The dynamic relationship between commercial real estate and stock markets

The universe of commercial real estate in the US was estimated to be worth $4.3 trillion in 2000 and accounted for approximately 10% of the total US investment market Conner, Hess and Li

BETWEEN COMMERCIAL REAL ESTATE AND STOCK MARKETS

YANG HAISHAN

(B.Eng (Hons), M.Sc., Chongqing)

A THESIS SUBMITTED FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

DEPARTMENT OF REAL ESTATE SCHOOL OF DESIGN AND ENVIRONMENT NATIONAL UNIVERSITY OF SINGAPORE

2003

I appreciate the dedication of the many individuals who helped me to complete this dissertation First and most foremost, I like to express my deepest gratitude to my academic advisor, Professor Liow Kim Hiang, for his tremendous guidance, encouragement, and extraordinary support throughout the whole process of this dissertation

I am also grateful to the rest of my dissertation committee, Dr Joseph Ooi Thian Leong and Dr Lum Sau Kim for their thoughtful input and guidance during the process of completing this dissertation In addition to my dissertation committee, I owe many thanks to Dr Tu Yong, who provided me the prompt feedback through emails and calls I am also grateful to Dr Zhang Xibin for his support as a good friend and as

a mentor

Moreover, I wish to extend to profound appreciation to my parents who have taught

me the importance of education and who have devoted their lives to supporting their children Finally, and most importantly, I am indebted to my girlfriend, Miss Sun Yang Without her constant encouragement, unceasing sacrifices, and emotional support, the completion of this dissertation could not have been possible

life has become clearer since she was called home to be with the Lord during my Ph.D study

LIST OF FIGURES

CHAPTER 1 INTRODUCTION

1.1 Background 1

1.2 Research Objectives 14

1.3 Research Questions 14

1.4 Significance of the Research 15

1.5 Research Data 17

1.6 Research Framework and Methodology - Commercial Real Estate Market, Stock Market, and Macroeconomy 18

1.7 Empirical Research Framework and Methodology 26

1.8 Scope of the Research 30

1.9 Outline of the Research 31

CHAPTER 2 LITERATURE REVIEW 2.1 Introduction 35

2.2 Relationships between Direct and Indirect Real Estate 35

2.2.1 Presence of a Pure Real Estate Factor in Indirect Real Estate 36

2.2.2 Causal Relationship Between Direct and Indirect Real Estate 43

2.2.3 Long-run Relationship Between Direct and Indirect Real Estate 50

2.3 Relationships between Real Estate and Stock Markets 53

2.3.1 Cross-correlation Analysis 54

2.3.2 Market Integration/Segmentation Analysis 55

2.3.3 Causal Relationship Analysis 60

2.3.4 Measurement Issues in Real Estate Return and Risk 62

2.4 The Macroeconomic Factors Influencing Real Estate and Stock Markets 65

2.4.1 Stock and the Macroeconomy 66

2.4.2 Real Estate and the Macroeconomy 72

2.5 Conclusion 75

3.2 Overview of Singapore 80

3.3 The Singapore Commercial Real Estate Market 82

3.4 The Singapore Stock Market 88

3.5 Listed Property Companies in Singapore 90

3.6 Review of the Market Performance 95

3.7 Review of Related Singapore Property Research – Commercial Real Estate and Property Companies 104

3.8 Concluding Remark 105

CHAPTER 4 UNIVARIATE LINEAR AND NONLINEAR ANALYSIS 4.1 Introduction 107

4.2 Descriptive Statistics 107

4.3 Univariate Linear Analysis 112

4.4 Univariate Nonlinear Analysis 116

4.5 Summary 121

CHAPTER 5 MARKET INTEGRATION BETWEEN COMMERCIAL REAL ESTATE AND STOCK MARKETS 5.1 Introduction 124

5.2 Research Scope and Hypotheses 125

5.3 Research Data 126

5.3.1 Office Price Index (PPIO) 126

5.3.2 Stock Market Index (SGXA) 127

5.3.3 Data Treatment for Synchronization 130

5.4 Research Methodology 130

5.4.1 Unit Root Test 133

5.4.2 Engle-Granger’s Cointegration Technique 140

5.4.3 Johansen Cointegration Technique 141

5.4.4 Fractional Cointegration Technique 143

5.5.2 Results of Cointegration Analysis 153

5.6 Summary 162

CHAPTER 6 CAUSAL RELATIONSHIP BETWEEN COMMERCIAL REAL ESTATE AND STOCK MARKETS

6.1 Introduction 165

6.2 Research Scope and Hypotheses 165

6.3 Research Data 167

6.4 Research Methodology 167

6.4.1 A Testing Procedure for Linear Granger Causality 167

6.4.2 A Testing Procedure for Nonlinear Granger Causality 169

6.4.2.1 Modified Baek and Brock Nonparametric Method 169

6.4.2.2 Bivariate Fractionally Integrated Error Correction Model 173

6.5 Results and Discussion 175

6.5.1 Results of Linear Granger Causality Tests 175

6.5.2 Results of Nonlinear Granger Causality Tests 178

6.6 Summary 183

CHAPTER 7 RELATIONSHIP BETWEEN COMMERCIAL REAL ESTATE AND STOCK MARKETS IN THE MACROECONOMY 7.1 Introduction 186

7.2 Research Scope and Hypotheses 187

7.3 Research Data 191

7.3.1 Gross Domestic Product 192

7.3.2 Interest Rate 193

7.3.3 Inflation 194

7.3.4 Exchange Rate 194

7.3.5 Data Treatment for Synchronization 194

7.4 Research Methodology 195

7.4.1 Cointegration Analysis 196

7.4.2 Vector Error Correction Model and Fractionally Integrated Error Correction Model 198

7.5.1.2 Results of Fractional Cointegration Analysis 208

7.5.2 Short-run Dynamics in a Macroeconomic Setting 210

7.5.2.1 Results of Vector Error Correction Model 210

7.5.2.2 Results of Fractionally Integrated Error Correction Model 213

7.5.3 Forecasting Performance 217

7.6 Summary 218

CHAPTER 8 CONCLUSION 8.1 Summary of Major Findings 222

8.2 Implication for Real Estate Investment 227

8.3 Research Contributions 229

8.4 Limitations of the Research 230

8.5 Recommendations for Further Study 231

BIBLIOGRAPHY

This thesis explores three major issues regarding dynamic relationship between commercial real estate and common stock markets: market integration, causal relationship, and long-run and short-run relations in the macroeconomic setting Various linear and nonlinear, parametric and nonparametric testing methodologies are applied to a Singapore data set over the 1975-2001 period in an integrated procedure

Before the relationship between the two markets is examined, we provide a comprehensive description of the distributional traits for each of the series throughout this dissertation In general, certain degree of nonlinear structure, other than linear fabric, is present in commercial real estate and some macroeconomic variables Thus, existing linear models would need to be enhanced to ensure that these nonlinear structures are accounted for in order to uncover the complex interrelation between commercial real estate and stock markets

The dynamic relationship between commercial real estate and stock markets is then addressed in three ways First, we examine integration/segmentation between the two markets While evidence of a long-run relationship for the two asset classes is found with Johansen cointegration tests, the standard residual-based cointegration techniques fail to provide any similar evidence We attribute the inconsistencies between the cointegration results to the possibility of fractional cointegration which allows for fractionally integrated equilibrium errors and thus encompasses a wide range of mean- reversion behavior Based on the parametric Geweke-Porter-Hudak (GPH) test, the empirical results do show that the equilibrium errors could be characterized by the fractionally integrated processes Therefore, the commercial real estate market is

Second, we consider linear and nonlinear causal relationships between the two markets In a linear Granger causality framework, the error-correction augmented VAR model shows that the commercial real estate market seems to informationally lead the stock market The results imply that the private real estate market contains information that can be impounded in the securitized market through insider trading However, the nonlinear Granger causality test based on the fractional cointegration indicates a significant bi-directional nonlinear causal relationship between the two markets The strong evidence of the feedback effect is consistent with the fractional cointegration results and provides support for the prior finding that the commercial real estate market does have an influence on the stock market, rather than simply acting as

a secondary absorber of information flow

Third, we examine the long-run and short-run relations between the two markets within the economic system including GDP, interest rate, inflation and exchange rate

If both corporate profits and rents were driven by economic fundamentals, then commercial real estate and stock prices should be expected to move together within the macroeconomic framework The results show that the two markets are linearly in long- run equilibrium within the economic system, and that after controlling for changes in the macroeconomy, the observed positive relation between the two markets weakens This implied that the correlation between the two markets is mainly due to their common reactions to fundamental economic activities In addition, the results indicate the presence of nonlinear fractional cointegration among commercial real estate, stock

than the general stock market Hence the commercial real estate market is significantly related to the key macroeconomic factors To unveil the implication of the nature of long-run relation for the short-term dynamics, we further investigate the forecasting performance of the fractionally integrated error correction model (FIECM) against the conventional vector error correction model (VECM) for the prediction of the commercial real estate price index Consistent with our expectation, incorporating fractional cointegration (if present) in an ECM does improve forecasting performance over the traditional ECMs

We do not intend to uncover the whole spectrum of the three selected issues in making inferences about the dynamic interactions between the commercial real estate and the common stock markets Within the sphere of mixed-asset allocation and corporate real estate management, the methods and evidence reported in this research can form the basis for further study in empirical literature related to the synthesis of modern economic and finance theories with real estate investment and management in the context of increasing interactions between real estate and broader capital markets

1.1 Advantage and Disadvantage of Direct and Indirect Real Estate

Investment 11

1.2 Overview of Research Data Requirements 17

2.1 Summary of Studies on the Presence of a Pure Real Estate Factor in Indirect Real Estate 37-38 2.2 Summary of Studies on the Lead-lag Relationship Between the Direct and the Indirect Real Estate Markets 45-46 2.3 Summary of Studies on the Long-run Relationship Between the Direct and Indirect Real Estate Markets 52

2.4 Summary of Studies on the Integration between Real Estate and Stock Markets with Asset-pricing Framework 57

2.5 Summary of Studies on the Integration between Real Estate and Stock Markets by Modeling the Relations Between Asset Prices 59

2.6 Summary of Studies on the Relationships Between Stock Market and Macroeconomic Variables 69-70 2.7 Summary of Studies on the Relationships Between Real Estate Market and Macroeconomic Variables 73

3.1 Singapore Economic and Financial Indictors: 2001 82

3.2 Ownership of the Major Office Buildings in the CBD Area 87

3.3 Market Capitalizations of the Top 15 SGX Mainboard Companies and the Contribution to the Market at the End of 2001 89

3.4 Market Capitalizations of Various Sectors in SGXA at the End of 2001 90

3.5 Listed Property Companies in Singapore (End of 2001) 92-94 3.6 Inter-asset Correlation Matrix and Return-risk Performances of the Commercial Real Estate and Stock Markets: 1975-2001 96

4.1 Characteristics of Series 108

4.2 Descriptive Statistics of Series (1980:1-2001:4) 111

4.3 ADF and PP Test for Presence of Unit Root (1980:1-2001:4) 113

4.6 Results of Rescaled Range Analysis 120

4.7 Results of BDS Analysis 121

5.1 Sectorial Reclassification of Companies Listed on the SGX 128

5.2 Mainboard Listed Companies Related to the Property Sector 129

5.3 Empirical Size of the GPH Test for Cointegration 150

5.4 Descriptive Statistics of the PPIO and SGXA 152

5.5 Perron Unit Root Test on Individual Indices for Known Breakpoints 153

5.6 Results of the ADF Test for Cointegration 154

5.7 Testing for Appropriate Lag-length in Johansen Cointegration 155

5.8 Results of Johansen Cointegration Test for Market Integration 156

5.9 Results of the GPH Test for Cointegration between Office and Stock 160

5.10 Summary of Results of Linear and Nonlinear Cointegration Test 162

6.1 Results of Linear Granger Causality 176

6.2 Results of Linear Granger Causality in Different Sub-periods 178

6.3 Results of Modified Baek and Brock Test 180

6.4 Results of Bivariate FIECM Nonlinear Granger Causality Test 182

6.5 Summary of Results of Linear and Nonlinear Granger Causality Tests (1975:1 –2001:4) 183

7.1 Comparison of New and Old Format for Presenting GDP by Industry 193

7.2 Synchronization Treatment for Macroeconomic Variables 195

7.3 Descriptive Statistics of Four Macroeconomic Variables 201

7.4 Unit Root Tests of Macroeconomic Variables 202

7.6 Results and Critical Values for the λmax and λtrace Test for

Relation in an Economic System 203

7.7 Results of Johansen Cointegration Test in a Bivariate System

(1980:1-1998:4) 207 7.8 Estimation Results of the ARFIMA (0, d, 1) Models 209 7.9 VECM Results for Commercial Real Estate, Stock and

Macroeconomic Variables 211 7.10 FIECM Results for Commercial Real Estate, Stock and

Macroeconomic Variables 214 7.11 Summary Table of Main Features of VECM and FIECM

Results 216 7.12 Comparison of Post-sample Forecasts of the Office Price Index 218

1.1 Distribution of Commercial Real Estate Space in Singapore 12

1.2 Distribution of Commercial Real Estate Capital Value in Singapore 12

1.3 Commercial Property Market, Stock Market and Macroeconomy: Theoretical Framework 20

1.4 Empirical Research Framework 27

3.1 Office Space Cumulative Supply and Occupancy Rates 83

3.2 Retail Space Cumulative Supply and Occupancy Rates 83

3.3 Geographic Distribution of Existing Office Stock 84

3.4 Price Movements of Commercial Real Estate and Stock Markets 97

3.5 Movements of the Macroeconomic Variables 98

3.6 Index Returns of Commercial Real Estate and Stock Markets 98

4.1 Singapore Office and Stock Price Indices (1980:1-2001:4) 109

4.2 Index Returns of Singapore Office and Stock Markets (1980:1-2001:4) 109

5.1 Correlogram of the LSGXA and DLSGXA, LPPIO-LSGXA Cointegrating Vector (1975:1 – 2001:4) 158

5.2 Correlogram of the LSGXA and DLSGXA, LPPIO-LSGXA Cointegrating Vector (1975:1 – 1987:3) 158

5.3 Correlogram of the LSGXA and DLSGXA, LPPIO-LSGXA Cointegrating Vector (1987:4 – 1996:1) 159

7.1 Differenced Logarithms of Macroeconomic Variables 201

7.2 Autocorrelation Function of the Equilibrium Error of Relationship in Economic System 209

1.7 Empirical Research Framework and Methodology

1.8 Scope of the Research

1.9 Outline of the Research

CHAPTER 1 INTRODUCTION

1.1 Background

Commercial real estate is an important asset in the economy It is different from other resources used by companies, and the relevant market also differs in many ways from other types of markets (Weimer, 1966) Commercial real estate is not only a factor of production but also an important investment A distinction is thus made in commercial real estate ownership between use decisions versus investment decisions At the basic level, commercial real estate

is an economic generator and contributor It serves the need of contemporary businesses, institutions, and individuals for an enclosed space in which their business activities are conducted (Roddewig, 1993) On the other hand, investors, regarding it as a capital asset, also drive the demand for commercial real estate space The dual functions of commercial real estate thus suggest that there are two distinct but interrelated real estate markets: the market for tenant space (space market) and the market for investment capital (the asset market) (Dipasquale and Wheaton, 1992)

The demand for commercial real estate space as a factor of production is linked to the conditions of the broad economy The demand for commercial real estate accommodation rests

on business firms’ production decisions Consequently the demand for commercial real estate

is a derived demand that is driven by the need for goods and services produced using commercial real estate space In this way, movement of the commercial real estate market would be influenced significantly by business firms’ decisions on resource allocation to commercial real estate space that are mainly based on the current and expected macroeconomic conditions.1 Therefore, commercial real estate is linked to the economy, and

1 The terms “real estate” and “property” are used interchangeably in this dissertation

As an asset, commercial real estate has received renewed attraction from many investment programs, especially at a time when the high-return boom years fade In the past decade, commercial real estate, in the equity side, has developed into an asset class with viable investment choices in two different forms:

(a) Direct commercial real estate market This is private investment in individual/joint ownership of physical commercial buildings such as office or shopping centers The main players in the direct real estate market are long-term institutional investors such as pension funds and insurance companies For example, about 45% of commercial real estate in the

UK is held by “investors’ and nearly half of this is held by institutional investors (Callender and Key, 1996).2

(b) Indirect commercial real estate market This is public investment in ownership of commercial real estate holdings Like stock investment, this type of securitized real estate vehicle enables retail investors to participate in the equity ownership The form of indirect equity commercial real estate varies from country to country but comprises two main classes The first and most general type of indirect real estate is listed property stock Some

of these “pure play” property companies are property developers and some are property investors.3 Most are involved in a combination of real estate investment, development and trading activities The second type of indirect real estate comprises Listed Property Trusts (LPTs) in Australia4 or US Real Estate Investment Trusts (REITs)5, both of which offer a share in an unlevered portfolio of real estate

Over the last decade, the market for indirect equity real estate has grown significantly In the

US, the capitalization of equity REITs market was only $8.7 billion in 1991 However, the market capitalization of the REIT industry has increased significantly since 1993 due to the extraordinary number of initial and secondary public offerings (Block, 1998) According to the National Association of Real Estate Investment Trusts (NAREIT) Statistics for the year 2002, approximately $118 billion was raised through initial and secondary public offering of REITs during the period 1993-2002 This trend continued in subsequent years primarily as a result of

2 The terms “private”, “unsecuritized” or “direct” real estate markets are used interchangeably Also applicable are the terms “public”, “securitized” or “indirect” real estate markets

3 Property developers are featured by property acquisition and trading mainly for development profit and property investors are featured by property acquisition and management aiming at obtaining capital gain and rent in the long run (Newell and Chau, 1996)

4 Property Trusts also exist in UK However, they are tradable only with primary issuer and at appraisal-based prices As there is no secondary market, such investments tend to have the characteristics of direct real estate (Chau, MacGergor and Schwann, 2001)

5 Most studies consider equity REITs rather than mortgage REITs An equity REIT acquires direct interests in real estate while a mortgage REIT purchases mortgage obligations backed by the underlying real estate (Chau, MacGergor and Schwann, 2001)

5

the increased institutional demand for REIT stocks By the end of 2002, publicly traded REITs’ total size had grown to $151.3 billion The LPTs in Australia are also significant From less than $5 million in the early of 1990s, the sector reached a market capitalization of $45.8 billion by the end of 2002 (Property Council of Australia)

In Singapore, listed property stock dominates the indirect real estate investment while equity REITs as an investment tool only emerged from 2002 Over the past ten years or so, the growth in the indirect real estate market (listed property stock) has been impressive The capitalization of securitized property sector has increased from S$1.99 billion at the end of

1986 to S$19.4 billion at the end of year 2001.6 Some of the listed property companies are primarily investment oriented and others are involved in a combination of real estate investment, development and trading locally and regionally.7 As for the REITs market, the first retail-based REIT (Capital Mall Trust) issued by a local property heavyweight CapitaLand was launched in July 2002 and the first industrial REIT (A-REIT) was launched by Ascendas Land in October 2002 Capital Mall Trust owns three major shopping malls in Singapore while A-REIT owns eight industrial properties Both REITs will makedistributions to unit holders out of rental income from these properties Another local property developer, Keppel Land, has also expressed interests in launching the first office-based REIT

Besides Singapore, REIT/LPT market in developed countries has seen its other Asian followers, mainly in Japan, Korea and Malaysia Two REITs in Japan (J-REITs) was listed in the Tokyo Stock Exchange in 2001 Both J-REITs raised a combined 130 billion yen (US$1.05 billion) from the capital market (Ting, 2002) Over the long term, Nikko Salomon Smith

6 The total size of property counter in Singapore covers the property companies incorporated and listed on SGX mainboard and traded in S$, which is in line with the criteria of SGX for property companies eligible to be included in the overall property stock price index, SGX All Property Index Furthermore, the market size for preferential shares is excluded

7 Some key examples of Singapore property companies are City Development, CapitaLand and KeppeLand

Barney Ltd estimates the J-REITs market will grow to 5 trillion yen (US$40 billion) in assets with an equity capitalization of 3 trillion yen (US$24 billion)

Two types of RETIs were introduced in Korea, Corporate Restructuring REITs (CR-REITs) and Ordinary REITs (O-REITs) CR-REITs are devised for companies to dispose of real estate for the purpose of corporate restructuring, whereas O-REITs are normal type of REITs which would invest in properties and distribute their income to investors The popularity of the two types of REITs in Korea varies among the institutional investors While CR-REITs have been well-received among the corporate community with about five consortiums having concluded Memorandums of Understanding to set up CR-REITs, the O-REITs have not been set up as expected Malaysia was the first Asian country in developing LPTs in 1989 Currently, 4 property trusts are listed on the Kuala Lumpur Stock Exchange Three of them were listed in

1989, and the fourth listed in 1997 The listed property trusts mainly focused on office properties in Kuala Lumpur Unlike the US and Australia, these property trusts are also able to cover significant levels of stocks and short-term investments in their total portfolios, with the level of non-real estate in the portfolio being as high as 40% in recent years (Newell et al, 2002)

Given that the real estate-related business builds blocks of the principal activities of securitized real estate, it is expected that the performance of the secutitized real estate be related to the performance of the underlying real estate market The linkage between the two markets from previous studies can be summarized as follows: 8

a) Both securitized real estate and its underlying real estate market share a common component reflecting real estate fundamental;

8 More details about the relationships between the securitized and unsecuritized real estate markets are available in

7

b) Securitized real estate performance leads the underlying real estate performance;

c) There is a long-term equilibrium relationship between securitized real estate and its underlying real estate market in the context of macroeconomy;

d) Both securitized and its underlying real estate markets may be closely integrated with other asset markets

However, indirect real estate, given its rapid development, is still small relative to its direct counterpart (at least) in dollar value In the US, where the indirect real estate market, i.e., the REIT market, has grown very fast in the past ten years, it still takes a small share in the pool of the commercial real estate The universe of commercial real estate in the US was estimated to

be worth $4.3 trillion in 2000 and accounted for approximately 10% of the total US investment market (Conner, Hess and Liang, 2001), whereas the NAREIT index (Equity) had a total market capitalization of about 3% of the total value of commercial real estate or $134 billion at the end of 2000 (www.nareit.com).9 In Singapore, as of the fourth quarter of 2001, using Jones Lang LaSalle (JLL)’s capital values and the Urban Redevelopment Authority’s (URA)10

published stock of commercial real estate in different areas across the island, the capital value

of the entire commercial real estate domain in Singapore, i.e., office buildings and shopping centers, was conservatively estimated to be worth about S$81 billion while the market size of the SGX All Property Index was S$19.4 billion at the end of same year This was less than 25% of the whole commercial real estate realm In South East Asia, approximately 11% of all institutional grade real estate is listed The public market real estate penetration in the global

9 In the US, REITs are classified into three sub-sectors based on investment types: equity REITs, mortgage REITs and hybrid REITs Usually equity REITs take the biggest share of total REITs market in terms of market capitalization Equity REITs account for 97.1% of US$138 billion of total REITs universe near the end of 2000

10 The URA is responsible for the strategic and physical planning of the country so as to optimally use Singapore scarce land resources It formulates and reviews concept plans, conducts local planning and administers the development control and urban conservation functions to regulate private and public sector’s property development proposals

context hovered around 12% in 2000 with Australia being the bellwether (55%) followed by North America (18%) and the UK (17%) (Steinert and Crowe, 2001) The ratio of indirect to direct commercial real estate in terms of capital value would be even smaller if we consider that the share of indirect commercial real estate may represent a minor part in the total value of indirect real estate market.11 Even on very optimistic assumptions in respect of the public equity real estate market’s growth, direct ownership of commercial real estate will dominate the total commercial real estate pie in the foreseeable future (McIntosh and Liang, 1998)

In addition to the relative size of direct and indirect real estate, differences in the characteristics of the two real estate investment vehicles in a mixed-asset portfolio further justify that direct real estate can contribute in the core real estate investment as an effective diversification tool For example, Craft (2001) recommends allocations to real estate of between 6% and 10% when including public real estate and 12% when considering private real estate In an asset/liability framework, Chun, Ciochetti and Shilling (2000) conclude that private real estate allocations should be between 12% to 16% and public real estate between 4% and 10% Furthermore, research on “Real Estate in the Investment Portfolio of a Pension Fund” carried out by ING Investment Management in 1998 found that in general, direct real estate has added value for the total investment portfolio of a pension fund while the indirect real estate has no added value for the portfolio Along this line, direct real estate should form the core real estate investment holdings while a minority investment in indirect real estate could be added to core direct real estate investments for tactical asset allocation purposes and

to increase overall liquidity (Spek and Doorn-Gröniger, 2001)

9

securities (bonds and treasure bills) and real estate.12 In the real estate investment arena, private or direct ownership makes up a dominating proportion For example, the ten largest Dutch pension funds invested, on average, 74% in direct real estate and 26% in indirect real estate in their real estate portfolios at the end of 2000 (Spek and Doorn-Gröniger, 2001).13Despite its obvious significance as a key economic generator in the economy and as an important component in the mixed-asset portfolio, direct commercial real estate is less understood than stock and fixed income securities, especially where its relationship with other two asset markets is concerned While sporadic studies appeared in exploring the relationship between the commercial real estate and stock markets by using securitized real estate as a proxy for the real estate market, the point is that there is inadequate evidence to support the belief that the indirect real estate market should necessarily behave in a similar manner as the underlying direct real estate market (Wang, 2001) The differences between the characteristics

of underlying direct real estate and indirect real estate investments can be briefly summarized

as follows:

a) The indirect real estate is a financial claim traded on stock exchange, which introduces some features of general stock market not present in the underlying direct real estate market In this sense, indirect real estate might not be viewed as an investment vehicle for pure real estate but as “hybrid financial assets that embody the economic characteristics of the underlying real estate market coupled with the volatility of the stock market” (Corgel, 1995)

b) The indirect real estate only represents a small portion of total real estate assets, both

12 Treasure bill (T-bill), usually in 3-month period, is termed as “cash” in investment community

13 The relative levels of direct and indirect real estate property in institutional investors may vary from the market with high public market real estate penetration to that with low public market real estate penetration, and could change over time For example, the ratio of private/public real estate investment of Intech, an managed fund in Australia which has the highest public market real estate penetration in 2000 (55%), varied from 1/3 in 1986 to 1/1.12 in 1995 to 1/3.25 in its real estate portfolios estate at the end of 2000 (Armytage, 2002) The public market real estate penetration in Continental Europe in 2000 stood at 3% (Steinert and Crowe, 2001)

in terms of space and dollar value, whereas the rest is dominated by direct real estate This situation would prevail in the foreseeable future

c) Besides the fundamentals in real estate market, the advisory and ownership structure, company management, accounting policies and taxation are all reflected in the performance of indirect real estate as well (Bradley, Capozza and Seguin, 1998; Friday, Sirmans and Conover, 1999; Capozza and Seguin, 2000) Furthermore, the indirect real estate is always geared All of them would result in a big variation from the underlying direct real estate in terms of the performance

d) Some indirect real estate vehicles such as listed property companies are not “pure players” in commercial real estate They make diversifications to include residential property and (or) hotel, even non-property business to some extent

Table 1.1 below summarizes the advantage and disadvantage of direct and indirect real estate investment

11

Table 1.1 Advantage and Disadvantage of Direct and Indirect Real Estate Investment

Accessibility

Not accessible to retail investors but the institutional investors due to large sums involved

Accessible to both institutional and retail investors due to small sums involved

Extent of real estate

exposure

High

• pure real estate exposure with fundamentals in real estate market reflected in the performance

• diversification to non-property business Availability &

duration of investment

performance

measurement

Source: Author’s compilation

Of the direct commercial real estate, the office market is very important For one thing, office space is essential and closely linked to the operation of modern businesses Hence it can be reasonably expected to be pro-cyclical with the whole economy In the long run, the growth of office space seems to move reasonably in tandem with the growth of office employment which largely depends on the economic structure and economic performance (Dipasquale and Wheaton, 1996) For another thing, office buildings constitute a significant share of the commercial real estate in terms of capital value and probably represent the best type of real estate to own in the long term for the institutional investors (Steele and Barry, 1993)

In particular, the office market in Singapore is quite significant Singapore is one of the Asia’s leading regional financial centers The financial sector, together with the insurance and business service sectors, are traditionally one of the economic drivers and the major users of

the office buildings The majorities of the office buildings in Singapore are held under single ownership by financial institutions or listed property companies for long-term investment Furthermore, the office market dominates the whole commercial market in Singapore in terms

of both stock and value.14 As of the fourth quarter of 2001, published statistics from the URA revealed that there was 100.7 million square feet (sf) of total commercial property space Of this, 67.3 million sf or 67% was office space, while the remaining 33.4 million sf or 33% was retail real estate space Of the total capital value of commercial property estimated at S$81 billion at the end of 2001, the office market accounted for 65% or S$52.65 billion while the retail market accounts for remaining 35% or S$28.35 billion Figures 1.1 and 1.2 depict the distribution of floor area and capital value of commercial real estate in Singapore in terms of property type, showing the leading role of the office sector in the commercial real estate market

Figure 1.1

Distribution of Commercial Real

Estate Space in Singapore

Office Space 67%

Retail Space 35%

Source: URA (2001)

Based on the above discussion, it is found that the office space market is a very important sector in the commercial real estate market Hence, in this dissertation, we focus on the office market in the commercial real estate area and explore its dynamic relationship with common stocks This is meaningful as both the office market and the stock market are two very

14 In the Singapore context, commercial real estate covers office and retail real estate

to number three while the U.S has risen to second place The report also identified Singapore

to be No 1 and No 2 in the areas of FDI / technology transfer and overall infrastructure quality respectively The commercial real estate market, in particular the office sector, has played an important role in enhancing the Republic’s status and attracting multinational companies to establish their regional headquarters In addition, unlike the appraisal-based NCREIF and IPD series commonly used for measuring the US and the UK real estate returns respectively, the Singapore commercial real estate price indices (office and retail) are generated from the transaction prices of properties in the Caveats lodged with the Registry of Titles and Deeds.16

Therefore, artificial smoothing would not play a role in our attempt From the above discussion, this study, though based on a specific country’s (i.e., Singapore) commercial real estate and stock markets, is expected to contribute significantly to the growing knowledge base

in the asset pricing literature with regard to understanding of market integration and price discovery between asset markets in the economy

15 Data would be explained in detail in Chapter Five

1.2 Research Objectives

The main objective of this dissertation is to comprehensively examine the dynamic relationship between the commercial real estate (office) and the general stock markets Specifically, we hope to

(a) Explore the market integration/segmentation between the commercial real estate and the

stock markets

(b) Examine the causal relationship between the commercial real estate and the stock markets

(c) Investigate the long-run and short-term relations between the commercial real estate and the

stock markets in the context of the macroeconomy

1.3 Research Questions

Specifically, this dissertation is conducted to address the following research questions:

Question 1: Is the commercial real estate market (office) integrated with the stock market?

Question 2: What is the causal relationship between the commercial real estate (office) and the

stock markets?

Question 3: Is there a long-run relationship between the commercial real estate (office) market,

the stock market and key macroeconomic variables? What are the implications of the nature of the long-run relationship, if found, for the short-term dynamics

16 Hong Kong is another dynamic Asia-Pacific property market that also offers transaction-based private real estate indices These indices are based on all transactions registered with the Land Registry and published by the Rating

15

between the commercial real estate (office) and the stock markets in the presence

of macroeconomic influences?

1.4 Significance of the Research

To the best of our knowledge, there is no published academic research conducted on the holistic time-series relationship between the commercial real estate (office) and the general stock markets Though some studies approach this issue with real estate stocks used as a proxy for the direct real estate (Lie et al, 1990; Ling and Naranjo, 1999; Wilson and Okunev, 1996; Okunev and Wilson, 1997), the significant differences in the institutional characteristics between the indirect and direct real estate markets result in their dissimilar behaviors Using the direct real estate data series allows a more factual investigation into the linkage between the two markets In addition, the fragmented manner in which previous studies were conducted entails a comprehensive examination of the dynamic relationship between the two asset markets Such research is necessary given the fact that commercial real estate, over the last decade, has developed a sophisticated relationship with all aspects of the capital markets and formed an integral component of contemporary institutional investors’ mixed-asset portfolios (Lend Lease Real Estate Investment, 2002; Prudential Real Estate Investor, 2003) Such study

is more relevant if we consider the increasing interaction and integration between real estate and broader capital markets in the foreseeable future

The research findings from this dissertation would be useful to both academics and practitioners For academics, the notion of the dynamic linkage between these two important asset markets has been central to many theoretical developments in both finance and real estate Specifically, the long-run equilibrium relationship between the asset markets would imply that deviations from each other tend to be eliminated over time and that the error correction mechanism is useful in reducing the predicted mean square error of the martingale

model over long forecast horizons In addition, the incorporation of fundamental economic activities into our analysis of the relationship between the commercial real estate and stock markets would provide useful insights into the underlying mechanism in which the commercial real estate market is correlated with the common stock market, which therefore allows us to capture more economic implications Finally, the examination of the causal relationship between the two asset markets may prove useful in learning more about the commercial real estate market through studying the process of price discovery between the commercial real estate and the general stock markets

For practitioners, the dynamic relationship between these two asset markets has important significance for portfolio diversification, strategic management of corporate commercial real estate, and market structure and policy-making From the portfolio management perspective, findings from this study would provide valuable insights for institutional investors who aspire

to optimize their portfolio returns from efficient asset allocation Meanwhile, this study would aid investors to forecast future performance from one market to another or using economic data for dynamic asset allocation strategies Furthermore, the examination of interactions between commercial real estate and stock markets would shed further light on and provide a better understanding regarding the role of the underlying commercial real estate asset in firms’ stock market performance Given that more non-property based companies, other than property companies, are investing significantly in commercial real estate, be it for operational, investment or development purpose, the degree of correlation between the two markets has important ramifications for dynamic allocation of corporate resources and strategic management of corporate commercial real estate

Finally, the finding of this study is significant for public policy issues dealing with market structure and regulatory impediments concerning the movements of capital among various asset markets based on the relevant macroeconomic contexts The exploration of extent to which the commercial real estate market are related to stock market would help the policy-

17

makers to better understand the structure and efficiency of capital market and to implement relevant measures to facilitate the market integration by technological innovation or increased securitization or curb the capital flows among various capital markets through imposing government regulatory barriers

1.5 Research Data

The broad-based nature of this research necessitates two types of data sets to be constructed Table 1.1 provides an overview of the data requirements for the research The first set of data includes the price indices of the office and the stock markets in the Singapore context for the period from 1975 to 2001 The office price index is compiled and provided by URA, whereas the price index of the general stock market is available from the Singapore Exchange Ltd (SGX) The second set of data includes the relevant macroeconomic variables It comprises the measurements of four economic activities from 1980 to 2001: Gross Domestic Product (GDP), interest rate, inflation and exchange rate, and can be obtained through the Singapore Department of Statistics

Table 1.2 Overview of Research Data Requirements

▪ Stock

▫ Office price index

▫ SGX all share index

URA/ Singapore Exchange

▫ Office price index

▫ SGX all share index

URA/ Singapore Exchange (c) (3)

▪ Office

▪ Stock

▪ Macroeconomy

▫ GDP

▫ 3-month Interbank rate

▫ Consumer price index

Data sources

▪ Stock

▫ Office price index

▫ SGX all share index

URA/ Singapore Exchange

1.6 Theoretical Framework - Commercial Real Estate Market, Stock Market, and Macroeconomy

Commercial real estate market has more than one face As we discussed before, it can be considered as a factor of production and refer to the market for the use of the physical space It can also be regarded as a capital asset and refers to the market for ownership interests in the property In this regard, commercial real estate has been put into the arena of the general capital market and has to compete with all other assets, stock in particular, for a place in investors’ portfolios Past research tended to focus on a separate analysis of each of these two markets (Fisher, 1992) In other words, the attention has been paid to understand how changes

in supply and demand affect equilibrium in either the space market or the real asset market This disconnection “neglects the growing role of real estate in the general capital asset markets where real assets are included in the institutions’ investment portfolio to diversify risk” (Sing, 2001) Furthermore, the interaction between the general asset market and macroeconomic conditions compounds the perplexity To have a holistic understanding of our theme – the dynamic relationship between the commercial real estate asset market and stock market – it is important to take into consideration the interaction among the commercial real estate space market, commercial real estate asset market, general capital market and macroeconomic situation

Figure 1.3 provides a stylized and generic picture of the dynamic relationship between commercial real estate asset market and stock market The objectives of the conceptual framework analysis are threefold:

(a) To provide the insights into the importance of, and linkages among three

markets – space, property asset and general capital asset within the national macroeconomic context

19

(b) To identify the linkage between the commercial real estate (asset) and

stock markets – market integration and causal relationship

(c) To relate the long-run movements of the commercial real estate (asset)

market and stock market to the macroeconomic conditions

1.6.1 Overview of linkages among three markets – space, property asset and general capital asset within the national macroeconomic context

1.6.1.1 Commercial Property Space Market

A healthy economic growth, as indicated by credit expansion and modest inflation, etc., usually boosts the corporate economy with business growth, increases in industrial output and more employment The expansion of corporate operations drives the end-user demand for commercial property space Other than business cycles, the secular change in the mix of urban economic structure, technologies and demographic patterns, occurring in the “real economy” and outside the realm of property space and property asset markets, have their own long-term effects on the demand for space This derived demand for space, given the existing supply of leasable space, competitively determines current rental rates (i.e., space market equilibrium) and the expected cash flows (NOIs) of a commercial property The inherent risk of the NOIs depends on the degree of uncertainty about future space market equilibrium (Archer and Ling, 1997)

In general, institutional investors’ strategic investment portfolios comprise stock, fixed income securities (bonds and treasure bills) and real estate In the capital market, real estate has to compete with other assets for a place in institutional investors’ portfolios According to modern portfolio theory, investors diversify their investment based on the risk-free rate and the variance and covariance of returns among the possible assets As investors bid for their own optimal mixed asset portfolio, their bidding simultaneously determines the current required risk premiums for investment according to their risk (variance and covariance of return among the possible capital assets in the past) profile

Although the general risk-free rate and the required risk premiums are seemingly determined

in the general capital market, the main driving forces behind the process come from the macroeconomic situation The risk-free rate is determined in the fixed-income security market

by the bid and ask prices of Treasure market investors around the world based on their projected global economic growth in general, yield (of bonds and T-bills) and inflation in particular As for required risk premiums, Arbitrage Pricing Theory (APT) proposes that there are several sources of macroeconomic risk and that most assets respond to fluctuations in these risk factors The example of these risk factors in the economy includes growth in industrial production, unexpected inflation, real short-term interest rates and difference between the returns on short-term and long-term government bonds Therefore, two general components from the capital market – risk-free rate and investor risk premiums requirements – enter into the expected rate of return of real estate investment and competing investment alternatives such as stocks, bonds

Commercial real estate is a special capital asset in the capital market given its dual roles in the economy On the one hand, it serves as a factor of production, which is linked to the space market On the other hand, it is also an important investment vehicle, which is linked to

general asset market As a consequence, the risk premium for real estate investment cannot be determined by considering only the riskiness of cash flows in the space market, or strictly from capital market In other words, the appropriate risk premium in the commercial real estate is jointly determined by the risk preference articulated in the general capital market and the specific risk profile in the real estate space market (mainly cash flows) priced in the capital market in relation to alternative investment opportunities

1.6.1.3 Commercial Property Asset Market

Commercial real estate asset market is the confluence between the general capital market and real estate space market The expected, or required rate of return (discount rates), real estate values, cap rate and construction feasibility are all determined in this market First, the required rate of return is determined by the interaction among risk-free rate, investor risk premiums requirements and the riskiness of the net rental income (NOI) to be required in the commercial real estate space market (Archer and Ling, 1997)

Secondly, real estate value is determined from the required rate of return and the forecasted NOI with the cap rate merely as the reciprocal of “P/E ratio” (property price/NOI multiple) Given the integration between commercial real estate asset market with other capital and financial asset markets such as stocks and bonds, the commercial real estate value would react

to changes in the general capital market even in the absence of a change in current or projected cash flow streams

Thirdly, the construction feasibility and pace of new construction are determined in the commercial real estate asset market If the current level of property prices is greater than the cost of new construction which depends on construction cost, land availability and cost and cost of capital, together with the macroeconomic shocks indicating the extent of bank and institutional lending, would determine the rate and the extent of real estate development to

capture the excess profits Though the real estate space market would reach equilibrium in the long run, the inelasticity of supply of space due to long construction time and rent level due to leasing term in the short run would prompt commercial real estate market to be highly cyclical before market-clearing real estate prices equaling to construction costs

1.6.2 Identify the linkage between the commercial real estate asset and stock markets

– market integration and causal relationship

Based on our conceptual framework, there are three channels for the linkage between the commercial real estate asset and the stock markets – market integration and causal relationship First, in the traditional neo-classical economics, the commercial real estate is treated as a factor

of production, a similar role as plant and machinery for business firms It is assumed that in equilibrium, the commercial real estate is supplied (or demolished) in reaction to changes in production technology and organization On the other hand, the stock market is a barometer of the macroeconomy and its performance reflects underlying corporate performance Corporate growth in profitability (usually associated with higher stock price) results in business expansion that further leads to rising rental level given increased demand for commercial real estate and inelastic short-run supply Rising rents generally lead to higher capital values in the commercial real estate asset market In recession, a reverse process happens Along this line, the performance of the stock market (corporate economy) has a significant influence on the movement of the commercial real estate asset market through the commercial real estate space market (factor of production) Under this scenario, the demand for real estate is a derived demand, and developers are likely to simply respond to price signals originated from shifts in occupational and investment demands

Second, commercial real estate is considered as an important investment asset other than a factor of production in a more sophisticated model (Corcoran, 1987; Fisher, 1992; Dipasquale and Wheaton, 1992; Fisher, Wilson and Wurtzebach, 1993) As a financial asset, the rental

income from the user market becomes the cash flow priced in the general capital market As a feedback, the commercial real estate price determined in the general capital market has an influence on construction feasibility and therefore space supply However, from a typically comparative static approach, the commercial real estate market is considered as responding to, rather than impacting, the outside environment (Lizieri and Satchell, 1997)

A third strand of relationship between the commercial real estate asset market and stock market involves the relative profitability between the built environment and industrial production from the urban social science literature, as suggested by Harvey’s original theory of capital switching in an advanced market economy (Harvey, 1982, 1985, 1989; Beauregard, 1994) From a Marxian perspective, Harvey’s theory considers the built environment (the second circuit of capital) as a relative self-governed kingdom over the sphere of corporate sector (the primary circuit of capital), which is reinforced by the treatment of commercial real estate as an investable asset He further reasons that, if the return falls in the industrial production sector, the capital is likely to switch to the second circuit (the real estate industry)

to pursue higher profit margins Together with other possible speculative and bank lending in real estate, the capital-switching model suggests that higher returns to commercial real estate would be linked to low returns to the corporate sector and vice verse In fact, recent years have witnessed the capital flows from a falling and volatile stock market to the commercial real estate market as a safety arena in the US and other developed countries In that sense, there is

an impact of commercial real estate on the overall market movement besides the conventional influence of corporate sector on built environment In other words, there is a feedback effect between the built environment and corporate sector This argument coincides with the modern portfolio theory to some extent Commercial real estate, as an investable asset, has to compete with all other assets for a place in investors’ portfolios Investors look through alternative asset classes to find the most appropriate combination of assets to include in their portfolio Each asset class is evaluated both for its contribution to portfolio performance and total portfolio risk

1.6.3 Relate the long-run movements of the commercial real estate (asset) market and

stock market to the macroeconomic conditions

As shown in Figure 1.3, the long-term movements of the commercial real estate asset market and the stock market both are driven by the macroeconomic conditions The influences of macroeconomic shocks on commercial real estate asset market appear in two ways First, in the real economy arena, the corporate economy, which determines the commercial space demand and rental rate given the existing supply, is underpinned by the general economic condition The construction market, which determines the potential supply of commercial space, is also influenced by such macroeconomic shocks as extent of bank and institutional lending The projected rental income is then used to determine the commercial real estate value based on required rate of return for commercial real estate investment Secondly, in general capital market, the sources of systematic risk in the general economy would be reflected in the required risk premium of commercial real estate investment which is an important determinant

of required rate of return

Likewise, the general economy impacts on stock market through two channels First, in the real economy context, the corporate economy is bolstered by the economic performance The corporate revenue is then become the cash flow priced in the stock market The influence of macroeconomic variables on the stock market in the general capital market context is similar to that in the commercial real estate case

Based on above analysis, the two asset markets are exposed to economic conditions and interact with each other within the economic system The general economy, through the corporate economy

in the real economy and the expected rate of return in the general capital market, is expected to tether the movements of the commercial real estate asset market and the stock market in the long run In other words, any long-term linkage between the two asset markets, if present, could be attributed to their common reaction to business-cycle variables to a great extent

1.7 Empirical Research Framework and Methodology

This study will start with a review of previous work in the relevant fields This will be followed by an investigation of the structure of the Singapore commercial real estate market, the general stock market and the macroeconomy, and a review of their historical dynamics These qualitative investigations and reviews are expected to provide some preliminary evidence on the relationship observed between the commercial real estate and the general stock markets and the underlying causal mechanisms Next, a thorough univariate analysis is conducted to provide a comprehensive description of the distributional traits of all the series utilized through this dissertation Then, three major issues regarding the relationship between the commercial real estate (office) and the general stock markets are explored: market integration/segmentation, causal relationship, long-run and short-term relationship in the economic system Figure 1.4 illustrates the empirical research framework The univariate analysis for the series, the three research issues, and the relevant econometric procedures are briefly explained below

Univariate Time Series Analysis

A thorough univariate linear and nonlinear analysis is conducted for all series used throughout this dissertation We first conduct the standard diagnostic tests on the series Then the linear

dynamics of the series are examined by use of the technique of ARIMA modeling, and the post-sample forecasts of the fitted ARIMA model for the commercial real estate is also

investigated to present a univariate linear prediction benchmark Next, the possible nonlinear structures in the “return” series are examined with the techniques in chaos theory