1 Introduction
Owning a house is a major life goal for many individuals and the housing market itself, in all its complexity, has long been an instrumental factor in performance of economies world-wide. As a consequence, housing markets around the world are constantly under scrutiny and their performance analysed. As seen from the recent financial crisis in 2007, house price fluctuations have a huge influence on economic performance. There has been increased scrutiny on the developments of asset prices following the subprime mortgage crisis in the US (Bjørnland and Jacobsen, 2010), not only in the US but throughout the world.
1.1 Housing in The United Kingdom
The housing market in the UK was not immune to this as shocks to both the mortgage and housing markets were incurred as a result of the financial crisis (Tse, Rodgers and Niklewski, 2014). The UK is a country that has experienced many housing shocks in the past and is currently experiencing crippling house prices and falling home ownership, from 72.5% in 2008 to 63.4% by the end of 2016. Areas within London and South East England are where housing is least affordable and with the introduction of government schemes such as ‘help to buy’ and ‘help to rent’, encouraging first time buyers, it is clear that homeownership and housing is high on the government agenda (S Wilcox, J Perry 2014).
Historically governments and central banks have employed fiscal and monetary policy, respectively, to control fluctuations in house prices, with the latter seen as the most effective. The way in which monetary policy can influence house prices is through what is known as the monetary transmission mechanism.
1.2 The Monetary Transmission Mechanism
Central banks can make use of monetary policy to influence asset prices and general economic performance. This process occurs through channels in what is known as the monetary transmission mechanism.
In the past, the UK government has made use of monetary policy to control house price fluctuations. Increasing interest rates, increases capital costs and mortgage rates causing the demand for housing to fall, resulting in falling house prices. This trend can be seen by looking at house price and interest rate fluctuations and the pattern they follow over time, it is clear they are inherently linked. There is an abundance of literature that analyses the role of housing in the monetary transmission mechanism and the causal link between the two.
However, following the financial crisis of 2007 Central Banks around the world have had to turn to unconventional monetary policy. In the UK interest rates were lowered to historically low levels becoming constrained by a zero-lower bound, this meant that unconventional monetary policy became the primary monetary policy tool and the Bank of England began to introduce a programme of quantitative easing (Signe Rosenberg 2018). With this in mind, there seems to be a scarce amount of literature that looks assesses the impact unconventional monetary policy has on house prices, especially in the UK.
1.3 Motivation for Research Question
Intrigued by the effect the 2007 Financial crash has had on economies world-wide, and the housing market in particular, in the first part of this paper I will look to contribute to current literature of the relationship between monetary policy and house prices in the UK by investigating:
‘An Analysis of the impact of unconventional monetary policy shocks on UK house prices following the Financial Crisis’
x
2 Literature Review
As previously discussed, the literature surrounding the relationship between monetary policy and housing markets, and the role of housing in the monetary transmission mechanism has been covered extensively, especially with respect to developed countries such as the US and UK.
The common consensus amongst the available literature is that contractionary monetary policy shocks have a negative impact on house prices, this is supported by what we know from economic theory. (Bjørnland, Hilde C., & Dag Henning Jacobsen, 2008) showed this for the top 8 OEDC countries in Europe, (Tse, C-B, Rodgers, T & Niklewski, J, 2014) look at the relationship in the UK and (Vargas-Silva, 2008) and (Jarocinski & Smets, 2008) look at the relationship within the U.S. Although these studies come to similar conclusions when assessing the link between monetary policy and house prices, very little look at the influence unconventional monetary policy has on house prices.
This literature review will analyse the current body of knowledge on the relationship between conventional monetary policy and house prices as well and unconventional monetary policy and house prices. Furthermore, because the available literature comes to a general consensus on the causal effect, the different methodologies for analysis used will also be discussed, as this is where the literature comes under some debate.
2.1 Economic Theory
The effects of house prices on an economy have long been studied by economists throughout history (Williams, 2011). Housing is a main source of wealth for most people, and therefore the price of houses can affect the wealth of the owner, influencing their consumption, which serves as a direct channel for economic performance as consumption is an instrumental factor in economic growth.
Economic theory tells us that there is a clear causal relationship between interest rates and house prices; an increase in said interest rates increases the cost of borrowing and capital, lowering the wealth of the homeowner causing a fall in demand for housing as well as restriction in supply, resulting in a fall in house prices (Williams, 2011). The opposite is true for a fall in interest rates, and there are many who attribute the unprecedented housing boom of the early 2000’s in the US and UK to consistently low base and mortgage interest rates. (Levin and Pryce 2009) drew on economic theory to argue that house price increases in the UK between 1996-2007 were driven by low real interest rates set by the central bank. Whereas, (Gilchrsit and Leahy 2002) dismiss economic theory arguing that monetary policy had little influence over asset prices during these years.
In much of the relevant literature the ‘home price puzzle’ (McCarthy and Peach 2002), (Vargas-Silva 2008) is used in argument when questioning the importance of economic theory in real life analysis. This puzzle refers to an increase in asset prices following a monetary policy shock, the opposite to what economic theory predicts. This ‘home price puzzle’ is evident in many empirical studies, which suggests there may be monetary policy channels other than interest rates that have a greater influence on the price of houses, or that economic theory in this case has little relevance in the complexity of the real world.
It is for this reason that economic theory is tested to try and uncover the extent to which it is holds, especially following an economic disaster like the crash of 2007. The most common form of analysis used within the literature, when assessing this theory are Vector Autoregressions.
2.2 VAR Analysis
Vector Autoregressions or VAR’s as they are more commonly know is a stochastic regression model that accounts for interdependencies between variables over multiple time series. This type of econometric analysis is common for analysing the effect of monetary policy on macroeconomic variables such a housing, particularly the use of structural Vector Autoregressions (Hilde C. Bjørnland, Dag Henning Jacobsen, 2008), this is consistent throughout the literature. They are popular within economic analysis because they are ‘flexible and simple models’ (Lütkepohl 2005) and have been employed more frequently in empirical studies following its advocation in the work of (Simms 1980) over standard simultaneous equation models.
Each variable within the model has its own equation and its own lagged values as well as the lagged value of the other variables. As such, the effect of monetary policy shocks on house prices can be assessed using this analysis and the main disparity when it comes to this analysis as we will see within the literature, is how the identification problem is overcome.
2.3 Conventional Monetary Policy
In order to maintain certain economic conditions within a country, a Central Bank will employ monetary policy measures such as raising or lowering interest rates to restrict or loosen money supply and restore economic conditions.
Prior to the Financial crisis the majority of the literature sought to assess the effect of monetary policy on macroeconomic variables in general and gain a better understanding of the monetary transmission mechanism (Uhlig 2005) (Bernanke and Blinder 1992) x(Christiano et al 1996). Early studies from (Bernanke and Blinder, 1992) looked at the effect of interest rate changes on macroeconomic variables, similar to (Uhlig, 2005) who assessed the effect of monetary policy on output. In both cases structural VAR analysis is employed to obtain results, and in both cases coming to very concrete conclusions in line with economic theory. These examples of literature set the foundations for studies that followed, with attention turning to housing following the crisis.
As previously mentioned, following the 2007 Financial crisis, asset price developments have become a primary focus, particularly amongst many central (Hilde C. Bjørnland, Dag Henning Jacobsen, 2008) and as such there has been an increase in the literature available around this period. (Vargas-Silva 2008) (Goodhart and Hoffman 2008) and (Jarocinski and Smets 2008) are some examples of the studies produced immediately succeeding the crisis. Each of these studies assess the impact of monetary policy shocks on the housing market but each make use of different VAR’s to obtain their results. It is clear that use of different identification schemes to overcome the identification problem in VAR analysis can have a significant impact on the results, and therefore it is important to review the different approaches to identification used within the literature.
(Vargas-Silva 2008) uses VAR analysis imposing sign restrictions on some of the variables, similar to that of (Uhilg 2005), whilst leaving the response of the house price variable open. This type of analysis is employed to “impose explicit theoretical restrictions in the estimation” to assigned priori assumptions made from economic theory to a number of variables. The results obtained show that contractionary monetary policy does have a negative impact on house prices in the US. Interestingly, when the results are compared to that of a standard Cholesky Decomposition, the response of the variable left unrestricted (housing activity) is much smaller than in the case of the Cholesky Decomposition. This study from (Vargas-Silva, 2008) indicates how the severity of a monetary policy shock on housing variables can differ depending on how a model is identified. This is crucial when it comes to the analysis of this study, as correct identification can determine the accuracy of the results.
Contrary to this, (Jarocinski and Smets 2008) conduct a similar analysis on the U.S but do not place any sign restrictions on the variables in their model. In this case they use a Bayesian VAR approach and obtain similar results to that of Vargas-Silva (2008), concluding that there is significant evidence that monetary policy shocks have an effect on housing starts and house prices in the US. However, their results suggest that the effects of a shock have a significantly longer lasting effect than those found by (Vargas-Silva, 2008) and that a negative impact on house prices can be shown without imposing sign restrictions on the house price variable. These two studies highlight the impact a difference in identification of a model can have on the results obtained. That is why it is important for this study that a sound identification procedure is followed as well as the inclusion of sensible variables in the model, so that accurate and convincing results can be obtained.
Another notable study from (Goodhart and Hoffman 2008) looks at the impact of monetary policy shocks on housing across 17 industrialized countries, hypothesising “that monetary shocks have stronger effects on house prices in times of house-price booms”. To achieve this, they opt for an augmented panel VAR analysis using dummy variables to represent house price booms across countries. The results of this study support those of (Jarocinski and Smets 2008) showing that the impact of monetary policy shocks on house prices are significant and in fact stronger during a housing boom. Given that the results of these two studies are in line with economic theory as well as each other, it would be tempting to assume that this negative relationship between monetary policy and house prices holds regardless of the country. However, as mentioned, there are empirical studies that expose a “house price puzzle”, (Bjønland and Jacobsen 2010) find this for Norway, Sweden and the UK and (Tan & Chen 2013) state that a puzzle is ‘particularly evident’ in China. This means the generalisation in line with economic theory cannot be supported as fact and there is need for further investigation. Furthermore, there is a clear gap in the literature, omitting the role of unconventional monetary policy, which has been widely employed around the world in the years following the crisis.
One study that goes some way in trying to assess whether the relationship between monetary policy and house prices has changed following the events of 2007 is (Tse, C-B, Rodgers, T & Niklewski, J 2014). This study recognises the impact the crisis may have had on the relationship between interest rates and house prices, taking a slightly different approach to see if the crisis has had an effect on the long-term relationship between house prices and interest rates. The conclusions drawn here is that there is evidence of a structural change in the relationship, however, this study still focuses on conventional monetary policy and interest rates in particular, omitting the influence unconventional policy measures may have in house price fluctuations following the crisis. As mentioned, this factor is extremely important because it became the main tool employed by the central bank to ease economic conditions in recent years and therefore needs to be assessed.
Having explored the literature surrounding the relationship in question, it seems there is a consensus amongst the majority of the literature, through different methods, that the causal relationship between monetary policy and house prices is negative. There have, in more recent years, been some studies that have investigated the effect of unconventional monetary policy on house prices, which is to be the main focus of our study, and these investigations have come to some interesting conclusions.
2.4 Unconventional Monetary Policy
Following periods of major economic crises as the one we saw in 2007, conventional economic policy can be rendered ineffective as Central banks have to constrain interest rates to a zero-lower bound, as took place in the UK. These low interest rates then can lead to liquidity scares as people decide to hoard their money rather than depositing in a bank.
Therefore, unconventional monetary policy, such as quantitative easing, is then necessary in order to inject money into and kick start an economy and encourage demand. Furthermore, these policy tools, conditional on their success in easing economic conditions, may become more prominent and used more frequently in the near future, as such, it is important and necessary to study how these unconventional shocks effect the housing market (Signe Rosenberg, 2018).
Again, the majority of the literature analysing unconventional monetary policy shocks look at its impact on macroeconomic indicators in general. (Peersman 2011) finds that a ‘25-basis point decline’ in the policy rate of the Euro area has a similar impact on economic activity as that of a ‘10 percent increase in the monetary base’, given that the policy rate is orthogonal to the monetary base. Similar findings have been achieved (Baumeister and Benati 2016) when looking at the US, Euro area, Japan and the UK, showing that a fall in the long-term bond yield spread, has a powerful effect on macroeconomic indicators such as output and inflation. These studies, however, do not analyse the impact on house prices, which adds to the motivation for research of this paper.
(Rahal 2016) postulates ‘that unconventional monetary policy can reduce real interest rates, reducing the user cost of housing, and therefore increasing the demand for (and price of) houses’. Adding to the study of (Goodhart and Hoffmann 2008) he uses Panel VAR with zero and sign restrictions to analyse 8 OEDC countries that have employed unconventional monetary policy, similar to that of (Baumeister and Benati 2016) and (Rosenberg 2018) who look at the US and Scandinavian countries, respectively. Conclusions drawn from these studies are that an unconventional monetary policy shock not only influences house prices, but also the supply of houses and the mortgage markets (Rahal 2016). The issue when focussing on a collection of countries is that conclusions drawn upon can often be general and not specific to certain countries. By focussing on a specific country such as the UK, where the dynamics of the housing market are very unique, more comprehensive conclusions can be draw from the results. Furthermore, many of these studies only focus on unconventional monetary policy and its effects rather than comparing it with conventional monetary policy, again, limiting the extent to which conclusions can be drawn.
However, (Rosenberg 2018) goes further in her study, building on (Rahal 2016) by comparing the effects of conventional and unconventional monetary policy on house prices to see if there is a significant difference. This analysis allows for a comparison between the two to obtain the relative impact of the two types of policy, and the conclusions drawn from this study is that the effect of unconventional monetary policy is similar to that of conventional but is more ‘sluggish’, however, there is a greater and more persistent peak in the impact of an unconventional policy shock on house prices than that of a conventional policy shock, which is interesting. These similar effects could, in fact, indicate that differences in policy do not have much effect on the overall impact of a monetary policy shock in general on house prices. Again, this study does not look at the UK where the structure and complexity of the housing market in very unique to that of Scandinavian countries.
The UK, like many other countries constrained its interest rates to a zero lower bound, following the crisis and introduced a programme of quantitative easing and as of August 2016, purchases of government bonds through this programme totalled at £465 billion (Bank of England), this is unconventional monetary policy that has never been seen before in the UK, but at the same time house price growth has been poor, despite the record low interest rates. This indicates that the effect of unconventional policy on house prices in the UK could be different to the other OECD countries analysed in the current literature, giving further motivation for this study.
In each study proxies are used to represent unconventional monetary policy shocks. Proxy variables are used in econometric analysis when the variable in question is unobservable or immeasurable. Unconventional monetary policy involves actions that influence the balance sheets of central banks (Peersman 2011), it is for this reason that proxies such as the 10-year bond yield spread and changes in the central banks total assets are used in the literature.
3 Methodology
3.1 Model
As outlined, VAR analysis is extremely prominent throughout the available literature. As such, I will be using this econometric method to conduct my analysis. As I will be looking into assessing the impact of unconventional monetary policy on housing, it is important to carefully consider which variables to include in my model and which proxies to use.
This study will look to employ a standard Structural Autoregressive Model with zero and sign restrictions as an identification scheme, similar to that of (Rosenberg 2018) and two lags will be used consistent with the majority of models within the literature. I will look to include six endogenous variables in my model including two proxy variables to represent conventional and unconventional monetary policy shocks. The variables to be included are:
• GDP per capita
• Real house price index
• Housing starts
• Short term interest rates set by the Bank of England (proxy)
• 10-year Government bond yield spread
GDP per capita is a common variable used throughout the literature in the study of monetary policy shocks , real house prices is the variable this study is interested in, housing starts representing the supply of housing, changes short term interest rate changes represent conventional monetary policy shocks and 10-year Government bond yield spread will be used as a proxy for unconventional monetary policy, similar to that of (Baumeister and Benati 2010).
3.2 Data
The data obtained for each variable is quarterly and seasonally adjusted, sourced from publicly available databases such as the ONS, nationwide and IMF. The data covers a period from Q1 1998 to Q4 2017, which is a time span that includes the build-up and aftermath of the crisis, which should be sufficient for this investigation.
4 Research Question and Conclusion
Having reviewed the extensive literature surrounding topic in question, there is a clear and important gap to be analysed, that is why this paper will focus on:
‘An Analysis of the impact of unconventional monetary policy shocks on UK house prices following the Financial Crisis’
This study will aim to establish the effect recently employed unconventional monetary policy measures in the UK have had on house prices and discover to what extent this effect differs from that of conventional monetary policy measures.
Having seen and experienced the detrimental effects the financial crisis had on the UK economy, and knowing that unconventional monetary policy has become an instrumental tool in influencing economic conditions, I feel it is vital to establish the relationship between unconventional monetary policy and house prices and to see if it differs from that of conventional policy. By employing SVAR methodology similar to that of Rosenberg and Rahal, this paper will uncover the extent to which house prices are influenced by the likes of quantitative easing to give an indication of what to expect from the housing market if unconventional monetary policy is continued into the future.