An analyses of including REITS in an investment portfolio

(1)

An analyses of including REITs in an investment portfolio 1

An analyses of including REITS in

an investment portfolio

--- Ferry Manshanden

Studentnumber:10339930 Supervisor: Martijn Dröes Finance: Real Estate

Abstract

REITs have become an important part of the investment portfolio’s of many investors. Previous literature has shown that REITs used to be a good diversification tool for an investment portfolio. Also the momentum factor of REITs used to have a predicting value on the returns. This paper shows that these characteristics of REITs have changed over the last decade. In the past , correlations of equity REITs and Mortgage REITs to the market have been relatively high. However it looks like the correlations will restore into their old pattern since the last few years showed a declining pattern with low correlations. Moreover the momentum factor in the Carhart four factor model showed for equity REITs and mortgage REITs insignificant values. This means that the momentum factor of equity REITs and mortgage REITs didn’t have any predicting value over the last decade. This was tested by a three month and six month momentum regressed by the Carhart four factor model.

(2)

Statement of originality

This statement certifies this paper is based on original research undertaken by Ferry Manshanden. Findings and ideas of others are referenced.

(3)

1. Introduction

Since the introduction of Real Estate Investment trusts or REITs in 1960 the United States REIT industry has been growing to a $1 trillion equity market and almost $2 trillion in real estate assets. Only very wealthy individuals used to be able to invest in real estate. But by creating REITs all kind of investors had the opportunity to make real estate investments. REITs are a type of security investing in real estate and are comparable to mutual funds. Because of the introduction of REITs it is possible for small investors to invest in large scale projects. There are two main types of REITs: The first type is equity REITs, which generate income by sales and rents of real estate. The second main type of REITs is mortgage REITs, those type of REITs generate income by the sales of mortgages.

Including REITs has quite some benefits over only including regular stocks in an investment portfolio. REITs have reliable dividend pay-outs because of it´s pay-out requirements. Moreover REITs are a natural protection against inflation and they are a liquid way of investing in real estate.1 Another benefit of REITs is that they show a low correlation with the overall market (Bley and Olson, 2003). This makes REITs a good tool to create a well-diversified portfolio. Some literature has been written about this topic. All of them found a declining pattern in the correlations between the REIT industry and the market. For example Waggle and Aggrawal (2006) found a correlation of 0.64 in the period of 1972 until 1987 which almost halved to a correlation of 0.36 in the period of 1988 until 2002. However there is much less known about the correlation between REITs and the market over the last decade. This study will analyse how the correlations has developed over the last ten years. Differences between equity REITs and mortgage REITs will be taken into account, as it is important for investors to know the characteristics of their investments.

A second area of focus will be whether the momentum can explain REIT returns. For this the research question of this paper will be: Can REIT returns be predicted by the momentum? Also there will be taken into account whether there are differences between equity REITs and mortgage REITs. Not much research has been done about this topic yet.

1

(5)

An analyses of including REITs in an investment portfolio 5 However the research which has been done states that REIT returns can be explained by the momentum. Finally a seasonality check will be done on the regression of the Carhart four factor model added with 11 month dummy variables. Research on including REITs in an investment portfolio is relevant since it is important for investors to know the characteristics of the investments they make. REITs are predicted to behave differently from most other stocks. The correlations of REITs to the market will be important because REITs can be used to reduce risk. To study whether REIT returns can be explained by the momentum is important as well. Since if the momentum can predict returns well, they can be used to predict future prices. Investors will know when to buy or sell REITs. This paper is meant to contribute on the knowledge about the behaviour of REITs in the aspect of correlations to the market and the returns. It differs from previous literature as the time period will be from January 2006 until December 2015. Data of the monthly REIT returns comes from the database of REIT.com. The other data comes from the database of Fama and French. Only the U.S. REIT market will be taken into account. Moreover, this study will in contrast to previous papers focus on the different results for equity REITs and mortgage REITs. As Equity REITs and mortgage REITs differ in the way of generating income, they also might show differences in the characteristics concerning diversification, momentum and seasonality. Previous papers took only equity REITs or the total REIT market into account.

2. Previous literature

2.1 Requirements

To be qualified as a REIT a firm must meet some requirements (Taylor and Bailey, 1963). The most important requirements are: (1) A REIT must pay out 90 percent of their net income to shareholders. (2) A REIT derives 75 percent of its gross income from rents or gains on the sales of real property or on shares of other real estate sources. (3) No more than 50 percent of the shares is permitted to be held by five or fewer investors. (4) A REIT needs to have a minimum of 100 shareholders. The REIT status will be beneficial on account of tax benefits.

2

(6)

An analyses of including REITs in an investment portfolio 6 2.2 REIT returns

Waggle and Aggrawal (2006) who did research on the correlation of returns of equity REITs to the market also found that equity REITs slightly outperform the market concerning returns. They found that in the period of 1972 until 2002 equity REITs had on average 1.1 percentage point higher annual return than the U.S. market. However this outperformance occurred mostly in the period before 1987. After that the market annual returns on average were slightly higher. Ruhmann, Bruce, Ritter and Davy (2015) who also found an outperformance of REITs compared to the market mentioned two factors which drive the outperformance. The first factor which drives outperformance of REITs on the market is that REITs have higher leverage. This higher leverage increases risk during market falls but also increases average returns. The second factor is that REITs invest more in non-core property sectors. This makes REITs more risky as well, but also drives the outperformance of higher returns on average.

2.3 Correlation with the market

Several studies has been done about comparing REITs to the market. Chandrashekaran (1999) has studied the correlations of the monthly returns of REITs and the S&P500. Chandrashekaran (1999) has calculated the correlations from 1975 until 1996 and also divided this in four subgroups to see whether there are some kind of patterns in correlations over time. For this other ratio’s and indexes were added to the data. Chandrashekaran (1999) found that from 1975-1979 the correlation between the REIT index and the S&P500 was 0.61. After a small increase to 0.79 from 1980-1984 the correlation kept on declining to 0.75 in the period of 1985-1989 and 0.48 in the period of 1990-1996. Similarly research of Case, Yang and Yildirim (2012) which also focuses on correlation between REIT and stocks found a declining pattern. This study reaches from 1972-2008 and is also divided in subgroups. This study has excluded REITS from the market index to avoid biased results. Case, Yang and Yilderim (2012) found a decline of the correlation until 2001. After that it started increasing again. Similarly, Bley and Olson (2003) came to the same conclusions. This study also took the differences between equity REITs and mortgage REITs into account. Bley and Olson (2003) found out that both type of REITs have a declining

(7)

An analyses of including REITs in an investment portfolio 7 pattern for the period from 1972-2001. They also found that the monthly returns of equity REITS are less correlated to the market as the monthly returns of mortgage REITs. This means equity REITs are a better tool to create a well-diversified investment portfolio than mortgage REITs. Same patterns for the correlation between equity REITs and large company stocks were found by Waggle and Agrrawal (2006). They found a correlation of 0.64 in the period of 1972-1987 which declined to a correlation of 0.36 in the period of 1988-2002. Waggle and Agrrawal (2006) also found that REITs outperformed the stocks over the whole period of 1972-2002 in average monthly returns with even a lower volatility. However these outperformance was mainly measured over the first period they measured which was from 1972-1987. This study focussed on finding the optimal allocation of equity REITs in an asset portfolio including equity REITs, large company stocks and long-term government bonds. They found that REIT returns are way more important than REIT-stock correlation for investors in making portfolio decisions. For investors of all risk levels had a one-percent change in the returns of REITs leaving the other returns unchanged much more effect on their decisions than a significant change of 0.1 in the correlation between REITs and stocks. On the other hand, Mull and Soenen (1997) claim that U.S. REITs were not a good tool to create a well-diversified portfolio over the period of 1985 until 1994. They compared stock return from the viewpoint of Canada, France, Germany, Italy, Japan, The United Kingdom and the United States to U.S. REITs. Even though this study claims in contrast to the other research that REITs show to much correlation to the stock market to reduce risk by portfolio diversification they also found that for some countries the correlations were rather low. Especially the correlation to the stock market for Japan and the United Kingdom were very low. Japan showed a correlation with the U.S. REITs of 0.11 and the United Kingdom a correlation of 0.12.

2.4 Momentum

There has been done way less research on the explaining value of the momentum on REIT returns. Chui, Titman and Wei (2003) who did research on the pattern between the momentum and the return in the period from 1984 until 2000 found a strong momentum effect in the post 1990 period. Especially for the REITs with a high market capitalization they found a strong effect of the momentum. Chui, Titman and Wei

(8)

An analyses of including REITs in an investment portfolio 8 (2003) also claimed REITs momentum effects have been stronger than other U.S. industries. Also, Derwall, Huij, Brounen and Marquiring (2009) found in their research that momentum is important in explaining REIT returns. They found that abnormal returns in REIT portfolios can be explained by momentum factors instead of managerial skills. This conclusion is found after using different models including the Carhart four factor model. This study was based on results of REITs reaching from 1980 until 2008. Bley and Olson (2003) also studied return patterns of equity REITs and Mortgage REITs. They found that investors can obtain abnormal high returns buying mortgage REITs immediately after the index moved up significantly. However the equity REITs shouldn’t be bought after an up move, they should be avoided for at least four months after the index moved up. When the index has made a significant decline investors can obtain abnormal returns for both equity REITs and mortgage REITs. They should buy it right after the decline and hold it for around six months. This has to do with mean reversion.

3. Data and methodology

The data for the return on equity REITs and mortgage REITs can both be found on REIT.com. These are monthly returns reaching from January 2006 until December 2015. The market return, risk free rate, small minus big factor and high minus low factor can all be found in the database of the website of Fama and French. Also these variables are on a monthly base. For the momentum two ranges will be compared. First a momentum with a 3 month lag and secondly a momentum with a 6 month lag. The market return stands for the average return of the market and risk free rate stands for a risk free interest rate. The market return subtracted by the risk free rate makes one of the variables for the Carhart four factor model. The market return by itself is used to calculate the correlations between the REIT returns and the market returns. Both the equity REITs as the mortgage REITs are taken into account, so the difference can be compared.

(9)

Table 1: Means, standard deviations, minimum and maximum

Mean Std. dev Min. Max. Equity REITs

Full set (obs=120) 0.877 7.356 -31.67 31.02

01-01-2006 until 31-04-2009 -0.561 10.555 -31.67 31.02 01-05-2009 until 31-08-2012 2.294 5.758 -10.96 14.40 01-09-2012 until 31-12-2015 0.899 4.049 -6.97 9.94

Mortgage REITs

Full set (obs=120) 0.068 5.590 -24.11 10.39 01-01-2006 until 31-04-2009 -1.682 7.724 -24.11 10.39 01-05-2009 until 31-08-2012 1.777 3.635 -7.11 9.71 01-09-2012 until 31-12-2015 0.107 4.04 -12.61 9.87

Market

Full set (obs=120) 0.701 4.448 -17.15 11.35 01-01-2006 until 31-04-2009 -0.526 5.246 -17.15 10.20 01-05-2009 until 31-08-2012 1.476 4.615 -7.88 11.35 01-09-2012 until 31-12-2015 1.155 3.028 -6.04 7.75

Table 1 shows that equity REITs outperform mortgage REITs and the market according to returns. However the risk was also higher than mortgage REITs and the market. That appears from the higher volatility´s. The outperformance and the higher risk correspond to the findings in earlier research. Mortgage REITs didn´t outperform the market. The volatility of mortgage REITs is slightly higher than the volatility of the market.

(10)

An analyses of including REITs in an investment portfolio 10 3.1 Hypotheses

The first hypotheses which will be tested in this paper is whether REITs are a good tool to diversify an investment portfolio. The null-hypotheses is: REITs are a good diversifier for an investment portfolio. If there is a low correlation the type of REIT will be a good diversifier. This means the null-hypotheses shouldn’t be rejected. A relatively high correlation would mean that the REIT isn´t a good diversifier for an investment portfolio. This means the null-hypotheses should be rejected.

Expectations based on previous literature is that there will not be much correlation between REIT returns and market returns. Especially for equity REITs a low correlation is expected. Also the correlation between REIT returns and market returns has had a declining pattern for decades, which also puts the expectations to a low correlation. However the latest literature showed a small up rise for the last few years. Based on this the results might show high correlation if the trend proceeds. If this hypothesis is not rejected, it can be concluded that REITs are a decent addition to a well-diversified portfolio.

The second hypothesis in this paper is whether the momentum can be used to predict REIT return. For this a lag of three months will be used:

H0: Momentum of three months is a key determinant of predicting REIT returns H1: Momentum of three months is not a key determinant of predicting REIT returns Also a longer period will be taken in to account to see whether a longer momentum period has more predicting value. For this a lag of six months will be used:

H0: Momentum of six months is a key determinant of predicting REIT returns H1: Momentum of six months is not a key determinant of predicting REIT returns The three month and the six month momentum are chosen because there might be a difference in explanatory value between a short lag and a long lag. There hasn’t been much literature written about the predicting value of the momentum concerning REIT returns. Nevertheless is predicted that REIT returns can be predicted by the momentum based on the previous literature. If this hypothesis is not rejected, it means that investors will know better when to add or remove REITs out of their

(11)

An analyses of including REITs in an investment portfolio 11 investment portfolio. It would mean that an investing strategy of buying REITs with a high momentum and selling REITs with a low momentum will lead to abnormal high returns.

3.2 Regression model

To see whether the momentum is a good tool to predict REIT returns the Carhart (1997) four factor model is used. This is an extension of the three factor model proposed by Fama and French(1993). The model takes the following form:

REIT returnt = 𝛽𝛽₀+𝛽𝛽₁(rmkt − rf)+β₂SMBt+𝛽𝛽₃HMLt+𝛽𝛽₄ΡREITMOMt-3+ 𝜀𝜀t

Where the REIT return stands for the monthly equity REIT returns or the monthly mortgage REIT return. rmkt stands for the monthly market return and rf is the risk free rate. SMB stands for small minus big and measures the historic excess returns of small caps over large caps. HML stands for high minus low and stands for the historic excess return of companies with high book to market ratio’s over companies with low book to market ratio’s. The momentum factor is taken as a three- or six month lag of the equity REIT return or mortgage REIT return.

4. Results

4.1 Correlations with the market

The first hypotheses in this paper states there is a low correlation between REITs and the market. The null hypotheses should be rejected. In table 3 is shown that for the equity REITs as the mortgage REITs the correlations are quiet high. Over the full period of the last decade the equity REITs showed a correlation of 0.75 and the mortgage REITs a correlation of 0.48.

Table 2: Correlation matrix

mrkt SMB HML Eq. REITs Mort. REITs

mrkt 1.000 - - - -

SMB 0.373 1.000 - - -

HML 0.315 0.178 1.000 - -

(12)

An analyses of including REITs in an investment portfolio 12 Mort. REITs 0.476 0.169 0.221 0.456 1.000

Table 3: Correlations to the market over time

Equity REITs Mortgage REITs Full set (obs=120) 0.7458 (0.0000) 0.4758 (0.0000) 01-01-2006 until 31-04-2009 (obs=40) 0.8280 (0.0000) 0. 4493 (0.0036) 01-05-2009 until 31-08-2012 (obs=40) 0.8325 (0.0000) 0.5674 (0.0001) 01-09-2012 until 31-12-2015 (obs=40) 0.2891 (0.0704) 0.3917 (0.0124)

For the equity REITs the correlation in the first two sub-period were almost equal around 0.83 and made a massive decrease to 0.29 in the last sub-period. The mortgage REITs started with an increase from 0.45 to 0.57, after which it also started to decrease to 0.39 in the last sub-period. Over the last few decades there was a certain stable decreasing pattern in the correlations between REITs and the market. Only the last few years of the available research showed a small increase. Based on these results the expectations were that correlation either would follow the downward trend or that the small up rise in the latest years would mean an upward trend again. This paper doesn´t show a stable continuing pattern in the correlations. Based on the first two sub-periods the null hypothesis should be rejected. The last-sub period on the other hand shows a low correlation which is in line with the last few decades. Perhaps the market was disturbed in the first two sub-periods.

In all periods the mortgage REITs are less correlated with the market than the equity REITs. This makes mortgage REITs to some extent better diversifiers for an investment portfolio. In an investment portfolio a mixture of different investments always reduces risk. So still equity REITs can be used in an investment portfolio to diversify it. However it appears from the results that REITs are not a particular tool to create a well-diversified portfolio anymore.

(13)

An analyses of including REITs in an investment portfolio 13 In contrast to expectations based on the past the momentum doesn’t show any explanatory value for returns of equity REITs and mortgage REITs. Therefore the null hypotheses relating to the momentum should be rejected for both types of REITs. 4.2.1 Equity REITs

Table 4: Regression Carhart four-factor model with equity REITs

Panel A Panel B Dependent Variables Equity REITs Coef. Std. Err.

t P-value Coef. Std. Err. t p-value

Mkt rf 1.058 0.110 9.63*** 0.000*** 1.043 0.1080 9.67*** 0.000*** SMB 0.268 0.207 1.29 0.198 0.199 0.210 0.95 0.344 HML 0.723 0.192 3.77*** 0.000*** 0.668 0.199 3.36*** 0.001*** MOM -0.011 0.061 -0.19 0.853 -0.091 0.063 -1.43 0.154 Cons. 0.246 0.442 0.56 0.579 0.340 0.063 0.76 0.447 R2 0.6137 0.6274 Obs. 117 114

*: significance at a 10 percent level, **: significance at a 5 percent level, ***: significance at a 1 percent level

In Table 3 can be seen that over the period from January 2006 until December 2015 the momentum didn´t show a pattern which can be used to predict Equity REIT returns. Panel A records the three month momentum and Panel B the six month momentum. Both the momentum on the 3-month interval and on the 6-month interval show a small negative effect on Equity REIT returns. The determinants are -0.011 and -0.091, which means that looking at the momentum a small change of the REIT return is expected in the opposite direction. However these determinants are not significant on the 10 percent level. Only the market return minus the risk free rate and the HML-factor show a significant explanatory value for Equity REIT returns. Based on these results predicted by the Carhart four factor model returns for equity REITs can´t be forecasted by simply looking at returns of the last 3 or last 6 months. This is contrary to what was found in previous literature where the results did show a

(14)

An analyses of including REITs in an investment portfolio 14 significant predicting value of the momentum. The R-Squared is high for both panel A and panel B. So the variability of the data fits well with the model.

4.2.2 Mortgage REITs

Table 5: Regression Carhart four-factor model with mortgage REITs

Panel A Panel B Dependent variable

Mortgage REITs

Coef. Std. Err. t P-value Coef. Std. Err. T p-value

Mkt rf 0.570 0.118 4.82*** 0.000*** 0.608 0.116 5.22*** 0.000*** SMB -0.078 0.221 -0.35 0.726 -0.105 0.223 -0.47 0.637 HML 0.145 0.202 0.72 0.475 0.092 0.208 0.44 0.660 MOM -0.060 0.085 0.70 0.485 -0.103 0.084 -1.24 0.219 Cons. -0.299 0.467 -0.64 0.524 -0.380 0.477 -0.80 0.428 R2 0.2387 0.2464 Obs. 117 114

Like the equity REITs, the momentum of the Mortgage REITs doesn’t have much explanatory value on the returns neither. The determinants of the momentum in this regression are -0.060 for the three month momentum and -0.103 for the six month momentum. These values are not significant on the 10 percent level as well. Therefore the momentum of three months or six months can´t predict mortgage REIT returns. This means that the results for the equity REITs and mortgage are quiet similar in this area. In this regression only the market return is significant on the 10 percent level. The SMB factor and HML factor don’t have a significant explanatory value on mortgage REIT returns. Also is the R-squared is low for both panels. This means that the data doesn´t fit the model well.

(15)

An analyses of including REITs in an investment portfolio 15 4.3 Seasonality

To check for robustness, also the calendar effects are tested. Eleven monthly dummy variables are added for the three month momentum model of Equity REITs and Mortgage REITs. The six month results are left out as the results are quiet similar to the three month results. The dummy variables reach from January until November. December was used as a base category. The regression model takes the following form:

REIT returnt = 𝛽𝛽₀+𝛽𝛽₁(rmkt − rf)+β₂SMBt+𝛽𝛽₃HMLt+𝛽𝛽₄ΡREITMOMt-3+τmonth +𝜀𝜀t

Below are the results. Panel A records the Equity results and Panel B records the Mortgage REITs results.

Table 6: Regression Carhart four factor model including monthly dummy variables

Panel A Panel B Dependent variables

Coef. Std. Err. t P-value Coef. Std. Err. T p-value January -3.380 1.783 -1.90* 0.061* 4.241 3.091 1.37 0.173 February 1.678 1.443 1.16 0.248 4.330 3.220 1.34 0.182 March -0.333 1.751 -0.19 0.850 5.311 3.140 1.69* 0.094* April 0.110 1.576 0.07 0.944 6.280 2.952 2.13** 0.036** May 0.232 2.313 0.10 0.920 5.484 3.123 1.76 0.082* June 1.304 1.568 0.83 0.407 7.147 3.321 2.15** 0.034** July August September October 0.846 3.819 -2.277 1.048 1.855 3.066 2.205 2.179 0.46 1.25 -1.03 0.48 0.649 0.216 0.304 0.632 4.386 5.241 3.988 5.860 3.376 3.699 3.449 3.242 1.30 1.42 1.16 1.81* 0.197 0.160 0.250 0.074* November -1.200 1.819 -0.66 0.511 4.671 3.376 1.38 0.170

(16)

Table 6 shows that for equity REITs barely any significant calendar effects are measured. Only January shows a negative effect on the 10 percent level. The mortgage REITs experienced more significant calendar effects. The months March, April, May, June and October all showed significant positive effects on the 5 percent level or on the ten percent level. The existence of significant calendar effects may play some role in momentum. However table 4 and 5 show there weren’t significant results for the momentum of REITs over the last decade. It is Expected that these calendar effects will disappear as investors will anticipate on this.

5.

Conclusion

Over the last decade REIT characteristics has changed a lot. The results of this paper show a large contradiction to previous written literature which was about REITs in earlier time periods. This research is different from earlier research as it separates the results for equity REITs and mortgage REITs where previous literature took equity REITs or the REIT market as a whole. But still results can be compared. Previous literature found an outperformance of REITs compared to the market during most time periods. In this paper is found that equity REITs still outperform the market based on excess returns and a lower volatility. However mortgage REITs didn´t outperform the market over the last decade. Previous papers also showed a low correlation between REITs and market and a declining pattern at most time periods. If REITs show little correlation to the market it means they can be used to reduce risk in an investment portfolio. Over the last decade the correlation between REITs and the market were rather high. Equity REITs showed a correlation of 0.7458 and mortgage REITs a correlation of 0.4758 on average over the period from 2006 until 2015. Remarkable is that mortgage REITs have been better tools to create a well-diversified portfolio than equity REITs over this time period. Previous literature stated that equity REITs had a lower correlation to the market than mortgage REITs. These outcomes are all contradicting the expectation based on previous literature. Nevertheless are results from the last sub period perfectly in line with the expectations. In the period reaching from September 2012 until December 2015 the

(17)

An analyses of including REITs in an investment portfolio 17 declining correlation pattern continued. Moreover were the correlations low and were the equity REITs less correlated to the market than mortgage REITs. A possible explanation for this is that the market was disturbed in the first two sub periods. The results from this paper also showed that there is no proof that REIT returns can be predicted by the momentum. A regression with the Carhart four factor model showed for equity REITs and mortgage REITs an insignificant momentum determinant. To test this a three month momentum and a six month momentum has been used. The seasonality check showed that for the equity REITs only a significant negative return was found for January. For the mortgage REITs significant positive returns were found for March, April, May, June and October.

Suggested is that the excess returns, volatility’s and correlation to the market for REITs will be researched on in future periods. Because it´s important for investors to know the characteristics of their investments. Concerning the momentum factor further research is required to test different models, different time period an different momentum factors in order to find a pattern. Since previous literature has found that the momentum had explaining value for predicting REIT returns, there is a possibility to find some results with explaining value in future research.

6.

References

Bley, J. and Olson, D ., 2003. An analyses of Relative Return Behavior: REITs vs. Stocks. EFMA 2003 Helsinki Meetings.

Carhart, M., 1997. On Persistence on Mutual Fund Performance. The Journal of Finance. Vol. 52(1): 57-82.

Case, B., Yang, Y and Yildirim, Y., 2012. Dynamic Correlations Among Asset Classes: REIT and Stock Returns. The Journal of Real Estate Finance and Economics. Vol. 44(3): 298-318.

Chandrashekaran, V., 1999. Time-Series Properties and Diversification Benefits of REIT Returns. Journal of Real Estate Research. Vol. 17: 91-112.

Chui, A., Titman, S and Wei, J., 2003. The Cross section of Expected REIT Returns. Real Estate Economics. Vol. 31: 451-479.

(18)

An analyses of including REITs in an investment portfolio 18 Derwall, J., Huij, J., Brounen, D and Marquiring, W., 2009. REIT Momentum and the Performance of Real Estate Mutual Funds. Financial Analysts Journal. Vol. 65(5): 24-34.

Fama, E and French, K., 1993. Common Risk Factors in the Returns on Stocks and Bonds. Journal of Financial Economics. Vol. 33: 3-56.

Mull, S. and Soenen, L., 1997. U.S. REITs as an Asset Class in International Investment Portfolios. Financial Analysts Journal. Vol. 53: 55-61.

REIT Financial Benefits.(n.d.). Retrieved from: https://www.reit.com/investing/reit-basics/reit-financial-benefits.

Ruhmann, S., Bruce, T., Ritter, M. and Davy, L., 2015. A Primer on U.S. Equity REITs and Their Role in an Institutional Investment Portfolio. NEPC. 1-14.

Taylor, G and Bailey, N., 1963. Real Estate Investment Trusts. Business Horizons. Vol.6(2): 71-80.

Waggle, D. and Agrrawal, P., 2006. The Stock-REIT Relationship and Optimal Asset Allocations. Journal of Real Estate Portfolio Management. Vol. 12: 209-221.

(19)

An analyses of including REITS in an investment portfolio