Evaluation of Real Estate Returns Across High Volatility Periods.doc

Document Sample
Evaluation of Real Estate Returns Across High Volatility Periods.doc Powered By Docstoc
					                       Excess Return and Risk Characteristics
                        of Asian Exchange-Listed Real Estate

                    John L. Glascock, Raymond W. So and Chiuling Lu1

                                      Current Version: July 12, 2012


          This research examines the behavior of excess returns of publicly listed real estate
          firms traded in Japan, Taiwan, Hong Kong, South Korea, Singapore and Thailand
          across different market conditions. Results indicate that, apart from Taiwan, publicly
          traded real estate firms in these markets do not exhibit excess returns behavior.
          Additionally, excepting South Korea, the risk characteristics of exchange listed real
          estate firms change across market conditions.

JEL Classification: F21

Key Words: Real estate, Market conditions, Asian markets, Risk characteristics

I. Introduction

The 1997 Asian financial crisis severely limited new investment in Asian economies. However,
before the crisis, international investors were attracted to the emerging markets of Asia in the
hope of finding underpriced real properties and/or securities, as well as for diversification
opportunities. Of key interest to academics and investment professionals is the question of how
well these economies, through their listed securities, are able to provide enhanced diversification
opportunities. Most prior research has concentrated on three issues: one, do markets move
together; two, can diversification be achieved without trading abroad; and three, can

  John L. Glascock is affiliated with George Washington University, Raymond So is affiliated with Chinese
University of Hong Kong, and Chiuling Lu is affiliated with Yuan Ze University (Taoyuan, Taiwan). Address
correspondence to John L. Glascock, Department of Finance, George Washington University, 2023 G Street, 540b
Lisner Hall, Washington, DC 20052, USA, Voice: (202) 994-7667, Fax: (202) 994-5014, Email:
The authors are grateful to Sabrina Lai, Christine Chung and Andrew Lau for their research assistance. So also
acknowledges the support from a Student Campus Work Scheme, Shaw College, the Chinese University of Hong
diversification be achieved by using closed-end country funds?2 However, in this research, we
concentrate on the excess returns and market-related risk behavior of a limited class of stocks:
real estate.
        Previous research on US real estate securities shows that a key attribute of real estate
investments is its low risk profile. Howe and Shilling (1990), Chan, Hendershott and Sanders
(1990) and Glascock and Hughes (1995) find that systematic risk for securitized real estates is
lower than market risk. This low risk profile is consistent with the observed low rates of return
that real estate firms demonstrate compared with market indices.
        A second interesting aspect of securitized real estate return is its risk behavior during
periods of high volatility. In terms of non-securitized real estate, White (1972) and Kau and
Sirmans (1979) document that Manhattan (New York) lands did not regain their pre-depression
values until 1970. Stock prices recovered much faster after the 1929 Great Depression than did
real estate prices. However, more recently, Mei (1999) indicates that the New York real estate
market recovered soon after the early 1990s crash, and that extraordinary returns were generated.
However, it appears that securitized real estate, at least recently, behaves differently and in a
manner that is consistent with defensive stocks3. This defensive behavior is consistent with the
work of Glascock (1991) where he finds that real estate betas become lower during recessions.
Also, Glascock, Michayluk and Neuhauser (2001) indicate that during the October 1997 stock
market decline, the value of real estate investment trusts (REIT) declined by only about one half
that of the overall market. Thus, we will consider the behavior of real estate listed securities in
emerging Asian countries and in Japan during varying market conditions.
        Compared with the real estate market in the US, Asia is characterized by land scarcity
and high population density, and thus real estate values are relatively high. There is a strong
desire to own real estate, and the recent history of accelerating economic prosperity is persuading
people that the price of real estate will not decline. This traditional belief has long convinced
most people that land ownership provided a relatively risk-free asset. This faith in real estate
seems to be evident also in the prices of securitized real estate. Miller (1998) indicates that the
credit risk that includes borrowing short-term dollars to invest in over-built local real estate was

 For each of these areas, see Karolyi and Stulz (1996), Errunza, Hogan and Hung (1999) and Bekaert and Urias
(1996) respectively.
at the heart of the Asian financial crisis. After the Asian financial crisis, real estate prices in
Asian economies plunged.
         We analyze the returns of listed real estate firms traded in Japan and five emerging Asian
markets (Taiwan, Hong Kong, South Korea, Singapore and Thailand). We examine whether or
not the risk characteristics of Asian real estate markets differ from those of the US. We also
examine the returns behavior of real estate in these Asian markets across differing market
         Our results show, in agreement with Glascock (1991), that there are no excess returns for
listed real estate firms across all markets except Taiwan. In addition, generally consistent with
Glascock (1991), the risk characteristics of exchange-listed real estate firms change with market
condition changes; however, different markets exhibit distinct variations. Additionally,
securitized real estate has not recovered in these Asian markets since the crisis.
         The rest of the paper is organized as follows. Section II provides a review of the literature
and the expectations; Section III contains a discussion of the sample and statistical procedures;
Section IV presents the results and Section V concludes the paper.

II. Expectations and Prior Research

Early research on real estate returns behavior tended to find excess returns for real estate 4. Many
of these works are limited owing to the use of appraisal data, and, more importantly, recent work
tends to find that real estate earns normal (risk-adjusted) returns5. Gyourko and Keim (1993) use
a set of samples that includes not only REITs but also general contractors, agents, managers, and
so on, to investigate the risk/return. They find that the returns of real estate-related firms are
dominated by a strong covariance with the stock market, and suggest that there is no excess
return behavior. Glascock (1991) finds that market conditions change REIT performance. During
up markets, REITs exhibit higher betas; and during down markets, they have lower betas.
However, on average, REITs under-perform the market on a nominal basis and earn fair returns

  This has been extensively reported in the popular press. See, for example, “REITs: the Antistocks,” Smart Money
(February 2002), 87. During 1998 and 1999, REIT values fell as the market continued to climb. However, in 2000,
REIT values increased as stocks declined significantly.
  See Sirmans and Sirmans (1986-87), who provide a review of the real estate return literature, indicating that most
studies find that real estate returns exceed those of stock and bond investments.
on a risk-adjusted basis. Such return behavior is supported by the work of Glascock and Hughes
(1995), who find that REIT betas are consistently below 1; 0.377 for the entire period (1972–
1991) and 0.343 annually. These findings—consistent with the prior work of Hendershott and
Sanders (1990), who examine equity REIT returns—suggest that REITs have a performance that
is explained by their betas. Additionally, they are consistent with considering securitized real
estate as a defensive type of investment. However, little is known about the behavior of foreign
securitized real estate.
         For emerging Asian markets, real estate plays a very crucial role in individuals’ investment
portfolios for several reasons. First, Asian capital markets were not well established until recently.
Therefore, not many securities or vehicles were available for investment. Second, real estate has
long been treated as a key risk-free asset for generations because of limited supply. Third, real estate
is viewed as the best collateral in the bank lending process. Accordingly, real estate became the most
favored investment target within Asia, and the real estate market is closely tied to capital markets.
Therefore, we expect the stock market and the real estate sector to be highly correlated within each
economy. In addition, we hypothesize that because Asian economies (excepting Japan) are still
developing with limited urban land supplies, real estate investment will still be attractive after the
Asian financial crisis. Therefore, there may be no structural change after the crisis.
         This research concentrates on publicly traded real estate-related firms from six Asian
markets. By using firms with publicly traded information, we have the advantage of a standard risk
measure (beta), whereas non-listed firms may have other risks (e.g., liquidity risk) that prevent the
beta from adequately representing risk. Furthermore, excess return behavior can be influenced by
market conditions (Klein and Rosenfeld, 1987). Thus, excess returns behavior found by previous
research may be due to risk shifting across market conditions. For example, Cumby and Glen (1990)
test the hypothesis that international mutual funds time their portfolios around market conditions,
using a similar methodology to that used here. Therefore, this paper explicitly considers alternative
market conditions in the tests for excess returns, to exclude any bias resulting from time

 See, for example, the work of Glascock and Davidson (1995), who analyze real estate and related firms for the period
1977 to 1986 and find no excess return using the Sharpe and Treynor indices.
III. Sample and Testing Procedure

The sample consists of real estate-related firms that are publicly traded in the stock markets of
Taiwan, Hong Kong, Singapore, South Korea, Japan and Thailand. The equally weighted return
of a portfolio of publicly traded real estate firms in the respective markets represents the monthly
real estate portfolio returns in each economy.
         The Taiwanese sample includes 34 companies listed on the Taiwan Security Exchange
(TSE) for the period from March 1982 to July 1999. These firms include builders, developers
and contractors who are assigned industry code 25. Monthly data are retrieved from the Taiwan
Economic Journal Data Bank (TEJDB). Monthly returns on the three-month interbank rate are
used to approximate the risk-free returns, and are collected from the Interbank Money Center.
Market returns are approximated by the Weighted Average Index of the Taiwan Stock Exchange
(also from TEJDB).
         The Hong Kong sample consists of 112 real estate firms listed on the Stock Exchange of
Hong Kong (SEHK). The SEHK classifies listed firms into the following categories: finance,
utilities, properties, consolidated enterprises, industrials, hotels and others. The SEHK's property
classification is used in identifying these real estate firms.
         The sample period is from June 1986 to March 1999, covering 13 years of monthly data.
Stock price data are collected from the Datastream. Since there were no government securities
(known as Exchange Fund Bills) in Hong Kong until the late 1980s, the three-month Euro–Hong
Kong Dollar Deposit average rate is used as the risk-free rate.6 Data on three-month Euro–Hong
Kong Dollar interest rate are collected from various issues of the Hong Kong Monthly Digest of
Statistics. Market returns are defined as the monthly return of the Hang Seng Index.

  Exchange Fund Bills are issued by the Hong Kong Monetary Authority (HKMA) and have maturities of less than
five years. The HKMA determines the coupon rate of the bill and the amount to be raised. Usually, the bills are
issued through auction, and buyers are mainly institutional investors who usually hold the bills until their maturities.
Thus, the secondary market for the bills is thin, and it is not practical to use the secondary market prices of the bills
to represent the risk-free rate in Hong Kong.
       Data on monthly returns of the Japanese real estate firms are also taken from the Datastream.
The Tokyo Stock Exchange classifies firms into 33 different industries. For this classification, we
identify 162 real estate-related firms and construction companies. For consistency with Taiwan and
Hong Kong, the average of three-month Euro–Yen deposit rates is used for the risk-free rate. In
addition, returns for the Japanese stock market are measured by the change of the Nikkei 225
Stock Index. Japanese interest rates and the Nikkei 225 Index data are provided by Datastream.
       The Korean Stock Exchange classifies listed firms into 40 different industries, ranging
from agricultural sectors, such as fishing, to heavy industry steel firms. Under this classification,
54 real estate-related firms are included in our study. Stock return data on these firms are taken
from the Datastream. Since interbank interest rates are not available, the Korean risk-free rate is
represented by the yield on government bonds. Monthly returns of the Korean stock market are
approximated by the monthly change in the Korea Composite Stock Price Index (KCSPI). Data
on the bond yield and the KCSPI are from the Datastream also.
       The Singapore Stock Exchange classifies firms into the following industry groupings:
industrial, finance, hotels, properties, plantation, tins and mines, sterling rubbers and plantations.
Twenty-two firms in the properties group are included in our sample. Stock return data are from
the Datastream database. The Singaporean risk-free rate is proxied by the three-month Euro-
Singapore Dollar deposit rate. Monthly returns of the Singaporean stock market are measured by
changes in the Strait Times Index. Data on the Strait Times Index and interest rates are also taken
from the Datastream.
       Finally, the Stock Exchange of Thailand (SET) has 31 industry classifications, ranging
from agribusiness to banking and investments. Using the SET definition of industries, 40
companies are included in the real estate portfolio. Stock returns on these real estate firms are
from Datastream. The Thai risk-free rate is proxied by the three-month interbank rate, and the
monthly change of the Stock Exchange of Thailand Index is used to measure stock market
performance. Datastream provides the data on interest rates and the market index. Table 1
provides a summary of the sample period, number of firms and other details of the data.
                                      [Insert Table 1 About Here]
       Japan is the only developed country that we study. The motivation for including Japan in
our sample is to have a benchmark for comparison, and to examine whether differences exist
between developing Asian areas and Japan, the significant developed country in the region.
Testing Procedure
Following the work of Merton (1981), Henriksson and Merton (1981), Fabozzi and Francis
(1977) and Cumby and Glen (1990), a dummy variable regression procedure is used to test for
changes in portfolio betas during bull and bear markets. In this work, an up market is defined as
one in which the market return exceeds the risk-free rate. Results are qualitatively the same if
other measures of an up market (e.g., leading indicators, market return exceeds historical average,
etc.) are used, but these are not reported.
        The basic model (in excess return form) is presented in Equation (1).

                   Rpt – Rft =  + u d + p (Rmt – Rft) + u d (Rmt – Rft) + t                    (1)

where Rpt is real estate portfolio return, Rft is the risk-free security return, Rmt is the market index
return, d is a dummy variable taking a value of one for an up market and zero otherwise, and t is
the random disturbance. The use of dummy variables for the alpha and beta during the up-market
periods allows us to test the hypothesis of no difference in excess returns and change in risk
across different market conditions. Equation (1) is estimated by ordinary least squares and the
results are qualitatively the same with or without White’s (1981) correction for heteroskedascity.
Results reported in the next section are based on those without White’s correction. Other
estimation methods, such as robust regression (allowing for outliers) and weighted least squares,
are used and the results do not change markedly as the estimation method changes.
        In Equation (1) p is the systematic risk of the real estate portfolio. Alpha () in Equation
(1) measures the “excess return” that cannot be explained by market movement (or the
systematic risk of the real estate portfolio). Theoretically, alpha should be zero because the
market will not price unsystematic risk as it can be diversified away. Coefficients u and u in
Equation (1) capture the effects of differences in market conditions on excess returns and
systematic risk, respectively. The null hypothesis to be tested is that there are no differences in
alphas or betas between the bear- and bull-market periods, i.e., u = 0 and u = 0. If u is greater
(smaller) than zero, it means that real estate firms have significant positive (negative) excess
returns in up markets.
       Similarly, if u is greater (smaller) than zero, it indicates that systematic risk of real estate
increases (decreases) in up markets.
       During the Asian Financial Crisis that began in July 1997, real estate prices plunged in
these Asian economies. For example, the real estate index in Taiwan decreased from 536 to 202
during the period from July 1997 to September 1999, and residential property prices in Hong
Kong in mid-1998 were down by 50 percent from their 1997 peak. To test whether or not the
Asian Turmoil created a structural change in the real estate markets of the region, we implement
the Chow (1960) test to detect any change in excess return behaviors. The data set is separated
into two subsets with July 1997 as the mid-point, and Equation (1) is run on these subsets of data.
The Chow test statistic (CHOW) is defined as:
                                     (SSE whole - SSE before - SSE after ) / K
                          CHOW                                                ,                    (2)
                                       (SSE before  SSE after) / (T - 2 K)

where SSEwhole, SSEbefore, SSEafter, are, respectively, the sums of squares of the regression in
Equation (1) for the whole period, the period before the financial crisis and the period after it. T
is the number of observations for the whole period, and K is the number of parameters to be
estimated, being four in this case. The null hypothesis is that there is no structural change before
and after the Asian Financial Turmoil, and CHOW is distributed as an F(K, T - 2K) variable.

IV. Empirical Results and Discussions

Descriptive Statistics and Market Conditions

                               [Insert Figure 1 and 2 About Here]
From Figures 1-1 to 1-6, we observe that the real estate markets in the six areas are closely
related to the capital market before the financial crisis. The capital markets were adversely
affected by this crisis, but they recovered quickly. However, real estate securities have not
recovered, especially in Thailand and Taiwan, and the spread between the equity market index
and the real estate markets index has increased. From Figure 2, we observe that although the US
real estate market plummeted during the 1997 crisis, it started to recover by the end of 1999.
                                     [Insert Table 2 About Here]

       Descriptive statistics of real estate returns are shown in Table 2. In this table, we observe
that the real estate returns of these six economies are positively correlated with their respective
markets, although the magnitudes of co-movements differ from each other, with the highest
being for the Taiwanese market (0.882) and the lowest being for the Korean market (0.594).
Additionally, Thai real estate shows the highest standard deviation (twice that of Japan),
indicating that the Thai real estate market is the most risky among these six economies. Japan
has the least variance and the lowest mean return. In addition, excepting South Korea, Japan’s
real estate market is the least correlated with its capital market, compared with the other
economies. For the US market, the mean monthly return of the NAREIT index is only 0.7
percent and its correlation with the S&P500 market index is 0.544. Obviously, among these six
Asian economies and the US market, we observe that the more mature the economy, the more
diversified the investment basket, and the less the correlation between the respective capital and
real estate markets. Consequently, the volatility—and therefore the risk and returns—decreases.
These preliminary statistics show that the real estate returns/risk behavior exhibits substantial
differences even within Asia, and generalization of previous evidence from the US and these
Asian economies may not be applicable to any individual economy.

                                     [Insert Table 3 About Here]

       The market conditions of the six Asian economies are presented in Table 3. Note that the
Hong Kong and Singaporean markets show the largest number of up-markets. These two
economies have more than 50 percent of their months designated as experiencing an up market.
For the Taiwanese and Japanese markets, around 50 percent of the time is classified as up-market.
Finally, the Korean and Thai markets show the least number of up markets (only about 44

Change in Risk and Excess Return Across Market Conditions
Our results of the dummy variable regressions are presented in Table 4. Table 4 also depicts the
significance of the parameters, as well as the explanatory power of the regression, which is
measured by the adjusted R-squared and F-statistic. There are several interesting observations.
First, four markets (Hong Kong, Japan, Singapore and Thailand) exhibit similar market
behaviors, while the behaviors of the South Korean and Taiwanese markets are different. Second,
the adjusted R-squareds for Equation (1) are high for each market (except South Korea),
indicating that Equation (1) can reasonably explain the variations in real estate returns. The
results are consistent with our hypothesis that the stock market and the real estate sector are
closely related.

                                      [Insert Table 4 About Here]

       For the Taiwanese market, the dummy variable coefficient u is negative and significant,
i.e., there are systematic risk changes across market conditions. The real estate portfolio beta
decreased in an up market and increased in down markets. This implies that real estate
investment becomes more risky when the market performs worse than the risk-free rate. This is
contrary to the finding of Glascock (1991) for US real estate firms: his sample showed a pro-
cyclical pattern of beta change. We observed that in Taiwan the beta decreases from 0.9617
during the bear market to 0.727 during the bull markets. This implies that a one percent increase
in the market risk premium will increase the expected risk premium on stock by only 0.727
percent, i.e., a 0.2347 percent decrease relative to bear market returns.
       The beta of the Taiwanese real estate portfolio decreases during prosperity and increases
during recession. This observation could be due to the liquidity change. During bull markets,
investors are more willing to buy or sell real estate because of the expectation of higher prices in
the near future. Consequently, the liquidity risk that is the most crucial in real estate transactions
decreases at the same time. With less risk, investors could be compensated by less return. On the
other hand, investors are reluctant to hold illiquid assets when the market is depressed.
Accordingly, risk compensation must be increased. Taiwan is the only market with a significant
u. This result implies that excess returns exist during bull markets. The adjusted R-square shows
that around 80 percent of the variation of real estate portfolio returns can be explained by our
model specification, and this is the highest of the six economies. Singapore also exhibits a
relatively high R-square value (80 percent). The supply of land is, of course, more limited in
Singapore than in most of the other economies studied, and to some extent real estate values
have not fallen as much in this market7. The third highest R-square value is for Hong Kong, an
island economy with very limited land resources.
           The adjusted R-square for South Korea indicates that the market explains only 36.51
percent of the variation in the real estate market. In addition, there are no excess returns and
there is no change in risk across market conditions. A similar situation is observed for Japan,
whose return on the real estate portfolio is less explained by the market (less than 60 percent)
compared with the other economies studied here.
           Empirical tests for Hong Kong, Japan, Singapore and Thailand provide results that are
consistent with Glascock (1991). First, real estate firms in these markets do not earn excess
returns in up markets, as the alphas for the up markets are not significant. Further, the
significant and positive beta coefficients for up markets indicate that when the markets rise, the
riskiness of real estate also rises. Nevertheless, the sensitivities of the up-market beta
coefficients are not uniform across the four economies. It varies from 0.4279 (Hong Kong) to
1.2796 (Thailand). This finding suggests that although real estate in these four economies does
not have excess returns during up markets, the risk propensity, and hence sensitivity, of real
estate to market conditions is market-specific. This result further suggests that local forces
(market-specific factors) have important impacts on the behaviors of the real estate returns of
the respective markets.
           Overall, our results support our no-excess return expectation and show that systematic
risk changes with market condition changes. However, the evidence also suggests that real
estate returns have different risk sensitivities and responses to respective market conditions.
Apparently, market-specific information still dominates the risk/return behavior of individual
real estate markets. This finding is intuitively appealing, because real estate is not a mobile asset,
and perfect substitutes are not readily available. The importance of local factors in determining
real estate return behaviors is consistent with these special features of real estate.

Impact of Asian Financial Crisis on Real Estate Returns
As discussed in the previous section, these Asian economies suffered significant declines in real
estate prices during the Asian Financial Crisis. To test whether this financial turmoil changed the

    See Figure 1-3 for the Singaporean case.
behavior of real estate returns, we separate the data into pre-crisis and post-crisis samples and then
run the Chow test as shown in Equation (2). The null hypothesis is that there is no structural change,
i.e., the real estate market behaves the same before and after the Asian Financial Crisis.

                                    [Insert Table 5 About Here]

        Results are presented in Table 5. Notice that the CHOW statistics are not significant,
indicating that there is no structural change in the returns behavior for these Asian markets.
Furthermore, this finding assures us that having an extreme market period in our sample does not
bias our empirical results.

V. Summary and Conclusions
In this research, we analyze exchange-listed real estate firms for six Asian economies (Taiwan,
Hong Kong, Japan, South Korea, Singapore and Thailand). A dummy variable test is used to
determine whether there are abnormal returns and risk shifts associated with various market
conditions. Generally speaking, the results indicate that no abnormal returns are associated with
changing market conditions in any market except for Taiwan. This is consistent with the results of
Glascock (1991) for US real estate securities. Market efficiency holds for the securitized real estate
market, at least for the US, Japan and the emerging markets (except for Taiwan).
        As for risk shifting, only South Korea shows no change in beta across different market
conditions. Unlike other economies, Taiwanese real estate firms show a decreased beta during up
markets. Findings for Singaporean, Hong Kong, Japanese and Thai firms are consistent with US
firms that exhibit beta increases in bull markets. Nevertheless, during down markets, all countries
show similar levels of risk (beta), below unity (except for South Korea, with 1.07). Betas vary from
0.68 in Hong Kong to 0.96 in Taiwan.
        No structural change is observed before and after the Asian Financial Turmoil for all
markets. Observed contrary beta behavior between the Taiwanese and the other economies in bull
markets may be caused by cultural difference—i.e., different investment behaviors and preferences.
Particular regulations concerning real estate transactions within each area, and different investment
environments, may also contribute to this observation.
Allen, P. and C. F. Sirmans, “An Analysis of Gains to Acquiring Firms' Shareholders: The
    Special Case of REITs.” Journal of Financial Economics 18 (1987), 175–184.

Bekaert, G. and M. Urias, “Diversification, Integration and Emerging Market Closed-End
   Funds.” Journal of Finance 51 (July 1996), 835–869.

Brueggeman, W. B., A. H. Chen, and T. G. Thibodeau, “Real Estate Investment Funds:
   Performance and Portfolio Consideration.” AREUEA Journal 12 (Fall 1984), 333–354.

Chan, K. C., P. Hendershott and A. Sanders, “Risk and Return on Real Estate Evidence from
   Equity REITs.” AREUEA Journal 18 (Winter 1990), 431–452.

Chang, E. C., and W. G. Lewellen, “Market Timing and Mutual Fund Investment Performance.”
   Journal of Business (January 1984), 57–72.

Cumby R. and J. Glen, “Evaluating the Performance of International Mutual Funds.” Journal of
  Finance 45 (1990), 497–521.

Chow, G., “Tests of Equality Between Sets of Coefficients in Two Linear Regressions.”
   Econometrica 28 (1960), 591–605.

Errunza, V., K. Hogan and M. Hung, “Can the Gains from International Diversification be
   Achieved without Trading Abroad?” Journal of Finance 54 (December 1999), 2075–2107.

Fabozzi, F. J., and J. C. Francis, “Stability Tests for Alphas and Betas over Bull and Bear Market
   Conditions.” Journal of Finance 32 (September 1977), 1093–1099.

Firstenberg, P. M., S. A. Ross and R. C. Zisler, “Real Estate: The Whole Story.” Journal of
    Portfolio Management 14 (Spring 1989), 22–34.

Geltner, G., “Bias in Appraisal-Based Returns.” AREUEA Journal 17 (Fall 1989), 338–352.

Giliberto, M., “A Note on the Use of Appraisal Data in Indexes of Performance Measurement.”
    AREUEA Journal 16 (Spring 1988), 77–83.

Glascock, J. L. Market Conditions, Risk, and Real Estate Portfolio Returns: Some Empirical
   Evidence. Journal of Real Estate Finance and Economics, Vol. 4, No. 4, December 1991,

Glascock, J. L., D. Michayluk, and K. Neuhauser. “The Riskiness of REITs Surrounding the
   October 1997 Stock Market Decline.” working paper, George Washington University,
   February 2001.

Gujarati, D., “Use of Dummy Variables in Testing for Equality Between Sets of Coefficients in
   Two Linear Regressions: A Note.” American Statistician (February 1970), 50–52.
Gyourko, J. and D. B. Keim, “Risk and Return in Real Estate: Evidence from a Real Estate
   Index.” Financial Analysts Journal 49 (September/October 1993, 39–46.

Hartzell, D. J., J. S. Hekman, and M. E. Miles, “Real Estate Returns and Inflation.” AREUEA
   Journal 15 (Spring 1987), 617–637.

Henriksson, R. D., and R. C. Merton, “On Market Timing and Investment Performance:
   Statistical Procedures for Evaluating Forecasting Skill.” Journal of Business (October 1981),

Howe, J. D. and J. D. Shilling, “REIT Advisor Performance.” AREUEA Journal 18 (Winter
  1990), 479–500.

Ibbotson, R. G. and L. B. Siegel, “Real Estate Returns: A Comparison with Other Investments.”
   AREUEA Journal 12 (Fall 1984), 219–242.

Karolyi, G. and R. Stulz, “Why Do Markets Move Together? An Investigation of U.S.–Japan
   Stock Return Co-movements.” Journal of Finance 51 (July 1996), 951–986.

Kau, J. and C. F. Sirmans, “Urban Land Value Functions and the Price Elasticity of Demand for
   Housing.” Journal of Urban Economics (February 1979), 112–121.

Klein, A., and J. Rosenfeld, “The Influence of Market Conditions on Event-Study Residuals.”
   Journal of Financial and Quantitative Analysis 22 (September 1987), 345–351.

Kuhle, J. L., “Portfolio Diversification and Return Benefits – Common Stock vs. Real Estate
   Investment Trusts (REITs).” Journal of Real Estate Research 2 (1987), 1–9.

Kuhle, J. L., C. H. Walther, and C. H. Wurtzebach, “The Financial Performance of Real Estate
   Investment Trusts.” Journal of Real Estate Research 1 (Fall 1986), 67–76.

Merton, R. C., “On Market Timing and Investment Performance, I. An Equilibrium Theory of
   Value for Market Forecasts.” Journal of Business (July 1981), 313–406.

Mei, Jianping, “Asian Real Estate in a Portfolio Context: Long-term Opportunities.” Real Estate
   Finance 15 (Winter 1999), 85–94.

Miller, Merton H., “The Current Southeast Asia Financial Crisis.” Pacific-Basin Finance Journal
   6 (1998), 225–233.

Sirmans, G. S., and C. F. Sirmans, “The Historical Perspective of Real Estate Returns.” Journal
   of Portfolio Management (Spring 1987), 2–31.

White, John R., “Have Central City Land Values Fallen?” Appraisal Journal (July 1972), 351–
White, H., “A Heteroscedasticity-Consistent Covariance Matrix Estimator and a Direct Test for
  Heteroscedasticity.” Econometrica 48 (1980), 149–170.
                                                                                                   Figure 1-1. Hong Kong

         18000                                                                                                                                                                                      30000
         12000                                                                                                                                                                                      20000

                                                                                                                                                                                                                         Real Estate

          6000                                                                                                                                                                                      10000
                0                                                                                                                                                                                   0
















                                                                                                              Year                                                                                                       Real Estate

                                                                                                          Figure 1-2. Japan

         45000                                                                                                                                                                                      3000
         30000                                                                                                                                                                                      2000

                                                                                                                                                                                                                  Real Estate

         15000                                                                                                                                                                                      1000
                0                                                                                                                                                                                   0















                                                                                                              Year                                                                                                  Real Estate

                                                                                                    Figure 1-3. Singapore

         3000                                                                                                                                                                                       700

         2500                                                                                                                                                                                       600

                                                                                                                                                                                                           Real Estate

         500                                                                                                                                                                                        100

           0                                                                                                                                                                                        0













                                                                                                            Year                                                                                                   Real Estate
                                                                                                                                                           Market                                                                                                          Market

                    88                                                                                                                                                                                                                         J-

                                                                                                                          J-                                                                                                                     86                                                         1200

                   -8                                                                                                       86
                     8                                                                                                    O                                                                                                                    F-
                    89                                                                                                        6                                                                                                                O
                  D                                                                                                       J-                                                                                                                    -8
                   -8                                                                                                       87                                                                                                                     7
                     9                                                                                                    A-                                                                                                                   J-
                  J-                                                                                                        88
                    90                                                                                                    J-                                                                                                                   F-
                  D                                                                                                         89                                                                                                                   89
                   -9                                                                                                                                                                                                                          O
                     0                                                                                                    O
                                                                                                                           -8                                                                                                                   -8
                    91                                                                                                        9                                                                                                                    9
                  D                                                                                                       J-                                                                                                                   J-
                   -9                                                                                                       90                                                                                                                   90
                     1                                                                                                    A-                                                                                                                   F-
                  J-                                                                                                        91                                                                                                                   91
                  D                                                                                                       J-                                                                                                                   O
                   -9                                                                                                       92                                                                                                                  -9
                     2                                                                                                    O                                                                                                                        1
                  J-                                                                                                       -9                                                                                                                  J-
                    93                                                                                                        2                                                                                                                  92
                  D                                                                                                       J-                                                                                                                   F-
                   -9                                                                                                       93                                                                                                                   93
                     3                                                                                                    A-                                                                                                                   O
                  J-                                                                                                        94                                                                                                                  -9
                    94                                                                                                                                                                                                                             3

                  D                                                                                                       J-

                   -9                                                                                                       95                                                                                                                 J-
                     4                                                                                                    O                                                                                                                      94
                  J-                                                                                                       -9
                    95                                                                                                        5                                                                                                                F-
                  D                                                                                                       J-
                   -9                                                                                                       96                                                                                                                 O
                     5                                                                                                                                                                                                                          -9
                                                                                                                          A-                                                                                                                       5
                                                                                                                                                                                                           Figure 1-5. Taiwan

                                                                                    Figure 1-6. Thailand
                  J-                                                                                                        97                                                                                                                 J-
                                                                                                                                                                                                                                                                                                                   Figure 1-4. South Korea

                                                                                                                          J-                                                                                                                     96
                  D                                                                                                         98
                   -9                                                                                                                                                                                                                          F-
                     6                                                                                                    O                                                                                                                      97
                  J-                                                                                                       -9
                    97                                                                                                        8                                                                                                                O
                                                                                                                          J-                                                                                                                    -9
                  D                                                                                                         99                                                                                                                     7
                     7                                                                                                    A-                                                                                                                   J-
                  J-                                                                                                        00
                    98                                                                                                                                                                                                                         F-
                  D                                                                                                       J-                                                                                                                     99
                   -9                                                                                                       01
                     8                                                                                                    O                                                                                                                    O
                                                                                                                           -0                                                                                                                   -9
                    99                                                                                                        1                                                                                                                    9




                                                                                                                                                         Real Estate                                                                                                     Real Estate

                                     Real Estate


                                                                                                           Real Estae
                                                                                                                                                                                                                                 Real Estate

Real Estate



                        M 8
                        M 0
                        M 2

                        M 4
                                                                                          Figure 2. USA

                        M 6
                        M 8

                                                 NAREIT Index
                                         Table 1

                              Description of Data Sample

                                   Number of
  Market        Sample Period     Sample Firms               Proxy of Risk-free rate
Hong Kong     1986:06 – 1999:03       112             3-Month Euro-HK Dollar Deposit Rate
   Japan      1986:06 – 1999:03       162                3-Month Euro-Yen Deposit Rate
 Singapore    1986:06 – 1999:03        22          3-Month Euro-Singapore Dollar Deposit Rate
South Korea   1986:06 – 1999:03        54                  Yield on Government Bonds
  Taiwan      1982:03 – 1999:07        34                3-Month Interbank Deposit Rate
 Thailand     1988:07 – 1999:03        40                3-Month Interbank Deposit Rate
                                          Table 2

     Descriptive Statistics of Monthly Real Estate Returns in the Six Asian Economies

                                    Descriptive Statistics                Correlation with
    Market              Mean             Median              Std. Dev.     Market Return
  Hong Kong             0.022*            0.020                0.106          0.830*
     Japan              0.010*            0.009                0.071          0.678*
   Singapore            0.015*            0.007                0.100          0.853*
  South Korea           0.019*            -0.004               0.143          0.594*
    Taiwan              0.012*            0.005                0.118          0.882*
   Thailand             0.024*            -0.014               0.184          0.759*
USA (NAREIT)**           0.007            0.004                0.034          0.544*

* Significant at 1%.
**We use the monthly change of the S&P 500 as the US market return.
                                            Table 3

                    Frequency of Market Conditions in Six Asian Markets

                   Frequency of Months in
    Market         Each Market Condition          Percentage of          Percentage of
                     Up          Down              Up Market             Down Market
Hong Kong             87           67                56.493                 43.506
Japan                 76           78                49.351                 50.649
Singapore             87           67                56.493                 43.506
South Korea           69           85                44.805                 55.195
Taiwan               107          100                51.691                 48.309
Thailand              58           71                44.961                 55.039

An up market is defined to exist when the market return exceeds the risk-free return. See Table 1
and the text for the definition of the risk-free rate.
                                               Table 4

Coefficients and Test Statistics for the Regression Equation for Real Estate Portfolio Firms
                   Rpt – Rft =  + u d + p (Rmt – Rft) + u d (Rmt – Rft) + t
                                         d=1: up market
                                       d=0: down market

  Market                            u            p             u          F-Stat.        Adj. R2
Hong Kong         -0.0329*        0.0156        0.6810*       0.4279*        116.115*        0.6930
                  (0.0086)       (0.0133)       (0.0685)      (0.1491)
Japan             -0.0371*        -0.0064       0.7561*       0.8231*         79.198*        0.5959
                  (0.0098)       (0.0145)       (0.1381)      (0.2342)
Singapore         -0.0413*        0.0011        0.7377*       1.2184*        194.618*        0.7915
                  (0.0077)       (0.0110)       (0.0734)      (0.1468)
Thailand           -0.0278        -0.0181       0.8226*       1.2796*         75.043*        0.6344
                  (0.0192)       (0.0293)       (0.1558)      (0.2772)
South Korea        -0.0045        -0.0046       1.0729*        0.2288         30.332*        0.3651
                  (0.0213)       (0.0342)       (0.2297)      (0.3937)
Taiwan             -0.0069       0.0226**       0.9617*       -0.2347*       267.013*        0.7948
                  (0.0076)       (0.0108)       (0.0724)      (0.0912)

F-Stat. is the F statistic for testing the significance of the regression model. Standard errors are in

* Significant at 1%
** Significant at 5%
                                             Table 5

              Chow Test for Structural Change of the Asian Financial Turmoil

           Market                        CHOW Statistic                Significance of CHOW
         Hong Kong                          1.539                               0.194
            Japan                           0.171                               0.953
          Singapore                         0.406                               0.804
         South Korea                        0.039                               0.997
           Taiwan                           1.104                               0.356
          Thailand                          0.467                               0.760

CHOW is the Chow (1960) test statistic for testing structural change. CHOW is distributed as an F(4,
T - 8) variable, where T is the number of observations.

shenreng9qgrg132 shenreng9qgrg132 http://