The Efficiency Analysis of Singapore Real Estate Investment

The Efficiency Analysis of Singapore Real Estate Investment Trusts(APEF 36) Hui. Chen Chiang Ming Chuan University, Taipei, Taiwan

Table of Contents Background Research Background and Objectives Literature Review Research Methods Findings Conclusion

One of the most influential city 3 rd highest annual GDP per capita SINGAPORE Strong financial centre High in HDI

Allowed to establish REITs in 1999 Threats: China and India Facing weak growth First successful launch in 2002 – Capita. Mall Trust Second largest Asian REIT • 28 REITs • 6 stapled trusts

Research Background & Objectives Motivation and Importance: Find efficiency level of REITs and suggest improvement Help shape government policies Strengthen SREIT market position Remain competitive Lack of research on Singapore REITs

Research Background & Objectives 1 st Objective: Find efficiency level of each S-REIT using Data Envelopment Analysis 2 nd Objective: Property-type effect 3 rd Objective: Geographical-diversification effect 4 th Objective: Geographical focus effect 5 th Objective: Size effect

Literature Review Data Envelopment Analysis Stochastic Frontier Analysis Non-parametric mathematical modeling Parametric regression technique Use linear combination of inputs and Based on functional form outputs of the sample firms Deterministic – attributes all deviations from optimal values as inefficiency Assumes stochastic relationship b/w inputs and outputs • Two error terms: efficiencies, noise Can handle multiple inputs and outputs For standard SFA, only handle one input with multiple outputs No need to assign weights to each input and output Weights are assigned to inputs Measures relative efficiency Measures optimal efficiency

Literature Review: Property-Type Property-type effect value-weighted premium, leverage ratio, G&A expenses, cash flow yield • Capozza and Lee (1995) REITs of the same property type have similar returns and movements • Mueller and Laposa (1996) Average return of office and industrial sector are consistently higher than their counterparts. Retail is also consistently lower. • Chen and Peiser (1999)

Literature Review: Diversification Reduction in technical efficiency and increased in scale efficiency with increased in property-type diversification. In overall, no impact. • Anderson, Fok, Springer, and Webb (2002) 64. 4% of transactions that do not diversify experience positive returns while 61. 2% that diversify experience negative returns • Campbell, Petrova, and Sirmans (2003) Does not have an impact on performance • Ro and Ziobrowski (2011), Topuz, Darrat, and Shelor (2005)

Literature Review: Size Larger REITs more efficient than smaller REITs for most years • Anderson, Fok, Springer, and Webb (2002) Merger and acquisitions led to diseconomies of scale • Topuz, Darrat, and Shelor (2005), Ambrose and Linneman (1998) Size effects value-weighted premium, leverage ratio, G&A expenses/Total Assets • Capozza and Lee (1995) Size is positively correlated with efficiency • Topuz and Isik (2009)

Research Method Quarterly report from REIT website Inputs: • Property Operating Expense • Managemen t Fees Listed in SGX from Q 12007 – Q 12015 Outputs: • Total Revenue • Total Assets 14 REITs, 462 REITs observations DEA regressed against factors Tobit Regression Analysis

Research Method Property. Type Geographical Diversification Retail Herfindahl Index Commercial Specialized Others Diversified Geographical Focus Singapore Size Small Medium Overseas Large

Research Results Name Property-type Others (Industrial and Ascendas REIT Business Space) Others Ascott Residence Trust (Hospitality/Residential) Cambridge Industrial Others (Industrial and Trust Business Space) Capita Commercial Trust Commercial Capita. Mall Trust Retail Capita. Retail China Trust Retail Market Focus Singapore Overseas (China) Singapore Overseas (China) First REIT Others (Healthcare) Overseas (Indonesia) Fortune REIT Retail Overseas (Hong Kong)

Research Results Name Property-type Market Focus Frasers Centrepoint Trust Retail Singapore Frasers Commercial Trust Commercial Singapore K-REIT Asia Commercial Singapore Mapletree Logistics Trust Starhill Global REIT Suntec REIT Others (Industrial and Business Space) Retail Commercial Singapore

Research Results 80, 000 Property Operating Expense (Input) Manager’s Fee (Input) In thousands SGD 70, 000 60, 000 50, 000 40, 000 30, 000 20, 000 Total Revenue (Output) 10, 000 0 2007 2008 2009 2010 2011 2012 2013 2014 2015 (Q 1) Year Source: From this research

Research Results Total Assets (Output) In thousands SGD 6, 000 5, 000 4, 000 3, 000 2, 000 1, 000 0 2007 2008 2009 2010 Year Source: From this research 2011 2012 2013 2014 2015 (Q 1)

Research Results Variable Average Standard Deviation Property Operating Expenses 15, 332 13, 864 60, 754 14 4, 381 3, 085 15, 504 551 55, 206 40, 141 173, 794 6, 003 3, 501, 320 2, 496, 828 10, 242, 098 269, 244 Manager’s Fees Total Revenue Total Assets Expressed in thousand SGD Maximum Minimum

Research Results REIT’s Name Ascendas REIT Ascott Residence Trust Cambridge Industrial Trust Capita Commercial Trust Capita. Mall Trust Capita. Retail China Trust First REIT Fortune REIT Frasers Centrepoint Trust Frasers Commercial Trust K-REIT Asia Mapletree Logistics Trust Starhill Global REIT Suntec REIT Total Mean 0. 77188 0. 86109 0. 83712 0. 90767 0. 72215 0. 7793 0. 93527 0. 71815 0. 66294 0. 59648 0. 58591 0. 67642 0. 70945 0. 53324 0. 73551 Standard Deviation 0. 134725 0. 065586 0. 186497 0. 06106 0. 024714 0. 016578 0. 04067 0. 046599 0. 019763 0. 04047 0. 067276 0. 016912 0. 053408 0. 048237 0. 138185

Research Results Operating Manager’s Fee Expense (as a % of Total Revenue) Pearson -. 077* -. 404** Correlation Significant (two 0. 049 0 *, **. Correlation is significant at the 0. 05 and 0. 01 level respectively tailed) Operating Expense and Manager’s Fee have (-) correlation with Efficiency!

Research Results ROA (based on Operating Profit) Debt Ratio (Total Liabilities/Total Assets) Pearson Correlation . 371** -. 155** Significant (twotailed) 0 0. 001 **. Correlation is significant at the 0. 01 level Source: From this research ROA has (+) correlation with Efficiency! Debt Ratio has (-) correlation with Efficiency!

Research Results Retail Commercial Others Mean Standard Deviation 0. 7184 0. 65583 0. 81636 0. 051122 0. 157688 0. 138633 F 65. 949 Sig. . 000* **. Mean Difference between groups is significant at the 0. 05 level (2 -tailed) Source: From this research

Research Results Mean Differences (J) (I – J) Sig. Commercial . 062574* . 000 Others -. 097958* . 000 Retail -. 062574* . 000 Others -. 160532* . 000 Retail . 097958* . 000 Commercial. 160532* **. Mean Difference is significant at the 0. 05 level (2 -tailed) . 000 (I) Retail Commercial Others > Retail > Commercial

Research Results Herfindahl Index Pearson Correlation 0. 013 Significant (two-tailed) 0. 78 Specialized Diversified NO IMPACT! Mean Standard Deviation t Sig. 0. 73901 0. 72737 0. 144256 0. 123035 0. 83 0. 407

Research Results Mean Standard Deviation t Sig. Overseas 0. 82345 0. 094039 Singapore 0. 70033 0. 137443 11. 047 . 000* *. Mean Difference between groups is significant at the 0. 05 level (2 -tailed) Focus on Overseas!

Research Results Mean Standard Deviation F Sig. Small 0. 79119 0. 137576 Medium 0. 69891 0. 105521 21. 062 . 000* Large 0. 71642 0. 150456 **. Mean Difference between groups is significant at the 0. 05 level (2 -tailed)

Research Results (I) Small Medium Large Mean Differences (J) (I – J) Sig. Medium . 092286* 0 Large . 074779* 0 Small -. 092286* 0 Large -0. 017506 0. 511 Small -. 074779* 0 Medium 0. 017506 **. Mean Difference is significant at the 0. 05 level (2 -tailed) Small > Medium = Large 0. 511

Research Results t Sig. Std. error Tolerance VIF 0. 027 34. 644 . 000* 0. 097 0. 012 8. 203 . 000* 0. 761 1. 314 -0. 742 0. 099 -7. 491 . 000* 0. 813 1. 229 -0. 079 0. 012 -6. 606 . 000* 0. 823 1. 215 Debt Ratio -0. 201 0. 073 Size – -0. 037 0. 011 Medium Operating Expense (as a -0. 114 0. 052 % of total revenue) **. significant at the 0. 05 level (2 -tailed) -2. 749 . 006* 0. 791 1. 264 -3. 467 . 001* 0. 981 1. 02 -2. 202 . 028* 0. 699 1. 43 Variable Constant Property type – Others Manager’s fee (as a % of total revenue) Geographical Concentration – Singapore Unstandardized Coefficient Estimated value of – B 0. 94 Collinearity Statistics

Conclusion Property operating expense and management fee have (-) correlation with efficiency level ROA has a (+) correlation with efficiency level Most efficient: FIRST REIT Least efficient: Suntec REIT Propertytype: “others” do best followed by retail Geographical diversificatio n: does not matter Geographical focus: overseas Debt level has a (-) correlation with efficiency level Least consistent: Cambridge Industrial Trust Size: smallsize REIT