2012). ‘Effect crisis op de woningmarkt dynamiek’, De Nederlandse Economie 2012, 154-164, aldaar 164.
Rosen, H. (1979). "Housing Decisions and the U.S. Income Tax: An Econometric Analysis." Journal of Public Economics, 11.
Teye, A. L., & Ahelegbey, D. F. (2017). Detecting spatial and temporal house price diffusion in the Netherlands: A Bayesian network approach. Regional Science and Urban
Economics, 65, 56-64.
Van Der Veer, J., & D. Schuiling (2005) The Amsterdam housing market and the role of Housing Associations, Journal of Housing and the Built Environment, Vol. 20, pp. 167-181.
Visser, P., & F. van Dam (2006). De prijs van de plek: woonomgeving en woningprijs, Rotterdam/Den Haag: NAI Uitgevers / Ruimtelijk Planbureau.
Page | 34
Visser, P., Van Dam, F., & Hooimeijer, P. (2008). Residential environment and spatial variation in house prices in the Netherlands. Tijdschrift voor economische en sociale
geografie, 99(3), 348-360.
Wit, E.R. de, P. Englund en M.K. Francke (2013). Price and transaction volume in the Dutch housing market. Regional Science and Urban Economics, 43(2), 220-241.
Yavas, A., & Yang, S. (1995). The strategic role of listing price in marketing real estate: theory and evidence. Real Estate Economics, 23(3), 347-368.
