Do the returns on securitized real estate resemble the returns of direct real estate in the long term?
Bachelor Thesis
Abstract
This thesis aims to explain to what extent the returns on securitized real estate resemble the returns on its underlying assets, which are direct real estate investments. To answer this question, a quantitative analysis has been conducted. The data used is the total annual returns of direct real estate, securitized real estate and the general stock market of the Netherlands, UK, and USA. If available, data from 1977 through 2014 is used. The returns, volatilities and correlations are calculated and compared for all three markets and countries. In addition, the risk adjusted cumulative returns are calculated to determine the effect of a long term investment strategy.
Author: C. Schouten
Supervisor: Prof. dr. P. van Gool FRICS
Date: 30-06-2015
Internship: SPF Beheer B.V.
Verklaring eigen werk
Hierbij verklaar ik, Casper Schouten, dat ik deze scriptie zelf geschreven heb en dat ik de volledige verantwoordelijkheid op me neem voor de inhoud ervan.
Ik bevestig dat de tekst en het werk dat in deze scriptie gepresenteerd wordt origineel is en dat ik geen gebruik heb gemaakt van andere bronnen dan die welke in de tekst en in de referenties worden genoemd.
De Faculteit Economie en Bedrijfskunde is alleen verantwoordelijk voor de begeleiding tot het inleveren van de scriptie, niet voor de inhoud.
Index
Introduction ...4 Methodology...5 Hypothesis ...5 Methodoly ...6 Data ...6Characteristics of direct real estate ...7
Strengths and weaknesses of direct real estate as an investment ...8
Characteristics of securitized real estate ...8
Strengths and weaknesses of securitized real estate as an investment ...9
Literature Summary ... 10 Results ... 11 Descriptive statistics ... 11 Dutch markets ... 11 USA markets ... 13 UK markets... 14
Comparing NL, USA and UK ... 15
Correlations ... 16
Dutch markets ... 16
USA markets ... 17
UK markets ... 18
Long Term Risk Adjusted Average Returns ... 19
Dutch markets ... 20
USA markets ... 21
UK markets ... 22
Results summarized, NL ... 23
Results summarized, USA ... 23
Results summarized, UK ... 24
Conclusion ... 24
Appendices ... 27
Introduction
Even though land and buildings have played an important role in both the ancient and modern world as a source of wealth, power, and economic strength, institutional investors such as pension funds have been reluctant to hold these less liquid and potentially
management-intensive assets. However, over the last 30 years, many institutional funds have discovered the advantages of investing in real estate. Investing in real estate was initially viewed as a manner to diversify the investment portfolio or reduce risk (Hudson-Wilson et all., 2003).
When real estate as an investment class became increasingly more popular, new ways to invest in real estate emerged. Public funds investing in direct real estate called Real Estate Investment Trusts (REIT’s) were then introduced (Van Gool et all. 2013). By investing in a Real Estate Investment Trust, the investor could invest in securitized real estate instead of buying their underlying assets. The question arises as to what extent these two ways of investing in real estate are similar. Since REIT’s are traded on the public market, they may embed stock market noise that is not related to the fundamentals driving real estate (Hoeslie & Oikrainen 2014). The returns on securitized real estate are therefore expected to be less stable than the returns on direct real estate. In the long term however, market sentiments are
expected to be corrected. Securitized real estate returns are expected to behave identically to their fundamental assets in the long term. If this occurs, then the question that comes to mind is if institutional investors, who have a long term investment horizon anyway, should invest in direct real estate at all.
This thesis aims to give an explanation to the question to what extent the returns on securitized real estate returns resemble the returns on direct real estate in the long term. To be able to do so, first a better understanding of the two markets is necessary. Therefore, this thesis initially provides an overview of the characteristics and positives and negatives of both direct and securitized real estate. Second, a literature summary is provided to review what the current and past literature tells us about the comparability of direct and securitized real estate. Third, a data analysis is conducted and the results are discussed. In the data analysis, the correlations and volatilities for both the short and long term are reviewed to truly understand the direct and securitized real estate markets. Finally, the short and long term risk adjusted returns are calculated to show if the risk adjusted returns on securitized real estate and direct real estate are comparable.
Methodology
In this paragraph the hypothesis, the methodology and the data used are discussed.
Hypothesis
This thesis aims to explain to what extent the returns on securitized real estate resemble the returns on direct real estate. Amongst the real estate investors, certain criteria, either supported by the literature or not, are popular. These criteria will first be stated and formulated as hypothesis second.
Securitized real estate is popular due to its liquidity and because in general, the long term average returns are assumed to be higher than the returns on direct real estate. However, critics argue that although this may be the case, the risk adjusted returns (Sharpe ratios) are lower. Critics also state that REITs are very volatile because they consider REITs as just another “stock” and therefore REITs will be highly correlated with the general stock market. REIT enthusiasts argue that this is only the case in the short term, and that this correlation is lower in the long term. Since pension funds have a long term investment horizon, the short term correlation are irrelevant. Although the fundamental assets are the same, REITs process market information quicker than direct real estate due to the smoothing and lagging in direct real estate returns (R. van den Goorbergh, 2015). In the long term, direct real estate is expected to correct for the market information also and therefore, in the long term, the cumulative returns on REITs and direct real estate should be strongly correlated.
The arguments that are quantified and formulated as hypothesis. The relevant hypothesis for each country will be discussed.
Hypothesis 1) Average securitized real estate returns are higher than the average returns on direct real estate.
Hypothesis 2) The contemporaneous risk adjusted returns on securitized real estate are lower than the contemporaneous risk adjusted returns on direct real estate.
Hypothesis 3) The returns on securitized real estate are more volatile than the returns on direct real estate.
Hypothesis 4) The long term risk adjusted returns on securitized real estate are better than the long team risk adjusted returns on direct real estate.
Hypothesis 5) The returns on securitized real estate are in the long term more correlated with the returns on direct real estate than in the short term.
Hypothesis 6) The returns on securitized real estate are in the long term less correlated with the returns on stock than the contemporaneous correlation.
Methodoly
Initially the current and past literature are reviewed. This review provides a better
understanding of the characteristics and the positives and negatives of both markets. The literature has also provided a general idea about the criteria that can be used to determine if the markets resemble one another. Second, A data analysis has been done to test this criteria. The data used is the annual returns on direct real estate, securitized real estate and the “risk free bond market”. Further explanation of the data used is provided in the “data” paragraph. Using the annual returns, the average returns, the volatilities, the contemporaneous correlations, Sharpe ratios can be calculated. To look at the long term, the cumulative 2, 3, 4 and 5 year returns are calculated. With these cumulative returns, the 2,3,4 and 5 year correlations can be calculated. Using the cumulative returns, the contemporaneous, 2 year, 3 year, 4 year, and 5 year Sharpe rations are calculated.
Data
All data used are annual total returns in local currency. The primary motivation for the use of annual returns is the limited availability of quarterly returns on direct real estate. When available, the time period used for this thesis is 1977-2014. A critical note on the direct real estate returns is that appraisal-based return indexes available are smoothed, understating both the true volatility of property returns and the covariance with property stocks (Brounen & Eicholtz 2003). Appraisers tend to partially rely on estimated values from previous periods, which creates aggregated series with high levels of first-order autocorrelation. This process smoothes the progress of the return series, and results in an inherent time lag. (Brounen & Eicholtz 2003).
The actual data used in this thesis (local currencies):
Risk free rate NL: Three month Euribor, European Central Bank Risk free rate USA: Three month Treasury Bill, Federal Reserve Risk free rate NL: Three month Treasury Bill, Bank of England
Direct real estate Netherlands: The ABN AMRO Real Estate Index (AAVI) (1977-1983) and the Investment Property Databank (ROZ/IPD) All Property Total Returns (1984-2014)
Public real estate Netherlands: GPR General Index NL (1977-2014) Stock returns Netherlands: AEX Index (1977-2014)
Direct real estate USA: NCREIF Total Return Index (1978-2014)
Public real estate USA: EPRA/NAREIT USA Total Return Index (1978-2014) Stock returns USA: S&P 500 Total Return Index (1978-2014)
Direct real estate UK: Investment Property Databank all property total return (1984-2014) Public real estate UK: GPR General Index UK (1984-2014)
Stock returns UK: FTSE 100 Total Return Index (1984-2014)
Characteristics of direct real estate
To understand to what extent REIT returns resemble the return of their underlying assets, a further understanding about the characteristics and positives and negatives of both direct real estate as REITs is necessary. Van Gool et all. (2013) provide an overview.
Characteristics of direct real estate:
By holding real estate, the investor participates both in an asset market and the rental market.
Since it is impossible to move buildings, the location is an important determinant of the value of the asset. This means that if there are any developments that cause the appeal of the location to decline, the value of the asset drops as well.
Every building is unique. This makes buying and selling buildings difficult since there are often few buyers and sellers compared to other markets like the bond market.
There is no such thing as “the real estate market”. There are only many small sub markets due to the location and heterogeneity.
There are no continuous price changes. This is because there are a small amount of transactions in the economical lifespan of a building. The information is mostly private which may cause market imperfections.
Direct real estate per unit is more costly compared to stocks or bonds. Therefore, to be able to diversify, the overall volume of the portfolio needs to be large.
Transaction costs are high. There are broker costs, contractual costs etc. Because of the lack of transparency in the real estate market, the costs of receiving the necessary information are high also.
Real estate is non-liquid.
The economical lifespan of land is infinite, the lifespan of a building is very long.
The production (construction) of real estate is time consuming.
Real estate faces a lot of governmental regulation.
Strengths and weaknesses of direct real estate as an investment
The characteristics of direct real estate create several strengths and weaknesses of investing in real estate, Van Gool et all. (2013) summarizes them. One of the major
advantages is the diversification benefit. The small or sometimes even negative correlation with stocks and bonds help in diversifying portfolio risk. The long term rental contracts involved with renting out a building are also a positive of investing in real estate because the contract causes the future cash flows to be stable. The rental contracts often have an
indexation clause allowing the building owner to raise the rental price every year by a certain percentage. Therefore, another strength of real estate is that investing in real estate allows an inflation hedge. History has shown that investing in direct real estate involves little risk and produce a decent return. The risk return ratio of direct real estate is thus another strength. The returns on real estate are not only dependent on market conditions, proper active management can increase the returns on real estate. The investor can thus to some extent improve its returns by active management. Real estate is in many countries fiscally treated different from other asset classes. Because buildings can be capitalized, there is a possibility of deducting depreciation expenses “brics“ can therefore be a fiscally attractive addition to a mixed asset portfolio.
There are not only advantages of investing in real estate, there are also some
weaknesses to consider. Property is very non-liquid and investing in property requires large amounts of capital. Investing in direct real estate is also costly in management. The direct real estate markets are not transparent and therefore, the performance is difficult to quantify.
Characteristics of securitized real estate
REITs are trusts that invest in real estate. The structure of public real estate funds differ from country to country. A more extensive explanation about American REITs is provided by Gentry et all (2004). They state that REITs initiate operations by raising capital from external markets and investing the capital in operating assets. The benefit of qualifying as a REIT is avoiding the double taxation of equity-financed investment. Unlike regular corporations, REITs receive an annual tax deduction for dividends paid out to shareholders. US REITs often distribute all of their taxable income to shareholders each year, which eliminates the corporate tax completely. To qualify as a REIT, among other things, a firm must meet certain asset and income tests that set minimum levels of real estate activity to prevent REITs from using their tax advantaged status in non-real-estate activities. REITs must
earn at least 75 percent of their income from real estate-related investments and 95 percent of their income from these sources as well as dividends, interest and gains from securities sales. Additionally, a minimum of 75 percent of their assets must be invested in real estate,
mortgages, REIT shares, government securities, or cash. While older REITs were often passive investors, several changes in tax rules in the late 1980s allowed REITs to actively manage their assets during the 1990’s (Gentry et all. 2004). Although some REITs invest in real estate mortgages, this thesis restricts focus to publicly traded equity REITs, which invest in rental properties.
Van Gool et all. 2013 states several characteristics for securitized real estate. Securitized real estate is, by definition, a liquid asset. Securitization also provides the possibility to invest in real estate with a limited amount of funds. Securitized real estate is traded on public markets and therefore more transparent than the private direct real estate market. Because of the transparency, the performance of securitized real estate is also more easily measured then the performance of direct real estate.
Strengths and weaknesses of securitized real estate as an investment
Securitized real estate has several strengths in favour of direct real estate. Expertise is assumed to be present at the fund’s management. An institutional investor that does not have all of the specific knowledge of real estate can profit from the expertise of the real estate investment trust’s management. Secondly, the investor can more easily buy and sell real estate because of the high liquidity of securitized real estate. Also, investors can enter the real estate markets using a small amount of funds. The investor simply buys a couple of shares in a REIT instead of buying actual property. The trade of shares in REITs is excluded from transaction taxes, just like ordinary shares. If the underlying property is transitioned from an owner, a transaction tax has to be paid. Another advantage of REITs is that the investor can benefit from economies of scale. These economies of scale exist because of the centralization of management expertise. A common trap in investing in direct real estate is that asset managers become too emotionally involved with the buildings they buy and sell. Investors investing in REITs do not suffer from that bias. By investing in REITs the investor has the possibility to generate higher returns than direct real estate. This is possible if the shares can be bought at a discount, which means that the REIT price is rated below the value of the underlying assets. This way the investor can profit from the appreciation of the share price in addition to the dividend yield. Another reason and advantage of REITs is that investors can decide to invest
in REITs using loan capital to increase the leverage effect to increase the return on equity. Lastly, a big strength is the transparency of the REIT markets and being able to benchmark using many available indices.
The use of leverage is also one of the disadvantages. Using leverage significantly increases the risk. Risk is not the only drawback though. Although liquidity is an advantage, the fact that the securitized real estate is liquid is because it is listed. Listed REITs move along with market sentiment and show comparable volatility to the general stock market. (van Gool, 2013)
Literature Summary
Given the characteristics, the strengths and weaknesses of both direct and securitized real estate, the extent to which securitized real estate resembles direct real estate can be discussed. An overview of the current and past literature is provided below to review on what earlier research concludes.
Several authors have investigated the co-integration of direct and listed real estate. If the two markets are co-integrated, then there are common factors that affect both returns and so the series’ will eventually adjust to equilibrium. Thus the perceived diversification benefits of REITs within optimal portfolios will be eroded away when the returns of direct real estate are also considered. Therefore, REITs would be operationally redundant in the long term for an investor who already owns direct property. (Lee, and Stevenson, 2004)
Gyourko & Kiem (1992) conducted research on to what capacity the lagged REIT returns predict the returns of the direct market after controlling for appraisal persistence. They found a significant result that the lagged REIT returns indeed can predict the direct real estate returns (Lee & Stevenson, 2004). F. Campeau (1994) found a long term relationship existing between the private and public real estate markets. In fact, he concluded that the markets are integrated (Lee & Stevenson, 2004). Glascock et all. (2000) also found a co-integration between the REIT and private real estate market (Lee & Stevenson, 2004). Clayton
&MacKinnon (2001) found that the sensitivity of REIT returns to private real estate showed a significant increase in the 1990s indicating that REITs are more integrated with private real estate than financial assets. These findings indicate that the public real estate market provides information about real estate performance that is subsequently impounded into the direct market and that the public market leads the private market. In addition, the studies
but that over the long-run, REIT returns equate to those of the private real estate market. Indicated that, when considering REIT returns and the returns from the direct market in the short-term both may have a place in optimal portfolios, but in the long-term one is a substitute for the other (Lee & Stevenson, 2004).
Hoeslie & Oikarinen (2012) also conducted research on the integration of the private and public real estate markets and state the following; the short-term co-movement between the securitized and direct real estate markets may be significantly diminished by the typically sluggish adjustment of direct real estate market prices to changes in the fundamentals. However, as in the long term both markets should adjust to shocks in the fundamentals and the impact of noise in securitized real estate prices should vanish, securitized real estate should strongly co-vary with the returns on a portfolio composed of equivalent direct real estate investments, since the fundamental asset is essentially the same in both markets. In line with this assumption, it has been established that over long horizons the linkages between indirect and direct real estate are substantially stronger than suggested by the simple
contemporaneous correlation figures (Giliberto, 1990; Geltner and Kluger, 1998; MacKinnon and Al Zaman, 2009; Oikarinen, Hoesli and Serrano, 2011).
Results
The results of the analysis in this thesis will be discussed for every market, an interpretation of the results will be provided and the hypothesis will either be rejected or not rejected. After the market specific analyses an overall analysis will be provided and the hypothesis will be discussed for the overall research.
Descriptive statistics
To provide a general overview on the different markets, the relevant descriptive statistics are summarized in graph 1-3 and table 1-3. The Dutch market is discussed first followed by the USA and the UK markets.
Dutch markets
The data for the NL markets was available from 1977, therefore the “starting” year for the NL data analysis is 1977.
Graph 1: Total Annual Returns NL Markets
Graph 1 shows the total returns on the risk free rate, direct real estate, listed real estate and the stock market in the Netherlands from 1977 until 2014.
Table 1: Descriptive Statistics NL Markets
The returns on Dutch listed real estate are 0.72% higher than the returns on Dutch direct real estate. This difference has a power of 0.405 using a two sample unequal variance T-test and is thus statistically not significant. Hypothesis one; Average securitized real estate returns are higher than the average returns on direct real estate, is therefore rejected for the Dutch market.
Graph one and table one show that the Dutch stock market is the most volatile followed by listed real estate. Stock and listed real estate are however more profitable than direct real estate. Direct real estate outperforms both listed real estate and the general stock market in terms of risk adjusted returns. This is partially because of the very high volatilities of stock and listed real estate compared to direct real estate with volatilities of 0.26, 0.18 and 0.06 respectively. The difference between the volatility of direct and listed real estate is statistically significant with 99% certainty using the two sample F-test. Direct and listed real
-0,60 -0,40 -0,20 0,00 0,20 0,40 0,60 0,80 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 2013 NL Direct RE NL Listed RE NL Stock
Descriptives NL Average Return Standard Deviation Average RF Rate Sharpe Ratio
NL Direct RE 8,37% 0,06 4,74% 0,66
NL Listed RE 9,09% 0,18 4,74% 0,25
estate markets do not appear to very similar, looking at the information provided by the descriptive statistics.
USA markets
The data for the USA markets was only available from 1978, therefore the “starting” year for the USA data analysis is 1978.
Graph 2: Total Annual Returns USA markets
Graph 2 shows the total annual returns on the listed real estate market, the direct real estate market and the general stock market in the USA from 1978-2014.
Table 2: Descriptive Statistics USA Markets
The returns on USA listed real estate are 4.95% higher than the returns on USA direct real estate. This difference has a power of 0.095 using a two sample unequal variance T-test and is thus statistically not significant with 95% certainty. Hypothesis one; Average securitized real estate returns are higher than the average returns on direct real estate, is therefore rejected for the USA market. Listed real estate in the USA is the most volatile and profitable asset class. The difference in volatility between listed and direct real estate is 0.14. This difference is statistically significant with 99% certainty using the two sample F-test. USA stock slightly
-0,5 -0,4 -0,3 -0,2 -0,1 0 0,1 0,2 0,3 0,4 0,5 0,6 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 USA Direct RE USA Listed RE USA Stock
Descriptives USA Average Return Standard Deviation Average RF Rate Sharpe Ratio
USA Direct RE 9,51% 0,08 4,90% 0,59
USA Listed RE 14,46% 0,21 4,90% 0,45
underperforms the average returns but is also 20% less volatile. In terms of risk adjusted returns the USA direct real estate market outperforms both listed real estate and stocks. Listed real estate is nearly 3 times as volatile as direct real estate. The direct and listed real estate markets do not show many similarities. The stock market and the listed real estate market do, however, show some resemblance.
UK markets
The data for the UK markets was only available from 1984, therefore the “starting” year for the UK data analysis is 1984.
Graph 3: Total Annual Returns UK markets
Graph 3 highlights the total annual returns on the listed real estate market, the direct real estate market and the general stock market in the UK from 1984-2014.
Table 3: Descriptive Statistics USA Markets
The table shows that listed real estate is the most volatile and profitable investment. The difference between the listed and direct real estate returns is 2.23%. This difference has a power of 0.33 and thus not statistically significant with 95% certainty. The difference between
-0,6 -0,4 -0,2 0 0,2 0,4 0,6 0,8 1 UK Direct RE UK Listed RE UK Stock
Descriptives UK Average Return Standard Deviation Average RF Rate Sharpe Ratio
UK Direct RE 9,71% 0,10 6,30% 0,33
UK Listed RE 11,94% 0,27 6,30% 0,21
the volatility of listed and direct real estate is 0.17. This difference is statistically significant with 99% certainty. Direct real estate would appear to be the best investment in terms of risk adjusted returns, closely followed by stocks. Direct and listed real estate do not appear to have many similarities based on the descriptive statistics.
Comparing NL, USA and UK
For the UK markets, data was only available from 1984. Therefore, to make sure the inter-market comparison is fair, all markets are compared from 1984-2014.
Graph 4: Cumulative annual returns 1984-2014
To better understand the overall performance, graph 4 shows the cumulative annual returns from 1984-2014. Graph 4 illustrates that USA Stock has performed the best, followed by UK Stock and USA Listed real estate. The graph also clearly shows the creation of and the burst of the real estate bubble during the period of 2001-2008. A critical note may be that
cumulative return performance are highly dependable on the period that is being reviewed. Choosing a different period can drastically change the outcome of the review on cumulative returns.
Table 4 Descriptive Statistics USA, UK and NL Markets
Table 4 provides the descriptive statistics of the Dutch, UK and NL markets. The second column of table 4 provides the average annual total return. This column tells us that in all 3 countries, the average returns on securitized real estate are higher than the average returns on direct real estate. This difference is however not statistically significant in all three countries (2 sample unequal variance T-test). Hypothesis one is therefore rejected for all 3 countries. The third column of table 4 provides the volatility of all three markets. The Dutch stocks returns stand out as the most volatile returns on stocks compared to USA Stocks and UK Stocks. For all three countries, direct real estate is (by far) the least volatile. On average, the returns on listed real estate are nearly three times more volatile than the direct real estate returns. Looking at column 5, The Dutch direct real estate market seems to be unique in terms of risk adjusted returns compared to the risk adjusted returns on securitized real estate. The difference in the Sharpe ratio of direct compared to securitized real estate in the Netherlands is 0.36. This is much larger than the difference in Sharpe ratio of direct and securitized real estate in the USA (-0.01) and UK (0.02).
Correlations
For every market, the contemporaneous, 2,3,4 and 5 year correlations are calculated. Every submarket and an overall analysis will be discussed.
Dutch markets
Table 5 provides the correlation figures of the Dutch market. Graph 5 shows the correlation figures in a bar chart.
Descriptive stat. Average return Standard deviation Average RF rate Sharpe ratio
NL Direct RE 7,87% 0,05 4,74% 0,66 NL Listed RE 8,17% 0,19 4,74% 0,18 NL Stock 12,44% 0,26 4,74% 0,30 USA Direct RE 8,32% 0,08 4,91% 0,44 USA Listed RE 11,75% 0,21 4,91% 0,33 USA Stock 12,63% 0,17 4,91% 0,45 UK Direct RE 9,71% 0,10 6,30% 0,33 UK Listed RE 11,94% 0,27 6,30% 0,21 UK Stock 11,11% 0,16 6,30% 0,31
Table 5
Graph 5
The analysis of the Dutch market does not show support for hypothesis five since a strong pattern showing an increase in correlation between direct real estate and the listed real estate cannot be detected. We do, however, see a clear trend of lower correlations between the listed market and the stock market if the holding period rises to 5 years. This can be explained by theory since in the long run, the returns on listed real estate should resemble the underlying assets. Stock market sentiment should therefore be filtered out in the long term. To test if the difference between the contemporaneous and the 5 year correlation is significant, the Fisher r-to-z transformation is used. The one tailed test of significance has a power of 0.0344. This means that hypothesis 6 cannot be rejected with 95% certainty.
USA markets
Table 6 provides the correlation figures of the USA market. Graph 6 shows the correlation figures in a bar chart.
Table 6: Correlations USA markets
Correlations NL NL Direct- NL Listed NL Listed - NL Stock NL Direct - NL Stock
Contemptuous 0,04 0,53 -0,01 2 year 0,01 0,60 -0,03 3 year 0,02 0,32 -0,01 4 year 0,06 0,25 0,00 5 year 0,07 0,14 -0,02 -0,10 0,00 0,10 0,20 0,30 0,40 0,50 0,60 0,70 1 2 3 4 5 NL Direct RE - NL Listed RE NL Listed RE - NL Stock NL Direct RE - NL Stock
Correlations USA USA Direct- USA Listed USA Listed - USA Stock USA Direct - USA Stock
Contemperous 0,15 0,35 0,12
2 year 0,29 0,25 0,24
3 year 0,40 0,17 0,30
4 year 0,45 0,16 0,28
Graph 6: Correlations USA markets
In contrast to the Dutch market, the USA market does provide evidence for hypothesis five since an increasing correlation between direct and listed real estate can easily be observed (0.15-0.53). The increase from 0.15 to 0.53 is, using the Fisher r-to-z transformation,
significant with a power of 0.0401. Hypothesis 5 can thus not be rejected with 95% certainty. The correlations between the listed real estate market and the stock market do seem to lessen as the holding period rises. However with a power of 0.1841, the difference is not statistically significant. This means that no evidence for hypothesis 6 was found. The correlations between the direct real estate market and the stock market do not provide any insight.
UK markets
The UK market is more difficult to evaluate since there is less data available. There is only data on listed real estate available from 1984 instead of 1977. The cumulative returns on listed real estate can be calculated, although the 2 year, 3 year, 4 year and 5 year correlations provide less than 30 observations. Table 7 provides the correlation figures of the UK market. Graph 7 shows the correlation figures in a bar chart.
Table 7: Correlations UK markets
Correlations UK UK Direct - UK Listed UK Listed - UK Stock UK Direct- UK Stock
Contemptuous 0,69 0,41 0,36
2 year 0,76 0,39 0,37
3 year 0,80 0,34 0,36
4 year 0,81 0,27 0,32
Graph 7: Correlations UK markets
The analysis of the UK market shows an increase in the correlation between the direct and the listed real estate market as the holding period increases. This difference is, however, only 0.12. The one tailed power of the difference is 0.1635 and therefore statistically
insignificant (using the Fisher r-to-z transformation). The analysis provides thus no evidence for hypothesis 5; the returns on securitized real estate are in the long term more correlated with the returns on direct real estate than the contemporaneous correlation. The correlation between the listed real estate returns and the returns on stock seem to lessen as the holding period increases. The difference between the contemporaneous correlation and the five year correlation is 0.23. This difference has a power of 0.1867(using the Fisher r-to-z
transformation) and is therefore not statistically significant.
Long Term Risk Adjusted Average Returns
To show the effect of investing with a long term horizon, the cumulative returns and their correlations have been discussed. However, returns are only one relevant characteristic of an investment, the volatility is just as important. If the returns on listed and direct real estate do resemble each other in the long term, then the risk adjusted returns should be comparable. In this paragraph, the question of whether or not a longer holding period has a smoothing effect on the volatility can be answered by reviewing the risk adjusted average cumulative returns.
If the longer holding period has a smoothing effect on volatility, then the risk adjusted returns should rise. Since listed real estate is very volatile, the risk adjusted returns on listed real estate should in the long term profit from the smoothing effect on volatility due to a
longer investment horizon. The risk adjusted returns on real estate should therefore mimic the risk adjusted returns on direct real estate in the long run. To calculate this effect, the
contemporaneous, 2,3,4 and 5 year cumulative returns have been calculated for each year. Every different holding period (1,2,3,4 or 5 years) have different volatilities. The average cumulative returns on real estate minus the average cumulative returns on risk free bonds provide the average cumulative excess returns. These excess returns divided by a holding period’s standard deviation produces the Sharpe ratio. If the smoothing effect on real estate on volatility is indeed present, then the average two year risk adjusted average cumulative returns on securitized real estate, should get closer to the average two year risk adjusted average cumulative returns on direct real estate. Consequently, the 3 year, 4 year and 5 year risk adjusted average cumulative returns on direct and listed real estate, should get closer and closer. The following three paragraphs will discuss if this effect is found in the analysis.
Dutch markets
Table 8 provides an overview of the behaviour of the average cumulative risk adjusted returns for the Dutch markets. Graph 8 shows a bar chart of these returns.
Table 8: Risk adjusted average cumulative returns for NL, given by the Sharpe ratio.
Graph 8: Risk adjusted average cumulative returns, given by the Sharpe ratio.
Sharpe ratio's Direct RE Listed RE Stock
1 year cumulative 0,66 0,25 0,37
2 year cumulative 0,72 0,13 0,38
3 year cumulative 0,77 0,09 0,43
4 year cumulative 0,82 0,35 0,60
The smoothing effect on volatility should be the largest in the most volatile markets. For the Dutch markets, the most volatile returns are on stock, followed by listed real estate and direct real estate (Table 4). Looking at the Dutch market, the five year smoothing effect on the volatility has yielded the direct real estate market an increase of 0.24 in the Sharpe ratio. The stock market shows an increase of 0.29 and the listed real estate market only increases with 0.16. This analysis shows that the smoothing effect is not nearly as strong for listed real estate as for general stock. The numbers show that the difference in the one year cumulative risk adjusted average returns between direct and listed real estate is 0.41. The difference of the five year cumulative risk adjusted average returns between direct and listed real estate is 0.59 in the Sharpe ratio. For the Dutch market, a longer holding period does not show a smaller, but a larger spread of the average risk adjusted returns in favour of the direct real estate market
USA markets
Table 9 provides an overview of the behaviour of the average cumulative risk adjusted returns for the USA markets. Graph 9 shows a bar chart of these returns.
Table 9: Risk adjusted average cumulative returns for the USA, given by the Sharpe ratio
Graph 9: Risk adjusted average cumulative returns for the USA, given by the Sharpe ratio
Looking at the market of the USA, the five year smoothing effect on the volatility has yielded the direct real estate market an increase of 0.22 in the Sharpe ratio. The stock market
Sharpe ratio's Direct RE Listed RE Stock
1 year cumulative 0,59 0,46 0,44
2 year cumulative 0,66 0,59 0,61
3 year cumulative 0,71 0,67 0,70
4 year cumulative 0,76 0,69 0,75
shows an increase of 0.35 and the listed real estate market shows an increase of 0.24. This analysis shows thus that the smoothing effect is the strongest for Stock, while listed real estate is more volatile. The gap between listed real estate and direct real estate does gets smaller. the one year average cumulative risk adjusted returns show a difference of 0.14 in the Sharpe ratio. The three year average cumulative risk adjusted returns only differ 0.04 in the Sharpe ratio. The five year average cumulative risk adjusted returns however differ 0.11 in the Sharpe ratio, indicating that Direct RE and Listed RE are more comparable in a three year holding strategy than a five year holding strategy.
UK markets
Table 10 provides an overview of the behaviour of the average cumulative risk adjusted returns for the UK markets. Graph 10 shows a bar chart of these returns.
Table 10: Risk adjusted average cumulative returns for the UK, given by the Sharpe ratio
Graph 10: Risk adjusted average cumulative returns for the UK, given by the Sharpe ratio
The 4 year smoothing effect on the volatility has yielded the direct real estate market an increase of 0.24 in the Sharpe ratio. The stock market shows an increase of 0.25 and the Sharpe ratio of listed real estate increased 0.10. This analysis shows thus that the smoothing effect is very weak for listed real estate. An even more interesting finding is that the 3, 4 and 5 year holding strategy have the exact same Sharpe ratio.
Sharpe ratio's Direct RE Listed RE Stock
1 year cumulative 0,33 0,21 0,31 2 year cumulative 0,39 0,29 0,43 3 year cumulative 0,45 0,31 0,48 4 year cumulative 0,52 0,31 0,52 5 year cumulative 0,57 0,31 0,56 0,00 0,10 0,20 0,30 0,40 0,50 0,60 1 2 3 4 5 UK Direct RE UK Listed RE UK Stock
The numbers show that the difference in the one year cumulative risk adjusted average returns between direct and listed real estate is 0.24. The five year cumulative risk adjusted average returns between direct and listed real estate is 0.26. A longer holding period does not show a smaller but a slightly larger spread of the average risk adjusted returns. For the UK market, based on the average risk adjusted cumulative returns, there is no evidence to state that a longer holding period closes the gap between direct and listed real estate.
Results summarized, NL
The returns on listed real estate are only 0,30% higher than the return on direct real estate. This difference is not statistically significant. The Dutch stock market is very volatile (0,26) compared to the listed (0,19) and direct real estate market (0,05). The difference
between the volatility of listed and direct real estate is found to be statistically significant. The smoothing effect on volatility is expected to be the largest in the stock market returns and correlations. Looking at the cumulative risk adjusted returns, we find this to be accurate. Stocks appear to profit the most (0,29) from the smoothing effect, followed by direct real estate (0,24). The listed real estate market profits (0.17). The gap between the risk adjusted direct real estate returns of direct and listed real estate does not seem to lower when the holding period increases. The results also do not show a significant increase in correlation between direct and listed real estate when the holding period increases. Listed and direct real estate thus, do not seem to have many similarities.
A significant effect is found on the correlations between the direct real estate market, which has a very low volatility of 0,06, and the stock market, which has a high level of volatility (0,26). The correlation drops from the contemporaneous correlation of 0.53 to the five year correlation of 0.12. This difference is found to be statistically significant. The Dutch market provides therefore no evidence to reject hypothesis 6.
Results summarized, USA
The returns on listed real estate are only 4.95% higher than the return on direct real estate. This difference is not statistically significant. Listed real estate in the USA is the most volatile and profitable asset class. The difference in volatility between listed and direct real estate is 0.14. This difference is found to be statistically significant. Because the listed real estate returns are the most volatile, the returns should benefit the most from the volatility smoothing of a long term investment strategy. The results show, us however, that stocks seem to profit more. The Sharpe ratio increases 0.35 for stock, 0.24 for listed real estate and with
0.22 for direct real estate. The gap between the Sharpe ratio’s also seems to lessen once the holding period increases from a one to a three year holding period. The gap in the Sharpe ratio is only 0.04 with a three year holding period. This might be an indication that as the holding period increases, the returns of direct and listed real estate seem to have a stronger
resemblance. The correlation figures support this. The correlation increases from 0.15 with a one year holding period to a correlation of 0.53 with a five year holding period. This
difference is found to be statistically significant. Using the descriptive statistics, the
correlation figures and the Sharpe ratio’s, the listed and the real estate returns seem to have similarities in the long term.
Results summarized, UK
In the UK, the listed real estate returns are by far the most volatile. Listed real estate, direct real estate and the stock market have volatilities of 0,27, 0,10 and 0,16 respectively. The listed real estate returns should therefore benefit the most from the volatility smoothing of a long term investment strategy. Looking at the risk corrected returns, stock profits (0,25) the most from the smoothing of volatility followed closely by direct real estate (0,24). Listed real estate profits only 0,10 from long term volatility smoothing. The gap between the listed and direct real estate returns widens from 0.12 to 0.26 in the Sharpe ratio. A long term investment strategy thus does not increase the similarities between the listed and direct real estate returns. The correlations between the listed and direct real estate market in the UK are very high. As the holding period increases the correlation between direct and listed real estate increases as well. However, the increase is not statistically significant. The listed and the direct real estate market do have some similarities. These similarities, however, do not appear to be dependent on the holding period, but on the market characteristics. Further research should determine what characteristics the listed and direct real estate markets in the UK have that make them seem similar.
Conclusion
This thesis aims to provide an explanation to the question to what extent the returns on
securitized real estate resemble the returns on direct real estate in the long term. In order to do this, first a better understanding of the two markets is provided by reviewing the
characteristics of direct and listed real estate. A literature summary is provided to introduce the subject and to provide criteria to test for resemblance between the direct and listed real estate markets. Hypothesis are formulated and tested using data analysis. The data used are
the annual total returns of stocks, direct real estate and listed real estate. The risk free rate is stated as the three month’s return on government securities of the three different countries.
For the data analysis, first the descriptive statistics are reviewed. Based on statistically non-significant differences in the returns on listed and direct real estate and the statistically significant difference in volatilities between the listed and direct real estate market, for all three countries, the direct and listed real estate returns do not seem to be very similar. To understand to what extend the listed and direct real estate market are correlated, first the cumulative returns for a 1,2,3,4 and 5 year holding period is calculated. The results are different for every county. For the Netherlands, the correlation between direct and listed real estate is minimal and does seems to increase when the holding period increases. No
statistically significant effect of the holding period is found. The correlation between direct and listed real estate in the USA does shows a statistically significant increase between the one and five year holding strategy. The UK is unique in the high (0,69) contemporaneous correlation compared to the Netherlands and USA, 0.04 and 0.15 respectively. A longer holding period does increase the correlation of direct and listed real estate returns, although the difference is not statistically significant. A longer holding period smoothens the average volatility. More volatile asset classes should profit the most from this effect. The smoothing of the volatility should, therefore, cause the Sharpe ratios of the direct and listed real estate returns to become increasingly similar when the holding period, and therefore the smoothing effect, increases. This effect is indeed found in the data analysis for the USA markets.
Increasingly similar Sharpe ratios caused by the smoothing effect on volatility due to a longer holding period were not found for Dutch and UK markets.
Using the descriptive statistics, the correlation figures and the Sharpe ratios, this thesis provides statistically significant evidence to state that the listed and the direct real estate returns in the USA markets in the long term are more similar than in the short term. This evidence is not found for the Dutch and UK markets.
The findings of this thesis study can support further research that should determine if the listed and the direct real estate market in the USA can be substitutes for one another. Further research is also necessary to determine why the UK is unique in the high
(contemporaneous) correlation between the returns on listed and direct real estate compared to the returns on the Dutch and USA market.
Bibliography
Bodie, Z. Kane, A. & Marcus, A.J. (2011). Investments and Portfolio Management. Ninth edition, P. 675-677.
Brounen D. & Eichholtz P. (2003) Property, Common Stock, and Property Shares. The
Journal of Portfolio Management, special edition 2013.
Campeau, F. (1994) A Microstructure Analysis of the Information on Securitized and Unsecuritized Commercial Real Estate Markets, Cambridge University Phd
Clayton, J. & MacKinnon, G. (2001) The Time-Varying Nature of the Link between REIT, Real Estate and Financial Returns, Journal of Real Estate Portfolio Management, 7, P. 43-54.
Gentry, W.M. Jones, C.M., Mayer, C.J.(2004). Do stock prices really reflect fundamental
values?: the case of REITs, NBER Working Paper No. 10850 P. 3-4
Glascock, J.L., Lu, C., So, R.W. (2000) Further Evidence on the Integration of REIT, Bond, and Stock Returns, Journal of Real Estate Finance and Economics, 20, P. 1-9
Gool, van P. (2013). Onroerend goed als belegging, Fifth edition.
Goorbergh, van den R. (2015). Van het zelfde laken een pak. VBA beleggingsprofessionals
journal. N.120 (2014) P.15
Gyourko, J. & Kiem, D. (1992) What Does the Stock Market Tell us about Real Estate Returns, Journal of the American Real Estate and Urban Economics Association, 20, P. 457-485
Hoeslie, M. & Oikarinen, E. (2012) Are REITs Real Estate? Evidence from International Sector Level Data. Research Paper Series N°12 – 15, Swiss Finance Institute
Hoeslie, M. & Oikarinen, E. (2014) Are Public and Private Asset Returns and Risks the Same? Evidence from Real Estate Data. Research Paper Series N°13 – 56
Hudson-Wilson, S (2003). Why real estate? An expanding role for institutional investors. The
Journal of portfolio management, special real estate issue, P. 12-25.
Senior, C. (2011). A meeting of two worlds. IP Real Estate magazine
Stevenson, S & Lee, S (2005) “The Case for REITs in the Mixed-Asset Portfolio in the Short and long run”. Journal of Real Estate Portfolio Management; Jan-Apr 2005, Vol. 11 issue 1, P. 2-3
Appendices
Appendix 1: Total Annual Returns
To ta l r e tu rn s R F ra te N L R F ra te U SA R F ra te U K EPR A N A R EIT U SA N C R EIF U SA U SA S P5 00 U K D ir e ct IPD U K L ist e d G PR G U K S to ck F TS E1 00 IPD N L G PR G N L St o ck N L A EX 1977 4 ,5 7 % 5, 13 % 13 ,0 8% -7 ,1 8% 24 ,6 8% 20 ,9 0% 7, 69 % 1978 6 ,5 4 % 6, 93 % -4 ,2 4% 16 ,1 1% 6, 56 % 16 ,0 2% 14 ,3 0% 4, 73 % 1979 9 ,4 7 % 9, 94 % 30 ,9 9% 20 ,4 6% 18 ,4 4% 8, 66 % 0, 90 % 15 ,4 1% 1980 1 0 ,6 9 % 11 ,2 2% 39 ,2 8% 18 ,0 9% 32 ,5 0% 14 ,5 8% 8, 40 % 27 ,2 2% 1981 1 1 ,6 3 % 14 ,3 0% 21 ,4 2% 16 ,6 2% -4 ,9 2% 14 ,9 8% 0, 80 % 5, 10 % -0 ,3 2% 1982 8 ,3 2 % 11 ,0 1% 31 ,0 3% 9, 43 % 21 ,5 5% 7, 52 % 5, 44 % 19 ,5 0% 24 ,7 3% 1983 5 ,6 2 % 8, 45 % 52 ,2 6% 13 ,1 3% 22 ,5 6% 7, 57 % 4, 00 % 23 ,0 0% 73 ,4 9% 1984 6 ,1 0 % 9, 61 % 9, 29 % 40 ,3 7% 13 ,8 4% 6, 27 % 8, 83 % 18 ,6 8% 23 ,5 6% 0, 87 % 13 ,4 7% 21 ,2 0% 1985 6 ,3 2 % 7, 49 % 11 ,5 6% -7 ,5 8% 11 ,2 4% 31 ,7 3% 8, 30 % 3, 42 % 14 ,6 1% 2, 44 % 5, 96 % 47 ,6 7% 1986 5 ,6 1 % 6, 04 % 10 ,3 5% -5 ,3 8% 8, 30 % 18 ,6 7% 11 ,2 7% 18 ,9 3% 18 ,8 6% 7, 53 % 7, 33 % -2 ,1 9% 1987 5 ,3 1 % 5, 72 % 9, 23 % -2 2, 02 % 8, 00 % 5, 25 % 25 ,9 7% 34 ,5 5% 6, 16 % 4, 74 % 5, 58 % -2 9, 71 % 1988 4 ,7 7 % 6, 45 % 9, 84 % 28 ,2 0% 9, 62 % 16 ,6 1% 29 ,5 1% 25 ,1 9% 10 ,1 5% 8, 33 % 13 ,0 6% 57 ,9 0% 1989 7 ,3 3 % 8, 11 % 13 ,2 9% 3, 90 % 7, 77 % 31 ,6 9% 15 ,3 9% 3, 53 % 41 ,5 5% 8, 54 % 8, 49 % 20 ,8 0% 1990 8 ,6 2 % 7, 55 % 14 ,0 9% -2 5, 32 % 2, 30 % -3 ,1 0% -8 ,4 1% -1 4, 18 % -6 ,6 1% 13 ,5 6% -2 7, 07 % -2 0, 32 % 1991 9 ,2 1 % 5, 61 % 10 ,8 2% 37 ,9 9% -5 ,5 9% 30 ,4 7% -3 ,1 0% -1 1, 04 % 22 ,0 2% 9, 86 % 6, 21 % 25 ,7 3% 1992 9, 29 % 3, 41 % 8, 91 % 16 ,3 8% -4 ,2 6% 7, 62 % -1 ,6 4% -1 1, 23 % 19 ,7 6% 8, 30 % -1 1, 59 % 6, 93 % 1993 6, 75 % 2, 98 % 5, 21 % 19 ,5 6% 1, 38 % 10 ,0 8% 20 ,1 9% 91 ,5 5% 25 ,1 9% 8, 67 % 40 ,2 2% 49 ,4 0% 1994 5 ,1 2 % 3, 99 % 5, 16 % 3, 40 % 6, 39 % 1, 32 % 11 ,8 7% -1 7, 64 % -6 ,5 1% 2, 61 % -1 4, 03 % 3, 70 % 1995 3 ,5 6 % 5, 52 % 6, 33 % 14 ,8 9% 7, 54 % 37 ,5 8% 3, 59 % 5, 54 % 25 ,9 7% 10 ,8 4% 1, 60 % 21 ,2 0% 1996 3 ,1 3 % 5, 02 % 5, 78 % 36 ,0 2% 10 ,3 1% 22 ,9 6% 10 ,0 3% 30 ,2 1% 16 ,8 6% 11 ,6 9% 19 ,4 6% 37 ,7 8% 1997 3 ,5 3 % 5, 05 % 6, 50 % 20 ,0 7% 13 ,9 1% 33 ,3 6% 16 ,7 6% 22 ,9 1% 28 ,6 8% 12 ,3 6% 14 ,9 0% 44 ,3 4% 1998 3 ,2 4 % 4, 73 % 6, 81 % -1 6, 96 % 16 ,2 4% 28 ,5 8% 11 ,7 6% -1 8, 11 % 17 ,4 7% 14 ,2 9% -1 ,3 2% 32 ,3 2% 1999 3 ,2 6 % 4, 51 % 5, 04 % -4 ,3 3% 11 ,3 6% 21 ,0 4% 14 ,5 0% 15 ,7 4% 20 ,5 9% 16 ,0 0% 6, 49 % 27 ,3 0% 2000 4 ,8 6 % 5, 76 % 5, 80 % 29 ,6 1% 12 ,2 4% -9 ,1 0% 10 ,4 5% 23 ,1 0% -8 ,2 3% 17 ,3 1% 8, 54 % -3 ,3 0% 2001 3 ,3 1 % 3, 67 % 4, 76 % 13 ,7 5% 7, 28 % -1 1, 89 % 6, 79 % -3 ,9 2% -1 4, 09 % 11 ,7 0% 3, 18 % -1 8, 71 % 2002 2 ,8 8 % 1, 66 % 3, 86 % 3, 59 % 6, 74 % -2 2, 10 % 9, 64 % -0 ,0 1% -2 2, 17 % 8, 70 % 11 ,6 5% -3 4, 49 % 2003 2 ,1 1 % 1, 03 % 3, 56 % 36 ,6 7% 8, 99 % 28 ,6 8% 10 ,8 5% 29 ,9 2% 17 ,8 9% 7, 10 % 20 ,7 8% 8, 44 % 2004 2 ,1 4 % 1, 23 % 4, 44 % 32 ,0 6% 14 ,4 8% 10 ,8 8% 18 ,3 3% 47 ,5 0% 11 ,2 5% 7, 70 % 38 ,8 9% 6, 50 % 2005 2 ,4 4 % 3, 01 % 4, 55 % 11 ,9 2% 20 ,0 6% 4, 91 % 19 ,1 0% 20 ,4 8% 20 ,7 8% 10 ,0 0% 17 ,0 0% 31 ,7 0% 2006 3 ,6 7 % 4, 68 % 4, 65 % 35 ,3 5% 16 ,5 9% 15 ,7 9% 18 ,1 0% 45 ,9 8% 14 ,4 3% 12 ,2 0% 42 ,1 8% 17 ,2 0% 2007 4 ,4 9 % 4, 64 % 5, 53 % -1 5, 59 % 15 ,8 4% 5, 49 % -3 ,4 2% -3 5, 02 % 7, 36 % 11 ,0 0% -1 0, 30 % 7, 50 % 2008 2 ,9 0 % 1, 59 % 4, 28 % -3 7, 56 % -6 ,4 6% -3 7, 00 % -2 2, 10 % -4 7, 39 % -2 8, 33 % 3, 40 % -3 9, 34 % -5 2, 32 % 2009 0 ,6 0 % 0, 14 % 4, 89 % 28 ,4 7% -1 6, 86 % 26 ,4 6% 3, 51 % 19 ,0 2% 27 ,3 3% 0, 00 % 41 ,0 6% 42 ,1 0% 2010 1 ,0 0 % 0, 13 % 4, 96 % 27 ,7 0% 13 ,1 1% 15 ,0 6% 15 ,0 9% 4, 49 % 12 ,6 2% 4, 20 % 14 ,7 8% 9, 20 % 2011 1 ,3 6 % 0, 03 % 4, 84 % 7, 54 % 14 ,2 6% 2, 11 % 7, 84 % -8 ,0 3% -2 ,1 8% 3, 80 % -2 3, 95 % -8 ,8 0% 2012 0 ,1 9 % 0, 05 % 0, 31 % 19 ,5 7% 10 ,5 4% 16 ,0 0% 3, 42 % 27 ,7 9% 9, 97 % 1, 20 % 8, 30 % 13 ,5 5% 2013 0 ,2 8 % 0, 07 % 0, 30 % 2, 46 % 10 ,9 9% 32 ,3 9% 10 ,7 3% 25 ,3 0% 18 ,6 6% 0, 50 % 9, 20 % 17 ,2 3% 2014 0 ,0 8 % 0, 05 % 0, 38 % 29 ,4 8% 11 ,8 1% 13 ,6 9% 17 ,8 0% 22 ,9 8% 0, 74 % 4, 40 % 22 ,6 1% 5, 65 %
