• No results found

Unconventional monetary policy and the Dutch housing market

N/A
N/A
Protected

Academic year: 2021

Share "Unconventional monetary policy and the Dutch housing market"

Copied!
44
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

1

Unconventional Monetary Policy and the Dutch Housing Market

Tim Hendriks

tjhendriks@outlook.com

10693793

Master thesis International Economics and Globalization

Supervisor: Prof. Dr. A.C.F.J. Houben

Second Supervisor: Dr. D.J.M. Veestraeten

University of Amsterdam

(2)

2

Statement of Originality

This document is written by Tim Hendriks 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.

Abstract

This thesis examines the impact on the housing market of the unconventional monetary policies used by the ECB. Unconventional monetary policy reduces interest rates along the yield curve and stimulates the allocation of capital to the housing market. In the short run, this leads to an increase in housing prices and a higher price-to-income ratio. The higher price-to-income ratio makes the housing market more sensitive to unexpected future shocks. Policymakers should use the period of low interest rates to accelerate the implementation of structural housing market reforms and macroprudential policies. This can reduce the cyclicality of the Dutch housing market, by preventing housing prices from rising too fast in the short run and reducing the risk of a downward price correction in the longer run.

(3)

3

Table of Contents

1. Introduction

5

2. Literature review

8

3. Unconventional monetary policy

11

3.1. Negative interest rates 11

3.2. Credit easing 12

3.3. Forward guidance

13

3.4. Quantitative easing 14

4. Transmission channels of unconventional monetary policy to the housing market

15

4.1 Negative interest rates 16

4.2. Credit easing 17

4.3. Forward guidance

17

4.4. Quantitative easing 18

4.5 Combined effects of unconventional monetary policy on mortgage yields

20

5. Effects of unconventional monetary policy on housing prices in the Netherlands

21

5.1 Methodology

21

5.2 Dataset 23

5.2.1. Housing market data 23

5.2.2. Long-term mortgage rate

24

5.2.3. Other demand variables

24

5.2.4. Macroprudential policies

25

5.3 Results 26

6. Potential risks to the housing market and financial stability in the future

34

7. Conclusion

38

(4)

4

9. References

40

(5)

5

1. Introduction

Since the financial crisis of 2008, the European Central Bank (ECB) has adopted a variety of unconventional monetary policy measures. In “normal” times, the ECB pursues its mandate of price stability by setting the nominal short-term interest rate. During the crisis, central banks needed more monetary policy tools to ensure price stability. This was due to the interest rate transmission mechanism becoming less efficient in influencing the consumption and investment decisions of firms and households. Early stages of unconventional monetary policy attempted to fix this transmission mechanism. Moreover, the nominal interest rate reached the effective lower bound. At the effective lower bound, it is hardly possible to stimulate the economy by further reducing the interest rate. Consequently, the ECB relied on unconventional monetary policy tools such as credit easing, quantitative easing and forward guidance.

This unconventional monetary policy took shape during three main phases of the recent financial crisis in the Eurozone. First, the ECB tried to combat the symptoms of the financial crisis. During this period, the interbank money market was barely functioning and the ECB took over a large share of the interbank market intermediation in the euro area by introducing a fixed rate full allotment (FRFA) tender procedure. FRFA helped to stabilize the banking sector by providing ample liquidity to the market. Furthermore, the ECB provided Longer-Term Refinancing Operations (LTROs) and expanded its list of eligible collateral. These new unconventional monetary policy tools were needed, because conventional monetary policy had become less effective in steering the financial conditions in the Eurozone.

During the second phase, the sovereign debt crisis, various European countries faced the risk of collapse of financial institutions. To prevent this from happening, governments were forced to intervene, which resulted in high levels of sovereign debt and rapidly increasing bond yields. The high bond yields of some southern European countries contained self-reinforcing dynamics and were thus a threat to the stability of the euro. In response, the ECB president Mario Draghi said that the ECB would do “whatever it takes to save the euro”. In addition, the ECB introduced sovereign bond purchasing programs, such as the Securities market program (SMP) and the Outright Monetary Transactions (OMT). These new unconventional monetary policy tools successfully repaired the monetary policy transmission mechanism and reduced the bond yields of distressed countries (Szczerbowicz, 2014).

(6)

6 The third phase, a period of low inflation and deflation

risk, resulted from the combined effects of the financial and sovereign debt crisis. This thesis focuses on the anti-deflation policies used by the ECB to combat this excessively low inflation. The necessity of unconventional monetary policy during this period is more controversial than during the global financial crisis and the sovereign debt crisis. At the end of 2014, inflation dropped below 0 % in the euro area. A main

reason for this low inflation was the sharp decline in oil prices. Some members of the Governing Council of the ECB saw a growing risk that low inflation could create a self-reinforcing negative spiral. This while the core inflation (inflation that excludes food and energy) was, and still is, relatively stable around 1%. The goal of monetary policy is to ensure price stability for goods and services. The ECB defines price stability as a year-on-year increase in the Harmonized Index of Consumer Prices of below, but close to, 2% over the medium term. To get the inflation rate close to target, the ECB introduced reinforcing unconventional monetary policy tools aimed at stimulating aggregate demand. Examples of these unconventional monetary policy tools are credit easing and quantitative easing.

There is a debate about whether these unconventional monetary policy tools are effective in achieving inflation targets. The ECB used unconventional monetary policy tools to counter the risks of deflation on the demand side of the economy. Since the introduction of these unconventional monetary policy tools, however, core inflation has hardly changed (Figure 1). Over the last years, inflation rates have been low in almost all advanced economies. One of the main reasons for this could be deflation driven by aggregate supply shocks. Supply-side deflation is the result of cost reductions gained through investments in new technologies and competitive price pressures from globalization (Bono et al, 2017). This lowers unit costs and puts downward pressure on output prices. Monetary policy tools aimed at stimulating aggregate demand might not be the right tools to combat such supply-side deflation (Beckworth, 2008).

At the same time, unconventional monetary policy tools can have serious negative side effects. Risks associated with these unconventional monetary policy tools are potential asset price bubbles, including housing price bubbles. Indeed, the global financial crisis has demonstrated that price stability does not guarantee financial stability. In the years before the financial crisis, inflation rates were low and stable while at the same time macrofinancial imbalances were building up. By the same token, the potential build-up of new housing price bubbles can be a threat to the financial stability of the Dutch economy. Financial stability can be defined in terms of the ability of the financial sector to manage risks, absorb

Figure 1: Core inflation (ECB, 2017)

0 0,5 1 1,5 2 2,5

apr-13 apr-14 apr-15 apr-16 apr-17

Core Inflation

(7)

7 shocks and facilitate economic processes. In this context, the debt to income ratio of households in the Netherlands is 218% of the Gross Domestic Product (GDP), one of the highest in the world (Appendix 1). Such high levels of household debt can put the financial stability of the Dutch economy at risk. Risks to financial stability have the potential to create widespread financial externalities, such as contagion and fire asset sales. Ultimately, this can result in negative outcomes and threaten sustainable economic growth.

This thesis aims to find an answer to the question what impact the ECB’s unconventional monetary policy tools to combat deflation have on the Dutch housing market. This is important, as an overpriced housing market could pose a threat to the Dutch economy. Section 2 provides a review of the existing literature on the effects of unconventional monetary policy on housing markets. Most of this literature relates to the United States, because the Federal Reserve introduced tools such as credit easing and quantitative easing earlier in 2009. Section 3 analyses the different unconventional monetary policy instruments used by the ECB. Section 4 focuses on the transmission channels of unconventional instruments to the housing market. Section 5 analyses the effects of unconventional monetary policy on the housing prices in the Netherlands by using a OLS time series regression model. Section 6 assesses the potential risks to financial stability. Section 7 concludes.

(8)

8

2. Literature review

Over the past decades, a large body of research has examined the impact of conventional monetary policy on housing prices and housing price bubbles. Dellepiane et al. (2013) compared the housing price bubbles of Spain and Ireland in the early 2000s in a case study. The economic structures and financial systems of Spain and Ireland are quite different, but both countries experienced massive housing price bubbles, which burst around 2008. These housing markets collapses naturally had a major impact on the economies of both countries. The different economic structures of Spain and Ireland made it possible to analyse the evolution of housing bubbles and to establish the factors that increase risks in the housing market. In their paper, Dellepiane et al. (2013) argue that three conditions led to the housing price bubbles in both countries. Firstly, both countries experienced a few decades of financial deregulation. This financial deregulation made it easier for households to get credit, which resulted in higher household debts and increased housing prices. Secondly, after Spain and Ireland entered the economic and monetary union, both countries experienced steep interest rate declines (Dellepiane et al., 2013). This cheap money fueled housing price increases of more than 300% and 420%, in Spain and Ireland respectively, between 1995 and 2007. Finally, Spain and Ireland both experienced a large inflow of international capital. This international lending was largely speculative. International investors expected housing prices to rise further, which resulted in a phase of speculative ‘mania’; Investors believed they had to buy before they were priced out of the market, and that the properties they bought could be sold for an even higher price in the future. When the housing market collapsed, international investors pulled their money out, which amplified the magnitude of the crash. Figure 2 shows that the ratio of change in property prices to change in earnings was the largest in Ireland and Spain, but also that this change was very large in the Netherlands. We can conclude that the

housing markets of small open economies are vulnerable to mobile international capital and cheap money.

0 0,5 1 1,5 2 2,5

DE AUS PRT ITA FIN UK

N OR SW E LUX GR EU 15 DK FRA BEL N TH SP IE

Ratio of change in property prices to change in earnings, 1995-2006

Figure 2: Ratio of change in property prices to change in earnings, 1995-2006 (Dellepiane et al. (2013))

(9)

9 Eickmeier & Hofmann (2013) establish that monetary policy in the United States has a large impact on housing prices. Their research suggests that monetary policy shocks increased the imbalances in the housing market in the years prior to the crisis. Looking back, they conclude that monetary policy, and especially the low level of policy rates, contributed to the unsustainable dynamics that were observed in the housing market between 2001 and 2006. When the housing bubble in the United States burst, this precipitated the global financial crisis.

The Federal Reserve introduced unconventional monetary policy tools such as quantitative easing and forward guidance to combat the global financial crisis. The main goal of these unconventional monetary policy tools was to lower the yield curve, and especially the long-term interest rate. A wide range of literature, including Krishnamurthy and Vissing-Jorgensen (2011), Swanson and Williams (2014) and Wright (2012) analyses the impact of unconventional monetary policy on the yield curve. There is broad consensus that unconventional monetary policy was effective in bringing down the medium- and longer-term interest rates. However, these papers focus only on the effects of unconventional monetary policy on the yield curve. A more fundamental question is what effects these lower yields have on the economy, and especially the housing markets.

There are two opposite views in the literature about the effects of unconventional monetary policy on the housing market. A part of the literature suggests that the large-scale asset purchase program of the Federal Reserve helped to stabilize the housing market in the United States. For example, Walentin (2014) shows in his paper that the asset purchases of the Federal Reserve in the mortgage market were successful in bringing down the mortgage spread. At the peak of the crisis banks were hardly providing loans and tried to deleverage their balance sheets. By providing liquidity to the market, the Federal Reserve helped to decrease mortgage spreads and increase the number of newly issued loans. Using a VAR model, Walentin (2014) find evidence that the lower mortgage spread helped stabilize housing prices, which had, in the short run, had a direct stabilizing effect on the economy of the United States.

Gabriel & Lutz (2015) used a structural factor-augment vector autoregressive (FAVAR) model to analyse the impact of quantitative easing on the housing market. Their paper confirms that expansionary monetary policies lowered the key housing market interest rates in the whole of the United States. However, the impact of unconventional monetary policies differed strongly across geographical regions. They establish that the unconventional monetary policy benefitted the weaker economic housing regions and stopped the decline in housing prices. Gabriel & Lutz (2015) find evidence that quantitative easing had the largest impact on the most speculative housing markets such as Florida and California. Altogether,

(10)

10 Gabriel & Lutz (2015) conclude that unconventional monetary policy can lower the distress in the housing market, but at the same time may increase risks in the most speculative markets.

Rahal (2017) studied the effects of unconventional monetary policy on the housing markets in different OECD countries, using a panel vector auto regression model. He finds evidence that unconventional monetary policy affects housing prices, peaking one or two years after a policy loosening. The recent recovery in housing prices in different OECD countries can be partly attributed to unconventional monetary policies (Rahal, 2017).

Summarizing the existing literature, one can conclude that unconventional monetary policy has an effect on the mid- and long-term interest rates and on housing prices. However, do these uncommon measures create new financial stability risks by contributing to housing price bubbles? Existing literature does not clearly establish the impact of unconventional monetary policy on financial stability in the longer run. It is important to note that monetary policy focusses on price stability in terms of consumer prices, and not on financial stability in terms of the financial sector’s contribution to the real economy. Price stability and financial stability are two different policy objectives for a central bank. The Tinbergen rule says that one should have at least as many instruments as objectives (Tinbergen, 1952). After the financial crisis, macro prudential policies were introduced to promote financial stability. It should be noted that monetary policy and macro prudential policy are both not directly aimed at the housing market, but both policy tools nonetheless have an impact on the housing market.

(11)

11 3.

Unconventional monetary policy

The ECB uses unconventional monetary policy to pursue its price stability objective. This unconventional monetary policy consists of four main mutually reinforcing instruments: negative interest rates, credit easing, quantitative easing and forward guidance. These instruments were introduced to combat distortions in the monetary policy transmission mechanism, which were a result of liquidity constraints, the effective lower bound, bank deleveraging and adverse self-reinforcing expectations. By repairing the monetary transmission mechanism, the ECB is able to transmit the policy decisions more evenly along the yield curve and across all Eurozone countries (Draghi, 2017). This chapter describes the four main unconventional monetary policy instruments used by the ECB.

3.1 Negative interest rates

The central bank acts as a bank for commercial banks. Commercials banks can borrow and deposit money at the central bank under certain conditions. The interest rates at which commercial banks borrow and deposit money are set by the central bank. The deposit facility rate is the interest rate that banks receive on funds that are deposited at the central bank. Before the crisis, the deposit facility was barely used. In response to the crisis, the deposits of commercial banks at the ECB increased sharply. Commercial banks were hardly lending money to each other and were looking for a safe place to deposit money. The deposit facility at the central bank is considered the safest place for banks to hold funds. As such, the deposit facility rate is the lower limit at which banks are willing to lend funds. The deposit facility rate is an important point of reference for financial markets.

Since June 2014 the ECB’s deposit facility rate has been negative. This means that commercial banks are actually being charged for depositing surplus liquidity at the ECB. When the central bank sets the nominal interest rate at a negative level, the nominal interest rate declines beneath the theoretical zero-lower bound. Negative nominal interest rates stimulate the economy by discouraging banks from depositing excess liquidity

at the central bank. The main reason for a negative Figure 3: EONIA, key ECB interest rates, and euro area bank lending rates on loans to NFCs

(12)

12 lower bound is the opportunity costs of holding money. While excess liquidity can be transformed into banknotes, thereby avoiding the negative interest rate, it is risky and cumbersome to pay for large transactions in cash, and storing a large amount of cash involves security and insurance costs. The effective lower bound is the point where a further reduction of the nominal interest rate ceases to provide aggregate stimulus to the economy. If negative interest rates become too low, deposits may be withdrawn and transferred into bank notes.

3.2 Credit easing

Credit easing involves an expansion focusing on the asset side of the balance sheet of a central bank. Credit easing is when a central bank lends money to commercial banks at longer maturities than normal open market operations against collateral. In credit easing the central bank does not hold the assets on its balance sheet, but increases its lending to banks. The yield curve is directly affected by the period for which indirect credit easing is conducted. Put differently, if the period of credit easing is longer, a larger part of the yield curve is affected. With credit easing, it is not the central bank but the banking system that endogenously determines the level of the monetary base. Banks decide how much extra liquidity they desire. If banks need extra liquidity it is available at the central bank in exchange for collateral.

In June 2014, the ECB announced the introduction of Targeted Longer-Term Refinancing Operations (TLTROs). TLTROs allow banks to borrow money from the ECB for a period up to four years for a fixed interest rate, and are designed to reduce the costs of funding. TLTROs are conditional; Banks can only borrow at deposit facility interest rates if the bank shows strong performance in loan origination to the real economy. TLTROs affect the balance sheet of bank in two complementary ways. First, TLTROs are relatively cheap sources of long-term funding, and by using TLTROs banks can replace more expensive sources of funding and extend the maturity of their liabilities. Second, TLTROs motivate banks to increase their asset portfolios, especially by lending to households and firms. TLTROs allow banks to borrow a multiple of their eligible lending, which provides liquidity to banks for asset expansion strategies.

(13)

13

3.3 Forward Guidance

Forward guidance is a verbal assurance from a central bank about its planned monetary policy measures. By providing forward guidance the central bank tries to influence the economic decisions of investors, firms and households by letting them know what to expect from monetary policy in the future. The central bank sends a clear message to the public what to expect from interest rates and the size of its balance sheet in the future. The central bank tells the public what it intends to do and what instruments it will use to achieve its policy goals. When investors, firms and households have a sense of where monetary policy is heading, this is assumed to increase confidence in their investing and spending decisions.

The term forward guidance is relatively new, but the idea that central banks can affect the expected interest rates by communicating their policy plans to the public is not. Various forms of forward guidance have been used since the mid-1990s (Brubakk et al., 2017). For example, the Reserve Bank of New Zealand (RBNZ) was one of the first central banks that provided a projection of future interest rates. The situation changed when the policy rates reached the lower bound. At this point, it was not possible for central banks to further ease the financial conditions. Central banks needed more tools and started to steer expectations about future interest rates. Hence, forward guidance became more important.

The main goal of forward guidance is to reduce uncertainty about adverse monetary policy surprises that could disrupt the financial markets and provoke significant asset price fluctuations. More specifically, forward guidance is a tool to affect the long-term interest rate. Forward guidance can have a Delphic and Odyssean element in it (Campbell et al., 2012). The Delphic element of forward guidance provides information about expected future policy, given the expected macroeconomic fundamentals. The central bank justifies its future monetary policy course by both the current state and the expected path of the economy. Most of the conditional forward guidance statements of the ECB have a Delphic element (Praet, 2013). The Odyssean element clarifies the monetary policy strategy of the central bank and is thus more binding in terms of future policy measures. The Odyssean element is a reflection of the statutes and the mandate of the central bank. In Greek mythology, Odysseus ordered his men to tie him to the mast of his ship so that he could resist temptation. The Odyssean element of forward guidance helps the central bank by anchoring itself to its strategy without losing sight of its general purpose as stated in its mandate.

One can distinguish between two dimensions of forward guidance. The first dimension is calendar-based forward guidance or state contingent forward guidance. Calendar-calendar-based forward guidance relates to monetary policy commitments during a certain period of time. State contingent forward guidance is a policy commitment conditional on economic conditions. The second dimension of forward guidance is

(14)

14 either qualitative or quantitative; quantitative forward guidance contains numerical values whereas quantitative forward guidance is expressed in more arbitrary and vague terms. For instance, the ECB may state that it will keep purchasing €60 billion of assets on a monthly basis until December 2017 (quantitative and calendar-based) or that it will retain the policy rate until labor market conditions improve sufficiently (qualitative and state- contingent)Borio and Zabai, 2016).

3.4 Quantitative easing

Quantitative easing is an unconventional monetary policy tool in which a central bank purchases assets, such as government securities, from banks and other financial institutions in order to bring more liquidity to the market. Quantitative easing has a direct impact on banks’ balance sheets and the availability of credit for households and firms. In times when the short-term interest rate is close to zero, the central bank can increase the money supply through quantitative easing to stimulate investments and consumption.

The current quantitative easing program of the ECB is the Asset Purchase Program (APP). The APP includes all asset purchase programs used by the ECB to address the potential risks of an excessive period of low inflation. The ECB started its APP in October 2014 with the introduction of the third Covered Bond Purchase Program (CBPP3), which was introduced to support the functioning of the monetary policy transmission mechanism in the Eurozone. One month later, in addition to CBPP3, the ECB introduced the

Asset-Backed Securities Purchase Program (ABSPP). The ABSPP provides new funding sources to banks. These new funding sources provide liquidity to banks, which stimulate them to issue new loans. This facilitates both the funding conditions and the transmission of monetary policy in the Eurozone.

In March 2015, the ECB expanded its APP by introducing the Public Sector Purchase Program (PSPP). Under the PSPP, the ECB started to purchase a monthly amount of €60 billion sovereign bonds and debt securities from governments and national agencies in the Eurozone. The securities and bonds purchased under the PSPP are subsequently made available for securities lending. This ensures collateral availability and liquidity in the market. The PSPP can lead to credit losses for the central bank and thus to risk reallocation in the euro area. For this reason, national central banks purchase the largest part of the bonds and securities from their home country. Only 20% of the purchased assets are subject to a regime of risk sharing (Deutsche Bundesbank, 2017).

In March 2016 the ECB announced a further expansion of its APP with the Corporate Sector Purchase Program (CSPP). With this announcement, the combined monthly amount purchased increased from €60 billion to €80 billion. The CSPP relates to investment-grade euro-denominated bonds issued by non-bank

(15)

15 corporations established in the euro area (DeMertzis and Wolff, 2016). This measure provides further monetary stimulus, specifically targeted at the corporate sector.

Finally, in April 2017 the ECB decided to downsized the APP from €80 billion a month to €60 billion a month in the light of reduced deflation risks. The APP is intended to continue until the end of December 2017, or beyond if necessary.

4. Transmission channels of unconventional monetary policy to the housing market

This chapter describes the transmission channels of the different unconventional monetary policy tools to the housing market. The effects of unconventional monetary policy tools are lagged, because it takes time for households and firms to adjust their economic behavior to changes in monetary policy. Figure 4 provides an overview of the main transmission channels of the different unconventional monetary policy tools to the housing market.

(16)

16

4.1 Negative interest rates

The reduction of nominal interest rates to negative levels is, to some extent, similar to the reduction of positive nominal interest rates. The lower nominal interest rate propagates through financial markets, influencing the interest rates in different parts of the financial sector. This has a direct effect on mortgages with short-term and variable interest rates. These lower mortgage interest rates make it cheaper to service a mortgage and easier to buy a new property. The negative interest rates should create a more favorable macroeconomic environment; hence, the financial situation of bank borrowers improves. The number of non-performing loans is negatively correlated with economic growth. When the financial situation of borrowers improves, it will be easier for them to get a mortgage.

The short-term interest rate is negatively correlated with risk taking. Low interest rates thus affect the decision-making process of investors. When investors face historically low interest rates on their savings accounts or investments, this leads to a search for yield. Investors might consider buying a property, as an investment, when the expected property price rise exceeds the applicable interest rate. This reaching for yield phenomenon is a consequence of low interest rates, and could increase the risks of different parts in the financial sector, including the housing market. Furthermore, negative interest rates on the deposit facility of the ECB create a situation in which banks prefer to lend money to firms and households rather than deposit their money at the ECB. This mechanism is also important in combination with other unconventional monetary policy tools. For example, when the ECB purchases assets from banks to provide extra liquidity to the market, banks can choose to lend this extra liquidity to the real economy or deposit the money at ECB. The negative interest rate makes it costly for banks to deposit money at the ECB and thus stimulates banks to lend it to the real economy. This makes it easier for households to get a mortgage.

But when nominal interest rates move into negative territory, there is an additional channel that affects the housing market. The introduction of negative interest rates has flattened the short and medium end of the yield curve. The medium end of the yield curve is a weighted average of all the possible future rate expectations. When the nominal interest rate reached the lower bound, markets expected that the nominal interest rate could only go up, not down. Because of this, the weighted average of future rate expectations was higher than the nominal interest rate. The introduction of negative interest rates removed this upward bias and flattened the short and medium end of the yield curve (Praet, 2016).

(17)

17

4.2 Credit easing

Credit easing affects the housing market’s two main transmission channels. First, credit easing relaxes the refinancing conditions for banks, which is expected to favor the borrowing conditions for firms and households. In the literature, this channel is called the direct pass-through channel (ECB Economic Bulletin, 2015). This channel is important in the case of TLTROs, which reduce the funding costs of banks and are designed to stimulate lending activities. Banks can only borrow at deposit facility interest rates if the bank shows strong performance in loan origination (Praet, 2016). This incentivizes banks to increase the supply of loans, which leads to more competition. More competition reduces the borrowing costs for the real economy. This should make it easier and cheaper for households to get a mortgage, which could, in turn, lead to an increase in housing prices.

Second, via the portfolio-rebalancing channel, which decreases the yields on a wide range of assets. TLTROs allow banks to borrow a multiple of their eligible lending. Banks can use this liquidity to buy assets such as private sector securities and government bonds. Furthermore, the TLTROs make it possible to replace maturing bank bonds with cheaper credit at the central bank. The holders of these maturing bank bonds are likely to rebalance their portfolios toward other assets until a new equilibrium is reached. In this new equilibrium, asset prices are expected to be higher and interest rates to be lower. The lower interest rates make it cheaper for households to get a mortgage and makes it more attractive to invest in housing market related assets.

4.3 Forward guidance

Forward guidance affects the long-term interest rate via the signaling channel. The ECB tries to influence market expectations about the future policy rate path. This has an impact on both future short-term and short-term interest rates. As mentioned previously, mortgages are typically contracts with long-term interest rates, particularly in the Netherlands. By reducing the long-long-term interest rates, the ECB makes it cheaper for households to get a mortgage.

Forward guidance reduces interest rate volatility and tries to prevent interest rate surprises that could disrupt the financial markets and provoke significant asset price fluctuations. This helps to avoid unrealistic and extreme outcomes, and creates a more stable macroeconomic environment. A stable macroeconomic environment increases the willingness of banks to provide loans, which makes it easier for households to get a mortgage and therewith fuels housing demand.

(18)

18

4.4 Quantitative easing

Quantitative easing affects the housing market via different transmission channels. First, the central bank purchases assets from banks and other financial institutions in order to provide liquidity to the market. When the received liquidity is not a perfect substitute for the asset sold, banks might rebalance their portfolios toward assets with longer maturities. The portfolio-rebalancing channel increases the demand for assets with longer maturities until a new equilibrium is reached. In this new equilibrium, assets prices are expected to be higher, implying lower long-term interest rates. The lower long-term interest rate should encourage savers to save less and to consume and invest more.

Second, when assets are purchased from counterparties outside the Eurozone, quantitative easing has a direct effect on the exchange rate. The exchange rate channel subsequently makes it is cheaper for investors from outside the Eurozone to buy a property in the Netherlands. For example, the EUR/USD exchange rate dropped from 1.4 at the peak in 2014 to 1.1 in May 2017. This implies a change in exchange rate of approximately 27%. This stimulates investors from outside the Eurozone to buy property in the Netherlands.

Third, quantitative easing affects the housing market through the risk-taking channel. Low interest rates can affect the decision-making process for savers and investors by decreasing the return on investments. To compensate for these low interest rates, investors will reach for yield and invest in riskier assets, which are characterized by higher returns. This reaching for yield phenomenon is a consequence of quantitative easing and can increase the risks in the housing market. Investors facing historically low interest rates on their savings accounts may consider buying property as an investment. Figure 5 shows a sharp rise in the total value of properties that are purchased without a mortgage.

(19)

19 Figure 5: Value of property purchases without mortgage in the Netherlands in billion euro (Kadaster, 2017)

Fourth, quantitative easing affects the housing market through the signaling channel. The sharp increase in the balance sheet of the central bank shows its commitment to its mandate. This has a direct effect on the market expectations on the short-term interest rates in the future. In addition, quantitative easing can increase the inflation expectations. The higher inflation expectations will reduce the real long-term interest rates, which could give an extra boost to consumption and investments.

0 1 2 3 4 5 6 7 8 9 20062007200820092010201120122013201420152016

Property purchases without mortgage

(20)

20 4.5 Combined effects of unconventional monetary policy on mortgage yields.

To achieve insight into the combined effects of different ECB monetary policy measures on long-term mortgage yields, it is useful to examine the monthly change. Figure 6 displays this monthly change in average long-term mortgage yield in the Netherlands.

Figure 6: Monthly change in average >10-year Dutch mortgage yield (DNB, 2017)

There are three points that stand out in Figure 6. First, in June 2014 the ECB announced a series of TLTROs, which should improve bank lending to the private sector in the euro area. Figure 6 shows the average long-term mortgage yield in the Netherlands decreased significantly after that announcement. The long-term mortgage yield decreased by about fifty basis points between the announcement of TLTROs in June 2014 and the start in September 2014. Second, in March 2015 the ECB expanded its APP by introducing the Public Sector Purchase Program (PSPP). Figure 6 shows that the start of PSPP had a significant effect on the average long-term mortgage yield in the Netherlands. In March 2015, the average long-term mortgage yield dropped by about eighteen basis points. Third, in March 2016 the ECB announced to expand its APP by introducing the Corporate Sector Purchase Program (CSPP). With this announcement, the combined monthly amount purchased increased from €60 billion to €80 billion. These measures reduced the long-term mortgage yield by about thirteen basis points.

It should be noted that these numbers are only an approximation of the effects of the unconventional monetary policy measures. As mentioned in the literature review, there is a broad consensus (Krishnamurthy and Vissing-Jorgensen, 2011; Swanson and Williams, 2014; Wright, 2012) that unconventional monetary policy reduces the medium- and longer-term interest rates.

-0,20 -0,15 -0,10 -0,05 0,00 0,05 0,10 0,15

jan-09 jan-10 jan-11 jan-12 jan-13 jan-14 jan-15 jan-16 jan-17

Monthly change in average >10 year Dutch mortgage yield

Start PSPP TLTROs

(21)

21

5. Effects of unconventional monetary policy on housing prices in the Netherlands

5.1 Methodology

This thesis examines the effects on the housing market of the unconventional monetary policy measures introduced to combat low inflation. The methodology is based on a time series regression model drawing on the relation between unconventional monetary policy and median transaction prices:

LogMTPRICEt = 𝛽0 + 𝛽1LogGDPt + 𝛽2MORTt-2 + 𝛽3CONFt + 𝛽4UNEMPt + 𝛽5LTVt + 𝛽6SCARt + 𝐷1CRISt + 𝜀t (1)

Where LogMTPRICEt is the logarithm of the median house transaction price in the Netherlands, GDPt represents the logarithm of GDP per capita, MORTt-2 is the lagged average long-term (>10 year) mortgage rate, CONFt is the consumer confidence, UNEMPt represents the unemployment rate, LTVt is the average loan-to-value ratio of first time buyers, SCARt is the housing scarcity defined as properties for sale divided by the number of transactions and CRISt is a crisis dummy from the period 2008 till 2013. 𝜀t is assumed to be i.i.d. over time, with finite variance 𝜎2.

The long-term mortgage rate is assumed to capture the stance of monetary policy, and is used to estimate the effect of the reinforcing unconventional monetary policy tools on the housing market. As mentioned previously, the effects of unconventional monetary policy tools are lagged, because it takes time for households and firms to adjust their economic behavior to changes in monetary policy. The length of the transmission lags from monetary policy to the housing market has been subject to discussion, as it is difficult to isolate lags with precision. Pétursson (2001) shows in his paper that it takes monetary policy roughly six months to affect domestic demand. In line with Pétursson, this thesis assumes that the average transmission lag from monetary policy takes six months to have an impact on the housing market.

At the same time monetary policy has an effect on other variables. Expansionary monetary policy increases the money supply. This increase in money supply leads to an increase in nominal output and is assumed to affects variables such as GDP per capita, consumer confidence and unemployment rate. In addition, the lower mortgage rates make it cheaper to buy a property, which increases the demand, the price and the number of transactions. An increase in the number of transactions leads, by definition, to a reduction of the housing scarcity in the Netherlands. Altogether, the predictors are assumed to be highly correlated.

(22)

22 To adress this problem of multicollinearity it is important to drop those predictors that are highly correlated. This results in a new time series regression model, drawing on the relation between unconventional monetary policy and median transaction prices:

LogMTPRICEt = 𝛽0 + 𝛽1 LogGDPt + 𝛽2MORTt-2 + 𝛽3LTVt + 𝜀t (2)

Where LogMTPRICEt is the logarithm of the median transaction price in the Netherlands, GDPt represents the logarithm of GDP per capita, MORTt-2 is the lagged average long-term (>10 year) mortgage rate, LTVt is the average loan-to-value ratio of first time buyers. 𝜀t is assumed to be i.i.d. over time, with finite variance 𝜎2.

But when does a rise in housing prices become too risky and unsustainable? This is difficult to predict, as there is no objective technique to measure a housing price bubble. Most housing market experts failed to recognize the latest housing price bubble. To facilitate an objective approximation, this thesis uses the (property) price-to-income ratio. This measure divides the median transaction price by the average GDP per capita. The price-to-income ratio of the Netherlands is higher than in most other developed countries.

The methodology is based on a time series regression model, drawing on the relation between unconventional monetary policy and price-to-income ratio.

PTIt = 𝛽0 + 𝛽1MORTt-2 + 𝛽2LTVt+ 𝛽3STOCKt + 𝜀t (3)

Where PTIt is the price-to-income ratio defined as the median transaction price divided by GDP per capita, MORTt-2 is the lagged average (>10 year) long-term mortgage rate, LTVt is the average loan-to-value ratio of first time buyers, STOCK is the total number of properties in the Netherlands. In contract to scarcity, housing stock is considered to be uncorrelated with monetary policy and thus more suitable for a regression with only three independent variables. 𝜀t is assumed to be i.i.d. over time, with finite variance 𝜎2.

The income ratio can provide a measure of housing affordability. A higher price-to-income ratio reduces the payment capacity. If the price-to-price-to-income increases to unusually high levels, properties may become too expensive as households have to spend an unsustainably large part of their income on housing.

(23)

23 A high property-price-to-income ratio does not, by itself, prove that there is a housing price bubble or that property prices are too high. The current low interest rate environment makes the high level of debt more affordable. However, a high price-to-income ratio makes the housing market more sensitive to unexpected shocks in the future.

5.2 Dataset

This study covers the period from 1999 to 2017 (18 years) using quarterly time series data. The dataset includes housing price data, mortgage rates and macroeconomic variables such as GDP per capita and the unemployment rate in the Netherlands. Table 1 provides an overview of the descriptive statistics of all variables of the dataset. The variables are: median transaction price, GDP per capita, mortgage rate, loan-to-value ratio, unemployment rate, consumer confidence and housing scarcity. This thesis uses data from the first quarter of 1999, as this was when the ECB’s monetary policy commenced. To standardize the different data sources into one dataset with similar time frequencies, all monthly data are converted into quarterly data.

Table 1: Descriptive Statistics

Variables OBS MEAN STD DEV MIN MAX

Dependent variable

Median transaction price 73 216.350 23.099 143.916 258.000

Independent variables GDP per capita 73 35.802 511 26.233 41.869 Mortgage rate 73 4,89 0,84 2,81 6,35 Loan-to-value 73 109 5,68 100 117 Unemployment rate 73 4,89 0,18 1,9 7,8 Consumer confidence 73 -10,58 2,15 -40 26 Housing Stock 73 7.075.820 356.019 6.522.362 7.686.178 Housing scarcity 73 10,74 0,84 3 30

5.2.1 Housing market data

The housing market data used for this thesis were provided by the Nederlandse Vereniging van Makelaars (NVM). The NVM is the largest association of real estate agents in the Netherlands, and

(24)

24 provides statistical information about the Dutch housing market. The dataset consists of quarterly data of dependent variables, such as the number of properties sold and the average transaction prices.

Developments in housing prices can be explained by supply and demand factors. Examples of supply factors are housing scarcity and housing stock (Case & Shiller, 2003). Housing scarcity data were also provided by the NVM and are defined as properties for sale divided by the number of transactions. In recent years housing scarcity has increased sharply in the Netherlands. In less than four years housing scarcity increased from 30 in the first quarter of 2013 to 6 in the last quarter of 2016. The problem with housing scarcity is that it is assumed to be highly correlated with monetary policy. The lower mortgage rates make it cheaper to buy a property, which increases the demand and the number of transactions. An increase in the number of transactions leads, by definition, to a reduction of the housing scarcity. Another supply factor that is included in the dataset is housing stock. Housing stock is defined as the total number of properties in the Netherlands; The data is provided by the Dutch statistical institution (CBS). Housing stock is chosen because it is a supply factor that is assumed to be uncorrelated with monetary policy. An increase in the housing stock is expected to reduce the average housing price in the Netherlands.

5.2.2 Long-term mortgage rate

The mortgage rate that is included in the dataset is the average long-term (>10 year) mortgage yield in the Netherlands. This long-term mortgage rate is used as a proxy to examine the effects of unconventional monetary policy on the housing market. As described in chapter four, the reinforcing unconventional monetary policy tools reduce the yields along the yield curve. Mortgages are typically medium- and long- term contracts. In

2016, 73% of the new mortgages in the Netherlands had a maturity longer than five years (DNB, 2016). Figure 7 shows that the different unconventional monetary policy tools contributed to a decline in the average long-term mortgage yield of new mortgages in the Netherlands from 4.71% in April 2014 to 2.82% in January 2017. 2 2,5 3 3,5 4 4,5 5 5,5 6 6,5 7 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017

Long-term mortgage rate

Figure 7: Average Long-term (>10 year) mortgage yield Netherlands (DNB, 2017)

(25)

25 5.2.3 Other Demand Variables

The dataset also includes a number of macroeconomic control variables that may affect housing prices. These economic indicators are provided by the Centraal Bureau voor de Statistiek (CBS). The CBS is a Dutch statistical institution that publishes statistical data about the Netherlands. The variables used in the dataset are GDP per capita, the unemployment rate and consumer confidence. First, GDP per capita is used as a measure for average income in the Netherlands. Income is often used as an indicator for a borrowers’ wealth and their ability to purchase property. Case and Shiller (2003) find a positive relation between income and housing prices covering fifty states in the United States, whereby higher income levels significantly increase housing prices. Second, unemployment rate is defined as the number of unemployed people as a percentage of the labor force in the Netherlands. Clearly when unemployment is rising, less people are able to buy a property. In addition, the fear of becoming unemployed may discourage employed people from buying a property. Third, consumer confidence is an economic indicator that measures the perception of consumers about the overall stance of the economy. Consumer confidence is a combination of the general economic climate and consumers’ willingness to purchase durable goods. The consumer confidence indicator is based on a survey of five questions about the last twelve months and the next twelve months to come. It can provide early signs of trend shifts in private consumption.

5.2.4 Macroprudential policies

Macroprudential policy focuses on reducing the procyclical behavior of the financial system, the interconnectedness of individual markets and financial institutions, and their common exposure to economics risks. Macroprudential policy may thus aim to lower high levels of private debt and strengthen the resilience of the financial system against unexpected shocks. After the crisis, macroprudential policies were introduced to tackle the risks associated with housing market booms more efficiently than fiscal and monetary policies. The two main macroprudential policy tools that were introduced in the Netherlands are reductions in the mortgage interest deduction and the loan-to-value ratio. Unfortunately, the time period is too short to include the mortgage interest deduction in the regression.

The loan-to-value ratio is a ratio of the maximum allowed mortgage against the value of the property. In the run-up to the financial crisis, the average loan-to-value ratio of first time buyers in the Netherlands increased from 100% in 1999 to 117% in 2009. In 2012 a statutory loan-to-value limit was set to 106%. Starting in 2012,the government gradually reduced the loan-to-value ratio by 1% per year to

(26)

26 reach 100% in 2018. This means that as of 2018, it will not be possible to get a mortgage that is higher than the value of the property.

5.3 Results

This section contains the results of econometric approaches described in the methodology. Table 2 presents the time series regression model, using the logarithm of the median transaction price as an indicator for the average housing price in the Netherlands.

Table 2: Time series regression model

Regressors I II III IV V

GDP per capita (log) 0,716*** 0,741***

(0,061) (0,064)

Mortgage rate -0,066*** -0,031***

(Lagged two quarters) (0,011) (0,008)

Loan-to-value 0,011*** 0,007*** 0,002 (0,001) Housing scarcity 0,004** -0,003*** 0,002 (0,001) Unemployment rate -0,017*** (0,005) Consumer confidence -0,001*** (0,000) Crisis dummy (2008-2013) -0,025 (0,016) Constant 4,776*** 12,619*** 11,034*** 12,234*** 4,042*** (0,641) (0,058) (0,209) (0,023) (0,648) Sample size 73 71 73 73 71 R2 0,65 0,31 0,33 0,07 0,93 Serial correlation 0 0 0 0 0

Mean VIF (Multicollinearity) 1 1 1 1 4,60

The dependent variable is the median transaction price (log). Note: Estimation method: OLS. Standard errors in parentheses: *<0.10; **p<0.05; ***p<0.01

(27)

27 The first four columns individually regress GDP per capita, mortgage rate, loan-to-value and housing scarcity on the median transaction price. In more detail, the regression in the first column shows that the positive relationship between the logarithm of GDP per capita and the logarithm of the median transaction price is 0.72 and significant at a level of 1%. In other words, an increase in GDP per capita by 1% increases the median transaction price by 0.72%. The second column shows that the lagged mortgage rate is, as expected, negatively correlated with the median transaction price and significant at the 1% level. The output indicates that a 100 basis points decrease in the mortgage rate, the proxy for unconventional monetary policy, leads to an increase in the median transaction price of 6.6% after six months. The loan-to-value ratio is individually positively correlated with the median transaction price. If the maximum loan-to-value ratio decreases by 1%, the median transaction price is expected to decrease by 1.1%. The regression in the fourth column shows a positive relationship between the scarcity and the logarithm of the median transaction price. However, this is not what one would expect, and this relation is considered to be biased and wrong. In the first four columns the explanatory power is relatively low because the regressions are done with only one variable; adding more control variables increases the relevance of the model.

The fifth column regresses all independent variables that are considered to be relevant on the median transaction price. The output shows that the R-squared is high with 0.93 and all the independent variables (except the crisis dummy) are significant at a level of 1%. After six months, a decrease in the mortgage rate by 100 basis points leads to an increase in the median transaction price of 3.1%. However, the negative relationship between consumer confidence and the median transaction price is not what one would expect, and this relation is considered to be biased and wrong. A possible reason for this might be the high degree of correlation between the variables, related to monetary policy. Expansionary monetary policy increases the money supply. This increase in money supply leads to an increase in nominal output and is assumed to affect variables such as GDP per capita, consumer confidence, and unemployment rate. The lower mortgage interest rates make it cheaper to buy property, which increases the demand and the number of transactions. An increase in the number of transactions leads, by definition, to a reduction of the housing scarcity in the Netherlands. Altogether, this model is expected to have a problem with multicollinearity. To get a sense of the correlation between the different variables, the Variance-Covariance Matrix is included in the appendix.

(28)

28 Table 3: Variance inflation factor (VIF) V is related to table 2 and I and II are related to table 4

Variable V I II Housing scarcity 7.73 2.81 GDP per capita 5.70 5.01 1.78 Crisis dummy (2008-2013) 5.56 Unemployment rate 5.14 Mortgage rate 4.07 3.22 1.85 Consumer confidence 2.08 Loan-to-value 1.94 1.38 1.17 Mean VIF 4.60 3.11 1.60

The variation inflation factor (VIF) is used to test for multicollinearity. As the name suggests, the VIF can test how much the variance is inflated because of collinearity. The variance, which is the square of the standard error, can be inflated upwards. As a rule of thumb, standard errors that are inflated more than twice may cause problems of multicollinearity, which means that VIFs that are larger than 4 can cause multicollinearity problems. Table 3 shows that the VIFs of housing scarcity, GDP per capita, crisis dummy, unemployment rate, mortgage rate and the mean VIF in the fifth column of Table 2 are all higher than 4. To correct for multicollinearity, highly correlated predictors, that are considered to be less important are removed from the model. The results are shown in Table 4.

(29)

29 Table 4: Time series regression model

Regressors I II

GDP per capita (log) 0,635*** 0,384***

(0,069) (0,047)

Mortgage rate -0,189** -0,044***

(Lagged two quarters) (0,008) (0,007)

Loan-to-value 0,009*** 0,010*** (0,001) (0,000) Housing scarcity -0,004*** (0,001) Unemployment rate Consumer confidence Crisis dummy (2008-2013) Constant 4,867*** 7,422*** (0,713) (0,493) Sample size 71 71 R2 0,90 0,87 Serial correlation 0 0

Mean VIF (Multicollinearity) 3,11 1,60

The dependent variable is the median transaction price (log). Note: Estimation method: OLS. Standard errors in parentheses: *<0.10; **p<0.05; ***p<0.01

The first column of Table 4 regresses the mortgage rate on the median transaction price, while GDP per capita, loan-to-value ratio and scarcity are included as control variables. It should be noted that the coefficient and standard error of the independent variable mortgage rate changes by including these control variables. By including the control variables, the mortgage is still negatively correlated with the median transaction price. The loan-to-value ratio is positively correlated with the median transaction price. This result is as expected; when the maximum allowed mortgage goes down one may expect that home buyers can get less money, resulting in lower median transaction prices. The regression shows a negative relationship between the scarcity and the median transaction price. But housing scarcity is

(30)

30 considered to be correlated with monetary policy because lower mortgage rates make it cheaper to buy a property, resulting in higher demand and more transactions. An increase in the number of transactions leads, by definition, to a reduction of the housing scarcity.

Column two is considered as the most relevant column of Table 4. First of all, the R-squared is high with 0.90. In addition, all the independent variables are significant at a level of 1%, and there are no problems with serial correlation and multicollinearity. The output of column four reveals that the logarithm of GDP per capita and the logarithm of the median transaction price are positively correlated. If the GDP per capita rises by 1%, the median transaction price is expected to rise by 0.39%. The long-term mortgage rate is negatively correlated with the median transaction price. A reduction in the long-term mortgage rate increases the demand for property through cheaper financing costs, resulting in an upward pressure on housing prices. The output suggests that a decrease of the long-term mortgage rate by 100 basis points leads to an increase in housing prices of 4.4% after six months lag. The loan-to-value ratio is positively and proportionally correlated with the median transaction price. In other words, if the maximum loan-to-value ratio decreases by 1%, the median transaction price is expected to decrease by the same percentage. A reduction in the loan-to-value ratio may thus be expected to have a positive effect on the financial stability in the Netherlands.

Since the ECB introduced unconventional monetary policy measures to combat low inflation, the long-term mortgage yield in the Netherlands has decreased by 189 basis points. As a result of the low mortgage rate, the median transaction prices in the Netherlands is estimated to have increased by approximately 8%. However, one may argue that a combination of the risk-taking channel and the exchange rate channel has also increased the value of domestic and international investments in the Dutch housing market. Due to this influence, the rise in housing prices caused by unconventional monetary policy is expected to be even higher than 8%.

(31)

31 But when does an increase in housing prices become too risky and unsustainable? This is difficult to predict, as there is no objective technique to measure a housing price bubble. Higher incomes are associated with higher property prices. To correct for income and to get a more objective approximation, the median transaction price is divided by the average GDP per capita. This price-to-income ratio can provide a measure of housing affordability. A higher price-to-income ratio reduces payment capacity. If the price-to-income increases significantly, housing may become too expensive, because households will have to spend an unsustainably large part of their income on housing.

Figure 8 shows the price-to-income ratio for the Netherlands between 1999 through 2017. The average transaction price was 6.5 times the average GDP per capita in 2008 when the housing bubble burst. During the crisis, the property price-to-income plummeted to a low of 5.2 in 2013. Since then, the price-to-income ratio has sharply increased to 6.1, because housing prices are currently rising three to four times faster than the average income.

Figure 8: Price-to-income ratio in the Netherlands 5,00 5,20 5,40 5,60 5,80 6,00 6,20 6,40 6,60 6,80 7,00 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017

Price-to-income ratio

(32)

32 Table 5: Time series regression model

Regressors I II

Mortgage rate -0,211***

(Lagged two quarters) (0,036)

Loan-to-value 0,040*** 0,026*** (0,008) (0,004) Housing stock ( x 10.000) -0,012*** (0,001) Scarcity Constant 1,672* 13,266*** (0,896) (0,890) Sample size 73 71 R2 0,25 0,84 Serial correlation 0 0

Mean VIF (Multicollinearity) 1 1,44

The dependent variable is Price-to-income ratio. Note: Estimation method: OLS. Standard errors in parentheses: *<0.10; **p<0.05; ***p<0.01

The first column of Table 5 suggests that the OLS regression result between the loan-to-value ratio and the price-to-income ratio is positive and significant at a significance level of 1%. This result is as expected; when the maximum allowed mortgage goes down one may expect that home buyers can get less money, resulting in a lower prices-to-come ratio.

The second column shows the effect of the long-term mortgage rate on the price-to-income ratio if we include the loan-to-value ratio and the housing stock in the Netherlands. In this OLS regression all independent variables have the expected sign, are significant at a level of 1% and there are no problems with serial correlation and multicollinearity. Moreover, with 0.84 the explanatory power of the OLS regression is quite large. The results of column two reveal that the long-term mortgage rate is, as expected, negatively correlated with the price-to-income ratio. A reduction in the long-term mortgage rate reduces the borrowing costs and makes higher levels of mortgage debt more affordable. If the long-term mortgage rate decreases by 100 basis points, the price-to-income ratio increases by 0.21. The loan-to-value ratio is positively correlated with price-to-income. Between 2012 and 2018 the loan-loan-to-value ratio is set to decrease with one percentage point per year. The OLS regression output indicates that a

(33)

33 reduction of one percentage point leads to a decrease in the price-to-income ratio of 0.03. Furthermore, the control variable housing stock has the expected coefficient and sign. Housing stock is assumed to be uncorrelated with monetary policy. If more properties are built and the housing stock increases, this leads to a reduction of the price-to-income ratio.

Since the ECB introduced unconventional monetary policy measures to combat low inflation, the long-term mortgage yield in the Netherlands decreased by 189 basis points. This lower long-term mortgage rate has had a significant impact on the price-to-income ratio. The model estimates that the lower long-term mortgage rate increased the to-income ratio by approximately 0.4. A higher price-to-income ratio does not, by itself, prove that there is a housing price bubble or that property prices are too high. The current low interest rate environment makes the high level of debt more affordable. But a high price-to-income ratio makes the housing market more sensitive to unexpected shocks in the future.

(34)

34

6. Potential risks to the housing market and financial stability in the future

The main vulnerabilities related to the housing market in the Netherlands are the high levels of household debt and the low mortgage collateralization (ESRB, 2016). This combination makes the housing market sensitive to cyclical fluctuations and amplifies the business cycle. If housing prices drop, a group of households will have negative equity because their mortgage exceeds the value of their home. This causes people to save more, and reduces consumption and investments. This could lead to a downward economic spiral, with lower housing prices and low levels of economic growth. Figure 9 shows the relation between economic growth and housing prices in the Netherlands.

Figure 9: Housing prices x1000 (vertical axis) and average economic growth in the Netherlands (CBS (2017) and NVM (2017))

Nicolais (2014) describes the conditions under which housing market bubbles have historically occurred. An improvement in prosperity or economic conditions increases the demand in the housing market. In the short run, the supply of housing is fixed, which means that the increased demand spurs higher housing prices. The increased demand and the higher prices stimulate the construction of new housing. But it takes several years before more housing is available. During this period the prices of existing properties continue to increase. This process can create a housing market environment with speculative expectations, in which prices and construction plans continue to rise. When housing prices stagnate, the supply of housing continues to increase because many properties are still under

€ 90 € 110 € 130 € 150 € 170 € 190 € 210 € 230 € 250 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017

Housing prices and economic growth the

Netherlands

(35)

35 construction. The housing market enters a period of ‘hyper supply’ in which the housing price bubble collapses. A collapsing housing price bubble almost always leads to a recession.

The situation described above is directly applicable to the Netherlands, and might be the biggest threat to the stability of the Dutch economy and housing market. The unconventional monetary policy tools of the ECB are reducing the main mortgage rates. The historically low mortgage rates are increasing the demand in the housing market. In the short run, the supply of housing is fixed. In fact, the Netherlands is currently already experiencing a housing shortage. The combination of low interest rates and housing shortages causes prices to rise. The increased demand and the higher prices stimulate the construction of new properties. But it takes time to construct larger numbers of new properties. In the meantime, the housing prices in the Netherlands continue to rise at a rate of 7.8% per annum(CBS, 2017).

The Planbureau voor de Leefomgeving (PBL) predicts that this housing shortage will last until 2025. After 2025 there will be a large housing market surplus due to the country’s demographics. After the Second World War, the Netherlands experienced a “baby boom”, i.e. a period with a very high birthrate. Figure 10 shows that the population of the Netherlands will grow by 6% from 17 million in 2017 to 18 million in 2040. This growth is entirely attributable to the longer life expectancy in the Netherlands. Figure 10 shows that the number of people older than 65 will grow by more than 50% from 3.1 million in 2017 to 4.8 million in 2040 (CBS, 2017).

Figure 10: Percentage change in population 2017-2040 (CBS, 2017)

The generation of 65 plus is the fastest growing group of homeowners. Nearly two out of three people between the age of 65 and 70 years old own their own home, as compared to only a third 25 years ago. Baby boomers are barely moving at the moment, which forms an obstruction in the housing market. The first baby boomers are now around 70. In the next fifteen years the supply of housing will explode,

-20 -10 0 10 20 30 40 50 60 total 0 - 20 20 - 65 > 65

Referenties

GERELATEERDE DOCUMENTEN

The model can be used to predict the impacts of scenarios (climate change, sea level rise, land use) and effects of policy actions on the occurrence of flooding events and

Als leerlingen door middel van observerend leren (X1) creatieve teksten schrijven (X2), waarbij de nadruk ligt op het toepassen van non-core-vocabulaire (X3), dan zullen zij

mensen hadden van de wereld Verklarende verbanden tussen oorzaken en gevolgen kunnen leggen en conclusies kunnen trekken Na het verband tussen de twee zaken te hebben

• Wireless sensor networks • Computer security • Transfer learning • Computer vision • Quality of experience • Smart grid Swarm Intelligence Static Complex Networks

Findings indicated that students were able to collectively advance the community’s discourse as they built on each others’ ideas, generated theories, questions and

The ritual dynamics of separation, transition and integration allow us to further scrutinise post-mortem relationships and, as I will argue, not simply to point to breaking

Surface roughness brings stress concentration point near the contact surface, possibly augmenting material anisotropic effects and are therefore detrimental for rolling bearing

Deze hebben te maken met: tweedeling van de studentenpopulatie, kosten, beperkte uitstralingseffect naar reguliere programma’s, motivatie van honoursstudenten ook