Regional house price differences are widening
Dutch Housing Market Quarterly
- In 2017, we expect house prices to rise by 7.6 % in the Netherlands, driven by the limited supply, household growth in the larger cities, the economic recovery and persistently low interest rates
- The slight rises in interest rates and worsening affordability mean that price rises in 2018 will be slightly lower, at 7.0%
- First-time buyers and young homeowners moving up the property ladder are experiencing greater obstacles in the owner-occupied housing market
- In anticipation of further price rises buyers are paying well above the initial asking price for homes in the larger cities, particularly Amsterdam. This behaviour can signify the formation of a price bubble
- In early 2018, house prices may reach the level of the previous national peak seen at the end of 2008, although prices may not recover until after 2019 in regions outside the Randstad
- Following a record number of sales in the first two quarters, for the remainder of 2017 we expect a stagnation in the growth of the number of transactions,leading to a total of 240,000 sales this year
The Dutch economy is currently running at full throttle, largely thanks to domestic consumption and exceptionally high consumer confidence. The same is also true of the Dutch housing market: the sharp rise in the number of sales (+19% in the second quarter) and robust growth of the price index for existing owner-occupied homes (PBK, +7.7% in the second quarter) reflect the strong tendency of Dutch households to move house and their willingness to buy.
We expect the number of sales to fall sharply during the second half of the year compared to the first half, possibly even falling compared to the second half of 2016. Prices will still continue to rise strongly during the coming eighteen months due to limited supply, low mortgage rates and the financial attractiveness of investing one’s own capital in houses. This last factor in particular increases the risk of a bubble developing in hot markets such as Amsterdam. Instead of households keeping their debt position low, they are using their own funds above all to buy houses over and above the asking price, in the implicit expectation that prices will continue to rise. At the end of this Dutch Housing Market Quarterly (and the appendant column) we will consider the question of whether the government should intervene in the current pricing processes and what would be the most effective options in this respect.
The market for existing owner-occupied homes
Number of sales
Since the number of houses for sale has fallen considerably since last year, one could be led to believe that the number of house sales is also falling fast. However, the number of transactions has again risen sharply, reflecting the strong tendency of Dutch households to move house and their willingness to buy. During the second quarter of 2017 the Land Registry recorded 58,370 sales (see Figure 1), some 16% more than a year earlier.
If we take the number of transactions in the first half of 2017 as the standard, demand for owner-occupied homes has never been as high as it is at present. Even in the first half of 2006 fewer homes changed hands (110,000 then compared to 114,000 now). Even so, it is worth realising that the number of households, and with it the number of potential buyers, has also risen sharply (from 7.14 to 7.72 million, representing an increase of 8%). As a result, the growth in the number of transactions is somewhat less spectacular than would appear at first sight.
For the second half of the year we expect a stagnation in the number of sales compared to the last six months of 2016. This leads us to expect around 240,000 sales in 2017 as a whole (see Figure 2). The number of transactions is likely to fall next year because the number of homes for sale has also fallen significantly (see, for example, Huizenzoeker.nl). Moreover, according to the Dutch Association of Real Estate Brokers (NVM) the number of signed purchase contracts fell in the second quarter (NVM, 2017). This figure anticipates the official transaction count by the Land Registry (see Box 1 for a further explanation).
Box 1: The difference between figures by the Land Registry and the NVM
The NVM records sales at the moment of signing the purchase contract. The Land Registry only records this when the sale has been executed by the civil-law notary. This happens on average two to three months later, which means that the NVM’s figures are ahead by two to three months. As a result, the NVM reports earlier than the Land Registry a rise or fall in the number of houses sold.
On 13 July the NVM published its figures on the housing market in the second quarter, which showed a year-on-year fall in the number of homes sold of around 0.8%. If we assume that the NVM’s figures forecast the recorded sales, we should then see in two to three months a fall in the housing market in the Land Registry’s figures too – see Figure A for the long-term relationship between these two variables. However, this conclusion is premature as several factors play a role.
For example, the NVM records only the homes sold by estate agents affiliated to the NVM, where their market share can fluctuate. During the last few quarters their market share has indeed fallen, which makes it more difficult to give an accurate estimate of the Land Registry figures for the coming three months. Moreover, the supply of houses for sale is severely underestimated. Compared, for example, to the figures produced by Huizenzoeker.nl, the NVM’s estimate for the Netherlands as a whole is lower by around 10,000 homes.
In short, if we bear in mind that it is difficult to estimate accurately how many homes will come on to the market in the short term and we do not know exactly what the NVM’s market share is, it is not possible to predict precisely the number of transactions recorded over three months on the basis of the NVM’s figures.
There are clear indications in the sales figures that first-time buyers, and probably also young homeowners moving up the property ladder, are experiencing more obstacles when buying a house. If we divide up buyers of existing owner-occupied homes by age, we see that buyers older than 35 years above all account for the largest proportion of growth in transactions (see Figure 3). Since the second quarter of 2015 the number of people aged 35 to 55 years and those older than 55 years who bought a house grew by 55 and 54% respectively. More transactions were also recorded among younger adults (<35 years) but this growth was significantly less at just 32%. The higher prices would therefore appear to benefit those moving up the property ladder, the main reason being that many of them have built up more equity in their previous house. This is also reflected in the transaction trends according to house type. While the number of transactions for apartments has risen by around 36% since the second quarter of 2015, the relative rise for other types of houses, particularly detached houses, was substantially higher (+50%). During the last twelve months in particular, we have seen only a slight rise in the number of apartments sold (see Figure 4).
Alongside the robust growth in the number of sales, prices for existing owner-occupied homes have also risen strongly. Compared to the second quarter of 2016, the house price index (Prijsindex Bestaande Koopwoningen – PBK) rose by 7.7% in the second quarter of 2017. This means that prices in the Netherlands are currently only 7.3% lower than the price peak seen in the third quarter of 2008 (see Figure 5). What is especially interesting here is that the speed of price rises is still accelerating. During the first half of 2017, each month saw a higher year-on-year rise than the preceding month (from +6.4% in January to +8.0% in June). The average sale price in the second quarter was roughly 260,000 euros, a rise of 21,000 euros in just one year.
In our previous Dutch Housing Market Quarterly we predicted that the house price index would rise by 6.5% in 2017 compared to 2016. Since the price rises in the first half of this year were substantially higher than we expected, we are adjusting our forecast upwards for the year as a whole. Partly due to the persistently low interest rates, the fall in the number of houses for sale and the considerable willingness of buyers in many regions to make offers well above the asking price, we expect the trend of strong price rises in the housing market to continue in the second half of the year too. Prices are expected to rise by 7.6% in 2017. In addition, we expect significantly higher prices in 2018 than previously forecast. Although the worsening affordability of owner-occupied homes is expected to slow growth somewhat, house prices may still end up 7% higher (see Figure 6). However, the new government’s possible new housing market policy and unforeseen events in the Netherlands and abroad make these forecasts for 2018 more uncertain.
Highlighted: regional differences in the housing market
The number of sales grew again in all provinces, but Utrecht and Noord-Holland in particular stand out compared to the rest of the country (see Figure 7). In Amsterdam, the number of transactions actually fell sharply. As the number of houses for sale is drying up, the fall in sales in the capital is now being followed by the city of Utrecht, with the result that sales in the provinces of Utrecht and Noord-Holland grew only slightly. Flevoland posted the highest growth by far in the past quarter, at 31%. In Gelderland, Limburg, Noord-Brabant, Zeeland and Overijssel too, relatively speaking many more houses were sold.
The fact that the number of sales in Flevoland and Gelderland is rising strongly can partly be explained by their position in the country, close to Utrecht and Amsterdam. In these cities, but also in the former ‘overflow’ towns of Haarlem and Amstelveen, it is becoming increasingly difficult for many people to find a suitable house, let alone one within their budget. Consequently, they are increasingly moving further afield to towns such as Almere and Arnhem.
Figure 8 shows clearly that the changes in house prices we have described for the country as a whole show marked regional differences. In Noord-Holland, Utrecht and Zuid-Holland we see year-on-year price rises well above the national average. On the other hand, the provinces of Overijssel, Drenthe, Limburg and particularly Zeeland show more modest price trends. Strikingly enough, Flevoland occupies a position among the provinces with the highest price rises. This too could be a sign that prices in the north of the Randstad are becoming too extreme for more and more people, or that they simply cannot find a house and therefore look further afield to Almere.
For each province and the four largest cities, Figure 9 shows how severe the price falls were during the crisis and the extent to which prices have now recovered. Particularly striking is that all the provinces suffered price falls of between 20 and 30%, with the exception of Zeeland, but the price recovery shows a much more varied picture.
Figure 10 visualises the sharp dividing line between the Randstad and the rest of the country. Driven by the robust price rises in urban centres such as Amsterdam and Haarlem, Noord-Holland is the first province in the Netherlands where average house prices are now higher than ever before. Without correcting for inflation, houses are now 1.8% more expensive than then. The house price index in Amsterdam is even almost 22% above its peak in 2008. Utrecht and Zuid-Holland are also experiencing price index levels reminiscent of 2008. In contrast, outside the Randstad a house still costs considerably less than nine years ago. In Friesland, Drenthe, Gelderland and Noord-Brabant in particular, houses are still much more affordable than they were at that time, with prices at least 12% below the peak of 2008. The fact that prices in Groningen and Limburg are closer to their peak than in their neighbouring provinces is largely due to the popularity of their capitals. In fact, as early as the end of 2016 average sale prices in the cities of Groningen and Maastricht had already reached the levels they were at before the crisis.
Now that Noord-Holland has passed its historical peak, the question arises as to when the other provinces will also do so. If we take the average rises in the price index over the last four quarters as a starting point, Utrecht is the first to follow Noord-Holland to a new price peak. This is already expected to occur in 2017. In Zuid-Holland and Flevoland too, price rises in the past quarters have been so high that it will not be long at all before homes will be more expensive than ever before. In all other provinces, this will only happen in 2019, and in Drenthe and Zeeland not even until 2020 (see Figure 11), at least if prices continue to rise at their current rate. A comparison of both maps shows clearly again the exceptional position of Zeeland: unlike other provinces, the range between the higher and lower prices are less extreme here.
Factors influencing market trends
The Dutch economy and consumer confidence
In our most recent Economic Update we describe how the Dutch economy is running at full throttle. Although the rise in wages is still only modest, consumer spending has risen strongly and this has had a decisive influence on economic growth. We also see this in the extremely high level of consumer confidence in the first half the year (see Figure 12).
This positive picture also extends to confidence among Dutch households in the owner-occupied housing market. The Eigen Huis Market Indicator of the Homeowners’ Association (Vereniging Eigen Huis - VEH) has stabilised over the last few months at 113 points. With the exception of declining affordability caused by rising house prices, actual and potential buyers are positive about the stability of the Dutch housing market (Boumeester & Lamain 2017). This translates directly into a strong willingness to buy and faster sales, where in many regions it has become the norm to pay above the asking price (see Koster & Rouwendaal, 2017).
Despite the rising prices of owner-occupied homes, current interest rates mean that housing costs for homeowners are kept relateively low levels. Figure 13 shows that mortgage rates have fallen since 2012 and since October 2016 have stabilised at around 2.4%. Alongside this, the low interest rates also make it an attractive option to move house. This allows those moving up the property ladder to pay off existing mortgages without penalties and then take out a mortgage at a lower interest rate. A second factor making owner-occupied homes more accessible has been the extension of the debt-to-income limits caused by adjusted mortgage norms by the National Institute for Family Finance Information (NIBUD) at the start of 2017 (Nibud, 2016). This has allowed two-income households to borrow larger amounts. A third factor is the increase in the gift tax exemption, which on 1 January 2017 was set permanently at EUR 100,000. This has improved the position on the owner-occupied housing market particularly of first-time buyers and young households moving up the property ladder with wealthy parents. All in all, despite the strong price rises, affordability of owner-occupied homes has not yet been a hampering factor. However, fears of worsening affordability in the future may have contributed to a persistently high growth in sales.
If we compare affordability in the owner-occupied housing sector with that of the rental sector, we see that the average owner-occupied home has become cheaper in recent years compared to the private rental sector. The causes of this are clear. In recent years renters have been unable to benefit from falling mortgage rates, but have experienced hefty rent rises. The feeling that you are throwing money down the drain if you do not buy a house plays a major role and drives up demand for owner-occupied homes even further. This is particularly the case among young couples and families with middle incomes: the public rental sector is closed to them, while the private rental sector is an expensive (or often too expensive) alternative to buying.
Rising demand for buy-to-let investments on the housing market is playing a major role. Recent studies by the Dutch Central Bank (2017) and the Land Registry (2017) show that in the local hotspots above all – the G4 cities but also other student cities within and outside the Randstad (e.g. Leiden, Groningen, Maastricht) – the demand for multiple property ownershiphas risen sharply since the crisis. Since these homes are generally rented out (often to their own children), the segment for owner-occupied homes is becoming smaller and prices are rising even faster. What’s more, buy-to-letters often buy these homes using their own funds and therefore do not need to meet any financing requirements. In practice this manifests itself as paying above the asking price, which can be financed in the long term by high rental incomes.
The rising demand for owner-occupied homes means greater shortages in the market for these homes as supply falls. By relating the supply of houses for sale to the transactions recorded by the Land Registry (with a delay of two months), Figure 14 shows that shortages have continued to intensify since mid-2013. In June 2017 there were only around six houses for sale for each transaction. There are once again strong regional differences, though. For example, the figure is lowest in Utrecht, Amsterdam, and Haarlem (around 2.5 homes per transaction) while the market in other regions is still relatively relaxed (such as in Zeeland, Limburg and Drenthe).
The fact that the number of new homes has not been able to keep pace with the growth in the number of households has been a deciding factor. A good (and extreme) example of this can be seen in the trends in Amsterdam: since 2011 the number of households has grown substantially compared to the number of homes (see Figure 15). For the Netherlands as a whole, the most recent forecast by Statistics Netherlands is a growth of an average of 62,500 households each year until 2020 (CBS, 2017). Since around 10,000 homes are demolished each year, a production of more than 70,000 homes is needed to prevent the housing shortage from getting any worse. The current forecast for the housing supply assumes only 55,000 new homes each year. But here too there are marked regional differences as population growth is strongly concentrated in certain localities. New homes are certainly not needed everywhere in the Netherlands.
Extreme price rises – what is the problem and what should we do about it?
Given our analysis of the trends in prices and sales, we would like to return at the end of this Housing Market Quarterly to the question of whether the Dutch market for owner-occupied homes finds itself in a credit-driven bubble again. In short: we believe that this is not yet the case. The fact that aggregate mortgage debt has risen over the last two years can largely be explained by the sharply rising number of sales. The tighter lending standards – particularly the lowering of the maximum permitted loan-to-value (LTV) to 100% in 2018 – and extra repayments are contributing to limited growth in mortgage credit and debt (see Figure 16 ). This was also the conclusion of a recent study by the Dutch Central Bank (2017).
The average LTV in the major cities has fallen and is now lower than in the rest of the country. The biggest price rises can therefore be explained by factors other than greater lending. The main ones are the economic recovery, strong confidence in the housing market and the economy, the relatively good affordability of owner-occupied homes compared to rented homes and the growing mismatch between the number of households and the housing stock. Even so, we believe that the risk of a bubble forming has risen at a local and regional level, not from greater lending, but due to the use of high levels of own capital by huse buyers. Most crucially, they are using savings and gift not to keep debt levels low, but to buy a house well in excess of the asking price. As such, they are anticipating strong price rises in the future, which in in our view is a potential sign of a bubble formation. The rise of buy-to-let investors is also an important point in this regard. They are often able to offer well above the asking price because no financing restrictions apply. After all, increasing demand among investors has been a sign of an overheated housing market in other countries.
It should be clear that if these trends continue and bubbles are formed in regional markets, this will pose a risk to the economy and financial stability (see, for example, WRR, 2016 or Jorda et al, 2016) because the greater the price rises in the coming periods, the greater the potential price falls afterwards. As a result, the risk of higher numbers and levels of households in negative equity and significant capital losses rises again.
Moreover, the high price rises we are currently seeing contribute to greater wealth inequality. It is becoming increasingly difficult for first-time buyers to find a suitable and affordable owner-occupied home without financial support. Given the present structure of the rental market where middle incomes pay very high rents and are hardly able to build up any savings, this leads in the long term to greater inequality between renters and buyers. Yet, volatile and regionally diverging house prices are also leading to a greater wealth inequalities between homeowners. This is due to the fact that the moment at which a house is bought in a boom-and-bust cycle as well as its location make a major difference in the potential increase in value of a house (see, for example, Hoenselaar & Mastrogiacomo, 2017).
Given our explanation of the underlying factors for the price rises, it would seem clear that the main challenge on the housing market is to bring supply and demand in better balance again. The key issue certainly is the construction of many more new homes. This applies particularly to the rental sector, so that it can once again be a genuine alternative for (younger) households on the housing market. However, limited capacity in the construction sector, a lack of locations ready for development and the fact that both property developers and municipalities benefit from shortages on the housing market contribute to delays in new building projects in the short run. In the long run, improved cooperation and coordination between local and national government, commercial parties and social partners is required to reach and implement a new construction agreement.
 It should also be mentioned here that it has now become the norm in many towns and cities to set a price of ‘offers above’ on homes. More research is needed to be able to estimate how this is related to a normal asking price.
 This is based on a revised estimate of our price model.
 At 100 points, the proportion of optimists and the proportion of pessimists in the housing market are equal.
 The Dutch Association of Real Estate Brokers (NVM) reports that homes offered through estate agents affiliated to it are sold within 63 days on average. In the first quarter of the year this was still an average of 77 days.
 The scope a household has for expenditure on the home is determined by deducting from the gross income the taxes and premiums payable and the other living costs. These other items of expenditure must remain affordable, also after taking out a mortgage. For 2017 it has been decided to use a debt-to-income limit for the highest income + 0.6 x the second income. In 2016 this factor was 0.5.
 The Scientific Council For Government Policy (Wetenschappelijke Raad voor het Regeringsbeleid - WRR) shows that the crisis on the housing market and high debt levels (many homeowners in negative equity) contributed substantially to the slow recovery of the Dutch economy until 2015. The reason is that in times of strong house price rises homeowners become over-optimistic, while the reverse happens in times of price falls.
 In our accompanying column we discuss a number of policy options, such as macroprudential policy, tax rises and the phasing out of subsidies, which may well slow down price rises in the short run, and how effective they could be.
The Dutch Housing Market Quarterly is a publication of RaboResearch Rabobank. The view presented in this publication has been based on data from sources we consider to be reliable. Among others. these include Macrobond, Land Registry, NVM, DNB, CPB and Statistics Netherlands. The date of completion is August, 10th, 2017.
This data has been carefully incorporated into our analyses. Rabobank accepts. however. no liability whatsoever should the data or prognoses presented in this publication contain any errors. The information concerned is of a general nature and is subject to change.
No rights may be derived from the information provided. Past results provide no guarantee for the future. Rabobank and all other providers of information contained in this brochure and on the websites to which it makes reference accept no liability whatsoever for the brochure’s content or for information provided on or via the websites.
The use of this publication in whole or in part is permitted only if accompanied by an acknowledgement of the source. The user of the information is responsible for any use of the information. The user is obliged to adhere to changes made by the Rabobank regarding the information’s use. Dutch law applies.
For more information. please call the RaboResearch secretariat on tel. +31 (0)30 – 216 6666 or send an email to ‘firstname.lastname@example.org’.
Christian Lennartz and Nic Vrieselaar
Menno Middeldorp, Head of RaboResearch Netherlands, Rabobank
© 2017 - Coöperatieve Rabobank U.A., Nederland