Dynamic Modelling of Real Estate Investment Ttrusts and Stock Markets

(1)

Dynamic modelling of real estate investment trusts and stock markets

Chien-Chiang Lee

a,

⁎

_{, Mei-Se Chien}

b

_{, Tsoyu Calvin Lin}

c

a

Department of Finance, National Sun Yat-sen University, Kaohsiung, 804, Taiwan

b

Department of Finance, National Kaohsiung University of Applied Sciences, Kaohsiung, Taiwan

c_{Department of Land Economics, National ChengChi University, Taipei, Taiwan}

a b s t r a c t

a r t i c l e i n f o

Article history: Accepted 20 November 2011 JEL classiﬁcation: C32 G12 G15 Keywords:

Real estate securitization Real estate investment trust Stock price

Causality

Generalized impulse response approach

Taiwan launched thefirst case of real estate securitization in 2005. The interrelationship between Taiwan Real Estate Investment Trusts (T-REITs) and the aggregate equity markets and segmented industries has drawn the interests of both investors and academia. This paper employs Toda and Yamamoto's (1995) pro-cedure and the generalized impulse response approach to uncover the extent and the magnitude of the rela-tionship between T-REITs and aggregate and segmented stock prices. We collected daily data of thefirst two issued T-REITs, Fubon No.1 and Cathay No. 1, from March 2005 to March 2010 and October 2005 to March 2010, respectively, to examine their causal relationships with aggregate stock markets, thefinancial sector, and the construction sector. The empirical results indicate that all variables have break points, reflecting shocks from the Subprime Mortgage Crisis or deregulation of the Qualified Domestic Institutional Investors (QDII) for Mainland Chinese to invest in Taiwan. We also discover that an individual T-REIT may lead or lag behind stock price indices due to its capitalization scale or business type. The transitory initial impacts of innovations in T-REITs on stock price indices are observed herein.

1. Introduction

Research studies have broadly explored the price, risk, and return of real estate markets, and the influential factors and interrelationships with other investment vehicles (Sing et al., 2006). Thefirst strand of re-lated research focuses on the in_{fluence of macroeconomic factors, such} as interest rates and inflation on real estate investment trusts (REITs,

hereafter).Gyourko and Keim (1992)found that important information

about property market fundamentals is impounded in REIT returns, es-pecially when they are adjusted to control for general market factors.

Liang and Webb (1995)further stressed that the market risk of mort-gage REITs (MREITs) is mostly interest rate risk, which is not diversified away.Mei and Gao (1995)also concluded that REITs’ returns could be explained by a function of fundamental economic variables. Most of the research in this strand indicated that the returns of REITs are in flu-enced by some macroeconomic factors, such as inflation, interest rates, and economic growth (Kim et al., 2007).

The second category of real estate literature engages in exploring

the interrelationship between REITs. Nelling and Gyourko (1998)

found that there is only weak evidence of predictability of REIT

returns based purely on past performance.Subrahmanyam (2007)

explored the existence of joint dynamics across the REIT and non-REIT sectors. Impulse response functions and Granger causality tests

indicate the existence of persistent liquidity spillovers running from

REITs to non-REITs. Moreover,Liow and Webb (2009)demonstrated

that the magnitude to which correlations are shown in international securitized real estate markets might largely be through the increas-ing integrated nature of the world real economy, rather than a result of the globalization of_{ﬁnancial markets.}

The third group of related studies attempts to discover the interre-lationship between real estate investment and other vehicles,

partic-ularly stock markets. He et al. (1996) examined the relationship

between REITs’ returns and bank stock returns. They found that

MREIT explains the risks and returns of bank stocks better than equity

REITs. Lizieri and Satchell (1997) explored that the wider equity

economy leads the real estate market in the short term, however, positive real estate returns may point to negative future returns in the rest of the economy in the long run.Ling et al. (2000)indicated that an out-of-sample prediction of excess returns on an equity REIT

has a lower power than an in-sample prediction.Hoesli and Camilo

(2007)showed in an international analysis that securitized real estate returns are positively associated with stock and real estate returns, but negatively related to bond returns. More related studies also have shown that REITs and the general stock market are integrated (Ambrose et al., 2007; Laopodis, 2009; Liu et al., 1990). Some studies investigates the impacts of movements or volatilities of stock prices

on REITs (Ambrose and Bian, 2010; Cotter and Stevenson, 2004,

2006; Devaney, 2001; Stevenson, 2002), but their conclusion regard-ing the direction of a causal relationship between real estate markets and stock markets tends to be weak.

⁎ Corresponding author. Tel.: +886 7 5252000x4825; fax: +886 7 5254899. E-mail address:cclee@cm.nsysu.edu.tw(C.-C. Lee).

Contents lists available atSciVerse ScienceDirect

Economic Modelling

(2)

In Taiwan the“Real Estate Securitization Statute” was promulgated in 2003. AsLee and Stevenson (2005)noted, REITs to some extent pro-vide a hybrid investment form, standing between equities and the ﬁxed-income sector. In addition, the asset maintains strong links with the direct real estate market. These inter-linkages provide such assets with unique characteristics. Real estate in Taiwan is extremely impor-tant for people's belief in the traditional notion of‘land is wealth.’1

With the introduction of this new mechanism, real estate can be trans-acted in the form of securities, which in turn increases its liquidity. In-dustries related to real estate in Taiwan have been eager to apply the securitization process for raising funds to liquidate real estate invest-ments. These REITs provide a new alternative for investment in the property market, which is traditionally considered as a secured, scarce, and precious but low-liquid vehicle for most Taiwanese investors. In the Taiwan context, people believe real estate is the most valuable asset, providing an interesting case for the empirical model of REITs.

Prior to the initial public offering (IPO), Taiwan REITs (T-REITs, hereafter) are required to go through an appraisal board meeting held by the government to evaluate the net asset value (NAV) for in-vestor protection. The appraisal meeting helps inin-vestors screen the fundamental value of the object's assets and also limits REITs’ appre-ciation potential. This special characteristic draws our interest to ex-amine the causal relationship of T-REITs with the equity markets,

including the overall stock markets, theﬁnancial sector stocks, and

the construction sector stocks.2The reason to examine the

relation-ship of T-REITs with both theﬁnancial and construction sectors is

that we want to discover what securitized real estate is like, or which of the following is related:ﬁnancial vehicles, traditional con-struction, or direct real property investments.

This paper applies the Granger causality test ofToda and Yamamoto (1995; hereafter TY)to draw conclusions with regard to the causal rela-tionships between T-REITs and weighted stock prices. The TY (1995)

method is a modiﬁed Wald (MWALD) test in the VAR framework. The

concept of causality is basically related to the topic of market integra-tion/segmentation. The hypothesis for determining which direction the causal relationshipflows is most likely from stocks to real estate, given the concept that the real estate market is not liquid and may be inactive in prices when compared to the stock market. These move-ments of asset prices may reflect the behaviors of arbitrageurs and in-vestment managers balancing their portfolios. This information may also benefit investors who consider different prices as a part of their

overall investment portfolios. Daily data of the ﬁrst two issued

T-REITs, Fubon No.1 and Cathay No.1, cover the periods March 2005-March 2010 and October 2005-2005-March 2010, respectively, and we exam-ine their causal relationships with the aggregate stock markets, and the ﬁnancial sector and construction sector in Taiwan. The beneﬁt from using the Granger causality test of TY (1995) is that it does not require knowledge as to the cointegration properties of the system and can be applied in the absence of cointegration (Lee and Chien, 2011).

The above method can also be used when stability and rank

require-ments are not satisﬁed. The model estimation procedures are indeed

quite simple, particularly in cases where there are more than two coin-tegrating vectors and when the OLS is valid. This paper also examines the transmission mechanism between these variables by applying the

generalized impulse response approach (GIRF) ofPesaran and Shin

(1998). The impulse response analysis can trace a variable's directional responses to a one standard deviation shock in another variable. This achieves both direct and indirect effects of innovations on a variable of interest, thus enabling us to comprehensively estimate the dynamic linkages between T-REITs and the Taiwan Weighted Stock Index.

The remaining part of this paper is organized as follows. Section 2 introduces the development of real estate securitization in Taiwan.

Section 3 brieﬂy describes the econometric methodology. Section 4

presents the empirical results. Section 5 offers the conclusion. 2. Real estate securitization in Taiwan

Up until 2009, a few countries in the Asia-Paciﬁc region have

implemented a real estate securitization mechanism, including Aus-tralia, Malaysia, Japan, Singapore, South Korea, Taiwan, Hong Kong, Philippines and India, while the U.K. and Germany in Europe are on their way towards following the mainstream. The REITs provide these countries with a new alternative for direct investment in the property market, which is traditionally considered as a secured, scarce, and precious, but low-liquid vehicle for most Asian investors. In Taiwan the“Real Estate Securitization Statute” was promulgated in 2005. There are two types of vehicles for real estate securitization in Taiwan: real estate investment trust (T-REIT) and real estate asset trust (T-REAT). T-REITs are similar to REITs in other countries and in-volve raising fundsfirst and then acquiring real estate targets. Like stocks, T-REITs have no specific investment period, while the structur-ing of T-REATs is very similar to that of asset securitization. T-REATs are established to_{first hold defined real estate with rental or operating} income for specific periods and then to raise funds in exchange for the particular properties. Capital gains are distributed to investors after the sale of the target properties for REATs. Both REITs and T-REATs are only allowed to operate as closed-end funds, and at least 75% of the funds must invest in existing properties or related rights that generate steady income as follows: asset-backed securities (ABS), bank deposits and acceptance, short-term commercial papers, and trea-sury bonds. Real estate development activities, although under discus-sion, are still prohibited for both portfolios of T-REITs and T-REATs.

Unlike the U.S. and Japan, Taiwan only allows the operation of an in-vestment trust (special-purpose trust) instead of an inin-vestment corpora-tion (special-purpose corporacorpora-tion). The minimum capital requirement for establishing a T-REIT is NT$1 billion and for a T-REAT is NT$300 million. (US$1=NT$30 in 2010 on average). For both T-REITs and T-REATs, any five certificate holders cannot own more than 50% of the total value of the issued certi_{ficates. The T-REIT is allowed to leverage financially for} the purpose of operation or dividend payout, but the maximum percent-age has not been regulated. The income of T-REITs is only required to be distributed within six months after everyfiscal year, but the ratio is not specifically regulated.

After the enactment of the“Real Estate Securitization Statute” in 2003, Taiwan introduced thefirst case of REIT to the public market in early 2005. Up through 2010, the accumulated market capitalization of T-REITs has reached US$1.8 billion, indicating the popular trend of real estate securitization for both investors and issuers in Taiwan. As stated above, T-REITs are required to go through an appraisal board meeting prior to IPO for evaluating the NAV for investor protection. The appraisal meeting helps investors screen the objective assets’ fun-damental value and also limits REITs' appreciation potential. Therefore, the pricefluctuation of T-REITs is smaller than those in the U.S. or in Japan. The low-risk and low-return characteristics of T-REITs also pro-vide investors a secured vehicle in Taiwan, together with the tax-free incentive for the dividends of T-REIT investors. The incentives above have resulted in over 60% of investors being individuals, according to Fubon No.1, thefirst T-REIT introduced to the markets.

The trend of real estate securitization indicates that REITs have suc-cessfully drawn investors' interests in this region, where real estate is 1

According to the census of Taiwan's Ministry of Interior in 2006, the average home ownership rate is over 87% and the average housing unit vacancy rate is 17.6%. These twoﬁgures are far above the average of other countries (Lee and Chien, 2011; Chen et al., 2011).

2

There are some paper discussing the volatility of REITs and stock prices. Most of these papers aim to utilize the GARCH model to analyze the volatility dynamics of REITs, includingDevaney (2001), Stevenson (2002), and Cotter and Stevenson (2004. 2006). A few of them utilize data on REITs to examine the role of stock market volatility on earnings management (Ambrose and Bian, 2010). However, the aim of this paper is to examine the characteristics of T-REITs that lead or lag the stock price index. We fo-cus on the price relationships of these two markets, not the volatility linkages, and the volatility in daily stock market index data is not discussed herein.

(3)

traditionally considered as an asset that appreciates in the long run. De-spite the popularity of the REIT mechanism in Asia, some unique char-acteristics of Asian REITs are worth noting. First, most REITs in Asia do not allow funds to invest in properties abroad. This restriction may lead these REITs to become less regional diversified and eventually en-counter signi_{ficant concentration risk and currency risk as the domestic} macro economy declines. Second, the economic scale of the Asian REITs, which are relatively smaller than those in the U.S., might face liquidity limitations in the future, especially as real estate markets stagnate or the rental or operating income streams shrink. Third, most of these countries have no regulations upon the minimum lease term of the properties for REIT investment. The uncertainty over continuity in rent-al contracts, especirent-ally for anchor tenants, may imply a significant vari-ation between the predicted and realized incomes, and consequently the forecast and fundamental NAVs of the invested properties. Fourth, some Asian countries do not provide tax advantages for investing in REITs, and some impose restrictions on REITs of minimum holding pe-riods for the properties. Relative to the experience in the U.S. and Aus-tralia, these features may lead to limited liquidity and marketability of REITs. International investors or researchers will need to take into ac-count the basic differences of REITs among various ac-countries for asses-sing potential risks.

3. Methodology

In line withToda and Yamamoto (1995), we consider the n-vector

time series Ztgenerated by the k-th order VAR model: Zt¼ Φ0þ Φ1tþ Φ2t 2 þ Φ3t 3 þ Π1Zt−1þ … þ ΠkZt−kþ Et; t ¼ 1; …T; ð1Þ

where Et~ N(0,Ω); Zt= (REITt', SPt'), where REITtand SPtrepresent T-REIT and stock price, respectively; and t represents a deterministic time trend. Economic hypotheses can be expressed as restrictions

on the coefﬁcients in the model in accordance with the following:

H0: FðπÞ ¼ 0;

where π= vec(P) is a vector of the parameters in Eq. (1); P =

(Π1…Πk); and F( . ) is a twice continuously differentiable m-vector valued function.

TY (1995) set up a simple procedure to facilitate testing for Grang-er non-causality in level VARs, which is estimated by the OLS with in-tegrated variables. The test essentially involves two stages. First, the augmented (k + d) VARs are estimated, where d is the maximal

order of integration. Second, we apply the Wald test to the_{ﬁrst k}

VAR coefﬁcient matrix to test for Granger causality. For testing the

null hypothesis, TY (1995) conﬁrmed that the Wald statistic

con-verges in distribution to anχ2

random variable with m degrees of freedom regardless of whether the process Ztis stationary, possibly around a linear trend, or whether it is cointegrated. This methodology minimizes the risks perhaps associated with misidentifying the or-ders of integration of the series, or the presence of cointegration, while additionally it minimizes the possibility of distorting the test size that frequently results from pretesting.

This paper applies the advanced generalized impulse response

tech-niques ofPesaran and Shin (1998)to examine the relationships

be-tween T-REITs and the aggregate equity markets and segmented industries. The results are expected to draw interests from investors and academia. We consider the two-dimensional VAR model as follows: Zt¼ A

Xp i¼1

ψiZt1þ εt; ð2Þ

where Ztis a (2× 1) vector of jointly determined endogenous variables; ψ1throughψpare (2× 2) coefﬁcient matrices; A is a vector of constants;

andεtis a (2× 1) vector of well-behaved disturbances with covariance Σ=σij(i, j = 1, 2). The generalized impulse response of Zt + nwith re-spect to a unit shock to the jth variable at time t is represented by (GnΣej)(σij)- 1, where Gn=ψ1Gn - 1+ψ2Gn - 2+⋯+ψpGn - p, n = 1, 2,…, G0= I, Gn= 0 for nb0, and ejis a (2× 1) selection vector with unity as its jth element and zero elsewhere.

Our methodology has two advantages over the standard impulse response analysis (Chen et al., forthcoming). First, it does not pre-suppose any ordering that has theoretical implications. Second, the methodology provides a meaningful interpretation of the initial im-pact of shocks - a feature that is missing in the traditional methodol-ogy, but might be important in the analysis of Taiwan Real Estate Investment Trusts where information is transmitted quickly. 4. Empirical investigations

We use two different stock prices of T-REITs' data, including Fubon No.1 and Cathay No.1 for empirical analysis. Properties in the portfolios of these two T-REITs are all located in Taipei City, and both T-REITs are ranked over twA. Fubon No. 1, market capitalization of US $0.25 billion, includes three ofﬁce buildings and one commercial build-ing that are 10 years old or newer. Cathay No. 1, market capitalization of US$0.43 billion, has two ofﬁce buildings and one hotel, all of which are over 40 years old. The sample period for Fubon No.1 is from March 10, 2005 to March 15, 2010. The sample data of Cathay No.1 covers the pe-riod from October 3, 2005 to March 15, 2010. Other data are daily prices for three different weighted indices: the Taiwan Weighted Stock Index (SPT), the Weighted Price Index of the Financial Sector (SPF), and the Weighted Price Index of the Construction Sector (SPC). All the data are collected from the Taiwan Economic Journal (TEJ) database. All of the variables used are in natural logarithms.

Figs. 1–5present the price movements of Fubon No.1, Cathay No.

1, and the SPT, SPF, and SPC series, respectively. Theﬁgures show

that these prices are non-stationary and exhibit similar patterns. In

Fig. 1 the price of Fubon No.1 shows more variation since March 2007 and seems to be breaking around April 2007 and June 2008. As to Cathay No. 1, its price also presents more variations after July 2007 and is breaking around April 2009.

4.1. Case 1— REIT of Fubon No.1

For a comparison, we apply the unit-root tests, including the ADF (Dickey and Fuller, 1979), the PP (Phillips and Perron, 1988), the KPSS (Kwiatkowski et al., 1992), the DF-GLS (Elliott et al., 1996), and the NP (Ng and Perron, 2001), from March 10, 2005 to March 15, 2010 to detect stationarity of these variables.Table 1presents the results of the classic unit-root tests. We follow the determining rule ofDoldado et al. (1990)to establish the appropriate model for the unit-root tests. Aside from this, we adopt the newly-developed Modiﬁed Akaike's

Infor-mation Criterion (MAIC) suggested byNg and Perron (2001)to select

the optimal number of lags based on the principle of parsimony. The re-sults inTable 1clearly indicate that all series are integrated of order one (I(1)) at the 5% level of signi_{ﬁcance for all unit-root tests without} struc-tural breaks.

All of the above unit-root tests are not appropriate for testing the stationarity of a series that encounters a structural change. Thus, we fur-ther take the structural break into account when employing the unit-root test. Ever since the renowned paper offered byPerron (1989),3

previous literature has been aware of the importance of allowing for a structural break when testing for a unit root. Interestingly, subsequent 3_{These are: Model A (the crash model), which allows for a one-time change in the}

intercept of the trend function; Model B (the changing growth model), which allows for a change in the slope; and Model C (the crash-cum-growth model), which allows for a change in both intercept and slope (please see a detailed discussion inLee and Strazicich, 2003; Perron, 1989; Smyth and Inder, 2004).

(4)

studies changed the test to allow for one unknown break point that is determined endogenously from the data. Without allowing for structur-al breaks, traditionstructur-al ADF type tests could cause a wrong decision when the null hypothesis is not rejected. To improve the default of ADF type tests without allowing for possible breaks in the series,Perron (1989)

proposed allowing for one known, or“exogenous,” structural break in

the ADF unit-root test. FollowingPerron (1989), Zivot and Andrews

(1992)offered to determine the break point“endogenously” from the data.

A structural break essentially corresponds to an intermittent shock with a permanent effect on the series (Hendry, 1996). The op-posite can also happen if the break occurs at the beginning of the sample (Leybourne et al., 1998). In order to take this possible regime

shift in the unit-root tests into account,Zivot and Andrews (1992,

hereafter ZA)developed a new category of tests that allow for a struc-tural break.

A potential problem common to the ADF-type endogenous break unit-root tests is that they derive their critical values assuming no

break(s) under the null. This assumption leads to size distortions in

the presence of a unit root with a break asNunes et al. (1997)

showed. As Lee and Strazicich (2004) suggested, if we lose the

power from ignoring one break, then it may let us extend that the query has a similar loss of power from ignoring two or more breaks in the one-break test (Smyth and Inder, 2004). To avoid problems of

bias and spurious rejections,Lee and Strazicich (2003)derived the

endogenous two-break LM unit-root test.

Lee and Strazicich (2003)held that the two-break LM unit-root test statistic can be estimated by regression according to the LM (score) principle as follows:

Δyt¼ η0ΔRtþ φ~St−1þ ut; ð3Þ

where ~St is a de-trended series, ~St¼ yt− ~Ψx−Rt~η; t ¼ 2; …; T; ~η is a

vector of coefﬁcients in the regression of ΔytonΔRt; ~Ψxis given by

y1−R1~η; and y1and R1denote theﬁrst observations of ytand Rt, re-spectively. Corrections for autocorrelated errors are accomplished

REIT (Fubon #1)

8 8.25 8.5 8.75 9 9.25 9.5 9.75 10 10.25 10.5 10.75 11 11.25 11.5 11.75 12 12.25 12.5 12.75 13 13.25 13.5 13.75 14 14.25 14.5 14.75 15 2005/3/10 2005/11/10 2006/7/10 2007/3/10 2007/11/10 2008/7/10 2009/3/10 2009/11/10

Fig. 1. Price movement of the REIT of Fubon No. 1(March 10, 2005–March 15, 2010).

REIT (Cathay #1)

7 7.25 7.5 7.75 8 8.25 8.5 8.75 9 9.25 9.5 9.75 10 10.25 10.5 10.75 11 11.25 11.5 2005/10/3 2006/5/3 2006/12/3 2007/7/3 2008/2/3 2008/9/3 2009/4/3 2009/11/3

(5)

via augmented termsΔ~St−j, j = 1,…, k, in Eq.(3), as with the ADF test.

For Model C, where Bjt=ΔDjtand BTjt=ΔDTjt* (j = 1, 2), the unit-root

null hypothesis is described byφ=0 (implying a unit root with two

breaks), and the LM test statistics are given by:

~τ ¼ t−statistic for the null hypothesis φ ¼ 0: ð4Þ

From this, the LM unit-root test can endogenously determine the two breaks by utilizing a grid search as follows:

LM_τ¼ inf_λ ~τ λð Þ: ð5Þ

There is a repeated procedure at each combination of break points, λj= TBj/T, j = 1, 2. As shown inLee and Strazicich (2003), the critical

values for Models A and C depend on the location of the breaks (λ).

Therefore, we utilize critical values that correspond to the location of the breaks.

We list the null and alternative hypotheses of Model C as: Null : yt¼ μ0þ d1B1tþ yt−1þ v1t

Alternative :yt¼ μ1þ γt þ d1D1tþ ω1DT1tþ yt−1þ v2t ð6Þ

However, if the vector of exogenous variables shows Rt= [1, t]′, then the DGP is the same as that shown in the no break LM unit-root test ofSchmidt and Phillips (1992).

Tables 2 and 3respectively present the results of the ZA (1992) tests andLee and Strazicich's (2003)unit-root tests. Both these two tests show that all series carry a unit root in the level and reject the null of“non-stationarity” in the ﬁrst difference. This insures the I(1) type series for all series considered. The results of the ZA test and

Lee and Strazicich's (2003)tests indicate that the estimated break

SPT

4000 5000 6000 7000 8000 9000 10000 2005/3/10 2005/11/10 2006/7/10 2007/3/10 2007/11/10 2008/7/10 2009/3/10 2009/11/10

Fig. 3. Movement of the weighted price index of Taiwan Stock Exchange (SPT) (March 10, 2005–March 15, 2010).

SPF

400 600 800 1,000 1,200 2005/3/10 2005/11/10 2006/7/10 2007/3/10 2007/11/10 2008/7/10 2009/3/10 2009/11/10

(6)

References

Ambrose, B., Bian, X., 2010. Stock market information and REIT earnings management. Journal of Real Estate Research 32 (1), 101–137.

Ambrose, B.W., Lee, D.W., Peek, J., 2007. Co-movement after joining an index: spillover of non-fundamental effects. Real Estate Economics 35 (1), 57–90.

Chen, P.F., Chien, M.S., Lee, C.C., 2011. Dynamic modeling of regional house price diffu-sion in Taiwan. Journal of Housing Economics. 20 (4), 315–332.

Clayton, J., MacKinnon, G., 2003. The relative importance of stock, bond, and real estate factors in explaining REIT returns. Journal of Real Estate Finance and Economics 27 (1), 39–60.

Cotter, J., Stevenson, S., 2004. Uncovering volatility dynamics in daily REIT returns. Working Paper, Centre for Real Estate Research, Smurﬁt School of Business, Uni-versity College Dublin.

Cotter, J., Stevenson, S., 2006. Multivariate modeling of daily REIT volatility. Journal of Real Estate Finance and Economics 32 (2), 305–325.

Devaney, M., 2001. Time varying risk premia for real estate investment trusts: a GARCH-M model. The Quarterly Review of Economics and Finance 41, 335–346. Dickey, D.A., Fuller, F.A., 1979. Distribution of the estimation for an autoregressive time

series with a unit root. Journal of the American Statistical Association 74, 427–431. Doldado, J., Jenkinson, J., Sosvilla-Rivero, S., 1990. Cointegration and unit-roots. Journal

of Economic Surveys 4, 249–273.

Elliott, G., Rothenberg, T.J., Stock, J.H., 1996. Efﬁcient tests for an autoregressive unit root. The Economic Record 77, 252–269.

Glascock, J., Lu, C., So, R., 2000. Further evidence on the integration of REIT, bond, and stock returns. Journal of Real Estate Finance and Economics 20 (2), 177–194. Gyourko, J., Keim, D., 1992. What does the stock market tell us about real estate returns?

Journal of the American Real Estate and Urban Economics Association 20, 457–485. He, L.T., Myer, F.C.N., Webb, J.R., 1996. The sensitivity of bank stock returns to real

es-tate. The Journal of Real Estate Finance and Economics 12 (2), 203–220. Hendry, D.F., 1996. A theory of co-breaking. Paper presented at the Time Series

Confer-ence organized by NBER-NSF.

Hoesli, M., Camilo, W., 2007. Securitized real estate and its link withﬁnancial assets and real estate: an international analysis. Journal of Real Estate Literature 15 (1), 59–84. Kim, J.W., Leatham, D.J., Bessler, D.A., 2007. REITs' dynamics under structural change

with unknown break points. Journal of Housing Economics 16, 37–58.

Kwiatkowski, P., Schmidt, P., Shin, Y., 1992. Testing the null hypothesis of stationary against the alternative of a unit root. Journal of Econometrics 54, 159–178. Laopodis, N., 2009. REITs, the stock market and economic activity. Journal of Property

Investment & Finance 27 (6), 563–578.

Lee, C.C., Chien, M.S., 2011. Empirical modelling of regional house prices and the ripple effect. Urban Studies 48 (10), 2029–2048.

Lee, S., Stevenson, S., 2005. The case for REITs in the mixed-asset portfolio in the short and long run. Journal of Real Estate Portfolio Management 11 (2), 55–80. Lee, J., Strazicich, M.C., 2003. Minimum LM unit root test with two structural breaks.

The Review of Economics and Statistics 85, 1082–1089.

Lee, J., Strazicich, M.C., 2004. Minimum LM unit root test with one structural break, De-partment of Economics. Appalachian State University Working Paper Series. Leybourne, S., Mills, T., Newbold, T., 1998. Spurious rejections by Dickey–Fuller tests in

the presence of a break under the null. Journal of Econometrics 53, 211–244.

Liang, Y., Webb, J., 1995. Pricing interest rate risk for mortgage REITs. Journal of Real Es-tate Research 10 (4), 461–469.

Ling, D., Naranjo, A., 1999. The integration of commercial real estate markets and stock markets. Real Estate Economics 27 (3), 483–515.

Ling, D., Naranjo, A., Ryngaert, M.D., 2000. The predictability of equity REIT returns: time variation and economic signiﬁcance. Journal of Real Estate Finance and Eco-nomics 202, 117–136.

Liow, K.H., Webb, J.R., 2009. Common factors in international securitized real estate markets. Review of Financial Economics 18, 80–89.

Liu, C.H., Hartzell, D.J., Greig, W., Grissom, T.V., 1990. The integration of the real estate market and the stock market: some preliminary evidence. Journal of Real Estate Fi-nance and Economics 3 (3), 261–282.

Lizieri, C., Satchell, C., 1997. Interactions between property and equity markets: an in-vestigation of linkages in the United Kingdom 1972–1992. Journal of Real Estate Fi-nance and Economics 15 (1), 11–26.

Mei, J., Gao, B., 1995. Price reversal, transaction costs, and arbitrage proﬁts in the real es-tate securities market. Journal of Real Eses-tate Finance and Economics 11, 153–165. Nelling, E., Gyourko, J., 1998. The predictability of equity REIT returns. Journal of Real

Estate Research 16 (3), 251–268.

Ng, S., Perron, P., 2001. Lag length selection and the construction of unit root tests with good size and power. Econometrica 69 (6), 1519–1554.

Nunes, L., Newbold, P., Kuan, C., 1997. Testing for unit roots with breaks: evidence on the great crash and the unit root hypothesis reconsidered. Oxford Bulletin of Eco-nomics and Statistics 59 (4), 435–438.

Okunev, J., Wilson, P., 1997. Using non-linear tests to examine integration between real estate and stock markets. Real Estate Economics 25 (3), 487–503.

Perron, P., 1989. The great crash, the oil price shock and the unit root hypothesis. Econ-ometrica 57 (6), 1361–1401.

Pesaran, M.H., Shin, Y., 1998. Generalized impulse response analysis in linear multivar-iate models. Economics Letters 58, 17–29.

Phillips, P., Perron, P., 1988. Testing for a unit root in time series regression. Biometrika 75, 335–346.

Qu, Z., Perron, P., 2007. Estimating and testing structural changes in multivariate re-gressions. Econometrica 75 (2), 459–502.

Schmidt, P., Phillips, P.C.B., 1992. LM tests for a unit root in the presence of determin-istic trends. Oxford Bulletin of Economics and Statdetermin-istics 54, 257–287.

Sing, T.F., Tsai, I.C., Chen, M.C., 2006. Price dynamics in public and private housing mar-kets in Singapore. Journal of Housing Economics 15 (4), 305–320.

Smyth, R., Inder, B., 2004. Is Chinese provincial real GDP per capita nonstationary? Ev-idence from multiple trend break unit root tests. China Economic Review 15, 1–24. Stevenson, S., 2002. An examination of volatility spillovers in REIT returns. Journal of

Real Estate Portfolio Management 8, 229–238.

Stock, J.M., 1994. Unit Roots, Structural Break and Trends. In: Engle, R.F., McFadden, D. (Eds.), Handbook of Econometrics. North Holland, pp. 2738–2841.

Subrahmanyam, A., 2007. Liquidity, return and order-ﬂow linkages between REITs and the stock market. Real Estate Economics 35 (3), 383–408.

Toda, H.Y., Yamamoto, T., 1995. Statistical inference in vector autoregression with pos-sible integrated processes. Journal of Econometrics 66, 225–250.

Zivot, E., Andrews, D., 1992. Further evidence on the great crash, the oil price shock, and the unit root hypothesis. Journal of Business and Economic Statistics 10, 251–270.