What is the effect of quantitative easing policy on US REITs returns?

(1)

What is the Effect of Quantitative Easing Policy on

US REITs Returns?

University of Amsterdam Faculty Economics and Business

BSc Economics and Business Specialization Finance and Organization

Author: Martijn Pasman

Student number: 10886737

Thesis supervisor: dhr. dr. M. Dröes

(2)

Abstract

This thesis examines how quantitative easing (QE) policy in the US impacted the REIT returns over the period of 2003-2017. Regression results show a significant negative relationship between the asset purchases made during the QE and REIT returns, controlling for movements in housing prices, GDP and real interest rate. This study contributes to the literature by showing that QE affects REITs through the portfolio rebalancing channel but not through the direct effect of lowering interest rate. It also shows that the effect differs between equity and mortgage REITs. Equity REITs are affected by QE, whereas MREITs are not as they are not seen as close substitutes for rebalancing by investors. The results of this study imply that investors can anticipate REIT return movements in response to QE policies and policy-makers need to take into account possible side-effects of QE on REIT returns.

Statement of Originality

This document is written by student Martijn Pasman who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used to create it. The Faculty of Economics and Business is responsible solely for the supervision of the completion of the work, not for the content.

(3)

1. Introduction

President of the United States, Dwight D. Eisenhower, created Real Estate Investment Trusts (REITs) in the Cigar Excise Tax Extension of 1960. Nowadays, around 80 million US citizens own REITs through their retirement savings or other investments funds. All REITs combined make up approximately $3 trillion in assets, which resulted into a $60 billion dividend income in year 2016 alone. REITs are seen as unique instruments, due to their diversification benefits for smaller investments. Since their creation they have grown rapidly and 32 of REITs are now members of the S&P500 index.1_{More than 300.000 properties in the US are owned by REITs,}

having a significant impact on the US economy. Because of the size and popularity of the REITs, it is important to know what kind of impact financial market announcements, policies and monetary shocks can have on REITs returns.

Previous studies investigated how REITs respond to interest rate and inflation rate changes (Mueller and Pauley, 1995) and further studies examined how conventional monetary policy instruments can impact REITs returns under various economic conditions (Chou and Chen, 2014). Other studies that examined unconventional monetary policies, such as quantitative easing2, have focused mainly on the channels of monetary policy transmission (Joyce et al. 2011), without investigating the possible side-effects on REITs returns. To the best of my knowledge, no previous study investigated how quantitative easing (QE) affects the REITs returns over the period of QE1, QE2 and QE3, which took place in the US after the aftermath of the financial crisis. Filling this literature gap, this study aims to answer the following research question:

- What is the effect of Quantitative Easing policy on US REITs returns?

There are two main types of REITs: equity REITs (EREITs) which invest in commercial real estate and mortgage REITs (MREITs) which invest in mortgages. To this end, study first examines total REITs returns and then splits the analysis into EREITs and MREITs returns separately. To investigate whether there is a difference between how QE1, QE2 and QE3 affect REITs returns, this study further includes the dummy variables for each QE period.

1_{Retrieved from https://www.reit.com/}

2_{Quantitative Easing refers to a monetary policy whereas central banks buy government securities and market securities with the aim to}

(5)

These specifications lead to the following sub-questions:

- Does the effect of QE policy on US REITs return differ for EREITs and MREITs?

- Does the effect of QE policy on US REITs returns differ for QE1, QE2 and QE3?

The study conducts the analysis by using monthly data on US REITs returns, monetary policy variables and US economy over the period 2003 - 2017. Accordingly, the study uses an OLS regression where REITs returns is the main dependent variable and QE size is the main independent variable. Real interest rate, house prices, GDP and stock prices are the control variables. The QE size is measured accordingly by the monthly changes in FED’s total assets, and additionally by FED’s total assets in proportion to the US debt and with a dummy for the increase in assets higher than 1%. Real interest rate is measured by using market yields on U.S. Treasury securities at a 10-year constant maturity corrected by inflation.

The results show a negative effect of QE policy on US REITs returns, significant to 1% level. In line with previous literature, interest rate and inflation do not have an impact on REITs’ returns, suggesting that QE is not affecting the US REITs through lowering the interest rate. On the contrary, findings show that QE, as an unconventional monetary policy, affects REITs returns on its own. As returns on the assets purchased by FED are decreasing with QE, investors are looking for substitutes, known as ‘portfolio rebalancing’ effect. Findings suggest that REITs can be seen as substitutes by US investors and as they rebalance their portfolios. Accordingly, the demand and price of REITs rises, and REITs returns fall. The results are equal for QE1, QE2 and QE3 as there is no significant difference between them.

The contribution of this study is two-fold. Firstly, it is the first study to link the quantitative easing policy to REITs’ returns, examining how unconventional policy such as QE can have an effect on the asset price of REITs as a financial instrument. Secondly, it explains how QE affects REITs through the portfolio rebalancing channel which was introduced by the previous research. The implications are that investors should anticipate the movements in REIT returns during the asset purchase periods. Policy-makers should be aware of the possible side-effects of QE on REITs’ returns, as decrease in returns may lead to lower spending behavior possibly slowing down the policy transmission to the economy.

(6)

The remainder of this thesis is organized as follows. Chapter 2 provides the literature review on REITs returns and effects of monetary policy, including the QE. Chapter 3 gives an overview of the data sample. Chapter 4 defines the methodology. Chapter 5 presents the results and their interpretation and Chapter 6 concludes.

2. Literature review

2.1 Definition and types of REITs

Real Estate Investment Trusts (REITs) were created in the 1960s in the US with two aims. Smaller private investors were given the opportunity to invest in US commercial real estate to diversify their portfolios, creating opportunities for real estate developers to attract new investors. There are very strict rules to be qualified for a REIT: 90% of a REIT’s taxable income must be distributed to shareholders in the form of dividends. Additionally, there must be at least 100 shareholders after the first year, the shares must be fully transferable and at least 75% of the total assets must be invested either in real estate assets or cash. Most REITs are listed on major stock exchanges and individual investors can purchase them through securities dealers (Semer, 2009).

Since their creation, REITs have been extensively studied in literature. Firstly, by targeting the real estate market, it was investigated if REITs need to be added to portfolios for diversification. Kuhle et al (2015) found equity REITs to provide diversification benefits for smaller investors which gives an opportunity to invest in commercial real estate which would otherwise be out of their league. Additionally, the relationship between the stock market and real estate market has been studied. Okunev et al. (1997) finds a unidirectional relationship between them. Conducting linear and nonlinear causality tests he also showed that nonlinear test results in an even stronger unidirectional relationship.

REITs can be divided into 2 groups; equity REITs and mortgage REITs. Equity REITs are most common and invest in direct commercial real estate. There have been numerous studies on the performance, the risks and the correlation on other assets of equity REITs. This type of REIT is often specialized in a certain area of commercial real estate like hotels, shopping malls or office buildings. The business model of the equity REITs is mainly focused on direct return through the rental income. Feng et al. (2011) found the majority of equity REITs able to pay

(7)

out large amounts of cash for the period 1993-2009. They note the adoption of operating partnership for a majority of the equity REITs. Chan et al. (2005) found the performance of equity REITs less risky than the stock performance in general and conclude that equity REITs returns are sensitive to changing interest rates.

Mortgage REITs are dealing with investing and owning of property mortgages. They loan money for mortgages or other types of real estate loans and buy existing mortgages. Their business model is focused on the net interest margin, the spread between the interest they earn and the cost for funding these loans. Mortgage REITs do not typically own real estate. Research by Liang and Webb (1995) on the pricing of interest-rate risk for MREITs suggests that MREITs should have a risk premium due non-diversifiable nature of interest-rate risk. They also found MREITs, due to their nature of their assets, to be interest-rate sensitive.

There is a debate on whether only one REIT class should be held in a portfolio or both equity and mortgage REITs should be included in the portfolio for diversification. Lee and Chiang (2004) compared equity REITs (EREITs) and mortgage REITs (MREITs) and concluded that both respond to economic and financial forces similarly, resulting in a possible substitutability. They also find that stock market has a strong impact on both REITs’ pricing. Later study by Zhang et al. (2016) investigated the substitutability of equity and mortgage REITs in real estate portfolios and found the driving economic factors of both REITs to be different, which led to a rejection of the substitutability hypothesis. According to Peterson and Hsieh (1997), MREITs shares underperform to EREITs by an average of 6,8% per year for the period 1976-1992 supporting the argument that they behave differently.

2.2 Quantitative Easing

The term quantitative easing was introduced by the Bank of Japan during their attempt to tackle the deflation which Japan suffered from in 2001. It is defined as ‘the intention to increase the monetary base by increasing the supply of reserves’. (Woodford, 2016). Joyce et al (2011) document that the FED also used QE policy by buying large quantities of agency debt and mortgage backed securities. The QE policy tends to be effective in boosting the economy by lowering long-term interest rate. However, the total effect and the duration of the effect, as well as channels through which QE works, can be diverse and sometimes uncertain.

(8)

In the US, the QE policy took place in the period of 2008 – 2011. According to Fratzscher and Straub (2016), it be defined by 3 periods: QE1, QE2 and QE3. Starting in November 2008, the FED introduced Quantitative Easing for the first time, as a policy tool named large-scale asset purchase programs (LSAP). LSAP aimed to help against the economic downturn caused by the global financial crisis. In November 2008, the LSAP announced the purchases of Mortgage Backed Securities (MSB) worth more than 1,000$ billion and Government Sponsored Enterprise (GSE) debt worth $200 billion. The end of the first round of purchases is 2010, meaning that the QE1 period can be defined from November 2008 to early 2010. Krishnamurthy and Vissing Jorgensen (2011) document in an event study, that QE focused mainly on Treasury rates as a target and those rates decreased mainly due safety effect.

The second period of QE, labelled as QE2, aimed to fulfil the goal of stimulating economic activity. Reaching this goal was attempted by lowering the long-term rates to support investment and by raising asset prices to stimulate demand. The Federal Open Market Committee (FOMC) re-invested MBS and principal payments of agency securities into long-term Treasury securities. In November 2010, purchases were extended, adding 600$ billion to the FED balance sheet, which made Treasury purchases the largest tool in QE policy. The QE2 period can be defined from November 2010 to mid 2011 (Fratzscher and Straub, 2016).

Krishnamurthy and Vissing Jorgensen (2011) discuss the similarities and differences for QE1 and QE2 in their paper. Evidence for both policies in inflation and signaling channel is found. Additionally, the demand for safe long-term assets is detected. Evidence for two channels is only found for QE2, which worked through the MBS prepayment channel and the corporate bond default risk channel.

The third and last period of QE, labelled as QE3, started in September 2012 announcing the purchase of $40 billions of MBS and $45 billion of Treasury bonds montly. Those purchases were slowed down between 2013 and 2014, meaning the QE3 period can be defined from September 2012 until late 2014 (Fratzscher and Straub, 2016)

(9)

2.3 Effect of interest rates and inflation on REITs’ returns

Numerous studies have been conducted on the effect of a changing interest rate and inflation on the returns of REITs. Early research from Mueller and Pauley (1995) compared monthly changes in REITs to monthly changes in interest rates (3-months Treasury bills, 10-year bonds and long-term government bonds) and changes in stock indexes (S&P 500 Price Index, the S&P 40 Utilities Index, the NAREIT Price Index and the Wilshire Real Estate Index) over the period 1972-1993. Their study shows that movements in REIT returns usually have a low correlation with the interest rates and a much higher correlation with stock price movements.

Chen and Tzang (1998) investigated the difference between long- and short-term interest rate changes for both equity and mortgage REITs. For the period 1980-1985 REITS was found to have a negative sensitivity to interest rates, for short and for long term interest rate changes. Most recent research from Shulman (2017) finds degree of interest rate sensitivity varies over time and actually switches direction. Despite recently high sensitivity, Shulman (2017) concludes that there is no long-run predictive rule for EREITs’ response to interest rates.

For the response to inflation, Simpson et al (2007) find an asymmetric response of equity REITs during the period 1981-2002. By looking at inflation and splitting it in positive and negative changes, they find mixed results as REITs returns rise for the both states of inflation. Lu and So (2001) do not find a causal relationship between REITs returns and inflation when further control variables are added to the regression. Payne (2003) further documents an insignificant effect of unexpected changes in the inflation rate on equity and mortgage REITs, leading to the conclusion that REITs returns could be affected by inflation only through the real activities or changes in federal rates, but not on its own.

2.4. Effect of monetary policy on REITs’ returns

Early research from Buetow & Johnson (2001), before the QE policy in the U.S., pointed out the difference in optimal portfolio composition among different monetary policies, claiming that investors need to anticipate and restructure their portfolios in response to FED’s actions. In their study, evidence is found for the significant effect of the FED’s monetary policy on real estate returns. The results suggested to fully invest in EREITs or T-BILLS depending on the

(10)

monetary policy at that moment. According to Buetow & Johnson (2001), real estate investors should monitor the monetary policy in the U.S. closely.

More recently, Chou and Chen (2014) investigated whether asymmetric effects on REITs return due to monetary policy in the US occur. In their research, boom and busts periods are defined and evidence for larger effects of policy actions on boom markets rather than on bust market are documented. Chen et al. (2010) divides time periods into bull, bear and volatile ones and investigates the effect of monetary policy on the returns of equity REITs. In the case of a bull market, REITs return are falling when changes in monetary policy occur. For changes in policy during bear and volatile periods, no sensitivity for EREITs return is found.

Various researchers split the effect of monetary policies into an expected an unexpected part. Brendin et al. (2007) looks solely at the effect of unanticipated changes in monetary policy on EREITs and finds strong responses in returns due to unexpected changes. In contrast, Chang (2011) looks into the effect on REITs for both expected and unexpected changes and finds that the monetary policy has a negative effect on REITs performance, which is stronger when there is a bust market rather than a boom market. Finally, Anderson (2012) compares the effect of monetary policy for both REITs and the S&P 500 and found REITs to be twice as sensitive for monetary shocks compared to the S&P 500 in periods of high-variance.

3. Data and descriptive statistics

3.1 Dependent variable

This section will focus on describing the data. Firstly, when looking at performance of REITs, their returns are used as a performance measurement. The data is on REITs returns for REITs in the US, both data for the returns of EREITs and MREITs together answer the main research question. Further data splits into returns for equity and mortgage REITs to answer the research sub-questions. The data is obtained from FSTE Nareit U.S. Real Estate Index Series on a monthly basis and consist of 180 data points from January 2003 until December 2017. The variables are displayed as total REITs return, EREIT return and MREIT return.

(11)

3.2 Independent variable

For measuring Quantitative Easing, different variables are used. Due to multicollinearity reasons between those variables, regression is done to see which of the 3 variables qualifies best for measuring Quantitative Easing. The first variable is based on the study of Lo Duca et al (2016), measuring QE as the proportion of FED’s total assets to the outstanding US debt. Both, the monthly data for FED’s total assets and the outstanding US debt are obtained from the Federal Reserve Bank of St. Louis.3_{This will be referred to as the Assets/Debt variable in}

the regressions. The second variable measuring QE is the monthly change in the FED’s total assets. As mentioned above, the monthly data of FED’s total assets is obtained from the Federal Reserve Bank of St. Louis and the changes with respect to the previous month are calculated. This will be referred to as the Change Assets variable in the regression. The third variable measuring QE is a dummy variable being 1 when the change in assets is larger than 1%.

3.3 Control variables

Housing, S&P 500, the long-term real interest rate and GDP are included as control variables. The control variable S&P 500 is the monthly change in returns of the S&P 500.4 There has been extensive research on the correlation of Stocks and REITS returns. Westerheide (2006) investigated the difference in returns for real estate securities and stocks. Evidence for the US found that real estate securities have been outperforming stock markets for the period 1990-2004, especially between 2001-2004. I expect the control variable S&P 500 to be positively related to REITs returns, since generally rising S&P 500 returns indicate economic growth and inflation which will led to higher rents leading to higher REITs returns.

The real interest rate variable is the market yield on U.S. Treasury securities at 10-year constant maturity and the inflation rate obtained from the Federal Reserve Databank. To transform the market yield from a 10-year constant maturity to a real interest rate, the inflation is subtracted from it. Taking the long-term real interest rate is more appropriate than the short-term, since REITs returns are assumed to be long-term investments. The relationship between REITs

3_{Data from}_{https://fred.stlouisfed.org/series/MVGFD027MNFRBDAL}_;_{https://fred.stlouisfed.org/series/WALCL} 4_{Data from:}_{https://www.investing.com/indices/us-spx-500-historical-data}

(12)

returns and S&P 500 is much more close than the relationship between REITs returns and the real interest rate, as illustrated in the Graph 1 and Graph 2.

The control variable Housing is the monthly change of house prices in the US. The data for the GDP is obtained from the database of the Federal Reserve Bank of St. Louis.5

Graph 1 - REITs returns and S&P 500 over time

Graph 2 - REITs returns and Real Interest rate over time

5_{Only quarterly available, 3month- average is taken for calculating monthly}

-40.00 -30.00 -20.00 -10.00 0.00 10.00 20.00 30.00

40.00

REITs and S&P 500

Total REIT return S&P 500

-40.00 -30.00 -20.00 -10.00 0.00 10.00 20.00 30.00

40.00

REITs and Real Interest

(13)

Table 1 - Descriptive Statistics (2003-2017)

N mean sd min max

EREITs return 180 1.082 6.457 -31.66 31.01

MREITs return 180 .513 5.565 -24.11 11.19

Total REITs return 180 1.038 6.113 -30.23 27.98

Assets Federal Reserve 180 2,432,563 1,483,061 721,326 4,507,150 % Change Assets 180 1.135 5.806 -8.499 71.14 US Federal Debt 180 13,730 4,803 6,941 20,969 Asset-to-Debt Ratio 180 159.8 53.78 87.82 240.44 % Change Housing Prices 180 .240 .565 -1.757 1.093 GDP Growth 180 .104 .222 -.489 .657

Real Interest Rate 180 1.307 1.068 -.74 3.2

S&P 500 180 .616 3.801 -16.94 10.77

Interest Rate 180 3.195 1.048 1.5 5.11

Inflation 180 1.887 .439 .6 2.9

Table 1 displays the descriptive statistics of the used variables. The mean of 1.082% of the EREIT return is higher than the 0.513% MREIT return. The volatility, measured by the standard deviation, is 6.457 for EREITs returns and 5.565 for MREITs return, indicating the higher returns and a higher volatility for EREITs returns. The total REITs returns are more similar to the EREITs returns statistics than to those of the MREITs. The minimal results of -31.66% and -24.11% for equity and mortgage REITs return have both been occurring during 2008, the year of the global financial crisis and the timeframe where QE1 was introduced. Furthermore, Table 1 shows that EREITs return on average outperformed the S&P500 over the 2003-2017 period, which expands the previous findings of Westerheide’s (2006).

However, the volatility of the EREITs return is higher compared to stocks, with larger minimum and maximum values, meaning EREITs have higher risks but also higher returns. The largest outlier of 71.14% change in assets was recorded in November 2008, the start of the QE1 policy in the US. Graph 3 shows the relationship of the change in assets over time and clearly indicates the periods of Quantitative Easing, which qualify change in assets as a suitable variable for measuring QE policy.

In 2008 a large increase in the assets of Federal Reserve is seen, which is be explained by the purchase of assets following the announcement of the first period of quantitative easing. In

(14)

March 2009, the FED extended the first period by buying more assets explaining the increase. In November 2010 the second period of QE started, causing an increase in federal reserves of roughly 20%. The third period of QE is also clearly shown in Graph 3.

Graph 3 - Total Assets of Federal Reserve

0 500000 1000000 1500000 2000000 2500000 3000000 3500000 4000000 4500000 5000000 2005 -01 2005 -06 2005 -11 2006 -04 2006 -09 2007 -02 2007 -07 2007 -12 2008 -05 2008 -10 2009 -03 2009 -08 2010 -01 2010 -06 2010 -11 2011 -04 2011 -09 2012 -02 2012 -07 2012 -12 2013 -05 2013 -10 2014 -03 2014 -08 2015 -01 2015 -06 2015 -11 2016 -04 2016 -09 2017 -02 2017 -07 2017 -12

(15)

Table 2 - Correlations of key variables

EREIT MREIT REITs Change

Assets Change Housing P S&P 500 EREIT 1 MREIT 0.494*** 1 REITs 0.997*** 0.548*** 1 Change Assets -0.332*** _-0.101 _-0.336*** ₁ Change Housing P 0.0566 0.0740 0.0639 -0.283*** ₁ S&P 500 0.713*** _0.468*** _0.728*** _-0.298*** _0.118 ₁ *_{p < 0.05,}**_{p < 0.01,}***_{p < 0.001}

Finally, Table 2 shows the highest correlation between EREITs and total REITs, indicating a larger effect of EREITs explaining the similarity between EREITs return and total REITs return. In accordance to the research of Mueller and Pauley (1995), a high correlation with stock prices movement is found. A high correlation defines a possible relevant relationship between variables and it is insightful to know if the relationships shown in Table 2 holds when the control variables are added to the regression.

4. Methodology

4.1 Hypothesis

During times of QE policy, central banks buy government securities to lower the interest rate. The FED is not buying assets like REITs, since they search for risk-free assets such as sovereign bonds. As shown in Graph 4, the effect of QE can occur through different channels. The channels work in different ways. The policy signaling and the confidence channels are affected only by announcement of the policy and not during the policy implementation. This is in contrast to the portfolio rebalancing channel, which is affected by the actual purchases. Both

(16)

signaling and confidence channels cause interest rate changes, but from the literature, it is known that changes in interest rates have no effect on REITs. So, I expect the actual purchases of QE to affect REIT returns through the rebalancing portfolio channel.

Accordingly, I expect investors to rebalance their portfolios to others investments, including investments in REITs, especially equity REITs, because they have more similarities to bonds. Meaning I expect the difference in results for EREITs and MREITs. Overall, I expect REITs to be affected through the portfolio rebalancing channel of Quantitative Easing. The increase of investments in REITs increases their prices and by economic theory we can expect their returns to decrease, leading to my main research hypothesis:

I expect the REITs return to decrease due to the effect of Quantitative Easing policy (through the rebalancing channel).

Graph 4- Transmittion channels of Quantitative Easing

(17)

4.2 Model and regression

This paper investigates the effect of QE policy on REITs returns. The base model includes REITs return as a dependent variable regressed on the following QE measures: the change in FED’s assets during QE, asset to debt ratio, and a dummy variable which is generated to account for periods where the change in assets is larger than 1%. To account for possible internal validity problems like omitted variable bias or simultaneous causality the control variables S&P500, Housing, Real interest and GDP are added. The full model, including the control variables, will look as follows:

REITsreturnt=ßo+ß1changeassetst+ß2asset/debtt+ß3dummyassets(change>1%)t+

ß2SP500t+ß3Housingt+ß4GDPt+ß5Realinterestt+ t (1)

FED’s assets change is accounted for the QE variable because it gives the best measure and therefore is included in the next models.

To answer the sub question, whether there is a different effect for EREITs and MREITs, 2 regressions are done. One including EREITs and MREITs separately as depending variables, resulting in the follow models:

EREITsreturnt=ßo+ß1changeassetst+ß2SP500t+ß3Housingt+ß4GDPt+ß5Realinterestt+t

(2) MREITsreturnt=ßo+ß1changeassetst+ß2SP500t+ß3Housingt+ß4GDPt+ß5Realinterestt+t

(3)

The last model indicates the periods QE1, QE2 and QE3 as dummy variables to answer whether there is a different effect for the different periods. The model looks as follows:

REITsreturnt=ßo+ß1Realinterestt+21SP500t+ß3Housingt+ß4GDPt+ß5Dummy_QE1t+ß6Du

mmy_QE2t+ ß7Dummy_QE3t+t

(18)

5. Results

Table 3 shows the main results. Accordingly, it shows the regressions for equation (1) for different QE measures and how each of them impacts REITs returns (columns 1 – 3). Column 4 present equation (4) to examine potential differences in the different QE periods.

Table 3 - Effect of QE on REITs returns and testing for difference in QE periods

Dependent variable: REITs return

Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

(1) (2) (3) (4)

VARIABLES Change Assets Asset Debt

Ratio

Dummy Measure

Dummies QE1 - QE3 Change Assets (QE measure) -0.154***

(0.0580)

Assets/Debt (QE measure) -0.00579

(0.00836)

Dummy_QE (QE measure) 0.125

(0.714)

Real interest 0.245 0.0444 0.248 0.183

(0.296) (0.423) (0.302) (0.317)

S&P 500 1.113*** 1.181*** 1.175*** 1.187***

(0.0857) (0.0841) (0.0844) (0.0848)

Change Housing Price -0.568 -0.163 -0.151 -0.0920

(0.581) (0.571) (0.580) (0.698) GDP Growth -0.723 -0.713 -0.669 -0.646 (1.408) (1.435) (1.436) (1.446) QE1 Dummy 0.206 (1.109) QE2 Dummy -0.688 (1.569) QE3 Dummy -0.925 (0.957) Constant 0.418 1.291 0.0601 0.296 (0.553) (1.809) (0.590) (0.658) Observations 180 180 180 180 R-squared 0.552 0.535 0.533 0.537

(19)

The asset to debt ratio indicates a low effect with a 1% increase in the asset debt ratio leading to a 0.006% decrease in REITs return. However, this result is not significant which make this relationship invalid. The problem with using this variable as QE measure could be that debt changes are affected due to other factors. Another variable measuring QE, the dummy variable, also shows insignificance results. For a dummy, it does not matter if it there is a 10% or 90% increase, but for the last variable measuring QE, Change in Assets, it does.

The significant coefficient on change in assets is negative, meaning a 1% rise in change in the FED’s assets causes a 0.154% decrease in the REITs return. The negative coefficient is explained by the economic theory of demand and supply. Due to the rebalancing effect of the QE policy, investors shift to REITs, leading to an increase in the price of REITs. An increase in the price lead to a decrease in the returns, as shown in the results. This measure is deemed the most appropriate because it captures the most variation in QE size. In other words, it uses the most information about the QE movements.

Specification 4 shows the results from the regression of the QE periods using dummy variables for each QE period. The QE measures are taken out because they measure the same phenomena as the separate dummies. The dummy variables are not significant, meaning that there is no difference for a specific QE period on REITs returns. Thus, this thesis shows a combined evidence from the 3 QE periods, and not the evidence for only a single period.

The coefficient for real interest is not significant, which strengthens the hypothesis stating that rebalancing effect and not lower interest rate has a negative effect on REIT during the QE. As expected, there is a positive relationship between REITs returns and S&P 500 returns. When there is a 1% rise in the S&P 500 returns, REITs returns will increase by 1.113%. The coefficient is significant at 1% level.

Finally, the R-squared of Table 3 for Change Assets, Assets Debt ratio and the dummy are 0.552, 0.535 and 0.533 respectively. Meaning respectively 55.2% 53.5% and 53.3% of the variation in the dependent variables can be explained by the variation in the explanatory variables. From Table 3 it can be concluded that Change in Assets is the best measure of QE policy on REITs returns. In the following regressions, Change in Assets is used as the variable measuring Quantitative Easing.

(20)

Table 4- Regression of QE on EREITs and MREITs separately

(1) (2) (3)

VARIABLES Total REITs EREITs MREITs

Change Assets -0.154*** -0.165*** 0.0481

(0.0580) (0.0627) (0.0694)

Change Housing Price -0.568 -0.654 0.251

(0.581) (0.628) (0.694) GDP Growth -0.723 -0.592 -1.433 (1.408) (1.522) (1.683) Real interest 0.245 0.302 -0.162 (0.296) (0.320) (0.354) S&P 500 1.113*** 1.149*** 0.707*** (0.0857) (0.0927) (0.102) Constant 0.418 0.386 0.322 (0.553) (0.597) (0.661) Observations 180 180 180 R-squared 0.552 0.530 0.226

Table 4 shows the results from equation (2) and equation (3). The effect of QE on EREITs is significant and the coefficient is -0.165, meaning a 1% increase in QE results in a 0.165% decrease in returns of equity REITs. This negative effect is larger for equity REITs separately, than for REITs together. On the other hand, a 1% increase in QE results in a 0.048% increase in MREITs returns, but this is insignificant, which means there is actual no relationship between them. A 1% increase in the real interest rate has different effect for equity and mortgage REITs returns. It results in a 0.302% increase in equity REITs returns and a 0.162% decrease in mortgage REITs returns. However, both of the coefficients are insignificant.

The R-squared for the equity REITs returns is 0.530, meaning 53% of the variance in equity REITs returns can be explained by the explanatory variables. For MREITs the R-squared is lower, 0.226, meaning only 22.6% of the variance in MREITs can be explained by the explanatory variables. In the Appendix Table 5 is added, where the real interest rate is divided into an interest rate part and an inflation part. Inflation is significant and the R-squared is higher for MREITs, indicating MREITs to be more sensitive to inflation than EREITs.

(21)

6. Conclusion

This paper has investigated whether Quantitative Easing policy has an impact on the returns of REITs in the US. By looking for variables capable to measure Quantitative Easing correctly and conducting the regressions on REITs return, this paper tests whether there was an effect over the total sample period of 2003-2017.

The results examine the effect for equity and mortgage REITs together and equity REITs separately. This paper shows that QE policy affects REITs through the rebalancing portfolio channel which is most significant for EREITs. In times of QE, investors rebalance their portfolios and equity REITs are seen as substitutes for sovereign bonds. The price of equity REITs rises and, by economic theory, their returns fall. The real interest rate does not have a significant effect on REITs returns, strengthening the conclusion that QE has an effect through the portfolio rebalancing channel and not via the direct interest rate effect. Lastly, this study finds no significant differences in the effect between the separate QE periods.

A limitation of this study is that only data from the US is included. Additionally, this thesis solely uses monthly data, whereas daily data could measure the effect more accurately. This study does not take into account the effect of other announcements and only focuses on the returns of REITs. For further research, daily data could be used, and further research can be extended with the data from European and Asian financial markets. Further research could also test for Hybrid REITs and investigate whether these results also apply for other assets and other unconventional monetary policies.

(22)

References

Ambrose, B. W., and P. Linneman, 2001. REIT organizational structure and operating characteristics. Journal of Real Estate Research, 21(3), 141-162.

Anderson, R., V. Boney and H. Guirguis, The Impact of Switching Regimes and Monetary Shocks: An Empirical Analysis of REITs, Journal of Real Estate Research, forthcoming, 2012.

Bredin, D., O’Reilly, G., and Stevenson, S., “Monetary shocks and REIT returns”. University College Dublin. School of Business. Centre for Financial Markets

Buetow G., and Johnson R., (2001) The Real Estate Asset Allocation Decision: Monetary Policy Implications. Journal of Real Estate Portfolio Management: 2001, Vol. 7, No. 3, pp. 215-223.

Chang, K. L., N. K. Chen and C. K. Y. Leung, Monetary Policy, Term Structure and Asset Return: Comparing REIT, Housing and Stock, Journal of Real Estate Finance and Economics, 2011, 43, 221–257.

Chan, S. H., W. K. Leung and W. Ko, Changes in REIT Structure and Stock Performance, Evidence from the Monday Stock Anomaly, Real Estate Economics, 2005, 33, 89–120.

Chang, C. Y., Chou, J. H., & Fung, H. G. (2012). Time dependent behavior of the Asian and the US REITs around the subprime crisis. Journal of Property Investment & Finance, 30(3), 282-303

Chen, M. C., C. L. Peng, S. D. Shyu and J. H. Zheng (2010), Market States and the Effect on Equity REIT Returns due to Changes in Monetary Policy Stance, Journal of Real Estate Finance and Economics, forthcoming, 2010.

Chen, K.C. and D. D. Tzang (1988). “Interest Rate Sensitivity of Real Estate Investment Trusts”, Journal of Real Estate Research, 3:3, 13–22.

(23)

Chou, Y. H., Chen, Y. C. (2014). Is the response of REIT returns to monetary policy asymmetric. Journal of Real Estate Research, 2014, vol. 36, issue 1, 109-136

Clayton, J. and G. Mackinnon (2003). “The Relative importance of Stocks, Bond and Real Estate Factors in Explaining REIT Returns”, Journal of Real Estate Finance and Economics, 27:1, 39-60.

Estrella, A., & Mishkin, F. S. (1998). Predicting US recessions: financial variables as leading indicators. Review of Economics and Statistics, 80(1), 45-61.

Feng, S. McKay Price, and C. Sirmans (2011) An overview of Equity Real Estate Nnvestments trusts (EREITs) : 1993–2009. Journal of Real Estate Literature: 2011, Vol. 19, No. 2, pp. 307-343.

Fratzscher, M., Lo Duca, M., & Straub, R. (2016). On the international spillovers of US quantitative easing. The Economic Journal.

Giliberto, M. S. (1990). Equity real estate investment trusts and real estate returns. Journal of Real Estate Research, 5(2), 259-263.

Hausken K., Ncube M. (2013) Transmission Channels for QE and Effects on Interest Rates. In: Quantitative Easing and Its Impact in the US, Japan, the UK and Europe. SpringerBriefs in Economics. Springer, New York, NY

Joyce, M. A. S., Lasosa, A., Stevens, I., and Tong, M. (2011). The financial market impact of Quantative Easing in the United Kingdom. International Journal of Central Banking, 7(3), pp. 113 – 161

Joyce, M., Miles, D., Scott, A., & D. Vayanos , (2012). ‘’Quantitative easing and unconventional monetary policy- an introduction. ‘’The economic journal, F271-F288

Krishnamurthy, Arvind, and Annette Vissing-Jorgensen. “The effects of quantitative easing on interest rates: channels and implications for policy”. No. w17555. National Bureau of Economic Research, 2011

(24)

Kuhle, J., Mohammed Al-Deehani, T., Mahmood, M. (2015). “Portfolio Diversification Benefits Using Real Estate Investment”. International Journal of Economics and Financial Issues Vol 5 Issue 4 2015

Liang, Y., W. McIntosh and J. R. Webb, Intertemporal Changes in the Riskiness of REITs, Journal of Real Estate Research, 1995, 10, 427–443.

Mueller, G. R. and K. R. Pauley (1995). “The Effect of Interest-rate Movements on Real Estate Investment Trusts”, Journal of Real Estate Research, 10:3, 319-326.

Lee, M.L., Chiang, K., and Craig Wisen (2004) Another Look at the Asymmetric REIT-Beta Puzzle. Journal of Real Estate Research: 2004, Vol. 26, No. 1, pp. 25-42.

Lu, C., and So, W. R., (2001) The relationship between REITs returns and inflation: A vector error correction approach. Review of Quantitative Finance and Accounting; Jan 2001; 16, 2; ProQuest pg. 103

Okunev, J. and P. J. Wilson. (1997). “Using Nonlinear Tests to Examine Integration Between Real Estate and Stock Markets”, Real Estate Economics, 25:3, 487–503.

Payne, J., (2003) Shocks to macroeconomic stata variables and the risk premium of REITs. Applied Economics Letters, 2003, 10, 671–677

Peterson, D., Hsieh, C., (1997) Do common risk factors in the Returns on Stocks and Bonds Explain Returns on REITs. Journal of Real Estate Economics: 1997, V25 pp. 321-345

Semer, S. L., (2009), A brief history of US REITs, Canadian Tax Journal, 57(4), 960–971.

Shulman, D., Giliberto, M., (2017) On the Interest Rate Sensitivity of REITs: Evidence from Twenty Years of Daily Data. Journal of Real Estate Portfolio Management: 2017, Vol. 23, No. 1, pp. 7-20.

(25)

Simpson, M.W., Ramchander, S., Webb, J.R. (2007) The Asymmetric Response of Equity REIT Returns to Inflation. Journal of Real Estate Finance and Economics, Vol. 34, p. 513-529

Swanson, Z., J. Theis, and K. M. Casey. (2002). “REIT Risk Premium Sensitivity and Interest Rates”, Journal of Real Estate Finance and Economics, 24:3, 319-330.

Westerheide, P., (2006) Cointegration of Real Estate Stocks and Reits with Common Stocks, Bonds and Consumer Price Inflation. ZEW - Centre for European Economic Research Discussion Paper No. 06-057.

(26)

Appendix

Table 5: Regression of QE on EREITs and MREITs separately with inflation added

(1) (2) (3)

VARIABLES Total REITs EREITs MREITs

Change Assets -0.154*** -0.165*** 0.0490

(0.0581) (0.0629) (0.0685)

Change Housing Price -0.574 -0.658 0.226

(0.582) (0.629) (0.686) GDP Growth -0.639 -0.531 -1.059 (1.417) (1.533) (1.671) Interest Rate 0.197 0.266 -0.379 (0.307) (0.333) (0.363) Inflation -0.647 -0.595 -1.625* (0.723) (0.782) (0.853) S&P 500 1.108*** 1.145*** 0.683*** (0.0863) (0.0934) (0.102) Constant 1.327 1.052 4.372** (1.594) (1.724) (1.879) Observations 180 180 180 R-squared 0.553 0.531 0.249

What is the effect of quantitative easing policy on US REITs returns?